A COMPUTER-BASED SYSTEM OF XRF ANALYSIS

AND ITS APPLICATION TO A GEOCHEMICAL STUDY

OF VULSINI VOLCANO, CENTRAL ITALY

by

Robin 1. Parker

Thesis submitted for the degree of Doctor of Philosophy

in the University of London

Department of Geology,
Imperial College of Science and Technology,
London, S.W.7. February 1979
 ABSTRACT

A series of interactive terminal-based computer programs for proces-

sing major and trace element data produced by X-ray fluorescence (XRF)
analysis of rock samples have been developed in conjunction with work
undertaken on multi-standard XRF calibration techniques. In major ele-
ment rock analysis using fused samples and standards, the construction
of standard calibration lines is facilitated by a knowledge of the back-
ground intensities for the elements analysed. As an alternative to measur-
ing the background "off-peak" or on blanks, a background iteration tech-
nique has been developed which results in reduced calibration errors and
reduced analytical time, A flexible trace element calibration method has
also been developed which allows emphasis to be placed on those stan-
dards considered to have the highest quality data.

The major and trace element analytical results may be fed directly
into a series of programs for statistical analysis , data plotting and the
calculation of various- parameters. The integrated and interactive nature
of the programs , combined with the use of sample grouping techniques ,
greatly facilitates the speed with which the data may be evaluated.

The analytical, data processing and data evaluation techniques

have been applied to the study of ignimbrite pumice and lava samples
from Vulsini volcano, central Italy. Certain of the pumice samples have
been affected by the alteration of leucite to analcime. A quantitative
X-ray diffraction (XRD) method was developed to correct the analyses for
this effect. The major and trace element variations in the Vulsini pum-
ices and lavas suggest a low pressure crystal fractionation process which
has resulted in a "High K" (K2O/Na2O > 2.0) evolutionary sequence:
leucite - basanite -o-leucitite/tephritic leucitite/leucite tephrite -1.leucite
phonolite/leucite trachyte --a-trachyte. A separate suite of "Low K" (K2O/
Na2O < 2.0) trachybasalt and trachyandesite lavas are also present.

Average major and trace element data for pumice samples from the
various ignimbrite eruptions suggest a cyclic magmatic evolution for the
volcano. Furthermore, pumice samples from certain ignimbrites show
distinct chemical trends indicating differentiation processes operating on
the scale of the individual ignimbrite magma.
 ACKNOWLEDGEMENTS

This project was undertaken while the author was in receipt of a

research assistantship in the Department of Geology, Imperial College.
The author wishes to acknowledge the support given to the project by
Prof. J. Sutton and Prof. R. Davis.

Sincere thanks are extended to Dr. G. Borley under whose super-

vision the project was carried out and whose encouragement and advice
at all stages in the work is much appreciated.

Fruitful discusssions on various aspects of the analytical work were

held with Messrs. P. Suddaby, M. Frost and P. Watkins. Mr. J. Willis
also provided assistance and advice with regard to major element analysis.
The XRF analytical work was carried out on a spectrometer purchased
with the aid of a Natural Environmental Research Council research grant.

Drs. R. Howarth and M. Thompson and Messrs. S. Earle and

J. Phillis provided useful help and advice on the computing and statistical
aspects of the work. The staff of the Imperial College Computer Centre
were a constant source of help. The computing work carried out in this
project was assisted in no small measure by the expanding range of soft-
ware and hardware facilities made available by the Computer Centre.

Dr. I. Gibson of Bedford College kindly made available the diffracto-

meter used in the X-ray diffraction work. Dr. G. Marriner provided advice
in the operation of the diffractometer. Drs. Gibson and Marriner also
made available the fusing and casting facilities used in the preparation
of the majority of the XRF standards and samples.

The author is .indebted to Dr. S. Sparks for suggesting the geochemical

study on Vulsini volcano and for providing an introduction to the field area
of this volcano. Stimulating discussions were held with Dr. Sparks and
Dr. G. Walker on the vulcanological aspects of the work, while Dr. R.
Thompson gave useful advice on geochemical aspects of the Vulsini study.
/continued...
 The University of London Central Research Fund kindly allocated
the author a grant for travel to Italy and for field expenses.

Mr. G. Bullen assisted in the X-ray fluorescence analysis of the

Vulsini samples while Mr. P. Watkins carried out the wet chemistry
determinations for Na20 and FeO in the samples.

Finally, the author is indebted to those who provided a refuge

from the desolate wastes and in particular to my wife, who also undertook
the onerous task of typing the thesis.
 CONTENTS

CHAPTER I INTRODUCTION 1

CHAPTER II MAJOR ELEMENT ANALYTICAL TECHNIQUES

2.1 Introduction
2.2 Summary of analytical methods 4
2.3 Pre-ignition of rock powders 8

214 . Multi-standard vs single standard calibrations 9

2.5 Calibration line construction 10

2.5.1 Background intensities 10
2.5.2 Quality of the calibration line 11
2.5.3 Backgrounds from blanks 14
2.5.4 Backgrounds by iteration 15

2.6 Data processing 16

2.7 Accuracy and precision 20
2.8 Discussion - 22

CHAPTER III TRACE ELEMENT ANALYTICAL TECHNIQUES 38

3.1 Whole rock pressed powder briquettes 38

3.1.1 Introduction 38
3.1.2 Standards 38
3.1.3 Analytical procedures 40
3.1.4 Data processing 44
3.1.5 Calibration errors and precision 45
3.2 Norrish fusion discs 48
3.2.1 Introduction 48
3.2.2 Sr and Rb analyses 48
3.2.3 Ba analyses 49
3.3 Discussion 51

CHAPTER IV GEOIC DATA PROCESSING SYSTEM 67

4.1 Introduction 67
4.2 Computing environment 69
4.3 First stage of GEOIC data processing 70
4.4 Second stage of GEOIC data processing 70
4.5 Third stage of GEOIC data processing 73
4.6 Discussion 75

CHAPTER V QUANTITATIVE DETERMINATION OF ANALCIME

IN PUMICE SAMPLES BY X-RAY DIFFRACTION 88
5.1 Introduction 88
5.2 Theoretical 89
5.3 Experimental 92
5.4 Results of standard calibrations and discussion 93

CHAPTER VI GEOCHEMISTRY OF PUMICE AND LAVA

SAMPLES FROM VULSINI VOLCANO 100
6.1 Introduction 100
6.1.1 General '100
6.1.2 Sample collection 103
6.1,3 Alteration of pumice and lava samples 104
6.2 Petrography of the pumice and lava samples 106
6.2.1 Pumice samples 106
6.2.2 Lava samples 108
6.3 Major and trace element geochemistry 111
6.3.1 Selection of lava analyses from the literature 112
6.3.2 Sample grouping 114
6.3.3 Geochemical variations 116
6.3.3.1 Introduction 116
6,3.3.2 Major element variations 119
6.3.3.3 Trace element variations 124
6,3.3.4 Major and trace element variations between
and within ignimbrite magmas 131
6.3.3.5 Discussion 132

CHAPTER VII CONCLUSIONS 181

7.1 Summary 181

7.1.1 Analytical and data processing techniques 181
7.1.2 Vulsini geochemistry 184
7.2 Future work 185
7.2.1 Analytical techniques 185
7.2.2 Post-analytical data processing 188
7.2.3 Vulsini geochemistry 188
APPENDIX A CORRECTION OF ANALYSES FOR SECONDARY
ALTERATION 190

APPENDIX B PETROGRAPHY OF THE ANALYSED PUMICE

SAMPLES 195

APPENDIX C PETROGRAPHY OF THE LAVA SAMPLES ANALYSED

IN THIS WORK 203

APPENDIX D VULSINI PUMICE AND LAVA ANALYSES 209

APPENDIX E Na20 AND FeO WET CHEMICAL ANALYSIS 223

REFERENCES 224

SUPPLEMENTARY MATERIAL THREE PRINTED PAMPHLETS

ENCLOSED IN POCKET INSIDE
REAR COVER
 LIST OF TABLES

Table 2.1 Major element analyses on Norrish discs.

Instrumental settings .. .. .. .. .. .. 23
2.2 Oxidation of FeO—.Fe2O3 at 850°C and 950°C
in silicate rock samples .. .. .. .. . .. .. 24
is 2.3 Determination of iterated background for TiO2 • 25
is 2.4 . Comparison of calibration data .. .. .. .. .. 26
" 2.5 Standards used in calibrations (Table 2.4) , . .. 27
2.6 Major element data processing - program START .. 28
" 2.7 Major element data processing - program MWT .. 29
2.8 Accuracy test on Norrish analyses using iterated
backgrounds for the calibration lines .. .. 33
" 2.9 Estimate of combined fusion and counting precision
for ignimbrite pumice sample 6502 .. .. .. 34
Table 3.1 Trace element analyses on pressed powder briquettes.
Instrumental settings .. . , . 0 .. .. .. .. 54
3.2 Trace element data processing -.program TRANS 55
3.3 Trace element data processing - program NUTRACE 56
a ss 3.4 Trace element calibration data for pressed powder
briquettes , . .. .. , . .. , . 57
"
3.5 Trace element analyses on pressed powder briquettes.
Detection limits .. .. .. .. . , .. . • 60
3.6 Trace element analyses on pressed powder briquettes.
Estimate of combined briquette preparation and
counting precision for ignimbrite pumice sample 6502 62
3.7 Trace element analyses on Norrish discs,
Instrumental settings .. .. .. .. . . , .. .. 63
" 3.8 Sr and Rb analyses on Norrish discs. Calibration
data .. .. .. .. .. • • , • 64
N It
3.9 65
Ba analyses on Norrish discs. Calibration data . ,
3.10 Sr, Rb and Ba analyses on Norrish discs.
Detection limits .. , . .. .. .. .. 65
3.11 Thorium in Tenerife pumices by XRF and neutron
activation (NA) analysis .. .. .. .. .. 66
Table 4.1 Program TAB: generation of data file for input into
norm calculation programs .. . . .. .. .. .. 77
" 4,2 Program AGNORMX: generation of calculated norm
data file for input into program TAB .. .. .. 77
 " 4.3 Program TAB: generation of line printer and data
base files for all samples .. .. .. .. .. 78
" 4.4 Program TAB: generation of line printer and data
base files for selected groups of samples .. •. 79
" 4.5 Program TAB: tabulated line printer output for
selected groups of samples .. .. .. .. 80
" 4.6 Program GEOPLOT: generation of binary and
ternary plots .. .. .. .. .. .. .. 81
" 4.7 Program STAT 1: generation of statistical data for
selected groups of samples .. .. .. .. .. 84
Table 5.1 Leucite and analcime analyses - all data on
moisture (H2O-) free basis .. • . .. .. 97
" 5.2 Calibration data for XRD determination of analcime 98
Table 6.1 Vulsini volcano.- ignimbrite eruption sequence .. 138
6.2 Ignimbrite pumice phenocryst concentrations .. 139
" 6.3 Lava phenocryst and groundmass mineralogy .. 140
6.4 Characteristics of Vulsini pumice groups • . .. 141
". 6.5 Characteristics of Vulsini lava groups .. .. .. 142
6.6 Excluded Vulsini analyses from the literature .. .. 143
6.7 Vulsini pumices: average data for ignimbrites and
pyroclastic falls .. .. .. .. .. .. .. 144
" 6.8 Vulsini lavas: average data for lava groups 145
" 6.9 Specific gravities of K-rich lavas and minerals .. 146
" 6.10 Distribution coefficients.. .. .. .. .. 147
Table A.1 XRD determination of analcime in ignimbrite
pumices and lava samples .... .. .. .. .. 192
" A.2 Correction of analysis of pumice sample 303 for
effect of alteration of leucite to analcime .. .. 193
" A.3 Determined weight per cent LOI (loss on ignition,
110°C—). 850°C) in pumice samples .... .. .. 194

Table D.1 Vulsini pumice analyses .. .. .. .. .. .. 210

" D.2 Vulsini lava analyses .. .. .. .. .. .. 216
 1

CHAPTER I

INTRODUCTION

This thesis has three themes: (i) the development of X-ray

fluorescence (XRF) and X-ray diffraction (XRD) analytical techniques for

the determination of element and mineral concentrations in geological

samples; (ii) the development of computer based methods for processing

and evaluating XRF geochemical data; and (iii) the application of (1) and

(ii) to the study of the geochemistry of pumice and lava samples from

Vulsini volcano, Italy.

In the XRF analysis of major elements much work in the literature

has been concerned with the development of firstly flux-fusion sample

preparation techniques and secondly methods for correcting for matrix

effects (Rose et al., 1963; Norrish and Hutton, 1969; Haukka and

Thomas, 1977). However, the calibration methods by which the samples

are analysed have not received as much attention. Chapter 2 is concerned

with a calibration technique that uses an iterated background method for

constructing standard calibration lines XRF techniques for the analysis

of trace elements on whole rock pressed powder briquettes are discussed

in Chapter 3. Particular attention is paid to the development of a flexible

Calibration method and to the evaluation of standard trace element concen-

trations recommended in the literature (Flanagan, 1973, 1976a and Abbey,

1973, 1975a). Trace element analysis (for Sr, Rb and Ba) on major element

fusion discs is also investigated. The final analytical technique, des-

cribed in Chapter 5, is a XRD method for quantitatively determining the

 2

concentration of the mineral analcime in certain Vulsini pumice and lava

samples. Analcime is present as an alteration product of leucite in these

samples and the determined XRD analcime values were required to correct

the sample major and trace element data for the effect of this alteration.

With regard to data processing techniques, Chapter 4 describes the

GEOIC system of interactive computer programs which are designed to be

executed from remote computer terminals. The advent of the digital elec-

tronic computer is causing an on-going revolution in business, industry,

government and not least of all in scientific research. A significant

stage in the development of this computer revolution was reached with the

introduction of powerful user orientated computer systems capable of pro-

viding both batch and interactive modes of data processing in conjunction

. with extensive on-line storage of programs and data files. The wide-

spread introduction of these systems since the early 1970's has provided

a computing environment which has yet to be fully developed and exploited.

The introduction in the mid 1970's of micro-processor based computer sys-

tems and intelligent terminals has provided further impetus towards dis-

tributed terminal-based computing systems in which interactive program-

ming is the rule rather than the exception. While batch systems for hand-

ling suites of geochemical data have been described (Till, 1977), the

GEOIC system represents a significant in depth study on geochemical data

processing and evaluation using interactive techniques.

The above analytical and computing methods have been used to study

the geochemistry of K-rich volcanic samples from Vulsini volcano (Chapter

6) . In addition to lava eruptions this volcano is characterised, like the

 3

other K-rich volcanoes in Italy, by the eruption of large volumes of

pumice-rich pyroclastic material. A considerable amount of geochemical

work has been done on Italian K-rich lavas (Rittmann, 1933; Schneider,

1965; Burri, 1968; Cundari and Le Maitre, 1970; Appleton, 1972;

Cundari and Mattias, 1974; Cundari, 1975; Ghiara and Lirer, 1976)

whereas very little work has been done on pumice samples from the asso-

ciated ignimbrite and air-fall eruptions. The work described in Chapter 6

represents the first detailed geochemical study of both lava and pumice

samples from one of these Italian volcanoes .

With regard to the Vulsini study the main thrust of the work has been

with respect to variations in the bulk sample major and trace element con-

centrations. Studies on the chemistry of the minerals from these samples

are considered to be outside the scope of the present thesis in view of the.

further time required for this work and in view of the considerable time and.

effort expended in the development of the XRF, XRD and computing tech-

niques discussed in Chapters 2 to 5.

 4

CHAPTER II

MAJOR ELEMENT ANALYTICAL TECHNIQUES

2.1 Introduction

Work was carried out on the determination by XRF spectrometry

of Si02, A1203, Ti02, Fe203 (total), Cr203, MnO, MgO, CaO,

K20, P2 O5 and NiO in fused silicate samples. The various factors

affecting the quality of major element XRF analysis of fused rock samples

have received a considerable amount of attention in the literature and a

variety of analytical schemes have been put forward (Rose et al., 1963;

Welday et al., 1964; Gunn, 1967a,b; Norrish and Hutton, 1969;

Fabbi, 1972; Harvey et al., 1973; Haukka and Thomas; 1977) . In

proposing these schemes of analysis a number of workers have used

single standard methods of calibration (e.g. Gunn, 1967a,b and Norrish

and Hutton, 1969) while others have recommended multi-standard calib-

ration techniques (e.g. Fabbi, 1972 and Harvey et al., 1973) . This

latter method has certain inherent advantages over the former and in this

chapter multi-standard calibration and sample analysis techniques will

be discussed with particular reference to the determination of background

intensities (Parker, 19786).

2.2 Summary of analytical methods

In this section the analytical methods adopted will be summarised.

Various aspects of the methods will be discussed in more detail in

subsequent sections.
 5

(1) Loss on ignition (LOI) values are determined by igniting

standard and sample powders at 850°C for 30 minutes. Higher tempera-

tures were found to cause certain glassy pumice samples and volatile

rich granites to become partially fused. The recommended concentrations

for the calibration standards are corrected for the effect of loss (or gain)

on ignition.

(2) The pre-ignited standard, blank and sample powders are

fused with a flux. The advantages of flux-fusion preparation of rocks

for XRF analysis have been covered in detail in the literature (Rose et al. ,

1963; Ingamells, 1970 and Harvey et al:, 1973). Two different methods

of sample fusion are available. Pure lithium tetraborate flux is used in

the first method of fusion, and the flux to sample ratio is normally 7 : 1.

The fusion melt is quenched to a glass and then ground and briquetted.

Successful analysis using this fusion technique requires careful attention

to the briquette preparation process. Factors such as the briquette press-

ing time and pressure should be standardised, and the surface of the press-

ing die should be maintained in a mirror smooth condition. The selected

grinding technique should ensure that the ground glass grain size distri-

bution is kept below 10 microns for all sample fusions (see Ingamells,

1970, p.327). These and other factors affecting this method are discus-

sed in some detail in Parker (1977).

The flux for the second method of fusion is the Norrish mixture of

lithium tetraborate, lithium carbonate and lanthanum oxide (47.0 : 36.6 :

16.3). The flux and sample powder weights are 1.5g. and 0.28g. res-

pectively, and 0.02g. of NaNO3 is also added to the fusion which is cast
 6

as a glass disc. This method has been described in detail by Norrish

and Hutton (1969) . NaNO3 is added as recommended by these workers

to ensure oxidising conditions. NaNO3 also assists in forming a homo-

geneous glass (Palme and jagoutz, 1977). Harvey et al. (1973) discuss

the Norrish method in detail and the plunger apparatus described by these

latter workers greatly facilitates the ease with which the glass discs may

be cast. Using this plunger "apparatus means that no special expertise

is required in preparing the glass discs and this removes the objections

to this method raised by Ingamells (1970) and Fabbi (1972) . Fabbi

(1972, p.237) considers that preparing ground glass briquettes is quicker

than casting glass discs. However, experience in our laboratory has

shown that the plunger apparatus enables satisfactory glass discs to be

cast on a routine basis at greater than twice the speed at which ground

glass briquettes can be prepared.

(3) The fused samples are analysed in a Philips PW 1212 XRF

spectrometer. The instrumental conditions are listed in Table 2.1.

Data processing is carried out using techniques based on Parker and

Willis (1977) . All count rate data are corrected for machine drift and

dead time. The count rate data from briquettes prepared by the first

method of fusion are corrected for matrix effects using the methods 'of

Gunn (1967a ,b). The matrix correction procedures described by Norrish

and Hutton (1969) are used to correct the count rate data from discs pre-

pared by the second method of fusion.

(4) In constructing the calibration line for a given element

a number of standards (usually from 6 to 14) , each prepared in duplicate

 7

or triplicate, are analysed. Each preparation is normally counted at

least twice. Matrix corrections are applied to the XRF intensities ,

rather than to the standard concentrations as done by Norrish and Hutton

(1969) and Harvey et al. (1973). The background count rate (determined

on blanks or by iteration) is subtracted from the peak count rate to give

the net peak intensity for each standard count. Mg backgrounds for

standards and samples are corrected for the effect of crystal (RbAP)

fluorescence using techniques described in Parker and Willis (1977) .

(5) The matrix corrected net peak intensities for each of the

standards used in the calibration are then summed and the standard con-

centrations are also summed. The slope factor of the calibration line

is then given by:

sum of standard concentrations

slope factor (SF) =
sum of net standard counts per second

(6) The count rate data for unknown samples are converted

into background corrected concentrations using the calibration slope

factors and the appropriate Gunn or Norrish matrix corrections.

(7) Na20 and FeO values for the unknown samples are deter-

mined by wet chemical techniques*. These data plus H2O- (110°C) and

LOI (850°C) are incorporated in the final sample analysis, and the XRF

oxide concentrations are adjusted accordingly.

(8) Variations in the weight of powder taken for Norrish stan-

dard or sample fusions are allowed, and the ratio of actual weight against

recommended weight is used to correct the concentrations assigned to the

* see Appendix E
 8

standards or samples. This facility allows considerable variation in

powder weight and this can be useful when reduced quantities of material

are available, as can be the case in mineral analysis. Norrish and

Hutton (1969, p.446) describe these procedures for their method and they

note that they have successfully analysed samples varying in weight by

0.25 to 2.0 times the recommended weight.

2.3 Pre-ignition of rock powders

Pre-ignition of rock powders has been adopted by a number of

workers (e.g. Compston et al., 1970; Willis et al., 1971; Borley, 1977;

Haukka and Thomas, 1977). The purpose of pre-ignition is to remove

volatiles and to oxidise FeO to Fe203 so that on fusion volatile loss and

oxygen gain should not occur. Volatile loss and/or oxygen gain result-

ing from ignition will alter the sample element concentrations and this

change may be calculated from the weight loss or gain on ignition. In

the ignition of silicate rocks at 850°C and 950°C the oxidation of FeO to

Fe203 is, however, incomplete as shown in Table 2.2 and Fig. 2.1.

Volatile loss may also be incomplete especially for minerals such as

topaz, epidote, staurolite, talc and scapolite (Riley, 1958). Extending

the ignition time beyond 30 minutes will assist the oxidation and volatili-

zation process but may result in excessive loss of volatile elements such

as Na. The data for ignition at 850°C in Table 2.2 and Fig. 2.1 show

that the residual FeO for the three samples with greater than 12% original

FeO is relatively constant at 5.4%. For samples with less than 12%

original FeO there is a trend (r = 0.801, slope = 0.561) of increasing resi-

dual FeO with increasing original FeO. The 950 °C data for samples
 9

containing less than 12% original FeO is similar to that for ignition at

850`C although as expected the residual FeO values are lower (Fig. 2.1).

The 950°C data for the three samples containing high original FeO contents

(> 12%) is lower and more variable than that for ignition at 850°C.

In calculating the volatile loss on ignition the measured weight

loss will be too low due to oxidation of Fe O to Fe203 (i.e. a weight gain)

If the original FeO in the sample is known then-the effect of this oxida-

tion should be included in the calculation of the volatile loss. However

if it is assumed, incorrectly inmost cases, that all of the FeO has been

oxidised then the calculated volatile loss will be too high. Determin-

ation of ignition volatile loss therefore requires knowledge of the residual

FeO in the ignited sample. In the absence of individual post-ignition

FeO values , the residual FeO in ignited silicate rock powders may be esti-

mated to the first approximation from the data presented in Table 2.2 and

Fig. 2.1. The correction of the observed sample volatile loss for FeO

oxidation will be discussed further in the sections on data processing and

the overall accuracy of the major element analytical procedures.

2.4 Multi-standard vs single standard calibrations

Multi-standard calibration techniques (e.g. Harvey et al., 1973)

offer a number of advantages over single standard methods (e.g. Norrish

and Hutton, 1969) . (1) The recommended concentrations for international

rock standards will contain errors (Fairburn et al., 1951; Ahrens, 1957;

Flanagan, 1969, 1975; Abbey, 1977) and these errors will tend to be

balanced out by the use of a number of standards. (ii) Sample

 10

preparation will introduce errors . These errors will tend to be balanced

out by the preparation of a number of standards. (iii) Standards that

fall significantly off a multi-standard calibration line are usually easily

identified and the reasons for their lack of calibration may be investigated

and/or the standard removed from the calibration. In a single standard

method there is no way of knowing if the recommended element concent-

rations for the standard are in serious error. (iv) A good calibration

line established for a range of standards will encourage confidence in the

various correction procedures used to calculate the final concentration vs

intensity data for the standards and samples. A single standard method

will not provide this check.

2.5 Calibration line construction

2.5.1 Background intensities

A feature of the method of calibration described in section 2.2 is

that the calibration line will automatically go through the origin. Because

of this any errors in the background count rate used to calculate the net

peak intensities will directly affect the quality of the calibration line.

Harvey et al. (1973) recommend that in their method of analysis the

background count rates for Norrish fusion discs should be determined off-

peak at convenient 219 settings for each element. For several reasons

backgrounds measured in this way have not been used in constructing

calibration lines as described above: (i) the total counting time for

samples and standards is significantly increased; (ii) for certain ele-

ments accurate background values are difficult to obtain due to the sloping
 11

nature of the background (an example is TiKoc which is on the shoulder

of the La Ln line - La is the heavy absorber in the Norrish fusion flux);

(iii) major element flux impurities will increase the effective backgrounds

for the elements concerned and backgrounds measured off-peak will

therefore be too low. The same effect will occur where there is pri-

mary X-ray beam spectral contamination (e.g. FeKd from Fe contamina-

tion in the target of tungsten or chromium X-ray tubes , see Norrish and

Hutton, 1969, p.437).

Norrish and Hutton (1969., p.443) recommend that background values

should be determined on blank . As an alternative to using blanks , an

iterative technique for determining background values is described in this

chapter. The "blank" and "iterative" methods for determining - backgrounds

do not suffer from the problems outlined above for the "off-peak" method.

2.52 Quality of the calibration line

In constructing calibration lines it is useful to have criteria for

evaluating the quality of any given calibration. The criteria adopted are

the average relative and average absolute errors of the standards with

respect to the calibration line. They are calculated as follows (Fig.2.2).

The net peak intensity for each individual standard is multiplied by the

slope factor (SF) to give the calculated (as opposed to recommended con-

centration for that standard on the calibration line. The difference bet-

ween the recommended and calculated concentrations is then the absolute

error for that standard. The relative error (as a percentage) is similarly

calculated. The absolute and relative errors for all the standards used

in the calibration are then averaged to produce average relative and

* corrected for dead time, machine drift and matrix effects

 12

average absolute errors for that calibration. Fabbi (1972) has used.a

similar approach in evaluating the analytical errors in his XRF method.

In the SF method of constructing the calibration line (see 2.2(5) )

the high concentration and therefore high intensity standards contribute

relatively more to the resulting slope factor as compared to the lower

concentration standards. This is desirable because of the better count-

ing statistics at the higher concentrations. Furthermore, the quality of

the analytical data for the lower concentration standards may not be as

good as that for the higher concentration standards. In their seminal

study on the variability of inter-laboratory analytical data for inter-

national rock standards, Fairburn et al. (1951) show that, for the major-

ity of oxides, there is an increase in the precision of the reported data

with an increase in the concentration level of the oxide being determined

(see also Ahrens , 1957; and Lister and Gallagher, 1970) .

Many workers use least squares regression techniques in const-

ructing X-ray fluorescence calibration lines. In these techniques all

the standards are weighted equally in the calculation of the regression

line. While procedures are available to weight the data points in a

least squares regression (Draper and Smith, 1966), it appears that such

procedures are seldom used as no references to weighted regressions

were found in the literature on XRF calibration techniques applied to

fused rock standards.

Draper and Smith (1966, p.80) discuss weighted least squares re-

gression and in particular the case where the variance of the dependent

variable y is proportional to the size of the corresponding independent

 13

variable x (i.e. VZ = V(y) = kx) . The standard deviation of XRF counts

is a = count, and hence the variance of the count is g2 = count.

In the calibration procedures described in section 2.2 above, the original

XRF counts (or intensities) measured on the fusion discs or briquettes

require correction for dead time, machine drift, backgrounds and matrix

effects, i.e. the variance of the original count is not exactly propor-

tional to the concentration x in the standard. This lack of proportion-

ality will be dependent on the size of these corrections. For those ele-

ments for which the corrections are relatively small the variance V(y) of

the original XRF count will be proportional to the first approximation, to

the concentration x of the element in the standard. Draper and Smith

show that for the case where the variance of y is proportional to x, the

slope b of the weighted regression line is:

b =
x

This method of regression is identical to the slope factor (SF) calib-

ration method discussed in section 2.2 above, except that the SF is in

fact = 1/b due to the way in which the nominal sample concentrations

are calculated, i.e. in program MW, Parker and Willis, 1977, the
sample XRF count rate data are multiplied by the SF and not divided by

the SF, which would be the case if the SF = b. With regard to size of

the corrections required for matrix effects, dead time, machine drift and

backgrounds, the following points may be noted. For normal silicate

rocks the matrix corrections are on average ± 3% (relative) for Norrish

fusions and ± 6% (relative) for the pure lithium tetraborate fusions (7 ; 1

dilution) (see Norrish and Hutton, 1969, and Parker, 1977). Dead time
 14

and machine drift corrections are normally less than 1 - 5% (relative),

while the background correction as a relative proportion of the total

count is on average less than 2 3% for elements such as Si, Al, Fe,

Ca and K.

A further point to be considered when using least squares regres-

sion methods is the underlying assumption that while the dependent

variable (i.e. the XRF count rates) may contain errors, the independent.

variable (i.e. the standard-concentrations) should not. This may not

be the case for XRF calibration lines based on fused rock standards.
The adopted standard concentrations for the disc or briquette will con-

tain errors both in terms of the "preferred" or "recommended" values for

these rock standards, and in terms of the errors associated with the

actual preparation of the standard (i.e. the errors associated with the

ignition, weighing, fusing, casting or pressing of the disc or briquette)

(see Till, 1974, p.99 and Madansky, 1959, for further discussions with

regard to regressions involving errors in both variables) .

2.5.3 Backgrounds from blanks

Fig. 2.3 shows a calibration line constructed for Ti02. The

standards and blanks in this calibration were prepared using the method

of Norrish and Hutton ,(19 69) . The relative errors are higher for the

lower concentration standards. This is considered to reflect the poorer

counting statistics, and may also reflect the quality of the recommended

data , at these lower concentrations. The standard data used in the

calibration are taken from Abbey (1973) and Flanagan (1973) . The dif-

ferences in the actual concentration data assigned to the standards in

 15

Fig. 2.3 as compared to the data presented by Abbey and Flanagan

reflect factors discussed in points (1) and (8) in section 2.2 above.

The matrix corrected background count rate value used in cons-

tructing the TiO2 calibration line in Fig. 2.3 was determined from the

average of two 5102 and four CaO blanks. Matrix corrected background

count rates determined on blanks prepared from different oxides may vary

by up to 10.per cent. As a result, Norrish and Hutton (1969) recommend

that blanks should be prepared from several different oxide compositions

and the results averaged. Furthermore, in order to guard against conta-

mination, each blank should be prepared at least in duplicate.

Further examples of calibration lines constructed using backgrounds

• determined from blanks are given in Parker and Willis (1977, p.154-157).

2.5.4 Backgrounds by iteration

Table 2.3 shows the determination of an iterated background for

TiO2 calibration line. The iteration procedure for a given calibration

line is carried out by initially setting the background count rate to zero

counts. per second (c.p.s.) followed by the calculation of the average

relative error (a .r.e.) and the average absolute error (a .a .e.). The
background is then incremented by delta c.p.s. and the calibration
line recalculated to produce a new a.r.e. Successive increments of
the background count rate lead to a progressively decreasing a .r.e.
until a minimum is reached, whereupon further increases in the back-
ground c.p.s. will lead to an increasing a.r.e. The optimum back-
ground count rate is therefore at the minimum a .r.e. Note that the
 16

changes in a.r.e. are smooth. As XRF data processing is normally

carried out with the aid of a digital computer, the iteration process

may conveniently be built into the calibration section of the data pro-

cessing program (Parker and Willis, 1977) . The iteration process may
be speeded up by applying a coarse (e.g. 5 c . p . s .) increment to begin

with, and then-a fine increment (e.g. 0.25 c.p.s.) in the region of the

a.r.e. minimum.

Table 2.4 lists major element calibration background count rates,

a.r.e.'s and a.a.e.'s, determined first by iteration and then from blanks .
Table 2.5 lists the standards used in -these calibrations. The iteration

method produces lower a . r. a .'s for all calibrations. With regard to the

a.a.e.'s MnO and K2O give the same values for both methods, while

the iteration method gives lower a .a . e .' s for all the remaining calib-

rations.

Further examples of calibration lines constructed using iterated

backgrounds are given in Parker and Willis (1977, p.165-167) .

2.6 Data processing

The raw analytical data is punched out on paper tape by the XRF

spectrometer as well as being listed on an "Addo" printer. This raw

data includes the sample and standard numbers as well as the counts,

counting times and spectrometer channel numbers for each element

analysed. The data tape is then read into the college computer system

and a permanent data file created. The inter-relationship between the

flow of data and the interactive programs CHECK, START, MWT, NFAC
 17

and GFAC used in processing the major element analyses is shown in

Fig. 4.1 (Chapter 4) .

Program NFAC and GFAC

These programs are used to generate Norrish and Gunn standard

files for input into programs START and MWT. These files contain

standard concentrations and matrix correction factors required in the

construction of the element calibration lines.

Program CHECK

This program checks the,raw XRF data file for standard and sam-

ple numbering inconsistencies and for spatial punch errors. The prog-

ram also interactively requests header information for the data file.

The program then writes out the data to a disc file for input into program

START.

Program START

Program START is a condensed and interactive version of programs

SORT and REORD described by Parker and Willis (1977) . An example of

the interactive terminal input and output from this program is given in

Table 2.6. Dead time and machine drift corrections are applied by the

program to the analytical data. The program monitors the size of the

machine drift corrections and if necessary a warning is written out at

the terminal. The program has been constructed to allow maximum flexi-

bility in the analytical sequence on the XRF spectrometer and as a result

the samples and standards may be analysed in any order. The sequence

of element analysis for any standard or sample may also be in any order.

The program sorts and re-arranges the analytical data into standard and
 18

sample data sets which are then written out to a disc file for input into

program MWT.

Program MWT

Program MWT is an expanded and interactive version of program

MW described by Parker and Willis (1977) . The interactive features

of this program provide for comprehensive terminal monitoring and list-

ing of , and interaction with, the analytical data during the process of

standard calibration. For example in a given analytical run the sum-

marised calibration data for all of the analysed elements may be listed

at the terminal (Table 2.7) . If this summarised data indicates that the

calibration for certain elements is suspect then the complete calibration

data may be listed for these elements or, if need be, for all of the ele-

ments. The listing of the complete calibration data for a given element.

(e.g. Si02 in Table 2.7) may show certain standards to be in error due

to bad machine drift or due to other causes such as mis-numbering of

the standard during analysis on the XRF spectrometer. - Another possi-

bility is that the "preferred" value for a given element in a standard may

be unreliable and/or the standard may be inhomogeneous with respect to

the element in question. The interactive features in program MWT allow

for these standards to be deleted and the calibration line to be reconstru-

cted and relisted at the terminal (Table 2.7) .

Other interactive features available for use during standard cali-

brations include options for calculating the background count rates by

iteration or from blanks if the latter have been analysed. This feature

provides the potential for rapid comparison of calibrations constructed

 19

using these two techniques. Further interactive options allow the

calibrations to be limited to standards- falling in certain ranges of

compositions. For example it may be desired to calibrate for Si02

with only granitic standards. In this case a lower level of say 68%

Si02 is set and all standards below this value are then automatically

excluded. Upper levels may also be set and; if desired, both lower

and upper levels may beset.

An interactive option allows the MgO calibration line to be split

into two calibrations , the first for low Mg0 standards and samples

(< 20.0% MgO) and the second for high concentrations (> 20.0% MgO).

It has been found that MgO standards prepared by fusion with pure

lithium tetraborate flux (section 2.2(2) above) do not produce good cali-

brations over the range of ultrabasic to acid standards. Better calib-

rations and sample analyses are obtained by splitting the standard cali-

brations as above. Standards prepared by the Norrish method (section

2.2(2) above) produce acceptable MgO calibrations over the ultrabasic

to acid range and splitting of the calibration line is not necessary when

analysing standards and samples .prepared by this method.

After the calibration data have been satisfactorily processed, the

program calculates the sample compositions. File dw (Fig.4.1) con-

tains for each sample wet chemistry Na20 and FeO data as well as sam-

ple descriptions and data for weight loss on drying at 110°C (= H2O-)

and loss on ignition at 8506 C (= LOI). These data are incorporated in

the final results. Program MWT also applies corrections for the effect

of incomplete ignition oxidation of FeO. The data in Fig. 2.1 are used
 20

to estimate the post-ignition residual FeO for each sample and this is

then used in the calculation. of the sample volatile loss on ignition.

Finally the program will list if desired the results of the analyses

at the terminal. The analyst may choose to list all or only selected

analyses as well as choosing one or two decimal places- (Table 2.7) .

The program generates a line printer file containing comprehensive cali-

bration and analytical data for the standards and samples. This file

may be dispatched to a line printer for printing. Program MWT also

generates a sample results file for input to program TAB (Chapter 4).

The interactive terminal based operation of the above programs,

combined with the file editing facilities available from the terminal,

provides considerable speed and flexibility in checking and correcting

calibration lines and sample data.

2.7 Accuracy and precision

In order to provide an estimate of the accuracy of the analytical

techniques described above for Norrish fusions , six standards were

analysed as unknown samples. This analytical run included the ultra-

basic standards NIM-P, NIM-D and DTS in the SiO2 calibration. The

data for each of the six standards analysed as unknown samples was

processed with that standard removed from all of the calibration lines

(the K-feldspar standard, British Chemical Standards No. 376, was not

used in the calibrations) . Duplicate preparations of each standard

were analysed (each preparation counted once) and the averaged results

are presented in Table 2.8.

 21

There is good agreement between the XRF values listed in Table

2.8 and the recommended (or preferred) values for the analysed stan-

dards. Note that the range of these igneous standards covers acid,

basic, ultrabasic and syenitic compositions. Norrish and Hutton

(1969) and Harvey et al. (1973) provide further data indicating that

the technique is capable of analysing a considerable range of compo-

sitions. The author of this thesis has successfully used the techni-

que to analyse P, Ca and Fe rich rocks from the sea-floor (Parker and

Siesser, 1972 and. Parker, 1975) .

The "recommended LOI" values listed in Table 2.8 are the sum of

the recommended H20+ plus CO2 values for each of the standards. The

determined LOI values were corrected for the effect of incomplete oxi-

dation of FeO to Fe203. Apart from H20+ and CO2, other volatiles

lost on ignition may include elements such as F, Cl and S; and this

will contribute to the variation in the determined LOI with respect to

the sum of the recommended H20+ plus CO2. There is nevertheless

reasonable agreement between these data for the standards analysed

in Table 2.8.

The major element XRF analyses presented in Table 2.8 and in

Chapter 6 are, with a few exceptions, the average analyses of dupli-

cate fusions for each sample. Each fusion disc was counted once.

In order to estimate the combined fusion and counting precision for

these analyses, sample 6502 was fused six times in duplicate and

each disc counted once. The averaged data for each pair of discs was

used to calculate the means, standard deviations and coefficient of

 22

variation for the six analyses (Table 2.9) . Duplicate counts and/or

longer counting times would improve the precision for oxides such as

MgO. Long counting times (greater than 100 seconds, Table 2.1) must

however be weighed against possible errors introduced by machine drift.

The data in Table 2.9 show that where an oxide is present in a concen-

tration greater than 1%, the percentage coefficient of variation is < 1%.

It should, however, be noted that the poor precision for MgO at the 0.6%

level indicates that this relationship would not hold for MgO at the 1%

level.

2.8 Discussion

The iterated background method produces superior calibrations as

compared to calibrations using backgrounds from blanks, and furthermore

the overall analysis time is reduced. Both the iteration and the blank

methods of background determination arepreferred to determining back-

ground "off-peak". This latter method suffers from a number of inherent

errors Wand significantly increases the analytical time.

The interactive options built into the major element program pro-

vide a flexible means of evaluating standard calibrations. Furthermore

the ability to list the calibration data and analyses at an interactive

terminal offers a significant time saving advantage over batch processing

of XRF data .

The accuracy test on the Norrish method of sample fusion and

matrix correction, combined with weighted calibration lines, and iterated

backgrounds, has produced satisfactory results.

Table 2.1 Major element analyses on Norrish discs—Instrumental settings

Mg(4) Ti Ca K Si Al P Fe Mn '. Cr Ni
. . .
Xray tube Cr Cr Cr Cr Cr Cr Cr W W W W

Kv 60 60 40 40 60 60 60 40 60 60 60'

mA 32 24 16 24 32 32 32 8 32 32 32

Crystal(1) ' RbAP LIP LiF LiF PET PET GE LiF LiF LiF LiF

Collimator(2) C F F F C C C F F F F

Counter(3) Fl Fl Fl Fl Fl Fl Fl Fl F1 Fl+So Fl+Sc

Counting time (secs) 100 40 40 40 F.0S5) 100 100 20 100 100 100

LiF = LiF200
C = coarse (480}pm), F = fine (160pm) collimator.
F1 = flow, Sc =.scintillation, vacuum path used for all elements.
Asymmetric window set on pulse height analyser for Mg.
F.C. = fixed count (300,000).
 24

Table 2.2

Oxidation of FeO-,Fe703 at 850° C and 950° C in silicate rock samples

Analyst: P. Watkins, Dept. of Geology, Imperial College.

FeO % determinations
literature* after 850°C after 950°C
sample value unignited for 30 mins . for 30 - mins
UC 311 - 26.66 5.49 2.24
UC 311 - 16.50 5.38 3.12

NIM-D 14.67 14.53 5.47 3.46

NIM-P 10,47- 8.94 6.39 5.18
BCR 9.05 8.52 3.59 2.76
MM-N 7.49 7.45 4.88 3.81
DTS 6.98 7.16 3.57 2,53
BR 6.60 6.54 2.19 1.78
DR-N 5.32 5.49 0.00 0.08
SY-3 - 3.87 2.85 2.35
SY-2 . - 3.57 2.90 2.31
AGV 2.08 2.05 0.37 0.23
NIM-G 1.30 1.30 0.11 0,02

NIM-L 1.04 1,08 0.74 0.81

91 - 1.40 0.60 0.38

Regression data for samples

with < 12 % unignited FeO

correlation coefficient for unignited

vs ignited FeO values 0.801 0.788
slope of linear regression 0.561 0.437
intercept of linear regression -0.333 -0,234
number of samples 12 12
I

*Abbey (1973, 1975)

 25

Table 2.3 Determination of iterated background for TiO2

• (iteration delta = 5 c . p. s . )

bkgd cps aae are

105 •013 3.74
110 •012 3.46
115 •011 3.18
120 .010 2.92
125 •.009 2.65
130 •i'7C8 2.38
135 •_ V6 2.12
140 • 007 1.86
145 • 006 1.64
150. • 006 1.42
155 •006 1.22
160 •005 1•06
165 • 005 • 94
170 • 006 •87 selected by iteration
175 • 006 •95
180 .007 1.18
185 •008 1.47
190 •009 1.75
195 • 010 2.04
200 •010 2.33
205 •011 2.63
210 •012 2.93
215 •013 3.23
220 •015 3.54
225. •016 3•85 determined from blanks
230 •017 4.16
235 •018 4.47
240 •019 4.79
245 • 020 5.10
250 •021 5.42
255 •022 5.75
260 • 023 6.07

bkgd cps = background counts per second

aae average absolute error

are = average relative error

Table 2.4 Comparison of calibration data
BKGD background
IT iteration
A.R.E. a average relative error (%)
A.A.E. average absolute error (wt.%)
C.P.S. counts per second

Si02 TiO2 A1203 Fe203 MnO MgO CaO K20 P205 Cr205 NiO
BKGD IT '
(c.p.s.) 49 172 16 61 144 44.0 81 27 59 665 146
BKGD BLANK
(c.p.s.) 27 201 22 49 144 47.5 96 29 51 654 141
A.R.E. IT 0.52 1.98 0.68 1.49 9.14 5.39 , 1.73 0.90 6.06 13.75 9.28
A.R.E. BLANK 0.53 4.50 3.66 1.60 9.17 8.77 1.93 0.92 8.66 32.48 11.04
A.A.E. IT 0.339 0.008 0.080 0.098 0.011 0.175 0.049 0.027 0.013 0.014 0.005
A.A.E. BLANK 0.343 0.011 0.090 0.102 0.011 0.190 0.054 0.027 0.018 0.019 0.006
NO. OF STDS . 12 12 8 14 6 12 12 12 6 4 6
RANGE OF STDS. 39- 0.05 - 0.26- 1.3- 0.03- 0.38- 0.68- 0.25-. 0.14- 0.01- 0.01-
(wt.%) 76 2.7 17.4 17.0 0.78 49.7 14.2 15.4 0.50. 3.6 0.31
Table 2.5 Standards used in calibrations (Table 2.4)

Si Ti Al Fe Mn Mg Ca K P Cr Ni'

GH + + + + . +
G-1 +• + +. + + + + +
G-2 + + + + + + + +
NIMG + ' + + + + +
GSP + + + + + + -F +:' +.
AGV + + + '+ + + + +
NIMS + + + + + + +
BCR + + + + + + + +
NIML + + + + + +
W-1 + + + + + + + + + +
NIMN + + + + + + + +
BR + + + + + + +
NIMP + + + + +
DTS + + + +
NIMD + + + + + +


28

Table 2.6 Major element data processing - program START

interactive terminal output and input comments

START OF PGM START

XRF DATA (FROM POM CHECK) ON TAPE 51

STDS DATA ON TAPE 52

** DRIFT ON REF 99 DL 482 EL 62 DRIFT 2.9 LUF 2.0 Program lists samples ex-
AFFECTS STD/BLK/SPL 140 130 120 ceeding the drift limit.

** DRIFT ON REF 99 DL 48. 6 EL 63 DRIFT 4.9 LUF 2.0

AFFECTS STD/BLK/SPL 140 130 120 "
Program lists total number of
NUMBER OF SPLS-BLKS-STDS IS 90 samples, blanks and standards
in XRF data file (Tape 51).

END. SUER SORT

Program lists reference sample
REFERENCE COUNT RATES:
0 0 9382. 0 6579. 0 count rates (used in normaliz-
8306. 0 1300. 9540. • 0
0 90. 619. 0 ing the XRF data for drift). .

13 BLANKS READ IN - Program lists number of blanks

and standards read in from the
107 STANDARDS READ IN • standard data file (Tape 52).
CONTINUE - Y/N
?v

21 SAMPLE DATA SETS WRITTEN TO TAPEI1

END SUBR REORDER AND PGM START

Full listings of the sorted and

NB - LINE PRINTER FILE. IS ON TAPE71 reordered data are on Tape 71
- DATA FILE FOR PGM MWT IS ON TAPE11 and Tape 11.
 29
Table 2.7 Major element data processing - program MWT
interactive terminal output and input comments
START OF PGH MWT -
XRF DATA (FROM PON START) ON TAPE 11
STDS DATA ON TAPE 52
WET CHEM, WEIGHINGSr DESCRIPTION DATA ON TAPE 53

HEADER DATA ON TAPE11 IS

WWP X
VULSINI PUMICE AND LAVA RUN
NORRISH METHOD - CAM74
NS
1.7
84.50
0000000000 4 50000
User selects no calibration line
CALIB LINE EDITING - Y/N MG CALIB SPLIT - Y/N editing and no Mg calibration
? nn line splitting.

LISTING OF CALIBRATION LINE DATA, OPTIONS:•

1 - SUMMARISED DATA ONLY, 2 - FULL DATA SELECTED ELEMENTS

3 - FULL DATA ALL ELEMENTS. 4 - NO LISTING ENTER 1,2,3, OR 4 User selects summarised calibra-
?1 tion line data .

OXIDE SLOPE BKG -PC CPS AV ABS ER AV REL ER FOM

SI02 .8526E-02 .36 42.75 .374 .68 .65 BI

1102 .3764E-03 .07 199.00 .007 .85 .65 BI
AL203 .1019E-01 .16 15.50 .079 .67 .69 BI
FE203 .2894E-02 .00 .25 .083 1.04 .94 BI
MNO .5204E-03 .07 138.25 .013 7.80 6.05 BI
CONTINUE Y/N
? y

MG COR FACS - CPS PER PERCENT OXIDE tI FE CR MN CA K NA NI

1.35 .06 -.01 .02 1.02 .56 0 0
MOO .6264E-01 3.13 50.00 .211 5.03 1.48 DI
CAO .1344E-02 .05 36.75 .047 2.08 1.26 BI
K20 .9681E-03 .02 20.25 .031 1.03 .87 BI
P205 .3096E-02 .18 58.50 .013 6.10 5.62 BI
ENTER '1 - PROCESS SPLES 2 - REPROCESS CALIB LINES 3 - EXIT User selects reprocessing of
?2 calibration line data.

HEADER DATA ON TAPE11 IS

WWP X
VULSINI PUMICE AND LAVA RUN
NORRISH METHOD - CAM74
NS
1.7
84.50
0000000000 4 50000

CALIB LINE EDITING - Y/N MG CALIB SPLIT - Y/N

? nn

LISTING OF CALIBRATION LINE DATA, OPTIONS!

User selects full calibration
1 - SUMMARISED DATA ONLY, 2 - FULL DATA SELECTED ELEMENTS
3 - FULL DATA ALL ELEMENTS. 4 - NO LISTING ENTER 1,2,3. OR 4 data for selected elements.
7

ENTER ELEMENTS FOR DESIRED CALIBRATION LINES AND END WITH 99

OPTIONS - SI TI AL FE CR MN MG CA NA K P NI EX EY EZ
User selects element Si (= Si02).
? si
? 99


30

Table 2.7 (continued)

interactive terminal output and input comments
5IO2 CALIB LINE DATA

STANDARD .AV STD C CALC C ABS ER REL ER COR CPS Full S102 calibration data listed.

8000 CAO/A 1 0 .254 0 0 29.8 0

8003 CAO/B 1 0 .260 0 0 30.5 0
8006 CAO/C 1 0 .248 0 0 29.0 ' 0
8007 CAO/C 1 0 .248 0 0 29.1 0
9086 DTS/C 1 40.475- 40.284 .191 .471 4724.9 0
9080 DTS/A. 1 40.576 40.377 .199 .489 4735.8 0
9083 DTS/B 1 40.677 ' 40.496 .181 .446 4749.6 0
9096 NIMP/C 1 50.752. 51.774 -1.022 -2.014 6072.5 0 Poor calibration for NIMP std.
9090 NIMP/A 1 50.770 51.522 -.752 -1.480 6042.8 0
9093 NIMP/B 1 50.806 52.129 -1.323 -2.605 6114.1 0

9053 W-1/D 1 52.548 51.951 .597 1.137 6093.2 0

9056 U-1/E 1 52.548 52.788 -.240 -.457 _ 6191.4 0
9050 W -1 /A 1 52.567 52.408 .159 .302 6146.8 0
9046 BCR/E 1 54.704 54.806 -.102 -.186 6428.0 0
9043 MAI 1 54.781 54.640 .141 .258 6408.6 0
9040 8CR/B 1 54.839 .54.839 -.000 -.000 6431.9 0
9030 AGV/A 1 60.092 60.282 -.190 -.316 7070.3 0
9033 AGV/B 1 60.113 60.410 -.297 -.495 7085.4 0
9036 AGV /C 1 60.220 59.434 .786 1.305 6970.9 0
9063 6SF/B 1 67.826 67.405 .341 .503 7915.1 0
9060 GSP /A 1 67.851 67.777 .074 .110 7949.4 0
9066 6SF /C 1 67.947 67.845 .102 .151 7957.3 0
9000 0-1 /A 1 72.969 72.399 .570 .781 8491.6 0
9006 0-1 /C 1 73.073 72.706 .367 .502 8527.5 0
9003 6-1/8 I 73.099 72.881 .218 .298 8548.1 0

FIGURE OF MERIT IS .65

SLOPE .8526E-02 BKG .364 AV A ER .374 AV R ER .681

ITERATED BKG CPS 42.75

CONTINUE Y/N

ENTER 1 - PROCESS SPLES 2 - REPROCESS CALIB LINES 3 - EXIT User selects reprocessing of
7 2 calibration line data.

HEADER DATA ON TAPE11 IS

WWP X
VULSINI PUMICE AND LAVA RUN
NORRISH METHOD - CAM74
NS
1.7
84.50 User selects calibration line
0000000000 4 50000
editing (and no Mg calibration
CALIB LINE EDITING - Y/N MG CALIB SPLIT - Y/N line splitting).
? un
BKG CPS BY ITERATION: 1 - INTER-ACTIVE 2 - IST MINIMUM 3 - NO
?2
DEL APLIC TO ALL CAL LINES - Y/N
? n

ENTER ELEMENTS FOR DESIRED CALIBRATION LINES AND END WITH 99

OPTIONS - SI TI AL FE CR MN MG CA NA K P NI EX EY EZ
User enters Si for S102 cali-
? si bration line editing.
? 99

EDIT CALIBRATION LINE DATA FOR SI

User selects standard deletion
ENTER I - Ur 2 - M. 3 - BKG. 4 - AFAC. 5 - DEL. 6 - ALL option.
?5

User enters std to be deleted

STD DEL. ENTER UP TO 20 NOS. PLUS A IF REQ. AND END WITH 9999
? 9090a (9090a = all NIMP standards
? 9999 deleted).

Table 2.7 (continued) 31

interactive terminal output and input comments

LISTING OF CALIBRATION LINE DATA, OPTIONS: User selects full calibration data
1 - SUMMARISED DATA ONLY, 2 - FULL DATA SELECTED ELEMENTS for selected elements.
3 - FULL DATA ALL ELEMENTS, 4 - NO LISTING ENTER 1r2r3, OR 4

? 2

ENTER ELEMENTS FOR DESIRED CALIBRATION LINES AND END WITH 99

OPTIONS - SI TI AL FE CR MN MG CA NA K P NI EX EY EZ

? si User selects element Si.

? 99

5102 CALIB LINE DATA

Full Si02 calibration data listed
STANDARD AV STD C CALC C ABS ER REL ER COR CPS (NIMP std excluded).
8000 CAO/A 1 0 .254 0 0 29.8 0
8003 CAO/B 1" 0 " .260 0 0 30.5 0
8006 CAD/C 1 0 .247 0 0 29.0 0
8007 CAD/C 1 0 .248 0 0 29.1 .0

9086 DTS/C 1 40.475 40.466' .009 .023 4748.6 0

9080 D15/A 1 40.576 40.559 .017 .043 4759.5 0
9083 DTS/0 1 40.677 40.677 .000 .000 4773.4 0
9053 W-1/D 1 52.548 52.126 .422 .804 6116.9 0
9056 W-1/E 1 52.548 52.963 -.415 -.790 6215.2 0
9050 W-1/A 1 52.567 52.583 -.016 -.031 6170.6 0
9046 BCR/E 1 54.704 54.979 -.275 -..503 6451.8 0
9043 BCR/D 1 54.781 54.814 -.033 -.059 6432.3 0
9040 8CR/B 1 54.839 55.013 -.174 -.317 6455.7 0
9030 AGV/A 1 60.092 60.453 -.361 -.600 7094.0 0
9033 AGV/B 1 60.113 60.581 -.468 -..779 7109.1 0
9036 AGV/C 1 60.220 59.605 .615 1.021 6994.6 0
9063 GSP/B 1 67.826 67.652 .174 .257 7938.9 0
9060 GSP/A 1 67.851 67.944 -.093 -.136 - 7973.1 0
9066 6SP/C •1 67.947 68.011 -.064 -.095 7981.1 0
9000 6-1/A 1 72.969 72.564 .405 .555 8515.3 0
9006 6-1/C 1 73.073 72.870 .203 .277 8551.3 0
9003 G-1/B 1 73.099 73.046 .053 .073 8571.8 0

FIGURE OF MERIT IS .36

Average absolute and average

SLOPE .8522E-02 BKG .162 AV A ER .211 AV R ER .353
relative errors are improved.
ITERATED BKG CPS 19.00

CONTINUE Y/N

.ENTER .1 - PROCESS SPLES 2 - REPROCESS CALIB LINES 3 - EXIT User selects sample processing.
9"1

END SUBR SLOPE

START OF SUBR CONC

FeO data present, weighing data
FEO INPUT - Y/N FREE FORMAT WTS INPUT - Y/N in fixed format.
? Wn

IGNITION OXIDATION OF FEO

RIGFAC = FACTOR FOR RESIDUAL FEO AFTER IGNITION

UPFEO = UPPER FEO LIMIT FOR RIGFAC
URFEO = RESIDUAL FEO APPLICABLE ABOVE UPFEO
RESID FED AFTER IGNIT AT 850 Ct RIGFAC = 0.56 r UPFEO = 12.0r URFEO = 5.4 User selects no change in FeO
CHANGE Y/N oxidation factors.
?n
START OF SUBR NADWT

34 WET CHEM NA VALUES

Wet chem. and description -
24 WET CHEM FED VALUES
weighing - data read in.
34 DESCRIPTION-WEIGHING PAIRS

Table 2 . 7 (continued) 32

interactive terminal output and input comments

LIST RESULTS - Y/N

?y User selects listing of

1 OR 2 DEC PTS - ENTER 1 OR 2 all sample analyses to
?1
one decimal place.
ALL OR SELECTED SAMPLES - ENTER A OR S
? a

RESULTS - UT PC OXIDES

SPL SI TI AL FE3 FE2 CR MN MG CA NA K P NI LOI SUM

4004 A 54.7 .5 19.8 4.1 .3 0 .1 .7 4.1 3.4 8.9 .1 . 0 1.7 99.5
4004 A 55.2 .5 19.8 4.2 .3 0 .1 .8 4.1 3.4 9.0 .1 0 1.7 100.2
6803 A 59.5 .4 18.1 1.7 1.4 0 .1 .6 2.7 2.8 10.0 .2 0 2.0 99.7
6803 A 59.7 .4 18.2 1.7 1.4 0 .1 .6 2.7 2.8 10.0 .2 0 2.0 100.0
6@05, A 59.3 .4 17,9 1,5 1.5 0 .1 .8 2.9 2.8 9.9 .2 0 2.2 99.5
6805 A 59.6 .4 18.2 1.6 1.5 0 .1 .5 2.9 2.8 9.9 .2 0 2.2 100.0 Results for samples
305 A 50.1 .8 18.3' 4.5 2.9 0 .1 2.9 8.4 5.6 2.0 .6 0 3.0 99.6 prepared in duplicate
305 A 50.0 .8 18.4 . 4.5 2.9 0 .1 3.1 8.4 5.6 2.0. .6 0 3.0 99.9 and triplicate (each
6904 A 58.4 .4 19.0 1.8 .6 0 .2 .2 2.2 4.5 8.3 .1 0 4.2 100.0 preparation analysed
6904 A 58.6 .4 19.2 1.8 .6 0 .2 .2 2.2 4.5 8.2 .1 0 4.2 100.4
once) .
6903 A 57.4 .5 18.6 2.9 0 0 .1 .3 2.5 4.0 8.6 .1 0 4.0 99.4
6903 A 57.6 .5 18.6 3.0 0 0 .1 .3 2'.5 3.9 8.7 .1 0 4.0 99.7
6903 A 57.8 .5 18.7 . 2.9 0 0 .1 .2 2.5 4.0 8.7 .1 0 4.0 100.0
4007 A 57.3 .4 19.2 1.9 1.0 0 .1 .4 3.1 4.5 8.2 .6 0 3.7 101.2
4007 A 56.9 .4 19.2 1.8 1.0 0 .1 .4 3.1 4.5 8.2 .6 0 3.7 100.7
6808 A 60.8 .4 18.4 2.9 0 0 .1 .5 2.4 3.2 9.6 .1 0 2.0 100.4
6808 A 60.8 .4 18.3 2.9 0 0 .1 .5 2.4 3.2 9.5 .2 0 2.0 100.4
303 A 49.4 .8 18.3 4.2 3.3 0 .1 3.1 8.5 5.5 1.8 1.1 0 3.3 99.7
303 A 49.3 .8 18.3 4.2 3.3
_ 0 .1..3.1 8.5 5.5 1.8 1.0 0 3.3 99.8
3712 A 57.5 .5 19.1 1.8 1.0 0 .1 .3 3.0 4.6 8.2 -.0 0 3.1 99.4
3712 A 57.3 .5 19.3 1.8 1.0 0 .1 .2 2.9 4.6 8.4 .0 0. 3.1 99.4
6908 A 57.9 .4.18.9 1.5 1.2 0 .2 .3 2.6 4.2 8.8 .0 0 3.5 99.6
6908 A 57.7 .5 18.9 1.5 1.2 0 .2 .4 2.6 4.2 8.8 .1 0 3.5 99.6
6541 A 60.5 .3 18.2 1.7 1.2 0 .1 .6 2.4 3.3 9.2 .1 0 2.6 100.5
6541 A 60.4 .4 18.1 1.8 1.2 0 .1 .6 2.4 3.3 9.3 .1 0 2.6 100.5
6541 A 59.2 .3 17.9 1.7 1.2 0 .1 .7 2.4 3.3 9.0 .1 0 2.6 98.8
805 A 59.2 .5 18.5 1.8 1.3 0 .1 .5 2.8 3.0 10.1 .1 0 1.9 99.8
805 A 59.5 .5 18.7 1.9 1.3 0 .1 .6 2.8 3.0 10:1 .1 0 1.9 100.6
3709 A 60.9 .3 18.3 1.4 1.1 0 .1 .2 2.1 4.2 8.6 .1 0 2.4 100.1
3709 A 61.0 .3 18.2 1.5 1.1 0 .1 .2 2.1 4.2 8.5 .1 0 2.4 100.1
6801 A 59.7 .4 18.1 2.0 1.2 . 0 .1 .8 2.7 2.9 9.9 .2 0 2.0 100.1
6801 A 59.7 .4 18.2 2.1 1.2 0 .1 .7 2.8 2.9 9.9 .2 0 2.0 100.2
•
3706 A 59.6 .3.18.3 1.2 1.1 0 .1 .3 2.1 3.6 8.0 .1 0 4.1 99.8
3706 A 58.4 .3 17.9 1.1 1.1 0 .1 .6 2.0 3.6 7.8 .1 0 4.1 98.1
3706 A 59.6 .3 18.4 1.2 1.1 0 .1 .5 2.1 3.6 8.0 .1 0 4.1 100.2
6101 A 49.9 .8 18.3 5.0 2.7 0 .1 3.0 8.3 6.4 1.1 .6 0 3.1 99.6
6101 A 50.4 .8 18.3 5.0 2.7 0..1 3.1 8.4 6.4 1.1 .6 0 3.1 100.3
6809 A 59.1 .4 18.2 3.3 0 0 .1 .8 2.8 2.9 9.6 .2 0 2.3 99.8
6809 A 59.4 .4 18.3 3.3 0 0 .1 .9 2.8 2.9 9.6 .2
• 0 2.3 100.3
804 A 59.5 .5 18.4 1.8 1.6 0 .1 .7 .2.8 2.8 9.9 .1 0 2.1 100.3
804 A 59.0 .5 18.3 1.9 1.6 0 .1 .6 2.8 2.8 10.1 .2 0 2.1 99.9

41 SAMPLES ANALYSED
2 TOTALS GT 100.7
1 TOTALS LT 98.7

END PGM MUT Full standard calib-

ration and sample
NB - LINE PRINTER FILE IS ON TAPE 72
- RESULTS FILE FOR PGM TAB IS ON TAPE 73 analyses on tape 72.
Table 2 .8 Accuracy test on Norrish analyses(8) using iterated backgrounds
for the calibration lines

((6)
G-1 (0 W-1 (Z) BR (a) NIM-S NIM-D tal K-FELr3 (a)
Det. Rec. Det. Rec. Rec. (2) Rec.
Det. Det. Rec. Det. Red. Det.

SI02 72.26 72.68 52.84 52.72 38.33 38.39 63.39 63.54 39.02 38-97 67.18 67.10
Ti02 .26 .26 1.08 1.07 2.65 2.61 .05 .05 .02 .02 .01 .01.
AL203 14.18 14.05 , 15.10 14.87 10.14 10.25 17.15 17.16 .35 .26 17.75 17.70
FE203 .89 .87 1.37 1.40 5.57 5.61 1.02 1.08 .79 .70 .08 .10
(3)FED .96 .96 8.73 8.73 6.60 6.60 .29 .29 14.67 14.67 0 0
CR203 0 0 0 0 0 ' 0 0 0 .44 .41 e e
MND .03 .03 .17 .17 .20 .20 .02 .01 .22 .21 .01 0
MOD .37 .38 6.50 6.63 13.53 13.35 .43 .48 43.43 43.68 .01 .03
CAO 1.30 1.39 11.02 10.98 13.78 13.87 .68 .68 .29 .26 .48 .54
(3)NA20 3.32 3.32 2.15 2.15 3.07 3.07 .41 .41 .06 .06 2.83 2.83
1(20 5.55 5.48 .67 .64 1.42 1.41 15.60 15.40 .01 .02 11.14 11.20
P2O5 .09 .09 .16 .14 1.00 1.05 ' .13 .14 0 .03 .03 0
NIO 0 0 0 0 0 0 0 0 .27' -29 . 0 0
1120- 0 0 0 0 0 0 0 0 0 0 0 0
(4)[AI .35 .41 .66 .59 3.23 3.17 .25 .33 .72 .68 .30 .35(})
7D1AL 99.53 99.92 100.42 100.09 99.50 '99.58 99.42 99.57 100.29 100.26 99.82 99.86
(s)FEao3 1. 96 1.14 11•07 11.11 Wrgo fa-98 1.31r f • 10- 17•09 f6 .97 •03 •10

(1) Flanagan (1973), recommended values on moisture (H2O-) free basis.

(2) Abbey (1973).
(3) FeO and Na2O = recommended values.
(4) LOI (Det.) = ignition loss at 850°C (corrected for FeO->Fe2Q3).
LOI (Rec.) = sum of recommended H20+ plus CO2.
(5) total iron.
(6) K-FELD = British Chemical Standards No. 376.
(7) recommended LOI as per British Chemical Standards certificate.
(8) average analyses from duplicate disc preparation (each disc counted once
 34

Table 2.9 Estimate of combined fusion and counting precision

for ignimbrite pumice sample 6502

Mean & (n=6) C.V.

SiO2 59.65 .366 0.61

TiO2 .347 .005 1.44

A1203 17.81 .159 0.89

Fe203(T) 2.99 .025 0.84
MnO .137 .005 3.65
MgO .57 .090 15.79
Ca0 • 2.40 .011 0.46
K20 9.27 .035 0.38
P205 .122 .004 3.28

= standard deviation: calculated from 6 analyses

of sample 6502. Each analysis = average of
two disc preparations counted once each.

8 x 100.
C.V. % = per cent coefficient of variation = mean
 Fig. 2.1 Oxidation of FeO -+ Fe203 at 850° C and 950° C in silicate rocks

7
0 6 •
°
tr, 5
E4 . -

ttl ° 3
M
•
'
..
r = 0.788 ) for • samples
•
O
w
0 2
1
40.

slope = 0.437 )
w
i. •
0

7
• 6 •
•
°° .•5
E4 •
•.
•
474
• 23
•~ •
r = 0.801 . ) for • samples
,°2 i
slope = 0.561 )
°' 1

0 5 10 15 20 25 30
Fe0% before ignition


Fig. 2.2 Calibration errors and slope factor

c1 - recommended concentration for

standard
. c2 - calculated

n - number of standards
A - lc1 - c21 - absolute error

average absolute error

n
std
' 100 • d/c1.- relative error (per cent)
I
I ( 1E(100 • d/c1)
( 1 average relative
I I n error (per cent)
I I

c1
- slope factor
Istd. c.p.s.
1 l
4nc

Conc.(wt. %) — )
Fig. 2.3 TiO2 calibration line using average background from blanks
STD = standard, RC = literature concentration, CC = calculated concentration, S = slope factor,
AE = absolute error, RE = relative error, CCPS = corrected counts per second, AAE = average absolute
error, ARE =.average relative error.
°
STO RC CC RE RE CCPS
0
o_ BCR/B 2.24 2.26 -0.02 -0.75 5788
8CR /B 2.24 2.24 0.00 0.09 5740
BCR/D 2.24 2.25 - 0.02 -1.09 5800
o BCR/0 2.24 2.25 - 0.01 -0.35 5758
0 8CR/E 2.23 2.25 - 0.02 -0.74 5772
m_ BCR/E 2.23 2.23 0.00 0.05 5727
O W-1/E 1.08 1.06 0 .02 1.89 2711
14-1/0 1.08 1.05 0 .03 2.70 2689
°
0 14-1 /A 1.08 1.07 0.01 1.12 2733

Z AGV /C 1.05 1.04 0.01 0.93 2672

RGV/B 1.05 1.05 -0.00 -0.06 2696
C..) RGV /A 1.05 1.05 -0.00 -0.33 2701
W
(no GSP /C 0.66 0.66 0 .01• 1.25 1684
ce • GSP/8 0.66 0.64 0 .02 3.52 1642
." GSP/R 0.66 0.64 0 .02 3.29 1646
w
G.

u7 0-1 /C 0.26 0.23 0.03 10.90 598

0-1/B 0.26 0.24 0 .02 7.41 622
z °. 0-1 /A 0.26 0.24 0 .02 9.27 607
u N NInP /C 0.20 0.1B 0.02 11.01 459
NICP/8 0.20 0.18 0.02 10.41 462
NInP/R 0.20 0.18 0.02 10.75 460
° N114G /C 0.09 0.09 0.00 5.03 222
°
NInG/8 0.09 0.09 0.01 5.52 220
cb.00 0.50 1.00 1.50 2.50 4IMG/A 0.09 0.11 -0.02 -19.95 280
PERCENT TI02
5102/8 0.00 0.09 0.00 0.00 242
5102 /A 0.00 0.10 0.00 0.00 244
S 0.00039 RAE 0.02 ARE 4.52
CAO /C 0.00 0.10 .0.00 0.00 .
254
CA0 /C 0.00 0.10 0.00 0.00 258
CA0/8 0.00 0.10 0.00 0.00 254
CA0 /A 0.00 0.10 0.00 0.00 258

*see section 2.5.3

 38

CHAPTER III

TRACE ELEMENT ANALYTICAL TECHNIQUES

3.1 Whole rock pressed powder briquettes

3.1.1 Introduction

Section 3.1 describes the computer-based techniques used in the

XRF. analysis of trace elements (Rb, Sr, Ba ,..Th, Y, Zr, V, Co, Cr,

Ni, Cr, Zn) in whole rock pressed powder briquettes. The briquettes

were prepared by pressing 5g. of -100 mesh rock powder in a hydraulic

press at 6 tons /sq. in. A backing of boric acid_ was used. The sam-

ple powders were ground in an agate "tema" mill. Standard trace ele-

ment briquettes had been previously prepared using ignited (850°C for

30 mins.) rock powder (G. Bullen, pers . communic .) . The unknown

sample briquettes were also prepared using ignited powders. The stan-

dard and sample trace element concentrations were corrected for the inc-

rease or decrease in concentration due to the ignition of the rock powder.

Table 3.1 lists the various instrumental variables selected in the.

analysis of the trace elements. Standards were analysed for Ba using

both W and Cr X-ray tubes, with coarse and fine collimators respectively,

in order to compare the quality of the results produced by these two

methods.

3.1.2 Standards

XRF analysis of trace elements in rock powders is critically

 39

dependent on the availability of high quality standard reference samp-

les. In this study the USGS standards W-1, G-1, G-2, AGV, GSP,

BCR, PCC and DTS were selected as the primary calibration standards..

The NIM standards (NIM-N, NIM-P, NIM -S, NIM-L, NIM-G and

NIM-D) were analysed aš secondary Standards. Important compilations

of the analytical data for the above standards are Fleischer (1969),

Flanagan (1969, 1973, 1976a) and Abbey (1973, 1975a). Table 3.4

lists the results of the compilations of Flanagan (1973, 1976a) and

Abbey (1973, 1975a) for the elements and standards examined in this

work. Inspection of the compilations indicates that the USGS stand-

ard data are likely to bē more reliable than those for the NIM standards

This is due to (a) the greater number of analytical results used in the

compilations for the USGS standards and (b) the greater agreement bet-

•ween the Abbey and Flanagan data for the USGS standards , as compared

to the NIM ,standards .

In comparing Flanagan and Abbey values for the USGS calibration

standards used in this work (Table 3.4) it may be noted that for certain

elements (e.g. Rb, Sr; Zn) there is good agreement while for other ele-

ments there are differences of up to 5 - 25% (e.g. Cr, V, Y) for certain

standards. Abbey and Flanagan have used different criteria in arriving

at the values found in their compilations, in particular with respect to

the rejection of outliers (see Flanagan, 1975 and Abbey, 1975b) .

Christie and Alfsen (1977) and Ellis et al. (1977) have proposed rela-

tively sophisticated statistical methods for calculating estimates of

standard values used in compilations. The adoption of either of their

methods will lead to changes in compilation trace element values.

 40

Examples given by these workers of recalculated values for BCR indi-

cate changes of up to 5% for elements such as V and 6% for Zn.

Another factor which should be considered when evaluating ana-

lytical data for international rock standards is the possibility that a

given standard may be inhomogeneous (Ridley et al., 1976; Engels

and Ingamells , 1977) . Standard inhomogeneity will mean that while

the values found in compilations may be meaningful in terms of the

average composition of the original bulk standard powder, these values

may not be such good 'estimates of the composition of the particular

standard sub-sample in use in an individual laboratory. The variation

in reported values for W-1 and G-1 in terms of sample inhomogeneity

or analytical errors is discussed by Vistelius (1970, 1971), Chayes

(1969, 1970) and Flanagan (1976b).

The standard compilation data and discussions -in Fleischer (1969) ,

Flanagan (1969, 1976a), Abbey (1977), Govindarajn and Roelandts (1977)

aft

and Rubeska (1977) indicate that there is considerable scope for improv-

ing the reliability of the preferred values for the international rock stan-

dards. Flanagan (1976a) notes that "there is nothing authoritative" in

the estimates in his compilation of USGS standards and that they are

what he considers to be "the most reasonable values at this time but

many analysts may wish to use preferred values of their choice".

3.1.3 Analytical procedures

A reference sample, giving good count rates is analysed with each

set of three sample or standard briquettes (the sample chamber of the

 41

Philips 1212 spectrometer holds four briquettes at a time) . The count

rates are corrected for the effects of dead time and background. Spec-

tral contamination from the analysing X-ray tube is corrected by using

a spectrosil glass sample (Leake et al., 1969) . Line overlap (e.g.

RbKp on YK) corrections are also applied (Leake et al., 1969; Leoni

and Saitta, 1976) . Mass absorption coefficients calculated from stan-

dard and sample major element compositions are used in correcting the

analyses for matrix effects.

The discussion in Chapter 2.4 on multi-standard vs single stan-

dard calibration in major element analysis also applies to trace element

analysis. The, multi-standard calibration technique described below is

based on calibrating the reference sample trace element concentrations

with respect to the data for a number of standards analysed as unknown

samples. This is done as follows.

The XRF intensity (corrected for dead time, matrix effects, etc . )

of the reference sample is related to the unknown samples or standards

by the ratio:

CR CS
(1)
IR IS

where CR = concentration of reference sample in ppm.

CS = concentration of unknown sample or

standard in ppm.

IR = intensity of reference sample in cps.

IS = intensity of standard or unknown sample

in cps.
 42

The concentration of a trace element in the sample or standard is

therefore :-
IS .
CS = CR — (2)
IR

In an analytical run an initial reference sample concentration is set

(= CR) for each element. For a particular trace element the CS values

for each of the standards analysed as unknown samples are then cal-

culated using equation (2) . The results, in conjunction with the

literature values for the standards , are next used-to determine if the

initial CR concentration assigned to the reference sample is high or

low. For each element the percentage by which the standard results

are in error is calculated (see below) and the initial reference sample

concentration CR is then corrected by the inverse of this percentage

error (PE) . When the corrected CR has been determined for each of

the analysed elements in the reference sample, the results for the un-

known samples are then calculated using equation (2) .

The calculation of the percentage error (PE) in the standards may

be undertaken in a number of ways.

Method I
std = n
Cstd lit -- Cstd ca1c
X 100
Cstd lit
PE std = 1 (3)

where Cstd lit = literature concentration of standard

11
Cstd calc = calculated
n = number of standards analysed
 43

The above method weights all the standards equally in calculat-

ing the percentage error.

Method II

std = n
Cstd lit - Cstd calc
PE = std = 1 (4)
100
std = n
Cstd lit
std = 1

This method places greater weight on the higher concentration

standards in calculating the percentage error. In this regard it is simi-

lar to the major element calibration technique. The count rate data for

the higher concentration standards will be more precise than that for the

lower concentration standards. Furthermore the literature data for these

higher concentration standards may well be of a higher quality as com-

pared to the lower concentration standards (see Chapter 2.5(2) and

Rubeska, 1977, p.19).

Method III

The uncertainty surrounding values for many trace elements in

rock standards evident in compilations and evaluations of standard anal-

yses (see section 3.1.2 above and Flanagan, 1969, 1976a, b; Abbey,

1977; Govindarajn and Roelandts, 1977; Rubeska, 1977) means that

analysts may wish to construct calibrations, fora given element, from

a restricted number of preferred standards. Restricted calibrations

based on a PE calculated from a sub-set of the analysed standards are

easily established using either equation (3) or (4) above. Unwanted

standards are simply excluded from the calculation of the PE.

 44

3.1.4 Data processing

The sample and standard numbers and the counting times and

counts measured by the XRF spectrometer, as well as the spectrometer.

channel numbers for each background and peak position, are punched

out on paper tape by the spectrometer. These data are also listed

on an "Addo" printer. The data tape is then read into the college

computer system and a permanent data file created. The data file is

processed using interactive programs TRANS and NUTRACE (Fig.4.1).

Program TRANS

This program initially checks the data for numbering inconsis-

tencies and spatial paper tape punch errors. After the file has been

satisfactorily checked the program reads from the interactive terminal

the analysts name and description of the data file, followed by inter- _

active requests for peak and background channel numbers for the anal-

ysed elements as well as common background numbers if required.

The program then transforms the data into a file suitable for input to

program NUTRACE. An example of the terminal input and output for

program TRANS is given in Table 3.2.

Program NUTRACE

This program is a modified version of the batch program )(TRACE

written by B.M. Gunn (University of Montreal) . The primary changes

to the program relate to the introduction of the multi-standard calib-

ration method and the conversion of the program to interactive terminal

operation.
 45

In processing the transformed trace element data the trace ele-

ment values assigned to the reference sample by the program are ini-

tially listed (Table 3.3) . These values may be interactively changed

once the PE correction factor has been calculated (3.1.3 above) . Tube

background factors (for correction of tube spectral contamination) are

then entered at the terminal. After reading in and sorting the trans-

formed trace element data the program requests line overlap correction

factors. The program then calculates the concentrations for the stan-

dards and samples; the results may be listed at the terminal and/or a

line printer results file dispatched for printing. The calculated results

for the standards are then compared with the preferred values for the

standards from the•literature (bearing in mind the discussions in 3.1.2

above) . If necessary new PE correction factors are calculated and the

program is then re-executed to re-process the data using the PE correc-

ted reference sample concentrations.

3.1.5 Calibration errors and precision

The accuracy of the analytical techniques will be assessed in

terms of the average relative errors between the calibration values for
the USGS standards and the averages of the combined Flanagan (1973,

1976a) and Abbey (1973,"1975a) values (AFA values) for these standards

(Table 3.4). It should be noted that as the standards have been used

to establish the trace element calibrations, the errors in the values

calculated for the standards should be considered as calibration errors.

True standard errors would only be obtained if the standards were run

as samples making no contribution to the calibration process.

 46

(1) Rb and Sr

There is excellent agreement between the calibration concent-

rations and the AFA values. The average relative error for these ele-

ments is n. 1%. Verdurmen (1977) has obtained a similar error level

in his XRF study on Sr and Rb analysis. The quality of the Sr and Rb

results listed in Table 3.4 is considered to reflect two important

factors;

(a) the good count rates available for these elements in the
calibration standards;

(b) reliable values for the trace element concentrations in the

standards. There is good agreement between the Flanagan (1973,

I976a) and Abbey (1973, 1975a) values. Flanagan (1976a), in discus-

sing the results for Sr and. Rb used in his compilations, notes that for

these two elements: "the averages by different methods (of analysis)

are in such good agreement that the final choices, influenced , greatly

by data obtained by some form of isotope-dilution technique, were

easily made".

(2) Ba/2, Y , Zr, Co and Ni,

There is generally good agreement between the calibration con-

centrations and the AFA values. The average relative errors for these
I
elements are in the range 3 - 5%.

(3) Ba/1, Th, V, Cr, Cu and Zn

There is fair agreement between the calibration concentrations

and the AFA values. The average relative error for these elements is

6 - 14%.
 47

The error ranges for elements in (2) and (3) above are considered

to be acceptable considering the uncertainties in the estimates of

Flanagan (1969, 1973, 1976a) and Abbey (1973, 1975a) (see 3.1.2

above) . The calibration Ba values in G-1 in Table 3.4 are low by

17 - 19% compared to the AFA value. However the compilation data

presented for Ba in G-1 by Flanagan (1969, 1976a) shows a consider-

able spread of values. Many workers consider that isotope dilution

analytical techniques are capable of producing data of the highest

quality (see 3.1.5(1) (b) above and discussions in Rubeska, 1977) and

recent work by Laeter and Hosier (1977) using this technique has resul-

ted in a value of 1055 ppm Ba in G-1. If G-1 Ba data are excluded

then the average relative errors for Ba are 2 - 3% in Table 3.4. The

determinations for V in G-1 and G-2, Cu in G-2 and Zn in PCC are all

> 20% in error with respect to the AFA_ values. If these standards are

excluded then the average relative errors for V, Cu and Zn in the remain-

ing standards lie in the range 4 - 6%.

Table 3.11 lists Th concentrations determined in pumice samples

by both XRF (as described in this -chapter) and neutron activation (NA)

techniques. The NA determinations of Th in these samples serves as

a useful check on the quality of the XRF data. Table 3.11 shows that

the average relative error* %1 is n, 7% which is the same as

the average calibration error in Table 3.4. Furthermore the average

(
difference ___2) between the two sets of data show that the XRF

values are on average only 0.3% low as compared to the NA values ,

indicating that there is negligible analytical bias between the two

methods.

* between the two sets of analyses

 48

The standard data presented in Table 3.4, and the analyses of

Vulsini samples presented in Chapter 6, are the result for each trace

element of a single count on a single rock powder briquette. In eva-

luating the data in Table 3.4 the standard trace element detection

limits presented in Table 3.5 should be borne in mind.

An estimate of the combined briquette preparation and counting

precision for the elements analysed in Vulsini sample 6502 is pre-

sented in Table 3.6,

3.2 Norrish fusions discs

3.2.1 Introduction

Trace element analyses for Sr, Rb and Ba were carried out on

Norrish fusion discs prepared in the major element analysis of leucite

and analcime separated from lava and pumice samples from Vulsini

volcano. The quantities of these minerals available for analysis were

relatively small (< 1 - 2g.) due to the difficulties of separation and

purification. The small amounts of material precluded the preparation

of pressed powder briquettes and as a result it was decided to use the

fusion discs for the trace element analyses.

3.2.2 Sr and Rb analyses

The instrumental settings are listed in Table 3.7 and calibration

and detection limit data are given in Tables 3.8 and 3.10.

All data calculations were carried out manually. The count data
 49

were inspected for drift and dead time effects and corrections applied

where necessary. Matrix corrections were not made as Norrish matrix

correction coefficients are not available for Sr and Rb. However the

dilution due to fusion and the presence of lanthanum oxide as a heavy

absorber in the flux means that matrix effects are considerably reduced

(Norrish and Hutton, 1969) . Background count rates were determined

on four Norrish fusion blanks (Si02 100%, Si02 70% - K2030%, Si02

70% - Ca0 30%, Si02 70% - Fe203 30%) and the results averaged. The

concentrations assigned to the standards are averages of values from

Abbey (1973, 1975a) and Flanagan (1973, 1976a). The ppm values and

the net-peak counts for the calibration standards (AGV, G-2, BR, GSP

and W-1) were summed and a calibration factor calculated for each ele-

ment. This calibration factor will give more weight to the higher pre-

cision count data available from the higher concentration standards (see

2.5.2 above) . This factor was then used to back calculate the cali-

bration concentrations for the standards (Table 3.8) . The relative

errors between the calibration and literature concentrations were then

calculated for the standards, followed by the average relative errors for

the two calibrations, The data in Table 3,8 shows that the calibration

errors for Sr and Rb on Norrish discs are N 4%. Although these error

levels are higher than those obtained for Sr and Rb on pressed powder

briquettes, they are nevertheless comparable to or better than the error

levels determined for the remaining trace elements discussed in section

3.1.5 above.

3.2.3 Ba analyses

The instrumental settings are listed in Table 3.7 and calibration

M
 50

and detection limit data are given in Tables 3.9 and 3.10.

The Ba L pi and Lp2 15 lines are unusable in the analysis of

Norrish fusion discs due to interference from strong La lines and the

low count rates for these Ba lines resulting from the fusion dilution.

The stronger Ba Lg(, line gives acceptable fusion disc count rates but

suffers from Ti interference when using the LiF200 analysing crys-

tal. The LiF220 crystal would give almost complete separation of

these two lines (Willis et al., 1969) but the 2 e angle for Ba Lk on

this crystal is ti 154.5° which is greater than the maximum possible

angle on the Philips 1212 spectrometer. However careful scanning of

the various standard fusion discs using Li F2 0 0 showed that the Ti Kg,

interference is not critical for TiO2 concentrations of < 0.5%. As the

TiO2 contents of the leucite and analcime samples were very low

(< 0.06%) and there are a number of suitable Ba standards with TiO2

concentrations of 5 0.5% (Table 3.10) , it was. decided to analyse the

samples and- selected standards using the Ba Leb line.

The analytical data were corrected for machine drift but not for

dead time as the count rates were. low (maximum = 525 cps). Matrix

corrections were applied using correction factors calculated for TiKQ.

(Norrish and Hutton, 1969) . The use of these Ti Ka, factors is accep-

table because of the close proximity of the Ba L4 and Ti Ko( wavelengths

and the absence of absorption edges of major elements between these

wavelengths (Willis et al., 1969) . Background counts were determined

on three Norrish fusion blanks (S102 100%, SiO2 70% - K2030% and

Na2O 50% - Al2O3 50%) . The matrix corrected results were averaged

V
 51

and subtracted from the matrix corrected counts for samples and stan-

dards. NIMS was used as the sole calibration standard in view of its

low TiO2 content (0.05%). A Ba concentration of 2495 ppm was assigned

to this standard (see data in Table 3.4). The net-peak intensity for

this standard was used to calculate a calibration factor from which the

Ba concentrations for the mineral samples and G-1 and G-2 were cal-

culated.

The determined G-2 concentration listed in Table 3.9 is very

close to the average of the Flanagan (1973, 1976a) and Abbey (1973,

1975a) values, even though this standard contains 0.5% TiO2. This

suggests that there is minimal interference from Ti K j, at this concent-

ration. The Ba value determined for G-T is low compared to the aver-

age of the Flanagan and Abbey data. However there is some doubt

with regard to the reliability of the Flanagan and Abbey values for Ba

in G-1 (see 3.1.5 above) . The Norrish disc value for Ba in G-1 sup-

ports the relatively low concentrations determined by the pressed pow-

der briquette method (Table 3.4) and in conjunction with the work of

Laeter and Hosie (1977) and Willis et al, (1969) provides further evi-

dence that the Flanagan and Abbey values for this standard are "too

high, or that this standard is inhomogeneous with respect to Ba.

3.3. Discussion

Major element analyses should approach 100% and this summa-

tion provides an important check (but not a guarantee) on the quality

of the analytical techniques. Trace element analyses cannot be

 52

checked in this way and confidence in accurate work is crucially

dependent on good quality concentration data for the calibration stan-

dards. The data for Sr and Rb in the USGS standards are considered

to be of a high quality (Flanagan, 1976a). It is however evident from

the literature that there is considerable scope for increasing the relia-

bility of the preferred values in many of the international rock standards.

If XRF counting errors are small then for many trace element analyses

the errors in the standard concentrations may well outweigh errors from

other sources such as those present in corrections for line overlap and/

or matrix effects. In view of this it is hoped that the advent of analy-

tical techniques such as isotope dilution mass spectrometry will lead

to a general improvement in the quality of preferred standard values.

Many workers use least squares regression techniques in estab-

lishing standard trace element calibrations. These techniques do not

weight the calibration line towards the higher quality data availabe from

the higher concentration standards (see 2.5.2 above) . Such weighting

is applied by the calibration technique described under Method II in

3.1.3 above. Furthermore the least squares regression technique assu-

mes that there will be no errors in the standard values. For most trace

element calibrations this will not be the case and there will therefore be

errors in both the independent and the dependent regression variables

(see 2.5.2 above) .

The Sr and Rb calibration data for the Norrish fusion discs indi-

cates that these discs may be used for the analysis of these elements

provided that an average calibration error of ni 4% is acceptable. In

 53

geochemical studies in which this applies the Norrish discs may there-

fore be used for both major element and Sr and Rb determinations. For

those samples containing < 0.5% TiO2 the Norrish discs may also be

used in the analysis of Ba. If no other trace element data are required

this will expedite the completion of the analytical work as there will be

no need for the preparation of pressed powder briquettes.

With regard to Ba analysis on either pressed powder briquettes-

or Norrish fusion discs it is regrettable that Ba Lot on Li P 220 is out-

side the 2 e range of commercially available spectrometers. Certain

workers (Strasheim and Brandt, 1968; Willis et al. , 1969) have modi-

fied their spectrometers to reach this angle (no 155° 2 e) and if this

feature were to become generally available on spectrometers it would

enable the higher count rates from the Ba L pG line to be utilized in

place of the weaker Ba IT' or 1432 ,15 lines.


Table 3.1 Trace element analyses on pressed powder briquettes — Instrumental settings

element Rb Sr Ba/1 Ba/2 Th Y Zr V Co Cr Ni Cu Zn

tube Mo Mo W Cr W Mo W W W W W W W
kv 90 90 60 60 90 90 60 60 60 60 60 60 60
ma 20 20 32 32 20 20 32 32 32 32 32 32 32
peak Km KoC 42115 Lp2,15 Lo(.. Koz KoC, KoC. KrX. KoC Koc K oc K o-
background +1.0 +1.10 +1.15 +1.25 +0.33 +0.9 +0.98 +3.21 +0.56 +1.30 +2.50 +0.50 +0.47
20 -1.0 -0.90 -1.00 -1.25 -0.33 -0.9 -1.02 -2.29 -0.44 -1.20 -2.50. -0.50 -0.47
Peak counting 10 10 100 100 100 20 10 40 40 40 40 40
time (sec.)
background 2x10 2x10 2,x40 2x40 2x40
counting time (sec. ) 2x20. 2)010 2x2O 2x20 2x20 2x20 2x40
peak overlap RbKp SrK~ TiKK FeKp VK~
__
tube spectral
interference / ,/

XRF spectrometer = Philips 1212.

Vacuum path, scintillation plus gas flow proportional counters and LiF220 analysing crystal used throughout.
Coarse collimator (480r) used for Ba/1 , fine collimator (160pm) used for Ba/2 and remaining elements.
 55

Table 3.2 Trace element data processing - program TRANS

interactive terminal output and input comments

START OF KM TRANS
TRACE ELEMENT DATA MUST BE ON TAPE 9

FIRST CHECKING PASS Trace element XRF data is checked

for errors.
EXIT - YIN
? R

COMMON BKGDS - Y/N

? u

TRANSFORM DATA - Y/N

? 'u

ENTER DESCRIP + NAME (2 SEP LINES) User enters data file title and/or
? trace element run
description and analysts name.
? J. bloass

ENTER NO OF ELEMENTS
? 3
User enters channel numbers
ENTER EL VS CHAN NOS FOR .3 ELEMENTS
? u 58 59 62 (bkgd 1 - peak - bkgd 2) for each
? sr 80 63 64
element.
7 rb Si 65 66

COMMON MOD CHANNEL NOS: ENTER CHANNEL. THAT

IS PRESENT. THEN CHANNEL TO BE ABDED - UP TO 10 EXTRA
CHANNELS AND END WITH -1, FORMAT IS (A2.X.A2) User enters common background
? 62 80
? 64 81 channel numbers.
? -1

END OF PGM TRANS

TRANSFORMED DATA IS ON TAPE 10

STOP '


56
Table 3.3 Trace element data processing - program NUTRACE
interactive terminal output and input comments

START OF POM NUTRACE

TAPE 5 = OUTPUT FILE FOR PGM TAB
TAPE 6 = OUTPUT FILE FOR LINE PRINTER
TAPE 8 = TRACE ELEMENT DATA INPUT FILE
TAPE 9 = SIDS AND ABS COEFT DATA INPUT FILE
TAPE 10 = SPL MAJOR CONCS DATA INPUT FILE

REF STD CR NI CU ZN ZR RB SR TH Y BA V CO
Trace element concent-
rations assigned by pro-
79999 445. 300. 200. 200. 300. 500. 500. 200. 200. 1500. 300. 200.
gram to reference standard,
ENTER ELEM NUMBERS OR -1 TO ESCAPE
? 2.3,5,6,7,9,10 User enters new trace
ENTER NEW TRACE PPMS (AS INTEGERS) element concentrations
? 353.335.489,727.1090.485,2507 for selected elements in
19999 445. 353. 335. 200. 489. 727. 1090. 200. 485. 2507. 300. 200. the reference standard.
ENTER TUBE BKGD FACTORS FOR CR, NI, CUr AND ZN User enters background
? 0 1.02 1.15 0
correction factors (used
NOMINAL DATA SORTED
in correcting for X-ray tube
spectral contamination).
BRIQUETTE PREPARED FROM IGNITED POWDER Y/N
~ y

SPL MAJ ELEM OXIDE DATA FROM POM TAB DATABASE FILE - Y/N

TEST WRITE OUT OF INPUT MAJ ELEM CONCS - YIN

? n
ENTER ZRFAC, YFAC. VFAC, COFAC. CRFAC User enters line overlap
0.099 0.250 0 0 0
correction factors.

LIST RESULTS - Y/N

?

TRACE ELEMENT RUN

J.BLOGGS

SPLE CR NI CU ZN ZR RB SR TH Y BA V CO Results for standards:

9000 0 22 13 0 220 216 253 0 13 987 0 0 9000 = G-1
9001 = G-2
9001 0 5 15 0 318 166 475 0 11 1810 0 0
9002 =AGV
9002 0 18 53 0 215 67 663 0 23 1236 0 0
9003 = BCR
9003 0 16 19 0 175 47 329 0 43 699 0 0
9004 = GSP
9004 0 12 .35 „ 0 514 255 234 0 31 1064 0 0 9005 =W-1
9005 0 72 116 0 86 21 188 0 25 149 0 0 9006 = DTS

1 0 0 0 0 0 0 0 9007 = PCC
9006 0 2663 9 0
0 9008 = NIMP
9007 0 2670 8 0 0 1 O O 0 0 0
9009 = NIML
9008 0 594 14 0 8 4 36 0 3 16 0 .. 0
9010 = NIMD
9009 0 5 13 0 526 188 355 0 20 378 0 0
9012 = NIMS
0 0 3 0 0 0 0 0 9013 = NIMN
9010 0 2148 5 0

9012 0 •5 23 0 13 530 61 0 0 2518 0 0 9014=NIMG

4 3 260 0 6 64 0 0
9013 0 107 15 0
0 0
9014 0 5 10 0 281 272 8 0 142 99

CONTINUE Y/N
?
0 19 641 0 0 Results for samples
i
0 347 346 0 120 110 328

83 158 14. 0 14 397 0 0

2 0 37 10 0
391 126 55 0 17 2158 0 0
3 0 38 2 0

LINE PRINTER FILE IS ON TAPE 6

RESULTS FILE FOR PGM TAB IS ON TAPE 5
STOP


Table 3.4 Trace element calibration data for pressed powder briquettes

Rb Sr Ba/1 13a/2 Th
F ave. calib . error F a ve. ca lib . error F ave. calib . error calib. error F ave.. calib. error
std. A ppm
Om PPm % A PPm' ppm II ppm 9E A Ppm ppm ppm % PPm % A PPm ! ppm PM %
r ~
21 190 160 2
W-1 21 21 0.0 190 191 0. 5 160 151 ' a 173 a 2 6 a
21 190 i 160 1 2
3 30 ; 675
BCR 97 47 47 0.0 330 330 0.0 .678 715 5.5 702 3.5 6 6 10 a
47 330 ~ 680 6
67 , 67 69 657 1208 •
AGV 3.0 659 646 2.0 1204 1154 4. 2 1190 1 .2 6 9 a
660 ' 1200 6
254 233 1300 104
GSP 252 256 1 .6 232 234 0.9 1300 1331 2. 4 1299' 0. 1 105 93 ,11.3
250 i 230 1300 105
220 250 1200 50
G-1 - - 218 0.9 - • - 251 0.4 - 997 16.9 976 18.7 - . - 53. i 6.0

1870
G-2 479 480 476 0 ,8 1860 1830 1 .6 1824 1 .9 24 24 25 j 4.2
480 1850 24

FCC

DTS

average relative error % 1.1 0.8 3.4*/6.1. 1. 7*/5.1 7.2

NIM-N I 100 88
I 110

NIM-P
550 2590
NIM-S 1 555 544 76 76 62 2495 2495
560 ; . 76 .I 2400
183 4480 498
NIM-L 192 185 4680
4580 4360
450
449 392
200

NIM-G 307 336 195 146

340 210

NIM-D i
1

F= Flanagan (1973, 1976a), A = Abbey (1973, 1975a).

a = std. excluded because . a counting error > 10%
* = average error with G-1 excluded.
 z

Table 3.4 (continued)

Y Zr V Co.

F ave. calib. error F ave. calib. error F ave. calib. error F ave. calib. . error
std. A PPm
PI= Ppm % A ppm ppm ppm % A ppm ppm ppm IE A ppm ppm ppm %
25 105 264
W-1 25 25 0.0 105 86 18.1 252 265 5.2 47 49 46 6.1
25 105 240 50
37 38
SCR 42 44 4.8 190 1 88 183 2.7 399 405 404 0.2 38 36 5.3
46 1B5 410 37

21 225 125
AGV 21 24 12.5 223 235 5.4 125 120 4.0 14 16 15 a
26 220 125 17
500 6
GSP 30 31 32 3.2 500 496 0.8 53 51 53 3.8 7 7 a
32 500 49 7
210. _
G-1 13 - 14 7.7 214 1.9 17 - 21 23.5 2 - 1 a

12 i 12 12 300 6
G-2 0.0 300 304 1.3 35 35 44 25.7 6 (27) b
12 300 34 6
112
PCC 31 29 6.5 111 107 3.6
31 11Q

DTS 13 12 9 a 134 141 5.2

135

average relative error % 4.7 5.0 3.9*/9.8 5.1

25 225
NIM-N 7 - 8 - 41 223 174 61 63 52
220

NIM-P 20 - 9 238 208 93 117

250 115
<5 6 30 - 9
NIM-S

NIM-L 30 - 17 79 78 68

100 -
NIM-G 175 280 261 •
260

141N1-17 42 44 33 173 269

2a~

F = Flanagan (1973, 1976a),. A Abbey 1973, 1975a).

a = standard excluded because ti counting error> 10%
b = gross error -' standard excluded.
* = ave. error with G -1 and G-2 excluded.
s

Table 3.4 (continued)

Cr N1 Cu Zn

std. ave. calib. error F ave. calib . error F ave. calib . error F ave, calib. error
F Pm ppm ppm
A PPm ppm % A 1313m PM % A PI= ppm % A Ppm ' Pi= PI= %
W-1 114 110 86
120
117 124 6.0 76
78
77 • 76 1. 3 110 115 4 .5 6 86 83 3.5
110 86
BCR 16 17 16 5.9 13' 15 13 a •19 19 20 5.3 120 120 118 1.7

AGV 12 14 a 18 18 0.0 63 62 54 12.9 84 84 83 1.2

1z 1;

1 19
GSP 3 14 a 11 13 a 353 34 3 2.9 98 98 104 6.1
13
45 -
G-1 1 - (25) b 13 - 13 a 49 8.2

5 85
G-2 6 7 a 12 12 17 41. 7 5 85 91 7.1
6 11 85
2730 2339 11 36
PCC
3010
2870 2669 7.0
2515
2427 2542 4. 7 11 11 11 a 36 36 45 25.0

4000 2269 45
DTS 4190 3753 10.4 2353 2520 7.1 7 7 2. 45 48
4379 2436 a
45 6.7

average relative error % 7.3 3.3 6.4*/13.5 4.9**/7.4

78
NIM-N 99 113 13 18 85 83 62
120 13
21966 470 17 100
NIM-P 23370 24875 520 602 14 16 22 100 . 109
24779 570 100
NIM -S 8
8 5 23 21 24 21 21 16
8 19 20
11
NIM-L is 18 17 11 3 15 12 280 300 422
15
N1M-G 11 11 6 15 14 14 60 60 51
10 13 60
2900 2120 8 90
NIM-D 2853 2946 1 2200 2129 9 7 90 92
2806 ) 2279 10 90

F = Flanagan (1973, 1976a), A Abbey ( 1973, 19754.

a = standard excluded because a counting error> 10%
b = gross error - standard excluded.
* = ave. error with G-2 excluded. . ** = ave. error with:FCC excluded.
 60

Table 3.5 Trace element analyses on pressed

powder briquettes - Detection limits

ave.
element std. conc., d.lim.,
ppm Ppm ppm

W-1 21 3.5
Rb G-1 221 2.1 3.0
NIM-S 556 3.3
NIM-S 76 3.5 _
Sr W-1 190 3.3 3.2
G-2 483 2.7

W-1 160 34.9

Ba/1 BCR 680' 31.7 31.5
G-2 1873 27.9
W-1 160 21.5
Ba/2 BCR 680 22.8 22.3
G-2 1873 22.7
BCR 6 5.0
Th. G-2 24 3.5 4.2
GSP 106 4.1
W-1 25 2.7
Y
BCR 42 2.8 _
2.9
W-1 105 7.3
NIM-G 280 6.7
6.0
DTS 12 3.2
AGV 126 3.3 3.4
BCR 406 3.8
GSP 7 3.2
W-1 49 5.5 4.1
DTS 134 3.5 .

AGV 12 2.4
Cr GSP 13 3.2 3.7
W-1 117 4.1
NIM-D 2833 5.1

Table continued /
 61

Table 3.5 (continued)

ave .•
element std. conc.', d .lim. , d .lim.,
ppm ppm ppm

AGV 18 2.8
Ni W-1 77 3.0 2.9
_ NIM-N 99 2.5
NIM-P 519 3.5

NIM-D 9 6.3
Cu GSP 35 3.9 4.9
W-1 110 4.4
NIM-S. 21 3.2
G-1 45 2.1 2.8
BCR 120 3.2

3Rb Rp - Rb
Detection limit = — - m=
m Tb std.conc.*

RP = peak counting rate

Rb _ = ' background counting rate
Tb , = background counting time

*Literature concentrations (average of Flanagan, 1973, 1976a

and Abbey, 1973, 1975a) adjusted for effect of rock powder
ignition.
 62

Table 3.6 Trace element analyses on pressed powder briquettes -

Estimate of combined briquette preparation and counting

precision for ignimbrite pumice sample 6502

element mean
. Alm PPm
C.V.
%
a
Rb 618 7.0 1.1

Sr 480 3.4 0.7

Ba/1 273 12.5 4.6
57.0 3.9 6.8
572 4.5 0.8
V 49.0 1.2 2.4
Co 3.0 1.9 63.3
Cr 5.4 1.3 24.1
Ni 6.4 1.1 17.2
Cu 3.4 2.5 73.5
Zn 74.8 0.8 1.1
Th 88.4 4.2 4.7

g = standard deviation; calculates from results of

5 pressed powder briquettes analysed once each.

C .V . % = per cent coefficient of variation

g x 100
mean
 63

Table 3.7 Sr, Rb and Ba analyses on Norrish discs —

Instrumental settings

element Sr Rb Ba

tube Mo

kv 90 • 90 60

ma 20 20 32

crystal LI F220 -LiF220 LiF 200

peak Kc< Ka L 04

counting . .40 40 100

time (sec.)

XRF spectrometer = Philips 1212.

Vacuum path, scintillation plus gas flow proportional

counters and fine collimator (160 um) used throughout.

r
 64

Table 3.8 Sr and Rb analyses on Norrish discs — Calibration data

Sr Rb

rel. rel.
std. conc.* calib. error conc; calib. error
ppm PPm PPm PPm

AGV 659 663 0.6 67 68 a

G-2 480 504 5.0 169 182 7.7

BR- 1350 1317 2.4 45 33

GSP 232 248 6.9 252 254 0.8

W-1 190 180 5.3 21 17 a

average relative error 4.0 4.3

NIMS 76 71 555 529

NIMG 307 330

NI ML 192 185

*Literature concentrations - average of Flanagan (1973; 1976a)

and Abbey (1973,
. 1975a) .

a = standard excluded because a counting error > 10%.

ti
 65

Table 3.9 Da analyses on Norrish discs — Calibration data

relative
std. conc .* determined error TiO2
ppm ppm % %

G-2 1860 1862 0.1 0.50

G-1 1200 1024 14.7 0.26

NIMS 2495 ** — .05

* Literature concentrations - average of Flanagan (1973, 1976a)

and Abbey (1973, 1975a).

** NIMS used as the sole calibration standard.

Table 3.10 Sr, Rb and Ba analyses on Norrish discs

Detection limits

Sr Rb Ba

d. lim.* 12.0 14.4 20.0

*Calculated from average background count rate determined on

blanks.

4
 G1)

Table 3.11 Thorium in Tenerife pumices by XRF and neutron

activation (NA) analysis

Analyst: j.A. Wolf, Dept. of Geology, Imperial College.

XRF NA
sample value value %
ppm ppm

GF8 37.0 34.1 8.5

GF10 35.5 34.0 4.4
GF12 30.0 29.1 3.1
GF13 24.6 25.2 - 2.3
EF1 17.4 18.2 4.4
EF2 21.2 19.5 8.7
EF3 18.3 19.0 - 3.7
EF4 18.8 16.4 14.6
CFI 28.8 25.8 10.4
CF2 22.9 29.2 -21.6
CF3 28.4 29.2 - 2.7
CF4 32.7 30.1 8.6
JF5 42.9 47.6 - 9.9
JF1 41.7 46.1 - 9.5
P4 29.9 31.1 - 3.9
P6 12.0 12.2 - 1.7
Pis 27.7 27.8 - 0.4
Gig 23.8 23.2 2.6
Gig 6 27.3 29.1 - 6.2

0 3/
= XRF conc - NA conc X 100
NA conc
127.2%
19 - 6.7%

Q%
n 194 = -0.3%
 fi 7

CHAPTER IV

GEOIC DATA PROCESSING SYSTEM

4.1 Introduction

GEOIC is a system for processing and evaluating major and trace ele-

ment data produced in the XRF analysis of rock samples. The aim has been

to design an integrated series of computer programs which interlink in terns

of their input and output data files. No reformatting of the data files is

required at any stage. All program options are interactive and the data pro-

cessing is carried out on remote computer terminals. The various options in

the programs are designed for maximum flexibility in the processing and eva-

luation of the analytical data. The precompiled programs and data files are

stored on disc in the central computer system and may be accessed at any

time from the terminals. Comprehensive output files generated by the pro-

grams may be directed to a line printer, while the output from a graphics

program which forms part of the system may be hard copied to microfilm; both

printed and microfilm output are suitable for direct inclusion in theses and

manuscripts. The GEOIC system accepts data for Si02, T102, A1203, Fe203,

FeO, Cr203, MnO, MgO, CaO, Na20, K20, P205, NiO, H2O-, H20+, V,

Cr, Co, Ni, Cu, Zn, Zr, Rb, Sr, Th, Y and Ba.

Data processing is carried out in three stages (Fig. 4.1).

(1) Processing of XRF data to produce major and trace element re-

sults (programs GFAC, NFAC, CHECK, START, MWT, TRANS, NUTRACE).

(2) Calculation of CIPW norms for each sample as well as collation

 68

and tabulation of major and trace element and norm data. A data base

file is generated at this stage (programs TAB, AGNORMX, LNORMX).

(3) Plotting and statistical analysis of data in the data base file

(programs GEOPLOT, STAT 1; plus STAT 2 and MIXER in future enhance-

ments of the system) .

GEOIC has been developed primarily with the needs of the igneous

geochemist/petrologist in mind and , in particular, for those studies

involving both major and trace element data for up to 150 rock samples.

This maximum is large enough to satisfy the majority of individual research

programs undertaken by igneous geochemists and petrologists, and it is

well suited to the requirements of Ph.D. thesis work. In order to accom-

modate larger research programs it is planned to increase to 300 the maxi-

mum number of samples that may be processed. The extra central memory

required for this will be obtained by the use of overlay techniques (Frisch

and Liddiard, 1976).

The pedigree of those programs in the GEOIC system (Fig. 4.1) which

contain code not written by the author is as follows. STAT 1 is a modified

version of a program written by Mr. S. Earle and Dr. R. Howarth. NUTRACE

is an extensively modified version of program XTRACE written by Dr. B. Gunn.

AGNORMX is a modified version of program AGNORM from Berkeley, Cali-

fornia (author unknown) . LNORMX is a modified version of a program writ-

ten by Dr. R. Le Maitre. Program MIXER was written by Dr. A. Duncan

and uses subroutine MXLNEQ ( Bryan et al., 1969), while STAT 2 was writ-

ten by Mr. S. Earle and Dr. R. Howarth.

 69

4.2 Computing environment

GEOIC is installed on the Imperial College computer system. This

is a dual mainframe configuration consisting of a CDC 174 (131K) and a

CDC 6500 (98K) interlinked via fast swap discs and extended core stor-

age (250K). The CDC 6500 is rated at 0.7 MIPS (million instructions per

second) while the CDC 174 is rated at 1.0 MIPS (data from S. Budd,

Imperial College Computer Centre, pers. communic.). A large disc (150

*
Mch) provides temporary storage for programs and data in execution while

a further 1500 Mch of disc space is available for permanent file storage.

Both computers run under the NOS 1.1 operating system. Much of the sys-

tem software for interlinking the dual mainframes was developed by the

Imperial College Computer Centre (ICCC). The 6500 normally processes

batch work while the 174 is usually devoted to interactive terminal proces-

sing under the Telex subsystem. Terminals operate at speeds of 100 -

300 baud . Either machine can support either mode (batch or Telex) or both

modes simultaneously and this feature helps to ensure a high level of relia-

bility in the Telex service. Interactive programs run under Telex may use

up to 25K of central memory, while batch programs may use up to 50K.

A third computer, a CDC 1700 (32K) is linked to the dual mainframes

and acts as a communications processor enabling up to 20 high speed (9600

baud) interactive graphic displays to be operated under Telex via the gra-

phics communication system developed by ICCC. ICCC have also deve-

loped a user orientated interactive graphics interface (i.e. the SIMPLE

package: Raby, 1976) for FORTRAN programs written by the computer user,

The provision by ICCC of a high level of software support in this field has

*
K = 1000 words central memory; Mch = million characters;
baud = binary digits per second
 70

greatly facilitated the development of interactive graphics programs such

as GEOPLOT (below) .

4.3 . First stage of GEOIC data processing

The major and trace element data processing using programs GFAC,

NFAC, CHECK, START, MWT, TRANS and NUTRACE has been described

in Chapters 2.6 and 3.1.4 above. The final results of the major and

trace element analyses are written out (by MWT and NUTRACE) to files

formatted for direct input into program TAB in the second stage of the data

processing.

4.4 Second stage of GEOIC data processing

Program TAB :

The purpose of this program is to bring together the various analy-

tical and normative data for a set of analyses as well as grouping, avera-

ging and sorting of the data.

The interactive features in the program allow the user to control the

data processing as follows:

(1) Deletion of any unwanted major element analyses.

(2) The replicate major element analyses for a given sample may

by averaged.

(3) Trace element analyses . may be linked to the appropriate major

element analyses for each sample.

(4) The analyses may be sorted into groups according to the eight
 71

character "group identifier" keywords present in the description of each

analysis (obtained from the dw file, Fig. 4.1). Average analyses are

calculated for each group. It is up to the geochemist to decide which

samples in a data set are to be grouped together by assigning common

group identifier keywords in the sample descriptions in the dw file.

Examples of group identifiers are as follows:

GRAN 1 = granite type 1

BAS 3 = basalt type 3

T BAS S = Tertiary basalts in Skye

PC ANOR = Precamprian anorthosites

IG A = ignimbrite eruption A

PF 2 = pyroclastic fall 2

(5) The analyses are reordered into ascending order with respect

to a preselected variable, which may be any oxide or the sample field

numbers. If the analyses have been sorted into groups then each

group will itself be reordered into ascending order for the variable in

question.

(6a) If CIPW norms are to be calculated, then the major and cer-

tain trace (Sr, Ba, Rb, Zr, Cr and Ni) element data are written out to

a file. This file serves as the input for the two norm calculation programs

(AGNORMX and LNORMX, below) . At this point program TAB ends and the

user then calculates the sample norms (Tables 4.1, 4.2) .

(6b) If the norms have already been calculated, then step (6a) is

omitted and the norms are read in from a scratch file (see below) and

linked to the appropriate samples (Table 4.3) .

 72

(7) Various major and trace element ratios are calculated for each

analysis.

(8) The major and trace element data together with the ratio and

norm data for each sample are written out to a line printer file using a

tabulation format. The group average data are also written out to the

line printer.

(9) The above procedures may be carried out for selected groups

of samples. These selected samples are extracted from the input data

file on the basis of their group identifier keywords (Table 4.4) Table

4.5 provides an example of tabulated output for the sample groups selec-

ted in Table 4.4.

(10) The sample major and trace element data and the norm data

may be written out to a data base file. Group average data are also

written to this data base file. The data base file serves as the input

file for the third stage of data processing (below).

Programs AGNORMX and LNORMX:

AGNORMX calculates CIPW norm data (including the components

of pl, di, by and ol) as well as various data for ternary diagrams, the

Thornton and Tuttle (1960) Differentiation Index (DI) and the Poldervaart
I
and Parker (1964) Crystallization Index (CI) . LNORMX also calculates

CIPW norms, but not the components of pl etc. It does however produce

line printer plots of various ternary diagrams as- well as recalculating

mineral analyses in terms of 'n' oxygen atoms. Both programs produce

line printer output as well as scratch files of sample norms for input into

program TAB (see above, (6b) and Table 4.2) .

 73

4.5 Third stage of GEOIC data processing

Program GEOPLOT:

This is designed to be run at an interactive graphics terminal

(e.g. Tektronix 4012/4014). It reads in the major and trace element

data, as well as the differentiation and crystallization index data, from

the data base file. The program enables the user to plot binary or ter-

nary diagrams for elements and oxides via a series of commands input

from the terminal keyboard. Certain element ratios and calculated vari-

ables may also be plotted. The chosen variables may be plotted for all

samples or the plotting may be restricted to specified group(s) within the

data base file. If desired only the averages of the sample groups in the

data base file may be plotted. In binary plots a linear correlation coef-

ficient calculated for the paired sample data. This coefficient is writ-

ten on to the plot diagram along with the number of data points plotted.

A further feature of the program is that the user may manually enter

binary or ternary data at the terminal keyboard for subsequent plotting.

This option is useful for plotting sample data which is not in the data base

file.

The main features of the interactive data base plotting sequence

undertaken by the user is as follows (Table 4.6):

(1) Binary or ternary plotting is selected.

(2) The variables to be plotted are selected: binary or ternary

plot labelling is automatically equated to the appropriate variable names

(3) The samples to be plotted are selected. These may be all

* Pearson product-moment
 74

samples, certain groups of samples, or only the group averages.

(4) The plot symbols are selected: each group of samples may be

plotted with a unique symbol (for up to 14 groups) or a single symbol may

be used for all groups.

(5) The resulting plot (Fig.4.2) includes a listing of the group

symbols used when plotting the data in separate groups.

(6) The plot may be hard copied to microfilm (a plot file is gene-

rated for subsequent plotting off-line) .

(7) The user may exit at this point or return to step (1) for

further plotting.

The ability to rapidly replot the selected variables in various com-

binations of groups is a powerful tool for investigating the chemical

characteristics of the samples under study (Fig. 4.2A,B). Where lin-

ear trends are evident the presence of the correlation coefficient on the

plot is important in providing an objective measure of the degree of. lin-

ear interrelation between the plotted variables (Fig. 4.2A,B).

Program STAT 1:

The program computes various statistical parameters for the sam-

ple data in the data base file. Mean, standard deviation, skewness and

kurtosis values are calculated for each element analysed as well as for

Differentiation and Crystallisation Indices. A correlation coefficient*

matrix is also calculated for the data. STAT 1 contains options which

allow the above computations to be carried out on selected groups (or

groups of groups) of samples from the data base file (Table 4.7). The

* Pearson product-moment
 75

computations may also be carried out on only the group averages for the

groups in the data base file. These options are similar to those descri-

bed for programs TAB and GEOPLOT.

Programs MIXER and STAT 2:

Future work will include linking the data base file with the inter-

active programs MIXER, for modelling igneous differentiation trends (c. f.

Bryan et al., 1969) , and STAT 2 which produces histograms of the data

for the analysed elements.

4.6 Discussion

In interactive computing it is essential that there is a reasonably

short response time between the start of a particular computation and its

completion. This response time will depend on factors such as the hard-

ware available, the quality of the operating system and the number of

other users logged onto the system.

CIPW norm calculations are widely performed on computers by earth

scientists and the time required for their computation may be used to pro-

vide a comparable estimate of the computing power available from the par-

ticular computing environment. Two tests were carried out under the

Telex system on the CDC 6500 computer on the calculation time required

for 125 norms using AGNORMX. The program was precompiled using the

MNF (Minnesota Fortran) compiler. In the first test 78 other users were

logged on and the real time taken was three minutes 45 seconds. The

central processor time was 15.6 seconds. A' second test, with 63 users

logged on, required one minute five seconds of real time and 15.5 seconds

of central processor time. The real and central processor times include
 76.

the generation, but not the printing, of the norm line printer file.

Till (1977) has described the HARDROCK package for geochemical

data processing. This package can process sets of greater than 150

samples and contains features not available in the present version of

GEOIC. HARDROCK is however a batch orientated system and uses a

line printer as a plotting medium. With regard to the GEOIC system,

the ability to do all computing, from initial major and trace element pro-

cessing through to final plotting and statistical analysis of the results,

on an interactive basis in a relatively short time (i.e. in less than an

afternoon) is a great advantage over batch systems. A prerequisite in

the rapid processing of the data is that there should be no or few errors

in the initial raw data; when errors are present the on-line editing and

correction of these errors from interactive terminals greatly expedites

the completion of the data processing.

The user selected "group identifier" keyword attached to each sam-

ple analysis facilitates the division of the samples into subgroups and

this adds considerable flexibility to the subsequent data processing and

interpretation. The interactive graphics and statistics programs com-

bined with the group identifiers provide powerful methods for studying the

inter-element characteristics of the data as a whole or in any combination

of subgroups.
 77
Table 4.1 Program TAB: generation of data file for input into norm calculation
programs
interactive terminal output and input comments
sfART OF PUM TAD
OXIDE DMA MUST BE ON TAPE 73
KEYWORD DATA ON TAPE 22
TRACE DATA ON TAPE 4
NORM DATA ON TAPE 7

**** ND - IF NO AVERAGING THEN NO ASCEND SORTING

AND NO CALCULATION OF RATIOS
ALSO NO OUTPUT FILE FOR NORM CAL C.

CO** ND - ALL NEGATIVE MAJ. ELEM. VALUES SET TO 0

Generation of data base file omitted
DATA BASE FILE - Y/N (NB ONLY FOR AVERAGED DATA) because norm data not yet available for
7 n inclusion in the data base file.

SPECIAL GROUP SELECTION Y/N Data processing carried out for all of
7 n the sample groups.

ENTER DELETION NUMBERS PLUS P (=PRIME) IF REQUIRED - END WITH 9999

This option provides for the deletion of
NAS DEL. NUNS MUST BE 4 DIGIT MACHINE A - NUMBERS
individual analyses if desired.
7 9999

194 ANALYSES READ IN

AVERAGING: 1 - NOr 2 - PRIMES`, 3 - PRIMES + GROUPS User selects averaging of samples

ENTER 1, 2r OR 3 and groups.
? 3
IS OUTPUT FILE FOR NORM CALCS REQUIRED - Y/N User selects generation of the data
7 y input file for the norm programs ,

108 PRIME A-NUMBERS FOUND

GROUP SORTING KEYWORDS: 1 - NOr 2 - FROM TML. 3 - FROM TAPE22 User selects input of group sorting
7 3 keywords from a previously created
Tape 22 file. (Option 2 requires that
GROUP SORTING KEYWORDS
the keywords be entered from tho
IO A
IG B terminal ),
IG C
IG [i-1
_ere-
SELECT ASCEND VARIABLE - FIELD NUMBER OR OXIDE
FN SI TI AL F3 F2 CR MN MG CA NA K P NI User selects sorting of samples into
00 01 02 03 04 05 06 07 08 09 10 11 12 13 ascending order of CaO concentration.
7 09
INPUT TRACE ELEM DATA Y/N INPUT NORM DATA Y/N User selects input of trace element
*NB* NO TRACE ELEM DATA ALLOWED IF NO MAJ CO NC AVERAGING data. The program will then link Zr, Sr, Ba
? yn Rb, Cr and Ni data to the major element
data for each sample in the norm cal-
WARNING - ALL ZERO VALUES EXCLUDED FROM GROUP AVERAGING culation input file (Tape 5, see below).

START OF SUBR PRINTER

*** NB - IF NO LINE PRINTER OUTPUT THEN

NO DATA BASE FILE

LINE PRINTER OUTPUT Y/N

7 n

END OF PGM TAB

FILE FOR NORM CALCULATIONS IS ON TAPE 5 Tape 5 is the input data file required by
programs AGNORMX or LNORID[ -
STOP

Table 4.2 Program AGNORMX: generation of calculated norm data file for
input into program TAB
comments
interactive terminal output and input

START OF PGM AGNORMX - INPUT DATA ON TAPE 5

First line = major element data.

NUMBER OF DATA LINES PER ANALYSIS = 2 Second line = trace element data.

131 NORMS CALCULATED

END OF POM AGNORMX - LINE PRINTER FILE ON TAPE 6 Line printer file (Tape 6) contains
- NORM FILE FOR PGM TAB ON TAPE 7 comprehensive listing of norm
calculations.
STOP
 78

Table 4.3 Program TAB: generation of line printer and data base files
for all samples

interactive terminal output and input comments

START OF PGM TAD
OXIDE DATA MUST DE ON TAPE 73
KEYWORD DATA ON TAPE 22
TRACE DATA ON TAPE 4
NORM DATA ON TAPE 7
**** ND - IF NO AVERAGING THEN NO ASCEND SORTING
AND NO CALCULATION OF RATIOS
ALSO NO OUTPUT FILE FOR NORM CALL.
**** NB - ALL NEGATIVE MAJ. ELEM. VALUES SET TO 0
DATA BASE FILE - Y/N (NB ONLY FOR AVERAGED DATA)
User selects generation of data baso file.
?
SPECIAL GROUP SELECTION Y/N
n
ENTER DELETION NUMBERS PLUS P (.PRIME) IF REQUIRED - END WITH 9999
NB: DEL. NUNS MUST BE 4 DIGIT MACHINE A - NUMBERS
7 9999
194 ANALYSES READ IN
AVERAGING: 1 - NO. 2 - PRIMES. 3 - PRIMES f GROUPS
ENTER 1. 2. OR 3
? 3
IS OUTPUT FILE FOR NORM CALCS REQUIRED - Y/N
? n Norms already calculated (Table s 4,1 and
4.2).
108 PRIME A-NUMBERS FOUND
' GROUP SORTING KEYWORDS: 1 - NO. 2 - FROM TML, 3 - FROM TAPE22
? 3
GROUP SORTING KEYWORDS
IG A
IG B
IG C
IG D-1
~ .
LL
irabG

SELECT ASCEND VARIABLE - FIELD NUMBER OR OXIDE

FN SI TI AL F3 F2 CR MN MG CA NA K. P NI
00 01 02 03 04 05 06 07 08 09 10 11 12 13
7 09
INPUT TRACE ELEM DATA Y/N INPUT NORM DATA Y/N User selects input of trace element and
*NB* NO TRACE ELEM DATA ALLOWED IF NO MAJ CONC AVERAGING
7 `!'! norm data.

WARNING - ALL ZERO VALUES EXCLUDED FROM GROUP AVERAGING

START OF SUBR PRINTER

*** NB - IF NO LINE PRINTER OUTPUT THEN

NO DATA BASE FILE
LINE PRINTER OUTPUT Y/N
?

END OF PGM TAB

LINE PRINTER FILE IS ON TAPE 23
Line printer and data base files contain
DATA BASE FILE IS ON TAPE 3 consolidated major, trace and norm data
for each sample.
STOP
 79
Table 4.4 Program TAB: generation of line printer and data base files
for selected groups of samples

interactive terminal output and input comments

START OF POM TAD

OXIDE DATA MUST DE ON TAPE 73
KEYWORD DATA ON TAPE 22
TRACE DATA ON TAPE 4
NORM DATA ON TAPE 7

**** ND - IF NO AVERAGING THEN NO ASCEND SORTING

AND NO CALCULATION OF RATIOS
ALSO NO OUTPUT FILE FOR NORM CALC.

**** NB - ALL NEGATIVE MAJ. ELEM. VALUES SET TO 0

DATA BASE FILE - Y/N (NB ONLY FOR AVERAGED DATA)
7
SPECIAL GROUP SELECTION Y/N User selects option for extracting, certain
7 9 groups of samples for processing_

ENTER GROUPING KEYWORDS - MAX OF 50 •

MAX CHARACTERS PER KEYWORD = 8, END WITH XX User enters identifier keywords for the
7 ie a-2 desired groups of samples.
T Pf-2
? xx

ENTER DELETION NUMBERS PLUS P (=PRIME) IF REQUIRED - END WITH 9999

NB: DEL. NUMS MUST BE 4 DIGIT MACHINE A - NUMBERS
7 9999

10 ANALYSES READ IN

AVERAGING: 1 - NO. 2 - PRIMES, 3 - PRIMES + GROUPS

ENTER 1. 2r OR 3
7 3
IS OUTPUT FILE FOR NORM CALCS REQUIRED - Y/N
7 n
5 PRIME A-NUMBERS FOUND
GROUP SORTING KEYWORDS
IG E-2
PF-2
•
SELECT ASCEND VARIABLE - FIELD NUMBER OR OXIDE
FN SI TI AL F3 F2 CR MN MG CA NA K P NI
00 01 02 03 04 05 06 07 08 09 10 11 12 13
.? 09

INPUT TRACE ELEM DATA Y/N INPUT NORM DATA Y/N

*NB* NO TRACE ELEM DATA ALLOWED IF NO MAJ CONC AVERAGING
? 99

WARNING - ALL ZERO VALUES EXCLUDED FROM GROUP AVERAGING

START OF SUBR PRINTER

*** NB - IF NO LINE PRINTER OUTPUT THEN

NO DATA BASE FILE
LINE PRINTER OUTPUT Y/N
7 9
END OF PGM TAB
The line printer and data base files con-
LINE PRINTER FILE IS ON TAPE 23 tain data for only those samples found in
groups IG E-2 and PF-2. An example of
DATA BASE FILE IS ON TAPE 3 the tabulated line printer output for these
two groups is given In Table 4.5
STOP
 80
Table 4.5 Program TAB: tabulated line printer output for selected
groups of samples
901 - 2911 - sample field numbers. Numbers > 900000 reserved for group averages (e.g. 900001
and 900002). IG E-2 and PF-2 are group identifiers. 901/AV 2--i the major element data for
sample 901 Is the average of duplicate analyses, etc. 900001/AV 3 -0. the data listed for number
900001 Is the average of three samples from group IG E-2, etc . The samples within the IG E-2
and PF-2 groups have been sorted according to increasing CeO content.

VULSINI PUMICES AND LAVAS

R. PARKER. IMPERIAL COLLEGE, 74/75/76

901 2501 4201 3701 2911 900001 900002

5102 40.45 40.11 40.92 . 59.62 60.36 9102 48.49 59.99

1IO2 .81 .75 .02 .32 .32 1IO2 .79 .32
AL203 18.39 17.61 17.60 18.27 17.95 AL203 17.07 18.11
FE203 5.92 4.69 4.88 1.03 1.25 FE203 5.16 1.14
EEO 2.26 2.66 3.07 1.31 1.13 FED 2.66 1.22
MNO .14 .17 .14 .15 .15 MNO .15 .15
MGO 3.40 2.86 3.57 .20 .26 MGO 3.28 .23
CAO 8.57 8.89 9.06 2.00 2.15 CAO 8.84 2.00
NA20 1.00 1.73 1.19 4.25 4.20 NA20 1.31 4.23
K20 8.08 8.18 7.79 8.11 8.46 K20 8.02 8.28
P205 .51 .91 .55 .07 .08 P205 .66 .08
H20- .65 1.08 .57 .46 .22 H20- .77 .34
LOI 1.41 1.61 1.18 3.53 3.18 L0I 1.40 3.36
TOTAL 99.59 99.25 99.34 99.32 99.71 TOTAL 99.40 99.53

F3/F2+F3 .72 .64 .61 .44 .53 F3/F2+F3 .66 . .40

K20/NA20 8.08 4.73 6.55 1.91 2.01 K20/NA20 6.12 1.96
MGO/K20 .42 .35 .46 .02 .03 MGO/K20 .41 .03
F34-F2/CA 1.01 .89 .95 1.33 1.23 F34-F2/CA .95 1.27
K20/CAO .94 .92 .86 4.05 3.93 K20/CAO .91 3.90
F3/F2+F3 = FE203/(FEO + FE203) F3/F2-4F3 = FE203/(FE0 i- FE203)
F3+F2/CA = (FE3+ + FE24)/CA2+ F34-F2/CA = (FE3+ + FE2+) /CA2 (•

V 218 221 210 37 35 V 216 36

CR 16 20 17 5 3 CR 18 4
CO 26 21 0 0 1 CO 24 1
NI 26 15 22 7 4 NI 21 6
CU 18 19 37 9 9 CU 25 9
ZN 80 86. 85 87 89 ZN 84 BB
ZR 321 330 316 709 716 ZR 322 713
RB 653 611 704 644 671 RB 656 650
SR 1461 1456 1535 195 245 SR 1484 220
TH 38 46 39 116 118 TH 41 117
Y 38 42 36 63 67 Y 39 65
BA 904 941 849 66 95 BA 898 81

•
K/RB 103. 111. 92. 105. 105. K/RB 101. 104.
RB/SR .447 .420 .459 3.303 2.739 RB/SR .442 2.991
1)1/K .0006 .0007 .0006 .0017 .0017 TH/K .0006 .0017

DI 50.19 53.62 49.85 83.51 85.07 DI 51.18 84.23

CI 37.17 31.39 37.09 7.56 6.43 CI 35.90 7.10

ZR .06 .07 .06 .14 .14 ZR .06 .14

OR 37.06 35.36 37.61 48.13 50.21 OR 36.36 49.14
PL 21.72 16.03 19.56 41.41 39.08 PL 19.09 40.11
(AB) 0 0 0 34.69 34.05 (AB) 0 34.25
(AN) 21.72 16.03 19.56 6.72 5.03 (AN) 19.09 5.87
LC 8.54 10.33 6.78 0 0 LC 8.82 0
NE 4.58 7.93 5.45 .69 .81 NE 6.00 .84
WO 0 1.19 0 0 .73 WO 0 .09
DI 14.10 15.53 17.58 2.41 2.89 DI 16.43 3.21
(WO) 7.56 8.32 9.37 1.18 1.45 (WO) 8.81 1.59
(EN) 6.54 7.12 7.76 .34 .65 (EN) 7.62 .57
(FS) 0 .09 .45 .89 .79 (FS) 0 1.05
OL 1.35 0 .84 .43 0 OL .39 0
(FO) 1.35 0 .79 .11 0 (FO) .39 0
(FA) 0 0 .05 .32 0 MT 6.77 1.65
MT 5.39 6.80 7.08 1.49 1.81 IL 1.50 .61
IL 1.54 1.42 1.56 .61 .61 HM .49 0
HM 2.20 0 0 0 0 AP 1.56 .19
AP 1.21 2.16 1.30 .17 .19

900001 /AV 3 IG E-2

901 /AV 2 IO E-2 PITIGLIANO 900002 /AV 2 PF-2
2501 /AV 2 IG E-2 PITIGLIANO
4201 /AV 2 IG E-2 ISCHIA
3701 /AV 2 PF-2 CANINO
2911 /AV 2 PF-2 SORANO
 81
Table 4.6 Program GEOPLOT: generation of binary and ternary plots
underlined data are terminal entries by user.

interactive terminal output and input comments

PGM GEOPLOT FOR PLOTTING GEOCHEMICAL DATA - MK 2 ,/ 4-78

DATA INPUT FOR UP TO 150 SAMPLES:

1 - DATA BASE FILE_ ON TAPE 3
2 - FREE FORMAT BINARY DATA FROM TERMINAL User selects input data on
3 - FREE FORMAT TERNARY DATA FROM TERMINAL Tape 3,
ENTER 1 2 OR 3: 7 1

B • BINARY OR T • TERNARY PLOT User selects binary plot of new

1 • NEU PLOTTING VARIABLES, 2 • NO CHANGE ? 81 variables;
ANALYTICAL VARIABLES ARE:
MAJ SI TI AL F3 F2 CR MN MG CA NA K P NI H- LI TO M1 M2 M3 M4
MA- 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 SI = major element oxide (x) etc.

TRA V CR CO NI CU ZN ZR RB SR TH Y BA T1 T2 T3
TA- 01 02 03 04 85 06 07 08 09 10 11 12 13 14 IS V = trace element (ppm) etc.

CALCULATED VARIABLES:
F3T = Fe2O3, F2T = Four
MAJ F3T F2T F2F3 NAK CAMG F2MN K/NA NA/K MI FI F2F3 = FeO+Fe203,
MC- 01 02 03 04 05 06 07 08 09 10 NAK = Na2O+K2O,
CAMG = CaO+MgO,
TRA K/RB K/BA RB/SR RB/BA BA/SR ZR/Y BA/ZR K/ZR TH/K K/SR
TC- 01 02 03 04 05 06 07 08 09 10 F2 MN = FeO+MnO,
TRA BA/RB CA/SR CA/Y TI/ZR CR/V MI = matte index,
IC- 11 12 13 14 15 FI = felsic index,

NORM DI CI
NC 01 02 User selects binary variables
ENTER B PLOT VARIABLES (EG MA-09 TA-12): ? MA-09 TA-09 (le CaO vs Sr).

PLOT 1-ALL SPLES, 2-CERTAIN GRPS, 3-AVE GRPS, 4-NO CHANGE User selects plotting of all
ENTER 1,2,3, OR 4 - FREE FMT ? 1 samples in the data base file.

DATA MAY BE PLOTTED WITH A SINGLE SYMBOL

OR WITH DIFFERENT SYMBOLS FOR UP TO 14 GROUPS.
MAX NUMBER OF GROUPS IS 30. GROUPS 15 TO 30
ARE PLOTTED WITH A VERTICAL DASH.
User selects group symbols for
SINGLE OR GROUP SYMBOLS, ENTER S OR G ? G the samples to be plotted.

0
DATA IS PLOTTED ON TERMINAL
SCREEN (Fig 4.2A).

HARDCOPY Y/N y User selects hardcopy option.


Table 4.6 (continued) 82

interactive terminal output and input comments

CONTINUE Y/N V

B • BINARY OR T • TERNARY PLOT

t • NEU PLOTTING VARIABLES, 2 • NO CHANGE 7 B2 User selects binary re-pl o t of
CaO vs Sr data, but only forcer-
PLOT 1-ALL SPLES, 2-CERTAIN GRPS, 3-AVE GRPS, 4-NO CHANGE taln groups of samples
ENTER 1,2,3, OR 4 - FREE FMT 7 2

ENTER 1-GRPS TO PLOT, 2-GRPS TO DELETE

ENTER 1 OR 2, FREE FMT 1 User enters keyword identifiers
of sample groups to be plotted.
ENTER PLOT GRPS, MAX 20, EXIT • 9999
7 IG A
? IG B
7 IG C
7 IG D-1
7 IG D-2
? IG E-1
? 9999

DATA MAY BE PLOTTED PITH A SINGLE SYMBOL

OR KITH DIFFERENT SYMBOLS FOR UP TO 14 GROUPS.
MAX NUMBER OF GROUPS IS 30. GROUPS 15 TO 30
ARE PLOTTED KITH A VERTICAL DASH.
User selects group symbols for
SINGLE OR GROUP SYMBOLS, ENTER S OR G ? G the samples to be plotted.

0 DATA IS PLOTTED ON TERMINAL

SCREEN (Fig 4.2B)

HARDCOPY Y/N
V

CONTINUE Y/N V

B • BINARY OR T • TERNARY PLOT User selects ternary plot of new

1 • NEU PLOTTING VARIABLES, 2 • NO CHANGE ? T1 variables.

ANALYTICAL VARIABLES ARE:

MAJ SI TI AL F3 F2 CR MN MG CA NA K P NI H- LI TO M1 M2 M3 M4
MA- 01 02 03 04 05 06 07 08 09 18 11 12 13 14 15 16 17 18 19 20
TRA V CR CO NI CU ZN ZR RB SR TH Y BA T1 T2 T3
TA- 01 02 03 04 OS 06 07 08 09 18 11 12 13 14 15
CALCULATED VARIABLES:
MAJ F3T F2T F2F3 NAK CAMG F2MN K/NA NA/K MI FI
MC- 01 02 03 04 05 06 8? 08 09 10
TRA K/RB K/BA RB/SR RB/BA BA/SR Z R/Y BA/ZR K/ZR TH/K K/SR
TC- 01 02 03 04 05 86 07 08 09 10
TRA BA/RB CA/SR CA/Y TI/ZR CR/V
TC- 11 12 13 14 1S
' NORM DI CI
NC 01 02
User selects ternary variables to
ENTER T PLOT VARIABLES :EG MA-05 MC-04 MA-08): MC-02 MC-04 MC-OS be plotted (ie FeOT vs Na2O+K2
vs CaO+MgO).
 83
Table 4.6 (continued)

interactive terminal output and input comment s

PLOT 1-ALL SPLES, 2-CERTAIN GRPS, 3-AVE GRPS. 4-NO CHANGE User selects plotting of all
ENTER 1,2,3, OR 4 - FREE FMT ? 1 samples.

DATA MAY BE PLOTTED WITH A SINGLE SYMBOL

OR WITH DIFFERENT SYMBOLS FOR UP TO 14 GROUPS.
MAX NUMBER OF GROUPS IS 30. GROUPS 15 TO 30
ARE PLOTTED WITH A VERTICAL DASH.

SINGLE OR GROUP SYMBOLS, ENTER S OR G ? S User selects single Plotting

symbol.

DATA IS PLOTTED ON TERMINAL

SCREEN (Fig 4.2C).

HARDCOPY YIN Y

CONTINUE YIN Y

B • BINARY OR T • TERNARY PLOT User selects ternary roplot of

1 • HEW PLOTTING VARIABLES, 2 • NO CHANGE ? T2 FeOT vs Na20+X20 vs CaO+MgO
PLOT 1-ALL SPLES, 2-CERTAIN GRPS, 3-AVE GRPS, 4-NO CHANGE User selects plotting of group
ENTER 1,2,3, OR 4 - FREE FMT ? 3 averages.

DATA MAY BE PLOTTED WITH A SINGLE SYMBOL

OR WITH DIFFERENT SYMBOLS FOR UP TO 14 GROUPS.
MAX NUMBER OF GROUPS IS 30. GROUPS 15 TO 30
ARE PLOTTED WITH A VERTICAL DASH.
User selects group plotting
SINGLE OR GROUP SYMBOLS, ENTER S OR G ? Si- symbols.

DATA IS PLOTTED ON TERMINAL

SCREEN (Fig 4.2D)

HARDCOPY YIN Y

CONTINUE YIN N
TO CREATE MICROFILM HARDCOPY: Program reminds user of post-
processing commands required
LIBFILE,PROC61 to create microfilm hardcopy of
-PROC61 plots.

STOP

11
Table 4.7 Program STAT 1: generation of statistical data for selected 84
groups of samples
interactive terminal output and input comments
POM STAT1 - MODIFIED FROM PGM OPTRAN (S.EARLE. AORG)
INPUT DATABASE FILE ON TAPE 3

VULSINI PUMICES AND LAVAS Program lists header Information from

R. PARKER. IMPERIAL COLLEGE. 74/75/76 data base tile.
WW 1.7 84.50 0000000000 4 50000
MTN
User omits optran options. (Thong options
OPTRAN OPTIONS Y/N provide the means for investigating the
7 N Box-Cox transformation Characferrisflcs of
the data (S,Earle, iK 41..b))
ENTER 1 C Cr 2- NO C C
7 1 User selects correlation coefficient
calculation.
VARIABLE DELETIONS = ENTER NUMBERS

SI TI AL F3 F2 CR MN MG CA NA K P NI H- LI TO M1 M2 M3 M4 User enters number equivalents of those

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 1B 19 20 elements for which data processing Is not
7 06 07 08 13 14 15 16 17 18 19 20 required.
V CR CO NI CU ZN ZR RB SR TH Y BA Ti 72 T3 T4 75 T6 77 TO
01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20
7 01 02 04 05 06 07 13 14 15 16 17 18 19 20
DI CI
01 02
7'02
FOLLOWING VARIABLES DELETED
C203
MNO
MOO
NIO
H20+
LOI

V
CR
NI
CU
714
ZR

CI

CALC FOR 1 - ALL SPLS. 2 - GBPS, 3 - AVE GRPS User selects data processing for only
7 2 certain groups of samples.

ENTER GAPS - MAX 30 - EXIT WITH 9999

? IG E-2 User enters keywords of desired groups
? PF-2 of samples.
7 9999
5 SAMPLES READ IN

TEST WRITE OUT OF INPUT DATA - Y/N

a N

If results are written to Tape 9 then no

RESULTS TO TAPE 9 Y/N terminal output. Tape 9 is a line
?N printer file. -

VULSINI PUMICES AND LAVAS

R. PARKER. IMPERIAL COLLEGE. 74/75/76

VARIABLE --MEAN-- --STD-- -SKEW- -KURT- --N--

SI02 53.09 6.309 .291 .756 5

1IO2 .6040 .2606 -.270 .756 5

A203 17.96 .3651 .055 .844 5

F203 3.554 2.254 -.197 .812 '5

:b
EEO 2.086 .8432 -.065 .888 5

CAO 6.134 3.710 -.288 .750 5



Table 4.7 (continued} 85
interactive terminal output and input
r comme nts
NA20 2.474 1.621 .235 .770 3
K20 8.124 .2396 .011 1.528 5
P205 .4240 .3547 .145 1.085 5
CO 16.00 13.23 -.324 .667 3

RD 656.6 34.30 .063 1.364 5

SR 978.4 693.3 -.209 .752 5

TH 71.40 41.75 .281 .753 5

Y 49.20 14.65 .266 .805 5

BA 571.0 449.1 -.282 .754 5
DI 64.45 18.18 .280 .757 5

ENTER NO. OF COLS IN MATRIX - MAX 15 FOR L P'TER, MAX 09 FOR TTY User enters selected column
? 09 width of the correlation co-
efficient matrix.

CORRELATION MATRIX

8IO2 TI02 A203 F203 FED CAD NA20 K20 P205

8IO2 1.0000
1IO2 -.9807 1.0000
A203 .3474 -.3297 1.0000
F203 -.9738 .9033 -.1874 1.0000
FED -.9290 .9374 -.5777 .8588 '1.0000
CAO -.9954 .9933 -.4083 .9682 .9511 1.0000
NA20 .9790 -.9967 .2628 -.9913 -.9101 -.9819 1.0000
K20 .6067 -.6503 .2115 -.5656 -.7603 -.6217 .6482 1.0000
P205 -.9114 .8496 -.5499 .8158 .8539 .9028 -.8144 -.3803 1.0000
CO -.9771 .9970 .2610 .9978 .9026 .9732 -.9996 -.9978 .7694
RD .0792 .0651 -.1338 .0037 .1370 -.0020 -.0861 -.4019 -.3492
SR -.9938 .9964 -.3907 .9742 .9472 .9994 -.9875 -.6258 .8880
TH .9934 -.9992 .3207 -.9850 -.9334 -.9947 .9953 .6473 -.8631
Y ' .9800 -.9940 .3064 -.9690 -.9469 -.9829 .9900 .7323 -.8243
DA -.9976 .9852 -.3632 .9750 .9182 .9946 -.9760
Iii .9924 -.9993 .3246 -.9822 -.9387 -.9942 .9948 -.5567 .9190
.6632 -.8595

CO RD SR TH Y BA DI
CO 1.0000
RD -.5992 1.0000
SR .9026 .0229 1.0000
TH -.9951 -.0259 -.9968 1.0000
Y -.9980 -.0869 -.9850 .9930 1.0000
BA .9739 -.0874 .9932 -.9899 -.9690 1.0000
DI -.9949 -.0361 -.9964 .9998 .9952 -.9876 1.0000

NUMBER OF SAMPLES USED IN CORRELATION CALCULATION

8IO2 1102 A203 F203 FED CAO NA20 K20 P205

5102 5.0000
TIO2 5.0000 5.0000
A203 5.0000 5.0000 5.0000
F203 5.0000 5.0000 5.0000 5.0000
FEO 5.0000 5.0000 5.0000 5.0000 5.0000
CAO 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
NA20 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
K20 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
P205 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
CO 3.0000 3.0000 3.0000 3.0000 3.0000 3.0000 3.0000 3.0000 3.0000
RB 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
SR 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
TH 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
Y 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
BA 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000
DI 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000

CO RB SR TH Y BA DI'
CO 5.0000
RD 3.0000 5.0000
SR 3.0000 5.0000 5.0000
TH 3.0000 5.0000 5.0000 5.0000
Y 3.0000 5.0000 5.0000 5.0000 5.0000
DA 3.0000 5.0000 5.0000 5.0000 5.0000 5.0000
DI 3.0000 5.0000 5.0000 5.0000 5.0000 5.0000 5.0000

User may return to the start

CONTINUE WITH 1HIS DATA SET Y/N of the program and reprocess
T N
the file if desired.

STOP
Fig. 4.1 GEOIC data processing system

Sigge 1
(Major element processing)

GFAC
rstd file
NFAC 1
f
Trace element processings
~

START MWT NUTRACE TRANS

Stage 2 AGNORMX
J tabulated
TAB )/J line printer
LNORMX 6- file

Stage 3 ' I
(data base _ _ -›,
MIXER
file

1 STAT2
STATT

o
o Fig. 4.2A Fig. 4 .2C Fig. 4.2 Program GEOPLOT:
R = 0.5907 N : 72 FEOT
binary and terna ry plots
N
z is ā ō generated in Table 4.6
¥y IGC •
O 1C 0-1 +
o 10 0-2 X
0 • X IG E-I O
ID- •• 6 Z Z - IG E-2 +
p IGF R
ei+r
x -v-+ x R. Z LP-1 Z
X0 LP-4 Y
o • • Z LP-5 X
o Yy x I LR-1 I
x _ LR-3 K
e x xZ LR-4 *
=— WII
x
0_
CL
x A - binary plot of all samples using group
ō
• x symbols for each sample.
CC o - + ++ +

83 B - binary plot of selected samples using

o ~ group symbols for each sample.
o &+
o +
.c- ternary plot of all samples using a
U •00 41.00 81.00 12.00 16.00 NR20*1(20 CRO+MGO single symbol for each sample.
CRO W T P C

R = correlation N = number of
o coefficient samples D- ternary plot of group averages using
o Fig. 4 .2B group symbols for the averages.
Fig. 4 .2 D
C R = 0.8665 N=36 FEOT
- • 10R 0 1G 0
4 IG 8 0 10 8 0
IGC s 1G s
O - 10 0-1 + 10 0-1 +
0 1G 0-2 X 10 0-2 x
0 IC E-1 O IG E-1 0
°I.
- ^ 1G E-2 +
O- A • IGF R
LP-1 Z
X LP-4 Y
Aa LP-5 X
0O p
0 LR-1 11
0 - LR-3 B
0 •
x •e LR-4 •
CL X
CL •
0
o O
0' •
cn,nJ O ` x Z ■
• 0 Y Y
• + • •
CI .
0
o ~ + + +
0 +
_ ..-
--1.60 2.40 31.20 4 '.00 4.80 NR20+K20 ..- CR0•MGO
CRO WTPC
 88

CHAPTER V

QUANTITATIVE DETERMINATION OF ANALCIME IN PUMICE SAMPLES

BY X-RAY DIFFRACTION

5.1 Introduction

This chapter is concerned with the quantitative determination of

analcime in certain pumice and lava samples collected from the Vulsini

volcano in central Italy. The volcanic rocks from this area are noted for

their high potassium values and the presence of leucite. Leucite readily

alters to analcime even at relatively low surface (25° C) temperatures

(Gupta and Fyfe, 1975). Chapter 6 discusses the results of the analy-

tical work on the samples and in order to correct the analyses for the

leucite-analcime alteration, the concentration of analcime in the samples

was required. The friable nature of the analcime in these samples results

in a significant loss of analcime on thin-sectioning the sample. This pre-

cluded the use of quantitative point-counting methods using optical micro-

scopy.

Table 5.1 lists three leucite and three analcime analyses. The

analcime samples were purified using a combination of hand picking,

sieving through -400 mesh (the analcime was very friable and could be

gently brushed through the mesh, leaving the other mineral components

behind) and electromagnetic separation. Leucite samples were purified

by hand picking and electromagnetic separation. X-ray diffraction (XRD)

scans were used to check the purity of the mineral separations. Appendix

A describes the method used in correcting the analyses for the effect of
 89

the leucite to analcime alteration. This correction uses the average leu-

cite and analcime analyses (Table 5.1) and the quantitative XRD determi-

nations of analcime in the samples Pian <r 2 , 174S a) .

Various techniques for quantitiative X-ray diffractometric determin-

ation of mineral concentrations have been discussed by Klug and Alexander

(1973). If an internal standard is not used then a knowledge of the mass

absorption coefficients of the samples and standards is required. These

coefficients may be determined directly on the powders (Leroux et al.,

1953; Williams, 1959; Norrish and Taylor, 1962; and Niskanen, 1964).

As an alternative to direct measurement, the absorption coefficients may

be calculated, provided that major-element concentration data is available

for the powders. While Klug and Alexander (1973) discuss in detail a

large number of quantitative XRD techniques, they do not include the method

of calculating the absorption correction. The purpose of this chapter is to

describe the application of this latter method to the Vulsini samples .

5.2 Theoretical

The XRD determinations were based on a calibration line construc-

ted from seven standards prepared by spiking a pumice sample with vary-

ing weight fractions of purified analcime. The required range of the cali-

bration line was 0 to 40 per cent analcime. The XRD intensities from the

standards were adjusted for absorption effects due to the change in the

total mass absorption coefficient resulting from the addition of the spiked

analcime. The basis for this mass absorption correction is given below.

The major element compositions of the standard powders were used

M
 90

to calculate total mass absorption coefficients using the formula:

(1)

whore = total mass absorption coefficient for the standard at a given

wavelength.

= mass absorption coefficient for pure element i in the standard

• at the given wavelength.

Wi = weight fradtion of element i in the standard.

The Al were calculated for CuKa, from an algorithm supplied by

K. Norrish to M.T. Frost (pers. corn.). Alternative sources for these

coefficients abound (Heinrich, 1966). The weight fraction of oxygen in

the powder was determined by difference and the absorption due to this

element was included in the calculation of /11.t .

The correction of the variation of At in the standards was as

follows. The observed XRD intensity (IOBS) for a given weight fraction

of mineral x in a standard of,t'is given by Klug and Alexander (1973):

Woe_ • K
(2)

where Y = instrumental constant.

-Wzc, = weight fraction of mineral x.

= density of mineral x.
CC-
 91

Now if we postulate a second, hypothetical standard containing the same

weight fraction of mineral x , but having a different total mass absorption

coefficientitz then the theoretical intensity (TTHEOR) that would be

observed in this standard will be:

wvc . K
1.1 H EOl2 (3)

If °Bs I At1 and z

t are known, then it follows. that:

-~- , #Icot Zoas Atf

~tz
(4)

Thus standards of dissimilar At may be compared by normalizing the

observed intensities of the standards with respect to a selected it

These normalized intensities will then produce a linear relationship when

compared to the weight fractions of mineral x in the standards. Note that

the At normalization may be carried out with respect to the At of any

one of the standards. The unknown samples may then be compared against

the normalized calibration, provided the sample At has been calculated

and used to normalize the sample intensity with respect to the calibration

line At . For large batches of samples a suitable computer program will

expedite the calculation of the sample ,~(t from the major element compo-

sition of each sample. /

 92

5.3 Experimental

A Philips X-ray diffractometer was used with a nickel filtered Cu

tube operating at 40 kv and 20 ma. The divergence and scatter slits were

1° and the receiving slit was 0.2mm. The X-ray intensities were mea-

sured on a scintillation counter having a linear response over the obser-

ved intensity range (max. observed intensity <700 c.p.s.).

All the standard and sample powders (0.5g aliquots) were hand

ground in an agate mortar to pass a -400 mesh nylon sieve. A test pow-

der, ground to -400 mesh, was found to give the strongest diffraction in- .

tensity for analcime when compared with coarser grindings (-300 and -170

mesh) . The standards were homogenized using a high speed shaker. By

far the greatest variation in the diffracted intensity in repeated analyses

of the same powder was found to be associated with the loading of the pow-

der in the diffractometer cavity-mount. In order to assist the production

of reproducible loadings , the cavity-mount was clamped upside down on a

glass plate using a small metal clamp attached to a wooden base. The

powder was then pressed into the cavity from the back and a small glass

slide applied as a backing. The glass-slide was held in place on the

back of the cavity-mount by "sticky" tape. The reproducibility of this

method of loading with be further discussed in the results section below.

0 0
XRD scans (7A - 1.5A) on the pumice powders produced sharp analy-

tical peaks and a generally low and flat background reflecting the dominantly
O
glassy matrix. Intensity measurements were made on the analcime 5.60A
0 0
line. The leucite lines 5.54A and 5.39A are possible sources of inter-

ference, but careful scanning in this region did not detect the presence of
 93

these lines in any of the pumice powders. Initially integrated intensities

were measured by counting the diffracted X-rays while scanning over the

peak. It was found, however, that by carefully setting the goniometer

on the 5.60A line and counting for a fixed time, that good calibration data

were produced. All the powders were counted in this manner for two con-

secutive periods of 40 seconds and the results averaged. The background

O
intensities were measured for 40 seconds either side of the 5.60A line

(normally ± 0.5 2e) , and the averaged results subtracted from the peak in-

tensities. Each standard powder was loaded and counted as above at

least in duplicate. The net-peak intensities from the loadings of a given

standard were finally averaged.

A reference sample was permanently mounted in a separate cavity-

mount and this sample was diffracted at N 2 hour intervals to provide data

to correct the peak minus background measurements for any machine drift.

The stability of the X-ray generator and counting circuits were such that

negligible drift occurred within this period.

5.4 Results of standard calibrations and discussion

The XRD calibration data for the seven standards are presented in

Table 5.2. In order to evaluate the reproducibility of the sample loading

technique, the standard deviation for the loadings of each analcime stan-

dard were computed and are listed in Table 5.2 with the corresponding rel-

ative standard deviation (as a percentage of mean counts) . These data

were not computed for standard 7 because it was only loaded twice. With

regard to the other standards , the number of loadings for each standard are
 94

not enough to allow completely reliable relative standard deviation (RSD)

data to be computed for each individual set of standard loadings. How-

ever, the narrow range of the computed RSD's , and the absence of large

changes in RSD with concentration, indicates that the mean RSD should be

significant. This mean value (2.07%) is similar to the best mean RSD of

2.0% reported by Niskanen (1964) for pure milled quartz loaded 10 times

in a rotating sample holder.

The factors that affect the variability of quantitative XRD data have

been discussed by Klug and Alexander (1973) and the reproducibility of the

standard loadings indicated in Table 5.2 is considered to reflect the follow-

ing factors: (i) fine grain size achieved by hand grinding all powders to

pass -400 mesh; (ii) the very poor cleavage exhibited by analcime meant

that preferred orientation effects resulting from loading were negligible or

absent; (iii) the ease of producing uniform loadings using the mounting

clamp described above. It may be noted that the reproducibility of the

loadings would have been further improved by the use of a rotating sample

holder and a larger primary beam (2°-4° ) . Furthermore, any preferred orien-

tation effects (often present with other minerals) could have been mitigated

by using a "rough" pressing surface (ground glass or filter paper, see

Nōrrish and Taylor, 1962, p.107) .

Table 5.2 lists the calculated total mass absorption coefficient and

the absorption correction factor (ACF) for each standard. This last factor

is simply the At of the individual standard ratioed against the At for

the first standard. The ACF's were then used to correct the observed stan-

dard intensities for the variations in At as per equation (4).

 95

The corrected intensitiess together with the percentage spiked anal-

cime in each standard were then used to compute a least squares regres-

sion line through the data. The slope (337.2 counts per 1% analcime)

and intercept (2478 counts) of this regression allowed the analcime con-

centration in the unspiked standard to be calculated (i.e. 5.67% anal-

cime) . From this the total analcime concentrations in the spiked stan-

dards were calculated and the results are listed in Table 5.2. Fig. 5.1

shows the standard concentrations plotted against intensities. The slope

of the calibration line and the corrected XRD intensities were then used to

back calculate the analcime concentrations in the standards. This allow-

ed the calculation of the individual standard absolute errors, and hence

the standard relative errors. Finally, the mean relative error (0.53%)

was calculated. In order to check the above calibration technique, a new

pumice sample ( At = 53.5) was spiked to produce four standards (4.77, .

9.77, 14.77 and 19.77 weight per cent analcime) . These standards were

mounted only in duplicate and the resulting calibration data computed as

above. This produced a mean relative error of 0.52%, with a relative

error range of 0.10% to 0.92%, for the four standards.

The mean relative errors (0.53% and 0.52%) for these calibration

lines indicate that with adequate care the accuracy of this technique of

mineral analysis can be brought within the one per cent relative error band.

The advent of rapid and accurate whole rock major element analysis

by X-ray fluorescence spectrometry combined with flux-fusion sample pre-

paration (e,g. the Norrish method: Norrish and Hutton, 1969; Harvey et

al. , 1973; Parker and Willis, 1977; Parker, 19786) has greatly facilitated
 96

the acquisition of major element geochemical data. In petrological

studies involving major element analyses as well as the quantitative

determination of sample mineral concentrations , the application of cal-.

culated mass absorption corrections to quantitative XRD intensities

should be considered. This approach provides an accurate and rapid

alternative to direct measurement of the mass absorption correction,

or to the use of internal standard techniques.

 97

Table 5.1 Leucite and analcime analyses •- all data on moisture (H2O-)
free basis

(for analytical methods see Chapters 2 and 3)

Leucite Analcime
2503 4601 4701 AVE 2502• 2802 6506 AVE
8102 55.24 55,41 55.41 55.35 54.78 55.05 55.05 54.96
T102 .05 .05 .05 .05 .06 .05 .06 .06
A1203 22.49 22.28 22.55 22.44 21.85 21.90 22.11 21.95
Fe203* .41 .42 .45 .43 .50 .43 .48 .47
MgO .44 .28 .38 . .37 .25 .21 .18 .21
Ca O .05 .06 .06 .06 .25 .50 .48 .41
Na20 .28 .36 .31 .32 11.92 12.19 11.97 12.03
1(20 20.17 20.32 20.49 20.33 .97 .27 .71 .65
Rb20 .23 .15 .16 .18 .31 .20 .29 .27
LOI }} 1.60 1.02 .62 1.08 9.32 9.20 9.54 9.35

Total 100.96 100.32 .100.48. 100.61 100.21 100.00 100.87 100.36

Number of ions on Number of ions on
basis of 6 oxygens basis of 6 oxygens
Si 2.014 2.032
Ti .001 .002N
Al .963 .957
Fe .012 1.020 .013 1.028
Mg .020 .012
Ca .002 .016
Na .023 .862
K .944 0.971 .031 0.899
Rb .004 .006
LOI/H20 . .131 1.155
2.99 3.00
Si+Al +Fe
K+Na+2Ca+2Mg+Rb 1.02 •95

* Total iron.
*i Loss on ignition, 850-C for 30 minutes.
Table 5.2 Calibration data for XRD determination of analcime

MEAN CORt TED

LOAD- S.D. R.S.D. SPIKED TOTAL CALC. ABSOL. REL.
STD P - B P- B
7. A.C.F. ANALCIME ANALCIME ANALCIME ERROR ERROR (%)
INGS C/40S (Z)
C/40 S C/40 S
(%) (%) (%)

1 4 2,458 47.0 1.91 57,77 1.000 2,458 0.00 5.67 5.62 0.05 0.88

2 4 4,762 137.2 2.88 56.53 0.979 4,662 5.00 10.67 10.66 0.01 0.09

3 6 7,241 85.9 1.19 55.36 0.958 6,937 10.00 15.67 15.87 0.20 1.28

4 6 9,626 193.8 2.01 54.04 0.935 9,000 15.00 20.67 20.59 0.08 0.39

5 4 13,563 296.2 2.18 52.10 0.902 12,234 22.50 28.17 27.98 0.19 0.67

6 7 17,992 402.9 2.24 50.25 0.870 15,653 30.00 35.67 35.80 0.13 0.36

7 2 21,887 - . 48.63 0.842 18,429 36.50 42.0 42.15 0.02 0.05

Mean 2.07% 0.53%

S.D. Standard deviation

R.S.D.(Z) Relative standard deviation as a percentage of mean loading intensity
A.C.F. Absorption correction factor
P-B Peak minus background
C/40 Counts per 40 seconds.
 99

Fig. 5.1 Analcime XRD calibration line

 100

CHAPTER VI

GEOCHEMISTRY OF PUMICE AND LAVA SAMPLES

FROM VULSINI VOLCANO

6.1 Introduction

6.1.1. General

Vulsini volcano lies at the northern end of the Roman volcanic region

in central Italy. This region (Fig. 6.1) and the Campanian volcanic region

to the south are noted for the large amounts of ignimbrite and pyroclastic

material, in addition to lavas, that have been erupted. - Vulsini is Qua-

ternary in age and the volcano is dominated by two adjacent calderas.

Bolsena caldera is occupied by Lake Bolsena , while Latera is dry except

for a small lake (Fig. 6.2). The volcano has a low profile with slopes 1°

to 1.5° on the outer flanks and 2° to 8° near the caldera rims (Sparks,

1975).

Nappi (1969) and. Sparks (1975) have outlined three phases in the

evolution of Vulsini volcano. In the first phase lavas and some pyroclas-

tics were produced. In the second phase large volumes of ignimbrite and

pyroclastic material were erupted and it was in this phase that the calderas

of Bolsena and Latera were formed. In the third phase pyroclastic cones

were built up on the rims and flanks of both calderas. During this last

phase lavas were also erupted within, and on the flanks of the Latera cal-

dera (Schneider, 1965) . K - Ar dates on some of these lavas give ages of

277,000 and 220,000 years (Schneider, 1965).

 101

The lavas , ignimbrites and pyroclastics erupted by the Roman and

Campanian volcanic regions are characterised by high concentrations of

potassium and associated elements (P, Ba, Zr, Sr, Rb, Th, U, Y) .

Potassium-rich volcanics, although relatively restricted in occurrence,

have been found in various parts of the world (Carmichael et al. , 1974) .

There is considerable interest in the origin and evolution of these K-rich

rocks and recent work on the Roman and Campanian volcanics includes

that of Savelli (1967), Cundari and Le Maitre (1970); Appleton (1972),

Cundari and Mattias (1974) , Cundari (1975) , Turi and Taylor (1976) , Cox

et al. (1976) , Thompson (1977) and Vollmer (1976, 1977) . However, all

these studies have been concerned with lava samples, and the associated

pyroclastic materials have received little attention.

Sparks (1974, 1975) has made detailed studies on the stratigraphy

and geology of the ignimbrites and pyroclastics erupted by Vulsini volcano.

Six major and several smaller ignimbrites and pyroclastic falls have been

identified. Sparks (1975) considarsthat these eruptions constitute at least

90% by volume of the products of this volcano. As a result of this work

on the stratigraphic and lithologic characteristics of the Vulsini ignimbrites

and pyroclastics it became possible to undertake a systematic geochemical

sampling program on these rocks. In sampling the ignimbrite flows, pumice

clasts were selected rather than bulk samples of pumice and enclosing ignim-

brite matrix. In ignimbrite eruptions pumice clasts are considered to be

representative of the original eruption magma as opposed to the bulk pumice

plus matrix mixture (Lipman, 1967; Walker, 1972) . Pumice samples were

also collected from two pyroclastic fall deposits described by Sparks (1975) .

The ignimbrite pumice samples were collected from unwelded flow units.
 102

Previous geochemical work on Vulsini ignimbrite pumices (Trigila

et al. , 1971 and Locardi and Mittempergher, 1967) suffer from the lack of

detailed stratigraphic control of the sampled ignimbrite eruptions (such as

provided by the work of. Sparks, 1975) as well as including analyses of

pumice samples uncorrected for the effect of leucite to analcime alteration

(Trigila et al., 1971) . The analytical work presented by Sparks (1975) ,

while benefitting from the detailed stratigraphic work undertaken, does

not take account of leucite to analcime alteration in certain of the samples

and is furthermore restricted in the number of samples analysed for each.

ignimbrite eruption.

A number of lava samples from Vulsini were also collected during

the ignimbrite pumice sampling program. In comparison with the paucity

of geochemical studies on Vulsini ignimbrites , a considerable amount of

work has been done on the lavas from this volcano (Washington, 1906;

Schneider, .1965; Mattias , 1965; De Fino and Mattias , 1965; Trigila ,

1969a, 1969b; Nappi, 1969; Appleton, 1970, 1972; Vollmer, 1975).

The purpose of the geochemical work undertaken in the present study

was firstly to investigate the geochemistry of the ignimbrite eruptions

which Sparks (1975) considers to be the dominant eruptive product of Vulsini

volcano. The second objective was to investigate the geochemical relation-

ship between the ignimbrites and lavas , with the aim of providing a more

complete picture of the geochemical evolution of the volcano. The third

objective in undertaking the Vulsini work was to provide a set of analyses

for use with the computer programs discussed in Chapters 2, 3 and 4.

This important aspect of the work provided a realistic data "test bed" for


103

developing, testing and integrating the programs used in the various data

processing stages, i.e.: raw data-----analytical results-->grouping and

tabulation .data base creation --graph plotting and statistical evalu-

ation.

6.1.2 Sample collection

The analysed lava and pumice samples were collected from locations

shown in Fig. 6.2. In this study the following sample numbering conven-

tion was adopted: each sample was assigned a four digit number in which

the leading two digits represent the sample locality while the trailing two

digits identify the actual samples collected from each location.. Examples

are as follows:-

805 fifth sample from location 8

2902 second sample from location 29

This numbering method allows for up to 99 localities with up to 99 samples

for each locality. The "all digital" nature of the method simplifies sub-

sequent data processing via computer programs (Chapter 4) as compared to

sample numbering systems involving both numbers and letters. In the pre-

sent study samples were collected from 71 localities. In order to accom-

modate literature samples within the above numbering system, it was deci-

ded to give these literature analyses sample numbers in the range 8000 -

9000.

Sparks (1975) has labelled the six main ignimbrite eruption as A, 8,

C, D, E and F. He has also identified a number of smaller eruptions and

pyroclastic falls (e.g. b, c, PF-2 and PF-4). In the present study ignim-
 104

brites b and c have been relabelled X and Y to avoid confusion with B and C.

Ignimbrite E was a. mixed magma eruption (Sparks , 1975) with a population

of light grey pumice samples (Si02 "J58.6%) contrasting with a population

of black pumice samples (Si02 ",48.5%). In the present study these

samples have been labelled E-1 and E-2 respectively. Table 6.1 lists

the eruption sequence and minimum eruption volume (Sparks, 1975) for

the six main ignimbrites and pyroclastic fall PF-2. The eruption sequence

for X, Y and PF-4 with respect to the other eruptions is unknown (Sparks,

1975).

6.1.3 Alteration of pumice and lava samples

Pumice samples are susceptible to alteration effects such as: (1)

vapour phase and hydrothermal crystallization of secondary minerals dur-

ing cooling after eruption; (2) hydration and palagonitization of the pum-

ice glass; (3) kaolinization of the pumice glass and (4) in the case of

leucite bearing samples the alteration of this mineral to analcime. Hand

specimen and thin section examination of the collected pumice samples

show that vapour phase and hydrothermal crystallization effects (e.g.

Keith and Muffler, 1978) are absent or insignificant. Certain samples

in thin section show various degrees of yellow-brown palagonitization

(Kerr, 1959, p.424) of the pumice glass, while other samples are noted

for patchy alteration of the glass to clay minerals. These samples were

discarded. Leucite bearing pumice samples show alteration of this miner-

al to analcime. The alteration is usually complete although remnant frag-

ments of leucite are sometimes found in the centres of the altered pheno-

crysts. One pumice sample was found carrying a few fresh leucite pheno-
 105

crysts although the majority are altered. Certain of the analysed lava

samples are also affected by a small degree of leucite to analcime alter-

ation. The pumice and lava analyses affected by this leucite to analcime

alteration were corrected using the techniques described in Chapter 5 and

in Appendix A.

Determination of LOI/H20+ in pumice samples from ignimbrites A, B,

C, D, E-1, X, Y and pyroclastic falls 2 and 4 gave results (after correc-

tion for leucite -3analcime alteration) in the range 2 - 4% (Table A.3),

whereas the average value for lavas of similar composition such as groups

LV-QTR, LV-VTR and LV-TRP is 1.05% (Table D.2). Ross and Smith
(1955) consider that most water present in pumice samples is due to hyd-

ration of the pumice glass and is probably.post magmatic. Because of

this Lipman et al. (1966) and Lipman. (1967) have. recalculated pyroclastic

analyses on a volatile free basis. In the present study it was decided

to recalculate the compositions of the above pumice samples with respect

to a 1.05% H2O level in order to facilitate the comparison of these pum-

ices with the lava groups LV-QTR, LV-VTR and LV-TRP.

E-2 and F pumice samples gave LOI/H20+ values comparable to the

similar LV-TEL and LV-L lavas and the data for these pumices have there-

fore not been recalculated.

Chemical alteration accompanying devitrification and hydration of

volcanic glass may be important in obscuring the original chemistry of the

magma. Noble (1970) and Scott (1971) have documented detailed changes

in alkali contents of ignimbrite (ash-flow) samples consequent on hydra-

tion and devitrification. However, the heterogeneous nature of ignimbrite

 106

eruptions (Walker, 1972; Sparks et al. , 1973) indicate that studies based

on whole ignimbrite samples should be treated with caution (Walker, 1972,

p.144). The chemical variations reported by Noble (1970) and Scott (1971)

may therefore not be entirely due to post emplacement alteration. This

also applies to the ignimbrite samples analysed by Noble (1967) and

Lipman (1965) .

With regard to the above, the homogeneous glassy lavas analysed by

Noble (1967), Lipman (1965) and Lipman et al. (1969) should however provide

reliable data on hydration and devitrification of volcanic glass. - These data

indicate that hydration of salic glasses commonly involves Na2O loss, and a

common but not invariable K2O gain. The data presented by Lipman (1965)

show that: (1) for devitrified glasses from salic lavas there is also

Na2O loss and K2O gain and (2) that hydration and devitrification do not

significantly alter the remaining oxides, except that the Fe2O3/FeO ratio

usually increases in devitrified glasses. These effects will be discussed

further with regard to the Vulsini pumice geochemistry in section 6.3.3.2

below.

6.2 Petrography of the pumice and lava samples

6.2.1 Pumice samples

In this section the main features of pumice samples are summarised.

Table 6.2 lists the modal phenocryst proportions found in a number of pum-

ice samples from the six main ignimbrites. More detailed descriptions of

the samples are given in Appendix B.

The pumice samples fall into two main groups. The first group
 107

(from ignimbrites A, B, C, D, E-1, X, Y and pyroclastic falls PF-2 and

PF-4) vary in colour from cream to dark grey-black. They are dominated

by phenocrysts of sanidine with lesser amounts of clinopyroxene, plagio-

clase, biotite, opaque oxides and apatite. The phenocrysts may form

glomeroporphyritic clots. Leucite altered to analcime is present in pum-

ices from ignimbrites B and C (Table A.1). Accessory apatite micro-

phenocrysts are present in all samples while accessory microphenocrysts

of sphene are present in pumice samples from ignimbrites B, D and Y.

Occasional phenocrysts of melilite (?) (see Appendix B) are present in

an E-1 pumice sample. The pumice glass in this first group of ignim-

brites varies from colourless to brown. The glass is generally clear

although in patches it may incipiently devitrified. Microlites and

spherulites of alkali-feldspar are common in some samples.

The second group of pumice samples (from ignimbrites E-2 and F)

are black and scoriaceous and carry phenocrysts of plagioclase, clino-

pyroxene, leucite-)analcime, sanidine, biotite, opaque oxides and

apatite. Clinopyroxene is the dominant phenocryst phase. One sample

was found carrying a few phenocrysts of fresh leucite. Phenocrysts of

plagioclase, clinopyroxene and opaque oxides may form glomeroporphy-

ritic clots. the XRD data in Table A.1 show high concentrations of anal-

cime in pumice samples from this group. Occasional phenocrysts of

leucite-)analcime are observed in thin section and in hand specimen but

the bulk of the analcime is considered to be present as the alteration

product of devitrified groundmass leucite as well as the alteration of leu-

cite microphenocryst quench "stars" (cf. Appleton, 1970, p.16) present

in the groundmass of the samples. The groundmass glass is noted for

 108

its brown-black colour and turbid devitrified character. One of the pum-

ice samples (101) , however, is noted for a deep brown and predominantly

glassy groundmass. The glass in this sample is only devitrified in pat-

ches.

6,2.2. Lava samples

Vulsini lava analyses , both those taken from the literature and

those analysed in this work, have been divided into a number of groups

in section 6.3.2 below. These groupings (Table 6.5) are based on (i)

normative mineralogy; (ii) K2O/Na2O ratio; and (iii) Thornton and Tuttle

(1960) differentiation index (DI) . In the present section the main petro-

graphic characteristics (from the literature and this work) of these lava

groups are summarised. Table 6.3 lists the modal phenocryst proportions

and groundmass phases found in samples from the various lava groups.
0
More detailed descriptions of the individual lavas collected and analysed

by the author are given in Appendix C.

LV-QTR group (Q trachytes)

Samples in this group are composed of phenocrysts of sanidine,

plagioclase, clinopyroxene, biotite and opaque oxides set in a trachytic

groundmass. Sanidine is the dominant phenocryst and ranges from 3 to

10mm in length, the larger crystals being found in a coarse grained macro-

phenocrystic variety. The groundmass is composed of orientated alkali-

feldspar laths as well as lesser amounts of prismatic pyroxene and sparse

laths of plagioclase.

The phenocrysts may build glomeroporphyritic clots.

 109

Normative Q distinguishes LV-QTR samples from the ne and/or of

normative LV-VTR group of lavas (below) .

LV-VTR group (ne and/or ol trachytes (or vulsinites: Washington, 1906))

These lavas are composed of phenocrysts of sanidine, plagioclase,

clinopyroxene and biotite set in a groundmass dominated by orientated

alkali-feldspar laths. Prismatic pyroxene, laths of plagioclase, opaque

oxides and accessory apatite are also present in the groundmass. The

phenocrysts may form glomeroporphyritic clots.

Lavas in this group are ne and/or ol normative. They are free of

modal leucite and this distinguishes them from the ne normative and modal

leucite bearing trachytes and phonolites of the. LV-TRP group (below) .

LV-TRP. group (leucite trachytes and tephritic leucite phonolites)

These- lavas are composed of phenocrysts of clinopyroxene, leucite,

plagioclase and occasional opaque oxides set in a groundmass of the above

minerals plus alkali-feldspar and accessory apatite. Leucite commonly

forms distinctive macrophenocrysts of 5 to 20mm in diameter. Alkali-

feldspar is the dominant groundmass phase. Phenocrysts of sanidine and/

or biotite are present in some samples. The phenocrysts may form glomero-

porphyritic clots .

LV-TEL group (leucite tephrites and tephritic leucitites)

These lavas contain phenocrysts of clinopyroxene and leucite. Oli-

vine phenocrysts are found in low DI samples while phenocrysts of plagio-

clase, sanidine and sometimes biotite are found in the higher DI samples.

The groundmass is composed of clinopyroxene, leucite and plagioclase

with accessory amounts of apatite and opaque oxides. Irregular patches

 110

of alkali-feldspar and biotite may also be present. Olivine and amphibole

are reported in the groundmass of some samples (Washington, 1906) .

LV-L group (leucitites and tephritic leucitites)

These lavas are dominated by clinopyroxene and leucite, forming .

phenocrysts and the bulk of the groundma.ss . Biotite and olivine pheno-

crysts are present in some samples. The groundmass carries accessory

plagioclase, biotite, apatite, opaque oxides and alkali feldspar. Ground-

mass amphibole (this work) and melilite (Washington, 1906) are reported

in small amounts in certain samples.

With regard to the lavas examined by the author, LV-L samples are

noted for relatively high concentrations of groundmass clinopyroxene, and

low concentrations of groundmass plagioclase, as compared to samples

from the LV-TEL group. Furthermore, phenocrysts of this latter mineral are

absent from the LV-L group. Sanidine phenocrysts are also absent.

LV-BSN group (leucite basanite)

This rock type is characterised by olivine and clinopyroxene pheno-

crysts set in a groundmass of these minerals and leucite. Irregular areas

of biotite and poikilitic plagioclase are also present in the groundmass as

well as accessory apatite and opaque oxides (Washington, 1906).

LV-TRAB group (trachyandesite and trachybasalt)

These lavas are composed of phenocrysts of clinopyroxene, olivine

and plagioclase set in a groundmass dominated by laths of plagioclase

and granular clinopyroxene. The groundmass also carries disseminated

opaque oxides. Alkali-feldspar may be present, both as phenocrysts and

in the groundmass . The phenocrysts may form glomeroporphyritic clots.

 111

LV-MTR (miscellaneous Q trachyte)

This leucite free trachyte is similar in phenocryst and groundmass

characteristics to the LV-VTR group but has significant normative differen-

ces (Table 6.5) . Furthermore, in comparison with the LV-VTR group the

biotite laths of this rock are noted for corroded/resorbed margins of granu-

lar opaque oxides. LV-MTR sanidine phenocrysts are also less frequent

and are noted for rounded and apparently resorbed margins.

6.3 Major and trace element geochemistry

Fifteen lava and 60 pumice samples collected by the author from

Vulsini were analysed for their major element composition using the tech-

niques described in Chapter 2. These analyses were corrected for the

effects of leucite to analcime alteration present in certain of the samples

(Chapter 5 and Appendix A) . In addition to the above analyses , fifty-

four major element analyses of Vulsini lavas were extracted from the lite-

rature (see below) for inclusion in the present study. The samples collec-

ted by the author were also analysed for Sr, Ba, Rb, Zr, Y, Th, V, Cr,.

Co, Ni, Cu and Zn using techniques described in Chapter 3. In addition

Sr, Ba, Rb, Zr, Y, Th and Zn data for eight lavas (Appleton, 1970) and

Sr, Rb and Th data for six lavas (Vollmer, 1975) have been included.

The data processing techniques described in Chapter 4 were used to

calculate CIPW norms , to tabulate the analytical and normative data and

to plot the data on microfilm.

 112

6.3.1 Selection of lava analyses from the literature

Analytical data for Vulsini lava samples were selected from the

following studies:

Washington (1906)

Schneider (1965)

Mattias (1965)

Trigila (1969a)

Trigila (1969b)

Appleton (1970)

Vollmer (1975)

Certain of the analyses given by the above workers have been excluded

(Table 6.6) from the present study for the following reasons:

(1) Analyses were excluded if the samples were affected by leu-

cite to analcime alteration. It is apparent from the petrographic descrip-

tions given in the above studies that this alteration has affected a number

of the analyses. A good example of this is provided by Mattias (1965,

Table 5, analysis 6). This sample is described by Mattias as having

suffered leucite to analcime alteration. The geochemical effect of this

is seen in the relatively high Na2O (4.75%) and H20+ (3.54%) values ,

and low K2O (0.73%) value, reported for this tephritic leucitite sample.

The analysis of another tephritic leucitite reported in the same table

(Mattias, 1965, Table 5, analysis 9) gives values for Na2O = 1.02%,

K2O = 7.17% and H20+ = 1.32%. According to the petrographic descrip-

tion given by Mattias , this latter sample is unaffected by leucite--+anal-

cime alteration. In comparing the above data , the leucite and analcime
 113

analyses given in Table 5.1 (this work) should be borne in mind.

(2) Samples that originated as "blocks" in tuffs were excluded

because of the uncertainties attached to their previous volcanic history

(i.e. are they perhaps wall rock samples and/or have they reacted in any

way with the magma that produced the tuff?) .

(3) Samples collected from "foam or froth" lavas were. excluded

as these lavas are considered to be welded pyroclastic rocks which have

suffered intense fumarolic alteration (Sparks, 1975, p.505).

(4) Two of the above studies on lavas have included analyses of

a pumice and an ignimbrite sample: These analyses have been excluded

because of the susceptibility of these samples to the alteration effects

discussed in section 6.1.3 above. The ignimbrite sample (Vollmer,

1975, Table 2, VLS-4) is described as being "kaolinized" . Furthermore

the heterogeneous nature of ignimbrite deposits means that they are un-

likely to be good samples of the original magma (see above) . There is

no petrographic evidence given for the freshness or otherwise of the

pumice sample (Appleton, 1970, Table E2, No. 16) .

The relatively high H2O data for certain Vulsini lava analyses pre-

sented by Nappi (1969) suggests that leucite--,analcime alteration has

occurred. In view of this, and in view of the lack of petrographic des-

criptions for the individual samples analysed by Nappi, it was decided

to exclude all of these samples from the present study. The data for

Vulsini lava analyses presented by Locardi and Mittempergher (1967)

were also excluded because of high H2O values and lack of petrographic

data. Pumice analyses presented by these latter workers and by De Fino

 114

and Mattias (1965) , Trigila et al (1971) and Sparks (1975) have also been

excluded because of sample alteration, especially of pumice glass to clay

minerals and leucite to analcime. Samples from the volcanoes at San

Venanzo (Mittempergher, 1965) and Monti Calvo (Trigila , 1966) have been

excluded as they are some distance from Vulsini (i.e..> 20km).

6.3.2 Sample grouping

The lava analyses selected from the literature and from the analy-

tical work undertaken in the present study have been grouped according

to three criteria:

(1) CIPW normative mineralogy

(2) Thornton and Tuttle (1960) differentiation index (DI)

(3) K2O/Na2O ratio

~► The results of the lava grouping are given in Table 6.5.

Table 6.4 summarises the normative, DI and K2O/Na2O charac-

teristics for the pumice samples from the ignimbrite and pyroclastic erup-

tions. The rock type names (Table 6.4) given to the pumice eruptions

were assigned on the basis of their similarity (i.e. norms, DI and K2O/

Na2O) as compared to the lava groups (Table 6.5). In making the com-

parison between the pumices and lavas, the phenocryst mineralogy of the

groups of samples was also taken into account (Tables 6.2 and 6.3).

In the LV-L leucitites 14 out of the 15 samples have normative

di + lc > 50% (in 6 samples it is > 60%) . The remaining sample has a

di + lc value of 46%. K-rich rocks with greater than 50% normative di +

lc are characteristic of mafic leucitites from Alban Hills (Thompson,1977)

 115

and Roccamonfina (Appleton, 1972) . The high di + lc values for the

Vulsini LV-L leucitites are therefore in keeping with the classification of

leucitites from other Italian volcanoes.

Appleton (1970, 1972) has divided samples from a number of Italian

volcanoes (Laziale (Alban Hills) , Phlegrean Fields , Roccamonfina , Somma ,

Vesuvius, Vulsini, Vico) into High K and Low K lava series depending on

their K20 vs . SiO2 contents. In the present work only the LV-TRAB group

belongs to Appleton's Low K series. All of the remaining Vulsini lava and

pumice groups belong to the High K series. Tables 6.3 and 6.5 indicate.

various characteristics of the High and Low K Vulsini lava samples , i.e.:

(1) the K2O/Na2O ratio for Low K lavas is < 2 whereas for High K

it is > 2;

(2) Low K samples are relatively enriched in normative plagio-

clase (ab + an > or) as compared to High K samples;

(3) the mineralogy of the Low K series is noted for phenocrysts

of both plagioclase and olivine. Furthermore the Low K series is free of

leucite while in High K samples of similar DI leucite is present.

In addition to the groupings discussed above, the pumice and lava

samples may be divided into two large groups based on their DI values.

There is a DI gap which separates low DI samples (DI < 63) from high DI

samples (DI > 72) . This gap may be due to insufficient samples or it may

reflect an underlying petrogenetic process. The gap between the high

and low DI groups forms a distinctive feature of the binary plots shown in

Fig. 6.5A - 6.25A. This gap may also be seen in the ternary K2O - Na2O -

CaO diagram (Fig. 6.4A) .

 116

6.3.3 Geochemical variations

6.3.3.1 Introduction

The average geochemical data for the Vulsini pumice and lava groups

are presented in Tables 6,7 and 6.8. Individual sample analyses (this

work and those taken from the literature) are presented in Appendix D,

Tables D.1 and D.2. In this section a qualitative model based on the

major and trace element variations will be proposed for the evolution of the

volcano. Major and trace element analyses of phenocryst and groundmass

phases are outside the scope of the present study (see Chapter 1). Pheno-

cryst and groundmass data are required for the development of quantitative

models for the observed variations in terms Of mass balance relationships

(e.g. Baker et al., 1977) (This line of work is suggested for future research

in the next chapter).

In discussing the major and trace element variations the role of leu-

cite as a fractionating phenocryst phase will be considered. The specific

gravity of leucite is relatively low (2.47 - 2.50) compared to lavas such

as leucite trachytes , leucite tephrites and leucitites (2.61 - 2.78) (Table

6.9) and this suggests that leucite phenocrysts in these lavas may sepa-

rate by flotation rather than by the more common process of crystal-sinking.

On the grounds of detailed petrographic, specific gravity and geometric-

structural evidence Buie (1941) considers that leucite flotation resulted in

strong enrichment of this mineral (later converted to pseudoleucite) in the

canopy of a bulbous shonkinite dyke in Montana. Sahama (1960) has

suggested that leucite (s.g. = 2.48) floated in certain central African lavas
(groundmass s.g. = 3.03) and was strongly concentrated in the top part of
 117

the magma.

In view of the above the author considers that leucite phenocrysts

in K-rich lavas (such as leucite trachytes , leucite tephrites and leuci-

tites) are more likely to fractionate by flotation than by sinking. Appleton

(1970) has come to the same conclusion.

With regard to the depth of the Vulsini magma chamber Thompson

(1977) considers that mafic potassic lavas such as leucitites and leu-

cite tephrites from Roccamonfina , Vesuvius and Vulsini volcanoes have

evolved their "pre-eruptive chemical composition at shallow levels in

the crust". Thompson (1977) has arrived at this conclusion on the

grounds of experimental melting studies and comparison of bulk compo-

sitions with that of the 1 - atmosphere diopside - leucite eutectic.

However in the case of the Vulsini samples the depth at which the magma

evolved must have been sufficient to ensure the stability of the biotite

phenocrysts found in the lava and pumice samples. In the pumice sam-

ples the biotite phenocrysts are fresh and unaltered, while in the lavas

the phenocrysts usually display resorbed margins suggesting disequilib-

rium. This latter effect may be due to relatively slow uprising and

extrusion of the lava magma thereby allowing time for biotite disequilib-

rium to develop in the low pressure near surface environment. Appleton

(1972, p.439) reports that rare, partially or almost completely resorbed

biotite phenocrysts are present in some of the Roccamonfina lavas. He

notes that "biotite would be rapidly resorbed with falling PH 2 O at the

relatively high temperatures of crystallization of potassic undersaturated

lavas (Luth, 1967)". The fresh euhedral state of the biotite found in the
 118

Vulsini pumice samples would be compatible with the nature of ignimbrite

eruptions , i.e. very rapid and explosive uprise of the magma from depth

followed by quenching of the pumice glass on exposure to air during

emplacement of the ignimbrite flow.

In the present work the Thornton and Tuttle (1960) differentiation

index (DI = normative Q + ab + or + ne + lc + kp) has been selected as

an index of magmatic evolution for plotting the oxide and trace element

data. Carmichael et al. (1974; p.48) note that this index has "a satis-

fying thermodynamic basis in that these normative components corres-

pond to minerals with low entropies of melting which demands

their concentration in low-temperature liquids".

The. Vulsini data for each oxide or element have been plotted on

binary diagrams (Fig. 6.5 - 6.44) as follows:

(a) All samples plotted (Fig. 6.5A - 6.25A) . In these

diagrams a single pumice plotting symbol has been used in
order to clarify the overall relationship of the pumice samples
to the lava groups. These latter groups have been plotted
with a unique symbol for each group.

(b) Only high DI (> 70) samples plotted (Fig 6.5B - 6.25B) .
In these diagrams the individual pumice (and lava) groups have
been plotted with unique symbols. These diagrams show the
relationships between the different high DI pumice groups , as
well as the relationship of the high DI lava groups to these
pumice groups. It should be noted that the expansion of the
DI scale in these diagrams shows more clearly the variations
in the high DI samples (e.g. Sr and K2O) as compared to the
combined high and law DI plots of all of the samples ((a) above) .
 119

(c) Selected high DI pumice groups plotted (Fig. 6.26 -

6.44) . These diagrams clarify the variations in pumice
samples from ignimbrites A, B and C as well as pyroclastic
fall 2. This latter pyroclastic fall underlies ignimbrite A
and is considered to represent the initial stage of the
ignimbrite A eruption (Sparks , 1975) .

Pumice samples from ignimbrites E-2 and F are noted for their

relatively low DI values (Table 6.8A: DI range = 46-54) as compared to

all the other pumice groups (DI values all > 70) . The chemical and petro-

graphic characteristics of E-2 and F pumices are very similar (Tables 6.2,

6.4 , 6.7 , D.1 and Appendix B) and they are thought to belong to the same

or closely related magma(s). There are no other pumice samples with DI

values < 70 and for this reason the E-2/F pumices form a clearly recog-

nisable group in the diagrams discussed in (a) above.

The Vulsini data have also been plotted on ternary diagrams (Figs.

6.3 and 6.4).. Group symbol identification as used in the binary plots

(see (a) and (b) above) have been applied to these ternary diagrams to

clarify the relationships between the various groups.

6.3.3.2 Major element variations

In the K2O + Na2O - Fe2O3T - MgO ternary diagram (Fig. 6.3A)

the trend of the samples shows some iron enrichment followed by strong

alkali enrichment. Fig. 6.3B and C show in more detail the alkali enrich-

ment trends for the individual samples from the high DI ignimbrites, pyro-

clastics and lavas. In the K2O - Na2O - CaO ternary diagram (Fig. 6.4A)

the overall trend towards the K2O apex reflects the strong enrichment of
 120

K20 over Na20 present in all of the samples. However Fig. 6.4B and C

show that within the high DI samples there is a change to Na20 enrich-

ment for certain groups of pumice samples. This reversal in the alkali

trend will be discussed in more detail below.

In the binary diagrams (Figs. 6.5 - 6.13) the High K series of sam-

ples show Si02, A1203, K20 and Na20 enrichment with increasing DI,

while Fe2O3T, . CaO, Mg0 and TiO2 are depleted. P205 increases and

then decreases. These major element trends are considered to indicate

the operation of a differentiation process based on low pressure crystal

fractionation, The rapid initial depletion of MgO suggests extensive

removal of olivine during the transition from leucite basanite to leucitite.

Further evolution of the magma to tephritic leucitites and leucite tephrites

is considered to be the result of separation of diopsidic clinopyroxene and

calcic plagioclase, as well as the separation of olivine in the more basic

samples. This would be compatible with the observed depletion trends

for CaO and MgO.

The Fe203T concentrations are initially rather scattered and show

only slight depletion until DI values of 60 are reached whereupon the

iron decreases rapidly, This suggests that removal of iron rich opaque

oxide phenocrysts (e.g. titanomagnetite, see below) becomes important

at this stage in the evolution of the magma.

Separation of relatively silica poor minerals such as olivine and

calcic plagioclase would be important in increasing the residual liquids

in Si02, thus producing the observed Si02 enrichment trend. Separation

of diopsidic clinopyroxene would tend to deplete the low DI residual

* total iron
 121

liquids in SiO2, but in these Vulsini samples this effect is apparently

outweighed by the enrichment resulting from the separation of olivine,

biotite, calcic plagioclase and opaque oxides such as titanomagnetite

(see below) .

Al2O3 would be depleted by removal of plagioclase whereas remo-

val of clinopyroxene and olivine would lead to the observed enrichment

trend. The overall trend of Na 20 and K2O enrichment (Figs. 6.7A and

6.8A) indicate separation of phases poor in alkalies such as olivine,

diopsidic clinopyroxene and calcic plagioclase. In the Na2O plot (Fig.

6.8A) certain of the low DI E-2/F pumice samples fall below the main lava

trend. These relatively low Na2O values may reflect the loss of Na from the

devitrified glass found in these pumices (see section 6.1.3 above) . The

high DI pumices are generally glassy thus indicating that Na loss due to

devitrification should not have seriously affected these samples.

The K2O values in the high DI (DI > 80) pumice samples indicate a

cross trend towards a relative decrease in this element (Fig. 6.-7B), where-

as Na2O shows an enrichment trend (Fig. 6.8B) . In section 6.1.3 it was

noted that hydration of pumice glass may lead to Na2O loss and K2O gain,

i.e. the opposite effect of the trends observed in these pumices. This

suggests that alkali movement due to hydration has not been a dominating

influence on these samples. If these alkali variations are not the result

of hydration effects, then it is considered that in these high DI (DI > 80)

pumices the observed trends may reflect the separation of K-feldspar

(sanidine) leading to depletion in K2O and residual enrichment in Na2O

(i.e. the orthoclase effect, Bailey and Schairer, 1964). This relationship
 122

can also be seen in the ternary Na20 - K20 - CaO diagram (Fig. 6.4) .

This interpretation is supported by the ubiquitous presence of sanidine

phenocrysts in the high DI pumice samples , suggesting that this mineral

was an important fractionating phase in these samples. Le Maitre (1962)

has noted a similar change in the direction of evolution of lavas from St.

Helena. These lavas (alkali-basalt --+trachyte series) when plotted in

a Na - K - Ca diagram also show an abrupt change towards Na enrichment

in the evolved trachytes (Le Maitre, 1962, p.1328 and Fig.6). The pre-

sence of abundant alkali-feldspar phenocrysts in these St. Helena late

state differentiates suggests that their Na enrichment may also be related

to the orthoclase effect of Bailey and Schairer (1964).

With regard to leucite fractionation, the role that these phenocrysts

will play in the evolution of the magma depends on whether it sinks, re-

mains in situ or floats upwards (see discussions above). If these pheno-

crysts were to sink then.K20 (and A1203, Si02) in the more evolved liquids

would be depleted. However the observed trends show enrichment of these

elements and this would be compatible with either an in situ behaviour or

the upward flotation of this K20 rich phase. When considering the crystal-

lization of sanidine in the high DI leucite trachytes and leucite phonolites,

the role of macrophenocrystic leucite may be of some importance. If flota-

tion of these phenocrysts occurs it is possible that they will reach elevated

and more salic levels in the magma which may subsequently become satu-

rated with respect to silica through continued gravity settling of mafic

phases. At this point the leucite phenocrysts will undergo reaction with

the magma to produce sanidine, according to the equation:

 123

5KA1Si2O6 + 3KA1Si308 + 5SiO2 —4- 8KAlSi308

leucite melt sanidine

The sanidine will form approximately twice as fast as the leucite goes

into solution (Turner and Verhoogen, 1960,

. p.107). This reaction not

only breaks down the leucite but also extracts Si (and K, Al) from the

melt. Assuming that the resultant sanidine separates by gravity sinking,

then the effect of this reaction is to buffer the residual magma against

further evolution towards salic derivatives. This may explain the pau-

city of samples with greater than 60% Si02. Only after all the leucite

has been converted to sanidine will the residual liquids be free to evolve

to higher Si02 levels by fractionation of mafic phases such as clinopyro-

xene, biotite, sphene, apatite and opaque oxides. Extensive production

of sanidine from leucite followed by gravity separation may also be related

to the observed depletion of K20 and residual enrichment of Na20 in the

high DI pumices as discussed above.

The P205 trend shows enrichment (low DI) and then depletion (high

DI) . This indicates initially a residual behaviour for this element until

fractionation of apatite phenocrysts becomes important in the more evolved

(high DI) stages of the magma.

The TiO2 trend, although showing scatter, is clearly towards deple-

tion in the high DI samples , suggesting that separation of a Ti-bearing

phase becomes important in these latter samples. Furthermore this TiO2

trend when combined with the distinct depletion of iron noticed in these

high DI samples would be compatible with fractionation of an opaque oxide

such as titanomagnetite. Fractionation of sphene, a phenocryst found in

 124

ignimbrite B, D and Y pumice samples , would also deplete residual liquids

Biotite phenocrysts are present in minor and variable amounts in

both the low and high DI samples. Separation of biotite would result in

the enrichment of SiO2 and depletion of Fe2O3T and TiO2 in the residual

liquids.

With regard to the Low K samples (LV-TRAB) the plots show that

these trachybasalts and trachyandesites form trends parallel to or within

the trends of the. High K - low DI samples. This suggests that the LV-

TRAB magma may have undergone similar fractionation processes as those

described for the low DI samples (but with the exception of leucite fractio-

nation, this mineral being absent in the LV-TRAB 'samples). In the SiO2

and K2O plots it will be seen that three of the LV-QTR lavas fall on the

extension of the LV-TRAB trend suggesting a possible connection between

these felspathoid free rocks. This relationship is not however apparent

in the plots for the other elements.

6..3.3.3 Trace element variations

The Sr, Ba and Rb variations will be discussed in terms of fraction-

ation of the observed phenocryst mineralogy and the distribution coef-

ficients (D = concentration of element in crystal/concentration of element

in liquid) taken from the literature for these phases (Table 6.10). For a

phenocryst. to deplete a trace element in the residual liquid, D must be

> 1.0. D values of < 1.0 will result in enrichment in the residual liquid.
 125

Distribution coefficients are sensitive to magma temperature,

pressure, oxygen fugacity and composition (Irving, 1978). The D data

listed in Table 6.10 can therefore only be regarded as a guide for the

Vulsini samples. Ideally interpretation of the trace element trends

should be based on the D values determined for the rocks in question.

However, as determination of such values for the Vulsini samples was

outside the scope of the present study, the data listed in Table 6.10

will be used. The fractionation model proposed in this sub-section is

therefore of a provisional nature and further work will be required to esta-

blish the Vulsini D values before a full interpretation can be made (see

recommendations for future research in the next chapter) . It may be noted

that Baker et al. (1977) have also used trace element distribution coef-

ficients taken from the literature in their work on fractionation models of

a basalt - benmoreite - trachyte suite.

The trace element plots (Figs. 6.14 - 6.16) will be considered first

with regard to the High K series of pumice and lava samples . The enrich-

ment trends for Sr and Ba shown by the low DI - High K samples indicate

that these elements were being concentrated in the residual liquids as

phenocryst fractionation proceeded. In these low DI rocks the dominant

phenocrysts are clinopyroxene and leucite and these phases have D < 1.0

for both Sr and Ba. Fractionation of these phases will therefore result in

Sr and Ba enrichment in the residual liquid. Olivine phenocrysts are also

present in the low DI samples and removal of this phase would also result

in enrichment of these elements. Plagioclase fractionation would tend to

enrich the residual liquid in Ba but deplete the liquid in Sr, whereas the D

data for biotite indicates that fractionation of this phase would deplete the
 126

residual liquid in Ba and enrich the liquid in Sr. The effect for Sr of

fractionation of these latter two phases would therefore tend to cancel out.

The high Ba D value reported for biotite indicates that removal of this latter

mineral would tend to outweigh the effect of plagioclase separation with

the result that Ba would be depleted. However the overall trends in these

low DI samples of increasing Sr and Ba suggest fractionation dominated •

by clinopyroxene, leucite and olivine phenocrysts and this is compatible

with the observed mineralogy, plagioclase and biotite being minor or

absent constituents in these samples.

In the Ba plot the low DI pumice samples fall below the lava samples

although still exhibiting an enrichment trend. These lower values may

possibly be due to Ba loss associated with devitrification of the pumice

glass (see 6.1.3 above) found in these ignimbrite E-2 and F samples.

The trends for Sr and Ba show. a marked change over from enrichment

in the low DI samples to rapid depletion in the high DI samples. In these

latter samples there is an order of magnitude decrease in Sr and Ba for a

relatively small increase in DI. The D values for Sr and Ba in sanidine

(Table 6.10) indicate partition of these elements into this mineral. In the

high DI samples sanidine is the dominant phenocryst in the more evolved

lavas and pumices and separation of sanidine would therefore lead to a

depletion trend in the residual liquid for Sr and Ba. Subsidiary amounts of

plagioclase and biotite phenocrysts are also present and fractionation of

these phases would tend to cancel out for Sr but result in Ba depletion as

noted above. Clinopyroxene and leucite fractionation would enrich the

residual liquid in Sr and Ba but as these phases are subsidiary and/or

 127

absent in the more evolved high DI samples , the overall trends are likely

to be controlled by sanidine (alkali feldspar) fractionation.

It is interesting to compare the trends of Sr and Ba depletion found

in the Vulsini high DI samples with data for alkaline lava suites from (i)

the mid-oceanic islands (P. Baker et al. , 1964; I. Baker, .1969; Zielinski,

1975; Le Maitre , 1962); (ii) the East African rift system (B. Baker et al. ,

1977; Barberi et al. , 1975) and (iii) Roccamonfina volcano, Italy (Appleton,

1970, 1972) . The data for the mid-oceanic islands and East African rift

system all show various degrees of Sr and Ba enrichment in the transition

basic—intermediate, followed by distinct depletion trends for these two

elements in the most evolved and differentiated rock types (trachytes ,

phonolites , pantellerites) . A characteristic .of these latter rocks is the

appearance of alkali-feldspar as an important phenocryst phase. P. Baker

et al. (1964, p.527), I. Baker (1969, p.1297) , B. Baker et al. (1977,

1).322) and Barberi et al. (1975, p.39) all link the observed Sr and Ba

depletions to the onset of alkali feldspar fractionation. Appleton (1972,

p.444) has linked Sr and Ba depletion, observed in certain Roccamonfina

late stage phonolites, to both alkali feldspar and plagioclase fractionation.

The modal data given by Appleton (1970, Table B-1) show that these phono-

lites carry phenocrysts of alkali feldspar, but negligible plagioclase pheno-

crysts. This suggests that the observed depletions are dominated by

fractionation of alkali feldspar.

The Rb plot (Fig. 6.17) shows considerable scatter for the low DI

pumice and lava samples and there is no overall trend. The scatter in the

pumice samples is considered to be due to variations in Rb concentration

 128

of analcime (alteration product of leucite) which is present in these pum-

ices from ignimbrites E-2 and F (Tables 5.1 and A.1) . Table 5.1 indi-

cates that Rb is . enriched in analcime relative to leucite and this may

reflect the relatively high Rb levels found in most of the pumice samples

as compared to the lava samples. However one of the pumice samples

has the lowest Rb level recorded for any of the analysed samples -and it

is apparent that alteration of leucite to analcime may not necessarily

result in enrichment of this element. The lava data is also scattered

and the reason for this is not clear. It may be noted that Rb data for
•Roccamonfina High K lavas (Appleton, 1972) are also relatively scattered.

This suggests that there may be some similarity in the processes control-

ling the Rb levels found in these low DI Vulsini and Roccamonfina lavas .

In the Vulsini High K -. high DI samples there is a trend of Rb enrich-

ment with.increasing DI (Fig. 6.16B) . While the Rb D values for biotite

are ambiguous, the data in Table 6.10 indicates that separation of sani-

dine, plagioclase and clinopyroxene will lead to enrichment of Rb in the

residual liquids. The dominance of sanidine phenocrysts in the high DI

samples suggests that the Rb enrichment trend is largely controlled by

sanidine fractionation as well as being aided by fractionation of plagio-

clase and clinopyroxene.

In volcanic suites from mid-oceanic islands and the East African

rift system, the data presented by P. Baker et al. (1964); I. Baker, 1969;

B. Baker et al. (1977) and Barbed et al. (1975) all show enrichment of Rb
in the late stage differentiates. This suggests that, as in the Vulsini

samples , Rb is behaving as a residual element. The data for Roccamonfina

 129

(Appleton, 1972) also show Rb enrichment in late stage phonolites.

In the Vulsini high DI samples the proposed model of dominant sani-

dine fractionation is therefore compatible with the trends shown by all

three of the above trace elements (Sr, Ba and Rb).. Furthermore it has

been suggested in the discussions on the major element variations (see

above) that in the high DI samples leucite may react to produce sanidine

(at approximately twice the rate at which leucite goes into solution) .

Extensive production of sanidine via this mechanism followed by fraction-

ation would therefore be consistent with the relatively steep trends obser-

ved for Sr, Ba and Rb.

The Zr, Y and Th diagrams (Figs. 6.17 - 6.19) all show enrichment

with increasing DI. This enrichment is in keeping with their "incompa-

tible" nature, i.e. exclusion from normal silicate minerals with the result

that they are concentrated in residual liquids (Ringwood, 1966 and Harris ,

1967). Fractionation of sphene, present in certain of the pumice samples

(ignimbrites B, D and Y) , will however modify the overall trends of enrich-

ment as this mineral will accept Y, Th and to a lesser extent Zr (Deer et

al. , 1966) .

Cr, Ni, V, Co and Cu all show depletion with increasing DI whereas

Zn shows enrichment in the low DI samples but no apparent trend in the

high DI samples (Figs. 6.20 - 6.25) . Fractionation of transition metal

ions is best considered in terms of Crystal Field Theory (Burns , 1970;

Carmichael et al., 1974) . This theory derives , from data on spinels

an Octahedral Site Preference Energy (OSPE) for predicting the take up of

the transition trace elements in phases such as pyroxene and olivine.

 130

For divalent transition metal ions the take up OSPE is as follows:

Ni> Cu> Co> Fe> Mn > Ca, Zn

while for trivalent:

Cr > Mn > V > Ti > Fe > Sc

The rapid depletion of Cr and Ni (Figs. 6.20 and 6.21) would be compati-

ble with their high OSPE values resulting in partitioning into early formed

clinopyroxene and olivine. Cu shows a greater rate of depletion compared

with Co and this would also be compatible with their relative OSPE values

The rate .'of depletion of V is lower than that of Cr and this is compatible

with their relative OSPE values. Ewart-et al. (1973) present data indi-

cating a high .distribution coefficient (D = 24) for V in titanomagnetite in

andesites and the V trend may possibly be controlled by fractionation of

this mineral. .V partitioning is however complicated by the sensitivity

of this element to magma oxygen fugacity (Irving, 1978) . Zn has a zero

OSPE and the low DI trend of enrichment suggests that it is behaving as

a residual element with little uptake by the crystallizing phases. It is

not clear what is the cause of the scattered Zn values ( and to a lesser

extent Cu) in the high DI samples. It may be noted that the Zn data

reported by Appleton (1972) are scattered for both basic (low DI) and more

evolved (high DI) K-rich lavas from the Roccamonfina volcano.

With regard to the trace element trends shown by the Low K lavas ,

there are insufficient data for reliable patterns to be identified. The

only exception to this is perhaps the Sr trend which indicates decreasing

Sr with increasing DI. Sr data is available for only four Low K lavas and
 131

further analyses are required to confirm this trend. Apart from the above

the Low K Cr and Ni data (for the single analysed sample) are relatively

high as compared to the High K samples of similar DI.

6.3.3.4 Major and trace element variations between and within

ignimbrite magmas

Table 6.1 lists the eruption sequence for the six main ignimbrites

and pyroclastic fall 2 (Sparks , 1975) . The average Si02 and DI values

listed in Table 6.1, and the average data listed in Table 6.7, indicate

a cyclic evolution of the ignimbrite compositions. Eruptions PF-2, A,

B and C are considered to reflect the first cycle of activity with the least

evolved magmas (ignimbrites B and C) being erupted last. The eruption

of the relatively evolved ignimbrite D magma indicates the onset of a

second cycle of activity (i.e. ignimbrites D, E and F). The D and E-1

pumices are more evolved (salic) than the E-2 pumices which are con-

sidered to be identical to the basic pumices found in ignimbrite F. The

presence of the E-1 and E-2 pumices in ignimbrite E indicate that this

eruption was a mixed magma eruption. This. may have resulted from the

salic E eruption entraining a portion of the basic ignimbrite F magma.

This latter ignimbrite was the last major eruption in the second cycle.

With regard to the E-1/E-2 eruption, mixed magma eruptions have been

described from various localities in other parts of the world (Wolf et al. ,

in press) .

Figs. 6.26 - 6.44 show the major and trace element variations in

pumice samples from pyroclastic fall PF-2 and ignimbrites A, B and C.

The DI varies from 76 to 86 in these samples and with increasing DI there

 132

are trends of increasing. SiO2 , Na2O and decreasing Fe2O3T (total iron) ,

MgO, CaO, TiO2 and P2O5. V, Ba and Sr decrease whereas Rb, Zr and

Y increase. The K2O plot (Fig. 6.28) shows an initial trend of enrich-

ment (up to DI 80) followed by a depletion trend (cf. Fig. 6.7B and

discussions in section 6.3.3.2 above) .

The enrichments and depletions with increasing DI shown in Figs.

6.26 - 6.44 suggest that differentiation processes similar to those dis-

cussed in sections 6.3.3.2 and 6.3.3.3 operated on the scale of the

individual ignimbrite magma. The variable compositions of ignimbrite B

and C pumice samples would be compatible with a zoned magma column

for each ignimbrite with the higher layers containing more evolved, and

the lower layers more basic, compositions. The very rapid and chaotic

nature of ignimbrite eruptions would cause pumice samples from different

parts of the original magma column to be mixed together so that the final

ignimbrite deposit would contain a range of pumice compositions at any

one locality (e.g. sample localities 14 and 11 : compare pumices 1401

and 1104 vs 1405 and 1102 in ignimbrite C) . Zoned ignimbrite magmas

from other volcanoes have been described by Smith and Bailey (1966) ,

Lipman et al. (1966) and Lipman (1967) .

6.3,3.5 Discussion

Previous studies on Vulsini volcano have involved the interpreta-

tion of data for limited numbers of samples and few of these studies

include trace element data. Furthermore the effect of leucite-->anal-

cime alteration has usually not been considered in interpreting the data.
 133

The present study has sought to bring together reliable data from previous

work and to combine it with new major and trace element data, particularly

with respect to both lava and pyroclastic samples. These latter samples

are volumetrically important in terms of the erupted products of Vulsini

volcano as well as in the eruptions of the other volcanoes in the Roman

and Campanian regions. Pyroclastic samples from these regions have

been generally ignored and the present work represents the first in depth

study of such samples.

The regularity of the trends for the High K sample series , both in,

the binary and in the ternary plots, and the serial nature of the sample

groupings within the trends , suggest that these volcanic rocks are related

by a systematic differentiation process that has led to the evolution of a

High K sequence of magmas corresponding to:

leucite basanite-+leucitite/tephritic leucitite/leucite tephrite—+

leucite phonolite/leucite trachyte -+trachyte .

The Low K trachybasalts and trachyandesites form separate trends

in the plots and this suggests that these lavas are not co-magmatic with

the main sequence referred to above.

The Vulsini major and trace element trends are considered to reflect

low pressure crystal fractionation processes operating on a parental K-

rich mafic basalt. The work of Appleton (1972) on the K-rich lavas from

Roccamonfina volcano resulted in similar major and trace element varia-

tions as those found in the Vulsini samples. Appleton also considers

that crystal fractionation controlled the evolution of the Roccamonfina

lavas to more salic types. Cundari and Le Maitre (1970) used a statistical
 134

Principal Latent Vector Variation method for evaluating the major element

trends in the Vulsini lava data then available, and these latter workers

showed that the :major element data were consistent with an evolutionary.

trend linking leucite basanite-3leucitite—leucite tephrite—)leucite

phonolite and trachyte. This conclusion is identical to that reached in

the 'present study based on both major and trace element data from new

sample analyses as well as from analyses taken from the literature.

With regard to the origin of the K-rich Italian lavas the limestone

assimilation hypothesis (Rittmann, 1933; Schneider, 1965; Marinelli

and Mittempergher, 1966; Burri, 1968) proposes that trachytic magma

assimilates dolomitic limestone followed by evolution to phonolitic leu-

cite tephrite and finally to tephritic leucitite. The present work, and the

work of Cundari and Le Maitre (1970) and Appleton (1970, 1972) indicates

that for these Italian rocks the evolutionary direction is from leucitite to

trachyte, i.e. the reverse of the assimilation hypothesis. Recent work

on element abundances (Savelli, 1967, 1968) and isotope ratios (Barbieri.

et al. , 1975; Turi and Taylor, 1976; Vollmer, 1976, 1977) also show that

significant limestone assimilation by the parental magma is improbable.

Pyroxenite assemblages (pyroxene ± biotite ± olivine) are found as

ejecta in the volcanic products of various K-rich Italian volcanoes and

Cundari and Le Maitre (1970) suggest a parental role for these pyroxenites.

They have shown that these rocks plot on the basic extensions of the

Vulsini and Somma - Vesuvius lava trends. Cundari and Le Maitre (1970)

suggest that subtraction of variants of the ejecta assemblages would

result in the observed evolutionary trend towards leucitites and leucite

 135

tephrites. These workers note that the Vulsini leucite basanite approaches

the chemical composition of the assumed parental assemblage. This would

be consistent with the Vulsini data (this work) which show that the leucite

basanite lava is the least evolved rock type present in the major element

trends.

In considering the origin of the parental magma there are two impor-

tant features of the Italian K-rich lavas which must be taken into account:

(1) the isotope enrichment in 87Sr and 180 relative to expected levels

for basalts derived from the mantle (Barbieri et al. , 1975; Vollmer, 1976;

Turi and Taylor, 1976); and (2) the enrichment of K and residual trace

elements (Savelli, 1967; Appleton, 1972; this work) relative to mantle

derived oceanic basalts. " The isotope enrichment has been interpreted

as reflecting contamination of the parental magma by a crustal component

rich in 87Sr and 180 (Barbieri et al. , 1975; Turi and Taylor, 1976) .

Vollmer (1976, 1967) has studied the Pb and Sr isotope distributions of

these K-rich lavas and considers that his data strongly suggests mixing

between crustal and mantle derived components.

The above interpretations of the Sr, 0 and Pb isotope data have

implied or invoked crustal contamination within the crust as the source

of the isotope enrichments (Barbieri et al., 1975; Turi and Taylor, 1976;

Vollmer, 1976). However, Thompson (1977) has suggested an alternative

based on lithospheric subduction. In this hypothesis rotation of the

Corsica - Sardinia microplate (Alvarez, 1972) caused eastward subduction

of the floor of the Tyrrhenian Sea beneath Italy. Thompson suggests that

if the sea-floor was overlain by thick argillaceous sediments derived from

 136

ancient sialic rocks , then the subducted material would be rich in both

180 and 87Sr and that in the deeper parts of the subducted zone partial

fusion would result in a liquid rich in K, Ba , Rb, 180 and 87Sr. This

melt would then migrate upwards and trigger further partial fusion and

magma generation in the overlying mantle wedge before finally reaching

the surface.

Ninkovitch and Hays (1972) have also proposed that subduction

is possibly responsible for the Ischia - Phlegrean fields - Somma -

Vesuvius volcanism. However these workers appeal to a subduction

zone descending in a north westerly direction under Sicily and Calabria

(Fig 6.1) , and relate this subduction zone to Europe/Africa plate-tectonic

motions . A similar north-west trending subduction zone has also been

postulated in this region by Barberi et al. (1973) . According to

Ninkovitch and Hays (1972) enrichment of K and associated elements is

related to the depth of the postulated subduction zone (earthquake foci =

300 - 400km) beneath the Ischia - Vesuvius volcanoes. They suggest

that these elements are "scavenged" by partial melts working their way

up through the overlying mantle wedge, and that the greater the depth to

the subduction zone the greater will be the amount of scavenging. In

support of this hypothesis they cite the relationship of increasing K2O

with increasing depth to subduction zone (earthquake foci) found for

Indo-Pacific volcanic arcs (Hatherton and Dickinson, 1969; Dickinson,

1975) . Ninkovitch and Hays (1972) do not make clear whether their

hypothesis may be extended to include the K-rich volcanoes (Alban Hills ,

Sabatini, Vulsini, etc.) found to the north of Ischia - Vesuvius.

 137

Barbera et al. (1973, p.5225) do not however favour the extension

of the Calabrian - Sicilian subduction zone as far north as Ischia -

Vesuvius and contend that extensive alkaline undersaturated volcanism,

as found along the west coast of Italy, is normally associated with rifted

continental structures (e.g. Rhine graben, East African rift system, Oslo

graben) and that in western Italy this volcanism also occurs "in con-

tinental collapsed areas".

It is evident that a better understanding of the tectonic complexities

of the macro- and micro-plate boundaries, relative plate motions, pos-

sible subduction zones and/or rift systems in this region of the Mediter-

ranean will greatly assist in developing (and eliminating) theories for the

origin of the K-rich Roman and Campanian magmas.

 138

Table 6.1 Vulsini volcano - ignimbrite eruption sequence

average average minimum eruption

ignim. SiO2 D.I. vol Km3 sequence

F. 49 52 . 0.3 7
E-1/E-2 59/49 80/51 0.8 6
D 60 85 1.2 5
C 58 81. 2.2 4
B 58 82 0.7 3

A 62 86 0.9 2

**PF2 61 86 - 1

*
Sparks (1975)
** PF = pyroclastic fall
 139

Table 6.2 Ignimbrite pumice phenocryst concentrations

volume % (tr = < 0.5%)*

ign spl no san+ pl . cpx bio op sph

3707 3 tr tr tr tr tr
3801 8 tr tr tr tr tr

3901 14 1 2 1 tr tr
5502 15 3 3 tr tr .tr

1401 3 tr tr tr tr
1102 7 2 3 tr tr tr

D 3715 15 2 3 tr 1 tr tr
4305 • 25 2 2 tr tr tr tr

E-1 2806 10 3 3 2 1- tr

E-2 2501 tr tr 3 tr tr

F 3002 tr tr 2 tr tr

san = sanidine, pl = plagioclase, cpx = clinopyroxene ,

bio = biotite, op = opaque, sph = sphene, ap = apatite.

*volume % determined by point counting.

 140

Table 6.3 Lava phenocryst and groundmass mineralogy*

volume % phenocrysts (tr = <0.5%) groundmass

no lc* san pl cpx ol bio op lc san p1 cpx o1 bio op ap

Qttrachytes: LV-QTR

8095 18 5 1. 1(?) 2
202 9 6 1 3 1

ne/olttrachytes: LV-VTR

8207 I 10 6 2 1

lcItrachytes and lc phonolites: LV-TRP

8206 16 1 3 2 + + + + +
8217 8 2 1 3 3 + + + + + + + +
4004 8 1 1 tr tr + + +- + +
4001 6. tr tr tr + + + + +
205 20 3 3 tr + + + + +
3401 2 4 3 tr 1 + + • + + + +

lc tephrites and tephritic leucitites: LV-TEL

8057 3 7 3 5 + + + +
8058 4 1 2 12 + + + _ +
2301 tr 1 2 tr + + + +
8053 4 1 2 4 + + + +
2302 4 3 1 + + + + + +
1301 tr 5 tr + + + + + +
6001 2 23 2 + + + + + +
6107 1 21 2 + + + + + +

tephritic leucitites and leucitites: LV-L

204 tr 3 + + + + ? + +
3201 2 10 + + + ? +
8215 -1 1 + + + + + +
8218 1 + + + + + + +

lc basanite: LV-BSN
8216 20 10

trachyandesite and trachybasalt: LV-TRAB

8063 1 3 10
6701 2 13 3
8092 3 10 2
8093 5 13 3
8060 1 2 11 2

miscellaneous trachyte: LV-MTR

401 tr 11 4 2 1

*data from the literature and this work

tQ, ne, ol = normative quartz, nepheline, olivine
if lc = modal leucite

*lc - leucite, san = sanidine, pl = plagioclase, cp( = cllnopyroxene,

of a olivine, bio = biotite, op = opaque oxide, ap = apatite
 Table 6.4 Characteristics of Vulsini pumice groups

ignimbrite/ no of normative characteristics

t
rock type
pyrocla stic DI t K20/Na20 saturation or t ab f an di
(Ic = modal leucite) samples
fall

PF-2 trachyte 2 86 2.0 ne, ± ol 50 35 6 3

A trachyte 5 86 2.1 ne, ± ol-i Q hy 51 34 .7 4

lc trachyte and 10 82 3.1 ne 60 14 7 3

B lc phonolite --

lc trachyte and 81 2.9 ne 58 16. 7 4

C 10
lc phonolite --
High K series

D trachyte 8 85 2.6 ne —0. 57 24 7 3

E-1 trachyte 3 80 3.4 ± ne, ±,hy, ± of 57 22 8 6

E-2 lc tephrite . 3 51 6.3 lc , ne , ± ol 36 0 19 16

and
F tephritic leucitite 7 52 4.6 lc, ne, ol 35 0 1.8 19

X 'trachyte 6 83 3.4 . ne, ± ol 59 23 7 5

Y trachyte 3 85 2.0 ne 52 27 7 2

PF-4 trachyte 3 85 2.6 ne 59 18 6 3

i averages
 •
Table 6.5 Characteristics of Vulsini lava groups

lava * no of normative characteristics

rock type . DI t K2O/Na2Ot
group samples saturation or i ab fi an t di t

LV-QTR •Q trachyte 4• 77 . 2.7 Q, hy 44 24 12 2

LV-VTR ne/ol trachyte 4 '75 3.1 ol, + ne, + hy 52 22 13 3
lc trachyte °'
High K series

LV-TRP tephritic 9 . 75 2.8 ne, ± ol 55 12 12 5

lc phonolite

lc tephrite +_ +
LV-TEL 25 46 3.2 ne, — ,
ol _ lc 37 • 0 20 22
tephritic leucitite — .—

tephritic leucitite > 20% lc

LV-L 15 39 4.3
leucitite ne , ± of 4 0 15 36

LV-BSN lc basanite 1 22 3.6 0l (15%)'

0.5 0 19 34
ne,ic

a4 a)
trachyandesite
LV-TRAB
.()
of ne
trachybasalt 11 47 1.6 ' , 25 21 19 19
a ----0- Q by

LV-MTR miscellaneous
Q trachyte
1 62 2.9 Q hy 40 20 17 7

*from the literature and this work, lc = modal leucite, Q = normative quartz, etc. t averages
LV-TRAB group includes 2 latites LV-TEL group includes 2 leucitites and a lc sommaite
LV-TRP " 2 lc tephrites LV-L 2 lc tephrites


143
Table 6.6 Excluded Vulsini analyses from the literature

Vollmer (1975., Table 2)

VLS-4 - kaolinized ignimbrite sample

VLS-12 - same sample as 2 in De Fino and Mattias (1965,Tabella I)
VEN - sample from San Venanzo, 20km N.E. of Vulsini volcano.

Appleton (1970, Table E2)

1 - block in foam lava

7 - block in tuffs/ashes
8 - foam lava
9 - inclusion in foam lava
16 - pumice from ash-cinder cone

Trigila (1969a, Quadro XII)

L3 - leucite to analcime alteration

A3 - n 11 11 tt
B1B - II - II 11 It

Trigila (1969b, Tabella II)

M2 - same sample as M2 in Trigila (1969a)

LA2 - leucite to analcime alteration
C1 1 - tt It If

1,3 - II It II II

sample L3 is the same sample as L3 in Trigila (1969a)

sample from Monti Calvo volcano, 20km from Vulsini.

Mattias (1965,Tabella V)

6 - leucite to analcime alteration

_ 11 It tt ft
7
_ tt It tt it
7a
II It It tt
8

Schneider (1965, p.394)

1 - block in tuff
2 _ It tt It

6 - II 11 It

11 - high H20+ content

12 - block in schlacken
 144

Table 6.7 Vulsini pumices: average data for ignimbrites and

pyroclastic falls
PF -2 A B C D E-1 E-2 F X Y PF-4
5102 61.41 61.56 57.72 57.78 59.69 59.26 48.54 49.29 60.02 59.51 58.58
TI02 .33 .32 .46 .47 .46 .49 .79 .80 .38 .42 .50
AL203 18.54 18.44 19.15 19.02 18.91 18.17 17.89 17.76 18.28 19.44 19.35
FE203 1.17 1.13 1.81 2.21 1.67 2.01 5.16 4.60 2.11 2.12 1.50
FEO 1.25 1.36 1.24 1.36 1.12 1.63 2.66 3.16 1.34 .95 1.26
MNO .15 .15 .13 .12 .13 .10 .15 .15 .11 .16 .14
MGO .24' .38 .43 .67 .41 .85 3.29 3.79 .75 .38 .35
CAO 2.12 2.15 3.27 3.27 2.60 3.36 8.84 8.99 2.80 2.48 2.60
NA20 4.32 4.17 3.30 3.39 3.69 2.82 1.29 1.69 2.93 4.33 3,86
K20 8.48 8.61 10.20 9.86 9.55 9.62 8.06 7.83 9.90 8.82 9.88
P205 .08 .08 .09 .13 .08 .18 .66 .55 .18 .08 .07
H20- .35 .27 .48 .18 .27 .17 .77 .42 .13 .29 .22
LOI 1.05 . 1.05 1.05 1.05 1.05 1.05 1.40 .85 1.05 1.05 1.05
TOTAL 99.49 99.67 99.33 99.51 99.63 99.71 99.50 99.88 99.98 100.03 99.36

F3/F2+F3 .48 .45 .59 .62 .60 .55 .66 .59 .61 .69 .54
K20/NA20 1.96 2.06 3.09 2.91 2.59 3.41 6.25 4.63 3.38 2.04 2.56
MGO/K20 .03 .04 .04 .07 .04 .09 .41 .48 .08 .04 .04
F3+F2/CA 1.28 1.31 1.02 1.18 1.17 1.20 .95 .94 1.34 1.32 1.17
K20/CAO 4.00 4.00 3.12 3.02 3.67 2.86 .91 .87 3.54 3.56 3.80
CAO/MGO 8.83 5.66 7.60 4.88 6.34 3.95 2.69 2.37 3.73 6.53 7.43
F3/F2+F3 = FE203/(FEO + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V 36 39 67 ' BO 60 80 216 222 60 57 74

CR 4 8 7 6 7 10 18 25 8 5 5
CO 1 1 3 3 2 3 24 29 3 z i
NI 6 . 6 6 5 6 7 21 24 7 6 6
CU 9 4 5 7 28 •21 25 36 15 23 23
ZN 88 83 80 81 86 70 84 83 73 107 95
ZR 713 664 619 534 601 416 322 314 450 756 581
RB 658 651 467 476 595 516 656 545 573 528 442
SR 220 279 607 1041 400 936 1484 1427 819 404 598
TH 117 106 113 128 117 77 41 38 70 163 117
Y 65 63 . 48 45 51 37 39 36 51 65 54
BA 81 94 264 791 127 472 898 913 558 287 65

K/RB 107. 110. 181. 172. 133. 155. 102. 119. 143. 139. 186.
RB/SR 2.991 2.333 .769 .457 1.487 .551 .442 • .382 .700 1.307 .739
TH/K .0017 .0015 .0013 .0016 .0015 .0010 .0006 .0006 .0009 .0022 .0014

DI 86.20 85.99 81.95 81.20 84.52 79.94 51.25 51.67 82.76 84.58 84.55
ZR .14 .13 .12 .11 .12 .08 .06 .06 .09 .15 .12
OR 50.32 51.09 60.43 58.42 56.63 57.02 36.25 34.74 58.69 52.29 58.53
PL 41.12 40.51 21.19 23.29 30.69 30.25 19.11 17.66 30.61 34.61 24.39
(AB) 35.07 34.45 13.95 15.81 23.96 21.82 0 0 23.21 27.14 18.17
(AN) 6.05 6.06 7.24 7.48 6.73 8.43 19.11 17.66. 7.40 7.47 6.22
LC 0 0 0 0 0 0 9.09 9.18 0 0 0
NE .80 .45 7.57 6.98 3.94 1.10 5.91 7.75 .86 5.15 7.85
WO .06 0 2.14 1.46 1.10 .04 0 0 0 .78 1.23
DI 3.29 3.50 2.81 3.82 2.50 5.90 16.41 18.86 4.50 2.04 2.84
(WO) 1.63 1.75 1.47 2.04 1.32 3.07 8.80 10.01 2.38 1.10 1.46
(EN) .60 .73 1.07 1.67 1.02 2.12 7.61 8.03 1.83 .95 .87
(FS) 1.06 1.02 .26 .12 .16 .71 0 .82 .29 0 .51
OL 0 .38 0 0 0 0 .41 1.10 .03 0 0
(FO) 0 .15 0 0 0 0 .41 .99 .02 0 0
(FA) . 0 .23 0 0 0 0 0 .11 0 0 0
MT 1.70 1.64 2.62 3.20 2.42 2.91. 6.77 6.67 3.06 2.37 2.17
IL .63 .61 .87 .89 .87 .93 1.50 1.52 .72 .80 .75
HM 0 0 0 0 0 0 .49 0 0 .49 0
AP .19 .19 .21 .31 .19 .43 1.56 1.30 .43 .19 .17

PF-2 = average of 2 analyses E-2 = average of 3 analyses

A .. 5 u F 7
B 10 X . 6 ..
C T " 3 ..
II 10
D " . 8 PF-4 " 3 ..
E-1 3
 •
▪

145

Table 6.8 Vulsini lavas: average data for lava groups

QTR VTR TRP TEL L BSN TRAB MTR

SI02 61.79 57.55 55.22 49.34 47.14 44.89 53.18 56.65
1102 .52 .61 .56 .85 .92 .95 .71 .86
AL203 17.20 19.10 19.73 17.25 15.04 12.73 16.16 17.55
FE203 2.29 3.20 3.24 4.36 3.53 3.31 2.52 2.02
FE0 2.49 1.36 1.25 3.26 4.25 4.35 3.53 4.33
MNO .12 .13 .14 .14 .14 - .12 .12
MGO 1.01 1.46 1.09 4.71 6.17 13.71 6.86 2.54
CAO 3.16- 3.82 3.91 10.03 12.50 12.95 8.97 5.55
NA20 2.81 2.86 3.31 1.96 1.55 1.02 2.67 2.38
K20 7.49 8.77 9.20 6.28 6.59 3.66 4.27 6.79
P205 .22 .24 .19 .42 .40 .23 .22 .34
H20- .10 .11 .60 .35 .36 .27 .20 .07
L0I/H20+ .75 .84 1.31 .90 1.14 1.59 .61 .43'
TOTAL 99.95 100.05 99.75 99.85 99.73 99.66 100.02 99.63
•
F3/F2+F3 .48 .70 .72 .57 .45 .43 .42 .32
K20/NA20 2.67 3.07 2.78 3.20 4.25 3.59 1.60 2.85
MGO/K20 .13 .17 .12 .75 .94 3.75 1,61 .37
F34-F2/CA 1.70 1.27 1.21 .83 .70 .67 .77 1.34
1(20/CAO 2.37 2.30 2.35 .63 .53 .28 .48 1.22
CAO/MGO 3.13 2.62 3.59 2.13 2.03 .94 1.31 2.19
F3/F2+F3 = FE203/(FE0 + FE203)
F31-F2/CA = (FE3f + FE2+)/CA2+

V 73 105 200 195 - 156 132

CR 5 - 7 84 42 - 233 8
CO 4 - 5 25 30 - 31 8
NI 5 - ' 6 44 70 - 111 9
CU 11 - 16 70 107 - 49 72
ZN 75 86 70 69 61 122
ZR 457 - 542 261 271 237 346
RB 469 388 • 410 485 538 - 330 391
SR 681 1218 1661 1373 1479 - 602 791
TH 90 74 138 . 47 56 - 58 53
Y . 47 - 49 39 40 . 35 39
BA 895 - 1732 1133 1188 - 592 1039

K/RB 133. 188. • 186. 107. 102. - 107. 144.

RB/SR .689 .319 .247 .353 .364 - .548 .494
TH/K .0014 .0010 .0018 .0009 .0010 .0016 .0009

DI 76.63 75.14 75.05 46.28 38.54 21.74 47.11 61.75

OZ 8.44 0 0 0 0 0 0 1.36
ZR .09 0 .11 .05 .05 0 .05 .07
OR 44.41 51.95 54.50 37.27 3.53 .48 25.34 40.25
PL 35.90 35.31 23.47 19.70 14.53 19.35 40.24 37.23
(AB) 23.78 21.99 11.73 .06 0 0 20.79 20.14-
(AN) 12.12 13.31 11.74 19.64 14.53 19.35 19.44 17.09
LC 0 0 0 0 27.91 16.58 0 0
NE 0 1.20 8.82 8.95 7.11 4.68 .98 0
DI 2.07 3.47 5.32 22.06 36.27 34.15 18.99 7.14
(WO) 1.05 1.86 2.85 11.71 19.03 18.12 9.99 3.63
(EN) .56 1.61 2.47 9.36 13.94 14.47 7.48 1.94
(FS) .46 0 0 .99 3.29 1.56 1.51 1.56
HY 3.54 0 0 0 0 0 0 7.90
(EN) 1.95 0 0 0 0 0 0 4.38
(FS) 1.59 0 0 0 0 0 0 3.52
OL . 0 .1.42 .17 1.85 1.26 15.43 8.23 0
(FO) 0 1.42 .17 1.66 1.00 13.79 6.73 0
(FA) 0 ' 0 0 .19 .26 1.64 1.50 0
MT 3.32 3.04 2.95 6.32 5.12 4.80 3.65 2.93
IL .99 1.16 1.01 1.61 1.75 1.80 1.35 1.63
HM 0 1.10 1.21 0 0 0 0 0
AP .52 .57 .62 .99 .95 .54 .52 .81

QTR = average of 4 analyse

VTR " 4 "

TRP 9 Il H.
TEL 26 major elements
. L " '16 " ///
BSN 1 ▪ I
TRAB 9 "
MTR " 1 " ./
 146

Table 6.9 Specific gravities of K-rich lavas1 and minerals2,3

lavas s.g. Si02 K20

trachyte • 2.44 60.8 6.3

(leucite free) 2.51 59.2 , 9.1
2.53 58.2 9.2
2.61 57.3 • 9.2
macrophenocrystic 2.55 55.9 10.5
leucite trachyte
leucite trachyte 2.61 55.2 8.5
2.65 55.9 8.8
leucite tephrite 2.66 50.4 9.4
2.70 47.5 7.5
leucitite 2.78 47.9 8.2
minerals
leucite 2.47 - 2.50 55.4 20.3

sanidine 2.56 - 2.62 63.6 - 67.3 7.1 - 12.1

1 lava data from Washington (1906, p. 23, 28, 31, 59, 47,
67, 97, 109, 118)
2 : leucite specific gravity data from Deer et al. (1966)
leucite Si02 and K20 data from Table 5.1
3 : sanidine data from Deer et al. (1966)
 147

Table 6.10 Distribution coefficients

D = wt. % element in crystal/wt. % element in liquid

Sanidine D Rock Reference
Rb .34 Ave. of 14, undersat. to 1
oversat.
.70 14 Italian ignimbrites , 2
oversat. to saturated.

.66 Italian rhyodacite 3

Sr . 1.20-2.80 trachytes + phonolites 4

3.87 Italian rhyodacite 3

Ba 1.17-8.95 trachytes + phonolites 4

6.12 Italian rhyodacite 3

Plagiocla se
Rb <.2 bas -and - dac*

Sr 3.27-4.31 trachytes + phonolites 4

1.18(1400.C)-3.06(1150 C) basalt to andesite 5

1.88 ave. trac-bas-and 5

1.27-2.84 bas -and -dac 3

Ba 0.72-1.09 trachytes + phonolites 4
5
0.16(1400 C)-0.56(1150 C) basalt - andesite
0.1.0 at <1060 C)
0.55 ave. trac-bas-and
0.05-0.59 bas -and -dac 3

Leucite
Rb 3.62 2 Italian lc lavas 6
3.01 12 " " 11

Sr 0.12 2 Italian lc lavas 6

0.02 12 " " " 7

0.39 2 Italian lc lavas

0.12 12 " " "

Clinopyroxene
Rb <.3 bas -and - dac 3

Sr 0.06 basalt 8
0.52 Italian rhyodacite 3

Ba . 0.13 Italian rhyodacite

Biotite
Rb 1.52 14 Italian ignimbrites 2
0.94 Italian rhyodacite 3

Sr 0.67 Italian rhyodacite 3

Ba 15.3 Italian rhyodacite

1- Noble and Hedge(1970) 5- Drake and Well (1975)

2- Dupuy(1968) 6- Hendersori(1965)
3- Philpotts and Schnetzler(1970) 7- Appleton(1970)
4- Berlin.and Ndndersun(1969) C- Berlin and Henderson(1968)

*bas = basalt, and = andesite, dac = dacite, trac = trachyte, lc = leucite

 148

Fig. 6.1 Italian potassic volcanoes

100 Km

Elba OVulsini Roman region

O Vico
©Sabatini
Rome
©Laziale-Alban Hills

©Roccamonf ina Campanian

Tyrrhenian ©Phlegrean Fields `; regio n
Ischia p ©Som ma -Vesuvius \
Sea

• Aeolian Is.d
• O

oPan telleria


.5
.Proceno • • Castel •
5 Km •12 Viscardo ••1
ardano
1B
.Acq pendente
3 .4
N
•
70 ■
ORVIETO
13 •■Canonica .
San Lorenzo Nuovo .
14. •Porano.•
Sorano •29 •30 r, , ~`
• •r~
Grotto i~ •2 ~
65.28 di Castro ./.55
. •1 ~`
•54 ‘

Gr acoli B oiS pa ■ Bols ena -32

.g
9 •26 ,~~~~• j •Lubriano
25 j ~~~ 60 • C a l d e r a 1i 34• Bagnoregio
Lake Murano ;
1I a Latera 1 61.
Caldera Lake Bolting i
i
1%

■Farnese • •. 69. Capodimonte

■∎`
•68 67 •Ischia at Castro
23 M ontefiascone
4.2 'i

.43

• Tessennano
• Fig.6.2 Vulsini Volcano -
Canino '40 sample localities
• Arlena di Castro ••••. — — Cotdaro magi!)
38• •397 41


150

FE203T
PF-2 +
IG R +
IG B +
IG C +
IG D +
IG E-1 +
IG E-2 +
IG F +
IG X +
IG Y +
PF-4 -}-

LV -DTR
LV -VTR X
LV-TRP t~
LV-TEL Z
LV -L Q
LV-BSN X
LV-TRRB e

A

NR20+K20 MGO

Fig. 6.3 Vulsini pumice and lava samples: Fe 2O3T(total iron) -

Na 20+K20 - MgO ternary diagram (Fig.6.3A above,

Fig.6.3B,C next page) .

 151

Fig. 6.3B,C

D
D

0g

PF- 2 ~P
]G R . X
]GB O
]G C
]G 0 0
IG E-]
IG x x
IG Y y
PF-4 X

LV-OTR B
LV-VTR S
LV-TRP

o*
o* 0 Y O
R X ~
 152

K20
PF-2 +
IG R +
IG B +
IG C +
IG O +
IG E-1 +
IG E-2 +
IG F +
IG X +
IG Y +
PF-4 +

LV-OTR
LV-VTR g
LV-TRP p
LV-TEL Z
LV-L p
LV-BSN

LV-TRRB

LV-MTR . ~

A
NR20 CRO

Fig. 6.4 Vulsini• pumice and lava samples: K20 - Na 20 - CaO

ternary diagram (Fig.6.4A above, Fig.6.4B,C next page).

• 0
153

* )Vs O
* X00 0
R Ō ® g
%
tg 4e
B
Y
O♦Y
O

PF-2 4a IG Y Y
IG R X PF-4
1G B 0
IG C LV—OTR 0
IG D O LV —VTR X
lG E-1 ® LV—TRP D
]G X X

x3ō

PF - 2 4•
IG R M
IG D
IG X X
IG Y Y
PF-4 X
 154

FIGURES 6.5 - 6.25

The following figures show the variation in Vulsini major and trace

elements with respect to the Thornton and Tuttle (1960) differentiation

index (DIFF IDX) . For each element the "A" diagram plots the data for

all samples, whereas the "B" diagram is restricted to high DIFF IDX sam-

ples (DIFF IDX > 70) . "B" diagrams are not plotted for Ni and Cr as for

these elements the high DIFF IDX samples are all uniformly low (see "A"

plots).
 155

Fig. 6.5 Si02

R = 0.9325 N - 129
PF - 2 +
IG A +
IG B +
IG C +
IG 0 +
IG E-1 +
IG E-2 +
IG F +
IG X +
IG Y +
DB ++ PF-4 +

e ►D LV-OTR
e D D. LV -VTR
A Z
A LV-TRP
A
A. LV-TEL
AAA
LV -L
g z LV-BSN

LV-TRAB A
22:f zz o + LV-MTR I
inzz OZ
ME' 0 m

oY
O M A
20.00 40.00
.. . 60.00 80.00 100.00.
DIFF IDX

R 0.6359 N = 67
I

X-<X® OGOX +
I PF-2
IG A
IG B
IG C .
I0 0
O
0 IG E-1
IG X
IG Y
PF-4
e
eo ItX 0 LV-OTR 9
X x LV -VTR' $
x LV-TRP D
0 y
exx0
x
® f •0
fX x
E B • 0oC0
a .

K 0 O
0
% t
p
D
D

B
I I I
76.00 80.00 84 .00 88.00
DIFF IDX

•

156

Fig. 6.6 A1203

R = 0.7907 N= 129
PF-2 +
10 A • +
IG B +
IG C +
0 IG 0 +
0 IG E-1 +
IG E-2 +
IG F +
IG X +
Li IG Y +
CL PF-4 +
1-0
3 . ►► LV-OTR 6
LV -VTR $
mN Z ►
+ LV-TRP
LV-TEL Z
4+ 4.+
LV -L 0
zz • + +A LV-BSN X
~D Z+
o
e .
z z LV-TRAB e
~_ • A A D
~ 2o e `t
LV-MTR I
pet„,
z` 6
y
0 0 z`
o
N 0
A
20.00 40.00 60.00 80.00 100.00
DIFF IDX

R - -0.0744 N = 67

-<x ® ouox+
PF-2
IG A
IG B
IG C
0 IG 0
0 1G E-1

N_
IG X
IG Y
PF-4

LV-OTR 0
►
► LV-VTR X
LV-TRP D
►

0 Q Y

O 00 Y ~# 0
s ► * 0 *E
► 8 = 0 0
0 O x
EB ® x coc x 5" r 0

0
0
co B
~ r i
'72 .00 76.00 80.00 84 .00 88.00
DIFF IDX
 -

157

Fig. 6.7 K2 O

R - 0.8680 N = 129
PF-2 +
IG R +
IG B +
IG C +
•IG D +
IG E-1 +
IG E-2 +
- lG F +
0 + IG X +
ZZ
1G Y +
t PF-4 +
a
~ o 0 z
LV-OTR B
3• cm LV -VTR X
z LV-TRP D
pM z LV-TEL Z
z LV -L p
O z z LV-BSN X

e LV-TRRB
zzzz z
p2
LV-MTR I
z ,Q
AAA
x
0 A
`1 0.00 40.00 60.00 80.00 10 0 .0 0
DIFF IDX

R - 0.1952 N = 67
PF-2 4
IG R X,
0 IG B 0
IG C seg
IG D O
0 0 . 0
e IC' E-1
0 x 148cx o lG X X
► * a . 00 Iii
IG Y Y
*
► PF-4 X
* x * 0
EB
LV-OTR 9 .
8 ► LV -VTR %
D ► EB LV-TRP D
B
► ►
•
s

0
0•
1 i T
72.00 76.00 80.00 84.00 88.00
DIFF IDX


158

Fig. 6.8 Na20

R - 0.8702 N = 129
PF-2 +
IG A +
IG B +
IG C +
1G D' +
'IG E-I +
IG E-2 +
IG F +
IG X +
IG Y +
If + PF-4
► +
++
LV-OTR B
e z It+
~ _ LV--VTR x
> Ze ► + +4F
+ LV-TRP
AAn
A
A
+
+
LV-TEL Z
ee2 LV -L p
4 z a
zz z LV - 85N X
° III
z
0 z++ z LV-TRAB e
in z zz °
LV-MTR I
vi ° +
°rim
+
Z z
et z
0
A
O + +
+ .

020.00 40.00 . 60.00 80.00 100.00

DIFF IDX

R = 0.6941 N = 67
t 't PF-2

-Cx®0 00 x+
IG A
IG B
IG C
IG D
IG E-1
IG X
IG Y
Y PF-4

LV-OTR 9
Y LV-VTR X
LV-TRP
• Y

*
0

0
* o
0 e
** x® x
B *%
0 at
44 x
0
63
B
•

72.00 76.00 80.00 84.00

1 88.00
DIFF IDX
 159

Fig. 6.9 Fe2O3T

O
R - -0.9397 N = 129
O PF-2 +
IG R +
IG B +
IG C +
z ern
p z IG 0 +
z Z . 4- IG E-1 +
aWl7 z + Z IG E-2 +
z zo +Z IG F +
$ IG X +
O e
1G Y +
PF - 4 +

LV-OTR B
LV -VTR. g
D
LV-TRP D
e LV-TEL g
LV-L 0
LV-BSN g

LV-īRRB

LV-MTR I

A
`\120.00 40.00 60.00 8b . 00 100.00
DIFF IDX

R = -0.9140 N = 67
PF-2

X-CX®040X+
IG R
IC B
IG C
IG D
IG E-1
IG X
IG Y
PF-4

LV-OTR 6
LV-VTR g
LV-TRP D

X
t
G

•
`\'72.00 76.00 80.00 84.00 88.00
DIFF IDX
 160

Fig . 6.10 MgO

0 R = -0.9486 N = 129
CD PF-2 +
IG A +
IG B +
lG C +
z IG - 0 +
0 1G E-1 +
IG E-2 +
N
IGF +
IG X +
IG Y +
PF - 4 +
CL
I- o LV-OTR
• a a z _ LV -VTR X
co LV-TRP p
a
LV-TEL 2
LV -L 0
z 4ō z LV-BSN g
a
z ~+
LV-TRAB

z +z LV-MTR I
ZI
ZZ

° A
I I
20 .00
° 40.00 6 0.00 80.00 100.00
DIFF IDX

R = -0.8461 N = 67
o"
PF-2 4•
N IGA X
IG B O
E IG C
► IG 0 O
s
IG E-1
IG X X
IG Y Y
a PF - 4 X

LV-OTR 9
od ►
LV -VTR X
EB LV-TRP

93
x
/24
CD Do
O
O

c52.00 76.00 810 .00 84.00 88.00

DIFF IDX
 161

Fig. 6.11 Ca O
R - -0.9865 N = 129
CO PF- 2 +
IG R +
IG 8 +
0 0 IG C +
0 10 D +
IG E-1 +
IG E-2 +
IG F +
IG . X +
IG Y +
PF-4 +
LV-OTR ®
LV-VTR X
LV-TRP
LV-TEL Z
LV -L
LV-65N E
LV-TRRB e
LV-MTR I
D +

0
0 A
0.00 40.00 6h.00 80.00 100.00
DIFF IDX

R _ -0.9076 N = 67
PF-2 4~
ID A Di
IG B 0
IG C.
e IG D
0 IG E-1 EB
0 IG x X
• IG Y Y
0 PF-4 X
►
0
LV-OTR 9
LV -VTR X
Es LV-TRP
0

0
CO B
`72.00 76.00 80.00 84 .00 88.00
DIFF IDX
 162

Fig . 6.12 TiO2

0 R = -0.7860 N = 129
co
I i PF-2 +
IG R +
LG B +
IG C +
IG 0 +
a IG E-1 + .
-a IG E-2 +
z IG F +
a a

IG X +
z IG Y +
a
PF - 4 +
Z (23
727 LV-OTR
a~zzr-_ LV -VTR B
Z eee LV-TRP
z z & At .
z
LV-TEL
e in
LV -L O
z
zz LV-BSN IS
e
a LV-TRRB e

LV=MTR I

0
O A
.20.00 40.00 60.00 80.00 100.00
DIFF IDX

R = -0.5788 N = 67
t i PF-2

X-CX®OGOX4
IG R
IG B
IG C
IG D
IG E-1
IG 'X
IG Y
PF-4

LV-OTR ®
LV -VTR g
LV-TRP D

0
b bRO R
o®° ~0 0
O O
O 00
b
S
x x
x xx X Y
X5(X X O

CV
i
°72.00 76.00 810.00 84.00 88.00
DIFF IDX
 163

Fig. 6 13

0 R = 0.7190 N.= 129

CO
PF-2 +
IG. A +
IG B +
IG C +
IG D +
IG E—I. +
IG E-2 +
IG F +
IG X +
IG Y +
PF-4 +

LV-OTR
_ LV —VTR X
LV—TRP
LV—TEL Z
LV —L 0
LV—BSN X .
z
• LV—TRAB A
0
LV—MTR I
0
AA,' A

ID

A
°20.00 40.00 60.00 8 '0 .00 1 00 .00
DIFF IDX

R = -0.7578 N = 67
PF-2 +
IG P X
1G B 0
JGC >rs
1G D
IG E-1 EB
1G• X X
IG Y Y
PF-4
z
x LV—OTR ®
0
LV —VTR X
03 LV—TRP D
X ® x
b D >x
b b •XX
LI) D
* D *
C\J O
CL D O O
D O X
D Y

0 O
xx
R XX •
0 0 O i ® O+ O
O X Y
Y •

°72.00 76.00 810.00 84 . 00 88.00

DI FF IDX
 164

Fig. 6.14 Sr

R = -0..5183 N = 90
• PF-2 +
IG R +
IG B +
1G C +
IG 0 +
IG.E-1 +
IG E-2 +
IG F +
IG X +
IG Y +
PF-4 +
LV-OTR
LV -VTR %
LV-TRP D
LV-TEL Z
LV -L D

LV-TRAB

LV-MTR I

0
0 A
°20.00 40.00 60.00 80.00 100.00
DI FF IDX

R = -0.869] N = 58
I l PF-2
IG R X
IG B 0
IG C ~3
IG D
• ► IG E-1
* * D
IG X X
IG Y Y
PF-4

S
• LV-OTR B
LV -VTR $
83
LV-TRP D
0
0 ox

i
O x XX

X
0 O
® g •
2(
• X X x.

0
0
i
c172.00 76.00 80.00 84.00 88.00
DI FF IDX
 165

Fig. 6.15 Ba
R = -0.5608 N - 84
PF-2 +
!GA +
IG B +
IG C +
1G D +
1G E-1 +
IG E-2 +
- 1G F +
1G - X +
1G Y +
PF-4 +
LV-OTR
LV-TRP
LV-TEL z
LV -L a
LV-TRAB

LV-MTR I

0
0
D
20 .00 40.00 60 .00 80.00 100.00
DIFF IDX

0 R - -0.8652 N = 56
N PF-2 •P
Cr)
IG A X
IG B O
IG C ~.
IG D
IG E-1
IG X " X
IG Y Y
PF-4 X

LV-OTR B
'.LV-TRP D
•

D
•
•

O
X
XX •
X
Y
so 4 x • Y
O 0 Jo
I I
072.00 76.00 80.00 84.00 88.00
DIFF IDX


166

Fig. 6.16 Rb

0 R = 0.1513 N = 89
0 PF-2 +
IG A +
IG B +
z 1G C +
0 1G D +
O IG E-1 +
IG E-2 +
_ IG F +
IG X +
1G Y +
PF-4 +
o +
+ 44 LV-OTR ®
++ _ LV -VTR g
0
LV-TRP D
Oz •LV-TEL g
'LV-L. 0
z Z rnD + .
z 2
LV - TRRB
z Z ++
z A LV-MTR I

eA
A
O

O
A
I - I. r
40.00 60.00 8b . 00 100 .0 0
DIFF IDX

R = 0.7344 N = 57
i PF-2 +
* e IG A DC
x IG B 0
x x x IGC fi
o e IG D O
x = 1G E-1
V IG X • x•
• IG Y Y
x x PF-4 R
x x °•
om 0• LV-OTR B
LV -VTR %
Y
LV-TRP D
Y
0 0

O
O •
CO ' •
•
•

I I
76.00 80.00 84.0 0 88.00
DIFF IDX
 •

167

Fig. 6.17 Zr
R = 0.8631 N = 83
PF-2 +
IC R +
10.8. +
IG C +
IG 0 +
IG E-1 +
IG E-2 +
orT IG F +
IG X +
IG Y +
D ► PF-4 +
4 411
.+ LV-OTR

D
LV-TRP
LV-TEL z
LV-L

LV-TRR8 e
7
# .+++ LV-MTR 1
a zz+ z*
z
e
ZZ 0 A
Z Z
O
O

0 I 1 1
20.00 40.00 60.00 8 0.00 10 0.00
DI FF IDX

R = 0.5693 N = S5 -
i PF-2 +
IC R X
IG B 0
IG C 4
Y
IG D
IG E-1 EB
IG X X
IG Y Y
0
PF-4 X

LV-OTR 9
LV-TRP
0
X 41

•
•
0 0 ®
$• m 0
•
X
XX C
93
93

O
O
B
1 T I
`"72.00 76.00 80.00 84.00 88.00
DIFF IDX
 168

Fig. 6.18 Y
R = 0.6444 N = 83
1 1 l PF-2 +
IG R , +
1G B +
IG C +
IG D +
IG E-1 +
IG E-2 +
IG F +
IG X +
IG Y _ +
PF-4 +
LV-OTR B
_ LV-TRP D
LV-TEL Z
LV-L
z
LV-TRRB
z

0 o z +a- LV-MTR
z
z +m + +
IZZo + +
z A
z+
O
0
0 1 0 T
A
'20.00 40.00 60.00 80.00 100.00
DIFF IDX.

R = 0.5908 55
N PF-2 +
Y IG R M
IG B O
IG C
• 1G 0 O
0
0 Y lG E-1 113
IG X x
N x x IG Y Y
CO 0 0
x PF-4 X
e
x x
0 LV-OTR B
x t LV-TRP D
x
0
00
0 Re
0
XX
•

0 0 •
0 *

O i
O D •

~ 1 m I
"72.00 76.00 80.00 84.00 88.00
DIFF IDX
 169

Fig. 6.19 Th

0 R = 0.7728 N = 89
N I I t PF-2 +
N IG P +
.IG B +
IG C +
IG D +
IG E-1 +
IG E-2 +
IG F +
IG X +
IG Y +
PF-4 +

LV-OTR m
LV -VTR g
LV-TRP D
LV-TEL g
LV - L D
G 44+
D ++ LV-TRRB
+•+
2 Z +
e LV-MTR I
z
Z e 1
0 ° ++
O 0 * ++
0 A
0 z
t+ A
N120.00 40.00 60.0o 8Ō.00 1 00 .00
DIFF IDX

0 R _ 0.2367 N = 57
N 1 PF-2 4.
N IG A X
IG 8 O
Y IG C
IG D'. O
IG E-1
IG X X
• IG Y Y
PF-4 K

Y •
LV-OTR B
• LV -VTR g
Y • LV-TRP D
p.

• 0

0 0 x
• • • X
• x

0 0 X
d
0 •
0 0
x Yx B
I 1
76.00 80.00 84.00 88.00
DIFF IDX
 170
Fig . 6.20 Ni
R = -0.8004 N - 76
1 1 PF-2 +
IG R +
IG B +
IG C +
IG 0 +
IG E-1 +
IG E-2 +
IG F +
e. IG X +
IG Y +
PF-4 +
z
LV-OTR

z LV-TRP D
LV-TEL
z
LV-L 0
0
LV - TRRB A

LV-MTR I

+ z

0
0 oa 'Not ++
1 1
020.00 40.00 60.00 80.00 100.00
DIFF IDX

Fig: 6.21 Cr

0 R = -0.5781 N = 76
1 1
N PF-2 +
Cr)
IG R +
IG B +
IG C +
0 z
0 IG D +
IG E-1 +
0 IG E-2 +
_
A I G .F +
IG X +
z IG Y +
PF-4 +

LV-OTR
_ LV-TRP D
LV-TEL Z
LV-L D
Of
U LV-TRRB e
0
0
LV-HTR I
o_
0
a
2++

0
°
co A
°20.00 40.00 60.00 810.00 100.00
DIFF IDX
 171

Fig. 6.22 V
R - -0.9340 N = 76
PF-2 +
IG A +
IG B +
IG C +
IG 0 +
IG E-1 +
IG E-2 +
++ Z IG F +
+ + ++ IG X +
+ z++ IG Y +
z PF-4 +

2 LV-OTR
_ LV--TRP D
A
LV-TEL z
LV-L 0
-LV-TRRB e

LV-MTR I

0
0
•
020.00 40.00 • 60.00 80.00 100.00
DIFF I-DX

R = -0.8860 N = 5S
O
PF-2 +
IG A X
t IG B O
► IG C 4 '
O 10. 0 O
0 • 10 E-1
IG X x
o_
rn IG Y Y
PF-4 R •

O
LV-OTR 9
LV-TRP D
xX

0
o •
•

X
X X
0
0
O
`172 .00 76.00 80.00 84.00 88.00
DIFF IDX
 172

Fig. 6.23 Co
R - -0.9522 N - 64
I i PF-2 +
IG R +
IG B +
lG C +
IG D +
]G E-1 +
10 E-2 +
IG F +
IG X +
IG Y +
PF-4 +

LV-OTR

LV-TRP D
LV-TEL
LV-L

LV-TRRB e

LV-MTR

I I
40.00 60.00 810.00
DIFF IDX

R = -0.6227 N = 44
PF-2 +
IG R X
1G B O
IG C >Q
IG D
0 IG E-1 93
0 IG X X
IG Y
PF-4 X

LV-OTR &
LV-TRP D
O
a- o
• ♦ O Y

• A 0 0 * O

CPC

M • X

I I w vT v
—72.00 76.00 80.00 84.00 88.00
. DIFF IDX
 173

Fig. 6.24 Cu
R - -0.7165 N = 76
1 1 PF-2 +
IG A +
IG B +
IG C +
IG D +
z IG E-1 +
JG E-2 +
0
IG F +
IG X +
JG Y +
z PF-4 +
0
LV-OTR
z LV-TRP D
LV-TEL Z
z 1
LV -L 0

LV-TRAB e

LV-MTR I

+ + ) + +++
+ D ++
6 +
O
+ ++
A
°20.00 40.00 60.00 80.00 100.00
DIFF IDX

R = 0.1035 N = 55
PF-2

X-< X ® O G OX+
1 1
IG A
IG B
IG C
0 IG D .
O
IG E-1
0 IG X
cv_ 1G Y
PF-4

LV-OTR A
LV-TRP

go
Y
O
ER
® Y 1K
D
x O
C7 •
~Y *4 +
O
O
0 0 "- 5~ x_ x Y
#« 0 •
1 1 1
°72.00 76.00 80.00 84 .00 88.00
DIFF IDX


174

Fig. 6.25 Zn

R = 0.4375 N = 83
PF-2 +
1G R +
IG B +
IG C . +
0 IG D +
0
IG E-1 +
1G E-2 +
IG F +
IG X +
IG Y +
PF-4 +

LV-OTR
_ LV-TRP D
LV-TEL Z
+ + + + ++
+ + LV -L EJ
++ + +44+
++ z
+ D + + LV-TRRB e
z D 0 +
0 ++ +T+ LV-MTR I
+4_
z +
e .
A ++
O
O
z
O
i r r
20.00 40.0 0 60.0 0 80.00 100.00
DIFF IDX

R = 0.2860 N = 55
PF-2 +
IGP X
1G B O
IG C
0 IG D
IG E-1 83
0 IG X X
IG Y
PF-4 X
Y
0 LV-OTR B
LV-TRP D
CL Y
CL e

z o
•
N-J p
V 0 4_
0
«
«
0 R M D(
i
CD_ « « 0 * X
CO p X e
X*
Co • X m e
®O X
0 93
X
0 FS
0 X
0
0
`72.00 76.00 8 .00 8. 0 0
0 88.00
DIFF IDX
 175

FIGURES 6.26 - .6.44

The following figures show the variation of major and trace elements

with respect to the Thornton and Tuttle (1960) differentiation index

(DIFF IDX) in pumice samples from Vulsini ignimbrites A, B and C and

pyroclastic fall 2. The data for Ni and Cr are not plotted as the data

for these elements are all uniformly low (see Table D.1).
 0
R = 0.8876 N = 27 N R = -0.5241 N = 27
PF-2 t PF -2 t
IGA X IG R x
x♦ IG B O
x IG 8 0
IG C 1G C •
0
o
o_
CO
o•
0
• 0 0 •O
•
0 •• 00 a
• •
00 • •
•
0
O O
Fig. 6.26 S102 Fig. 6.28 IC20
O 0
•
•
•
•

r c
'j 72.00 76.00 80 .00 84.00 88.00 ° 72.00 76.00 8Ō.00 84 .00 88.00
DIFF IDX DIFF IDX
R = -0.1200 N= 27 R = 0.8638 N=27
PF-2 t PF-2 t
1G R x 1G R X
IG B 0 1G 8 0
1G C * IC C *

rn_ 0

0
• 0
• 0
O 0 0
•
0 • 0
• •
• •
• •• 0
Fig. 6.27 A1203 Fig. 6.29 Na 20
• •
•
• 0
0 •
0
O x •
0
x 0
0
0
xx •
m i i r
`72_00 76 .00 ab.00 84.00 88.00 c\/72 .00 76.00 80.00 84 .00 88.00
DIFF IDX DIFF IOX
 R = -0.9338 N = 27 R N = 27
PF -2 t PF-2 t
IG A X IC A. X
10 B O IG 8 0
•IC C * IG C •

••

0 •
0 Fig. 6.30 Fe2O3T Fig. 6.32 CaO
0

0•

O 0 - o

0 X
Mx
0 R X

N72 .00 76.00 8b.00 84.00 88.00 76 .00 80.00 88.00

D IFF IDX DIFF IDX
0 R = -0.8305 N. 27 R = -0.6975 N = 27
CD
PF-2 t PF-2 t
IC A X 1C A x
10 B 0 IC 8 .
IC C •
• IC C •
•
0
N •
• 0s 0
• • 0
O 0 0
O 0
• 0
•
•
•
o •
Fig. 6.31 MgO Fig. 6.33 TiO
o 0 • X 2
X
•
a
00
x
0 •X
O O •,
0
0
072.00 76.00 80.00 84.00 88.00 °72.00 76.00 80.00 84.00 88.00
DIFF IDX D IFF IDX
 0
R = -0.8004 N = 27 _ -0.8630 N = 27
'
PF-2 4• O PF - 2 t
N
10 A x' 1G A x
•
108 0. 16 8 0
• • IG C • • . IG C •

1.17 • •
a 0
t
0
•
•
0
X . •
• XX a
0 0• 0 S Fig. 6.34 P 0 Fig. 6.36 Ba
0 2 5
0
•

0 - •
0
0 ° 6
0 x
072 .00 76.00 80.00 84.00 88.00 072.00 76.00 80.00 84.00 86.00
DIFF IDX DIFF IDX
0 0
R = -0.9018 N = 27 R = 0.8305 N=27
PF-2 t 0
CO 0 PF-2 t
10 A x N 1G ii x
• 10 8 0
IG B 0
IC C •
• • IC C *
0
O

• •
•
•
a •
a

a• Fig. 6.35 Sr
• o 0 Fig. 6.37 Rb
0 0 8
0 0 0

0 •
0
0
o
0 •
•
0 o •
~72. 00 76 .00 ab .00 88.00 "72.00 76.00 80.00 84.00 88.00
DIFF IDX DIFF IDX
 R = 0.7885 N = 27 R = 0.4502 N= 27
.0
PF-2 t 0 PF-2 +
IG R x N
IC R x
10 B O
0 • •
IC C •
10 B 0
IG C *
0 0
0 0
° °
0
• 0
co_
K
x x
•
0

•
•

Fig. 6.38 Zr 1-
0 Fig. 6.40 Th
0
•
ō_
CO
•
O
• ° °
• • 0
0
D •
°72.00 76.00 80.00 84.00 88.00 72.00 76.00 8b.00 84.0O 88.00
DIFF IDX DIFF IDX
0 0
0 R = 0.7917 N = 27 •
R -0.9173 N = 27
0 PF-2
N PF-2 +
IO R X
•
10 0 0 • IG R x
• IG 8 0
• 10 C * IO C *
0
O 0
X x 0 •
x
• 0)
°
°
• 0 0 •
•
00
0 °

° 0
0
• 0 ° %•
••
Fig. 6.39 Y Fig. 6.41 V
• 0 •
0 0
0 •
K
• K K
0 • 0 101
• 0 •
O
`'7 2. 00 76.00 80.00 84.00 88.00 R72.00 76.00 80.00 84.00 88.00
DIFF IDX DIFF IDX
 R = -0.0397 N = 27 R = 0.4145' N = 27
ID PF-2 4 PF-2 •
IG R x 1G R x
1G B O 1G B O
1G C * 0 IG C • •
0
0
610 0 •

• X x
ON •
• •
0 •
3 o • • • •
•
• to
CL x
0 •
• •

• 00 x Fig. 6.42 Cu • Fig. 6.44 Zn

• • 0 0 MK
O 0
0 0 0 •
0
0

O
0
072.00 716.00 810.00 84 .00 88.00 76.00 80.00 84.00 88.00
DIFF IDX DIFF IDX
0 R = -0.8009 N = 23
0
PF-2 •
rn 1C R x
10 8 0
IG C *

• 0

C7
U • 0 0 • Fig. 6.43 . Co

• •• x
0
0
?o-b e M 1
"12_00 75.00 60.00 84.00 88.00
DIFF IDX
 181

CHAPTER VII

CONCLUSIONS

7.1 Summary

7.1,1. Analytical and data processing techniques

Major element analytical techniques using XRF spectrometry on

fused rock samples were investigated with particular reference to the

methods for determining the background count rate for the analysed ele-

ments. An iterative method was developed for determining backgrounds

and comparative data show that this method produces superior calibrations

as compared to backgrounds determined from blanks. Backgrounds deter-

mined"off peak"may be inaccurate due to interference from elements pre-

sent in the flux and in the X-ray tube target as well as increasing the

analytical time. A range of ultrabasic, basic, intermediate and acid

standards were used and the iterated background technique gave the fol-

lowing average calibration errors (relative) for the standards: < 1%- for

Si02, A1203 and K20; 1-2% for Ti02, Fe2O3T (total iron) and CaO;

5-6% for MgO and P205; and 9-14% for MnO, NiO and Cr203.

XRF trace element analytical techniques were investigated for

whole rock pressed powder briquettes and for the fusion discs used in

major element analysis . The analysis of pressed: powder briquettes gave

average standard calibration errors of sc 1% for Sr and Rb; 3-5% for Ba (Cr

tube), Y, Zr, Co and Ni; and 6-14% for Ba (W tube), Th, V, Cr, Cu and

Zn. Comparison of the contrasting values published for the same trace
 182

element in the literature on international rock standards indicates that

if XRF counting errors are small, then the errors in the standard values

recommended in the literature may well outweigh residual errors present

after correction for line overlap, spectral contamination and/or matrix

effects.

The Sr and Rb calibration data for Norrish fusion discs indicates

that the latter may be used for the analysis of these elements provided that

an average calibration error of 4% (relative) is acceptable. In studies in

which this applies the Norrish discs may therefore be used for both major

element and Sr and Rb determinations. For samples containing < 0.5%

TiO2 the Norrish discs may also be used in the analysis of Ba .

An integrated suite of interactive computer programs was developed

for major and trace element data reduction and evaluation. The various

interactive options in the programs are designed for maximum 'flexibility

in processing the data from remote computer terminals. The options in

the major element program provide for selective or comprehensive listing

of, and interaction with, the analytical data during standard calibration.

After the calibration process has been completed the sample analyses may

be listed at the terminal. The interactive options in the trace element

program provide a flexible method for establishing the standard calibrations,

after which the results may be listed at the terminal.

The major and trace element results are then linked together by a

tabulation program which creates an input file for CIFW norm programs.

The tabulation program will also sort the analyses into groups according

to an eight character "group identifier" keyword attached to each analysis.

 183

Average analyses are also calculated for each group. The tabulation

program links the major, trace and normative data to produce a data base

file as well as a tabulated output file for dispatch to a line printer.

The data base file may then be used as the input to an interactive

graphics program which enables the user to plot binary or ternary diagrams

via a series of commands entered from a graphics terminal. The data may be

plotted for all samples or the plotting may be restricted to specified group(s)

within the data base file. If desired only the averages of the groups may

be plotted. The samples in each group may be plotted with a unique

symbol for up to 14 groups. The group identifiers plus symbols are listed

alongside the plot. In binary diagrams a correlation coefficient is cal-

culated for the paired data and this is written on to the diagram along with

the number of points plotted.

The data base file also serves as the input to an interactive statis-

tics program which calculates various parameters (e.g. mean, standard

deviation, skewness and kurtosis) as well as calculating a correlation

coefficient matrix for the data. These computations may be carried out

for selected group(s) or group averages from the data base file.

The group identifier keyword attached by the user to each sample

analysis facilitates the division of the suite of samples under study into

groups and subgroups. This important feature adds considerable flexi-

bility in the processing and evaluation of the data via the tabulation,

graphics and statistics programs.

A quantitative XRD method was developed for determining analcime

 184

concentrations in pumice (and lava) samples . The calibration technique

uses mass absorption factors, calculated from major element data, to

correct the standard and sample XRD intensities for absorption effects

resulting from compositional variations. If major element data are avail-

able, the application of calculated mass absorption corrections to quanti-

tative XRD intensities provides an accurate and rapid alternative to direct

measurement of the mass absorption correction, or to the use of internal

standard techniques.

7.1.2 Vulsini geochemistry

A suite of 60 pumice and 15 lava samples from Vulsini volcano were

analysed for major and trace elements using the XRF techniques summarised

above. XRD techniques (above) were used to correct the analyses for the

effects of leucite to analcime alteration present in certain pumice- and lava

samples. The pumice samples are from ignimbrite and air fall deposits

which are thought to represent a large proportion of the eruptive products

of this volcano. Previous geochemical work in the literature has con-

centrated on major element studies on the Vulsini lavas and litte attention

has been paid to the pumice deposits. Fifty-four lava analyses selected

from the literature have been included in the present study.

The major and trace element variations in the pumice and lava sam-

ples are considered to indicate the operation of a differentiation process,

based on low pressure crystal fractionation, which has led to the evolu-

tion of potassium rich "High K" (K2O/Na2O > 2.0) magmas in the sequence:

leucite basanite—+leucitite/tephritic leucitite/leucite tephrite-4leucite

 185

phonolite/leucite trachyte—).trachyte. The pumice samples are predomi-

nantly salic leucite phonolites/leucite trachytes and trachytes. A group

of basic tephritic leucitite/leucite tephrite pumice samples are however

also present. A subsidiary suite of "Low K" (K2O/Na2O < 2.0) trachy-

basalt and trachyandesite lavas are considered to have evolved separately

from the main sequence of High K lavas and pumices.

Geochemical variations in pumice samples from certain of the ignim-

brites indicate that differentiation processes operated on the scale of the

individual ignimbrite magma. It is suggested that prior to eruption this

resulted in a zoned magma column with higher and more differentiated

layers overlying less evolved compositions at-lower levels.

It is evident that previous geochemical studies on Vulsini volcano

have not taken due account of the voluminous salic differentiates typified

by the phonolite and trachyte pumice samples. Furthermore it is consi-

dered that geochemical studies on the other Italian K-rich volcanoes (e.g.
Vico, Sabatini, Alban Hills, Roccamonfina, Vesuvius) will be incomplete

if the large volumes of ignimbrite and pyroclastic material, associated with

these volcanoes, are excluded frōm analysis and evaluation.

7.2 Future work

7.2.1 Analytical techniques

Major element analysis

At present, for those studies in which Cr and Ni are trace elements,

major element analysis is broken into three separate runs (or passes) for
 186

the samples through the spectrometer: (i) Cr tube run A: Si, Al, Ti, M

Ca., K (analysing crystals = PET, RbAP, LiF200); (ii) Cr tube run B: P

(analysing crystal = Ge); (iii) W tube run: Fe, Mn (analysing crystal =

LiF200) (see Table 2.1). The Philips 1212 spectrometer only accommo-

dates three crystals at any one time, hence the need for a second run on

the Cr tube to do P. If all the major elements could be determined in a

single pass of the spectrometer this would greatly facilitate the analy-

tical process as the XRF data could then be processed immediately and

the results evaluated. Recent calibration tests have shown that Fe may

be satisfactorily determined on the Cr tube instead of the W tube. Mn

calibrations on the Cr tube are however affected by the proximity of the

MnKKc and CrKp (from the X-ray tube) lines. The dispersion of the

LiF200 analysing crystal does not adequately separate these two peaks.

Tests have shown that the higher resolution LiF2 20. crystal may be used

for Mn calibrations even though there is a relatively high background from

the CrKp peak. However this procedure requires a crystal change thus

precluding a single pass analysis. As an alternative to changing over

the crystals, it may be possible to calibrate for Mn by using the LiF200

crystal and a secondary collimator in front of the scintillation counter

(Jenkins and de Vries, 1967, p.33). With the flow counter switched off,

the use of secondary collimation will improve the resolution of the LiF200

crystal and provided that (i) the . measured intensities are not reduced to

unacceptable levels and (ii) that goniometer reproducibility is not a prob-

lem, then Mn calibrations on Norrish major element discs should be feas-

ible using the Cr tube.

 187

A further alternative would be to treat Mn as a trace element and

carry out the analysis on whole rock pressed powder briquettes, along

with the other trace elements. In most igneous rocks MnO is present

at concentrations < 0.3 - 0.5% and its contribution to the calculation of

matrix corrections in a major element analysis is therefore small enough

to be disregarded (Norrish and Hutton, 1969, p.446).

Determination of Na2O on Norrish fusion discs requires that NaNO3,

added .to the fusion mixture to promote oxidizing conditions , is replaced

by LiNO3. The acquisition of a TIAP analysing crystal to replace the

RbAP -crystal will increase the count rates available from both NaKpc and

MgK vc . The use of LiNO3'and a T1AP crystal will give satisfactory Na2O

analyses on the fusion discs (K. Norrish, pers. communic.).

P Kg is interfered by second order Ca Kp lines when using analy-

sing crystals such as PET; the Ge crystal by contrast has almost no second

order reflections and this crystal is normally used for determining P K.

However this requires a crystal change, precluding a single pass analysis

for P as well as all the other elements. It may be possible to carry out

P analysis on the PET crystal if corrections for the second order Ca inter-

ference are applied. Techniques similar to those used to correct Mg Koc

for interference from crystal fluorescence (Chapter 2) could be used to

make these corrections. The calibration errors are likely to be worse

compared to Ge crystal calibrations, but this may be acceptable in view

of the increased speed of analysis. For those studies particularly con-

cerned with the highest accuracy in P analysis, the samples could of

course be analysed in a separate run using the Ge crystal.

 188

Trace element analysis

Expansion of the range of trace elements which may be analysed is

desirable; data processing for elements such as Nb and Pb would form

useful enhancements to the trace element program.

7.2.2 Post-analytical data processing

The following developments are suggested:-

(1) Increase the number of samples that may be processed by

programs TAB, GEOPLOT, etc. by the use of overlay techniques where

necessary.

(2) Enhance GEOPLOT program to plot data base samples in the

ternary diagram representing the experimental system Si02 - NaAlSiO4 -

KAlSiO4 (petrogeny's residua system: Bowen, 1937; Hamilton and

Mackenzie, 1965) .

(3) Integrate STAT2 and MIXER programs (Fig. 4.1) with the data

base file and the system of grouping keywords used to subdivide the data

base samples.

7.2.3 Vulsini geochemistry

The following lines of further work on the Vulsini lava and pumice

samples are suggested:-

(1) Carry out microprobe analyses of phenocrysts from pumice

and lava samples, as well as pumice groundmass glass. Compare

compositional variations found in these phases with the evolutionary

 189

model postulated from the whole rock data.

(2) Carry out whole rock analyses for rare earth elements and Nb;

compare these data with the postulated evolutionary model.

(3) Analyse leucite basanite samples for all trace elements, inclu-

ding (2) above, and compare with the postulated model. These data will

be of interest because of the possible parental role for Vulsini leucite

basanites (this work and Cundari and Le Maitre, 1970) .

(4) If microprobe analysis of Sr, Rb and Ba in feldspars and pumice

glass proves to be difficult then separation of these phases will allow

analysis using XRF techniques. The results will enable crystal/liquid

distribution coefficients to be estimated, thus providing a check. on the

proposal that feldspar fractionation controls the observed trends for these

trace elements in high DI samples.

(5) Investigate mass balance fractionation models using program

MIXER combined with whole rock analyses and the phenocryst compositions

determined in (1) above (cf. Baker et al., 1977).

(6) Plot high DI (i.e. salic rich) samples in the SiO2 - NaA1SiO4 -

KAiSiO4 ternary diagram. This plot should assist in developing theories

c
for the late stage evolution of the Vulsini trachytes and phonolites.
 190

APPENDIX A

CORRECTION OF ANALYSES FOR SECONDARY ALTERATION

A.1 Correction of major element analyses for alteration of leucite

to analcime

Table A.1 lists the analcime concentrations determined in Vulsini

pumice and lava samples using the XRD technique discussed in Chapter 5.

Table A.2 illustrates the method used in correcting the analysis of sample

303 for the effect of leucite to analcime alteration. Calculations using

the chemical formulae presented in Table 5.1 show that the molecular

weights for the two minerals are close (leucite = 220.5, analcime = 223.0).

In view of the similarity of the molecular weights it was decided to recal-

culate the sample analyses on a one to one basis with respect to the deter-

mined analcime and leucite compositions (Table A.2).

A.2 Correction of trace element analyses for alteration of leucite to

analcime

XRF scans indicated that Sr and Rb were present in the leucite and
l

analcime minerals analysed in Chapter 5. These minerals were therefore

analysed for these two elements using techniques described in Chapter 3.2.2.

The average determined Sr values were: leucite = 31 ppm; analcime = 130ppm,

while the average Rb values were: leucite = 1624 ppm; analcime = 2466 ppm.

These values, in conjunction with the XRD analcime concentrations, were

used to correct the Sr and Rb data for the Vulsini pumice and lava samples

analysed in this work. The correction procedures were identical to those

described above for the major elements.

 191

Analyses for Ba in the leucite and analcime phases were also

carried out using techniques described in Chapter 3.2.3. The average

Ba values were: leucite = 86 ppm;. analcime = 43 ppm. In view of the

*
relatively poor analytical precision for Ba at these concentrations it was

decided not to alter the Vulsini pumice and lava Ba data.

A.3 Correction of pumice analyses for glass hydration

In Chapter 6.1.3 the effect of secondary hydration of pumice glass

is discussed. Because of this alteration the high DI pumice analyses

(Table D.1) have been normalized to 1.05% LOI (loss on ignition) which .

is the average LOI/H20+ value for lavas of similar overall composition.

Table A.3 lists the pre-normalization LOI values for the pumice samples.

Note that these latter values have been corrected for the effect of leucite

to analcime alteration (see A.1 above) .

* 3 a counting error fts 30 ppm

 192

Table A.1 XRD determinations of analcime in ignimbrite pumices

and lava samples -

sample % analcime sample % analcime

Ignimbrite B Ignimbrite F
3713 8.05 3002 28.05
4008 7.18 303 32.75
3901 14.44 3016 28.95
3712 9.67 6101 37.70
3902 1.48 101 16.20
5502 5.83 305 32.34
4102 5.53 3003 29.06
4007 7.93
4101 8.12
5501 9.01.

Ignimbrite C Lavas
1102 2.88 4001 2.37
1101 4.77 205 0.90
2915 4.00 204 0.49
1406 6.72 6107 0.46
7003 0.78 4004 1.17
1405 6.20 3401 1.08
6001 0.62

Ignimbrite E-2
2501 30.23
901 36.91
4201 32.68
 193

Table A.2.. Correction of analysis of pumice sample 303 for effect

of alteration of leucite to analcime

analcime in 303 = 32.75%

ie . wt. fraction = 0.3275 = K

uncorrected - analcime + leucite corrected

analysis mineral % K mineral xK analysis
303 analysis analysis* of 303

A B . C D**

S102 49.13 54.96 17.999 55.35 18.127 49.26

TiO2 .78 .06 .020 .05 .016 .78
A1203 17.96 21.95 7.189 ' 22.44 7.349 18.12
Fe2O3 4.20 .47- .154 .43 .141 4.19
MgO 3.07 .21 .069 .37 .121 3.12
Ca O 8.50 .41 .134 .06 .020 8.39
Na2O 5.68 12,03 3.940 .32 .105 1.84
K2O 1.82 .65 .213 20.33 6.658 8.27
LOI 3.49 9.35 3.062 1.08 .354 0.78

see Table 5.1

** corrected analysis D = A - B + C
Table A..3 Determined weight per cent LOI (loss on ignition, 110° C-4850° C)
in pumice samples*

sample LOI. sample LOI sample LOI sample LOI

ignimbrite A ignimbrite C ignimbrite E-1 pyroclastic fall PF-2

3707 3.25 1401 1.80 2804 2.19 3701 3.53
4103 2.70 1104 1.72 2806 2.13 2911 3.18
3709 2.44 1402 2.51 2805 2.14
3801 2.63 1103 2.69
3708 3.19 7003 2.49
2915 2.18
' ignimbrite X pyroclastic fall PF-4
1406 2.00
1405 1.83 6502 2.61 . 2402 3.55
1101 2.10 6803 1.99 2403 3.28
ignimbrite B 1102 2.15 6801 2.02 2401 3.91
6809 2.26
3901 2.22 ignimbrite D 6805 2.21
3712 2.28 2910 2.28
4101 2.50 501 3.21
3713 2.56 3715 3.12
4007 3.09 4301 2.82
3902 2.90 3006 3.69 ignimbrite Y
5501 2.66 4305 2.90
4008 1.95 805 1.88 6904 4.19
4102 2.68 804 2.10 6903 3.98
5502 2.60 2906 4.28 6908 3.51

*After correction for leucite to analcime alteration

 195

APPENDIX B

PETROGRAPHY OF THE ANALYSED PUMICE SAMPLES

Ignimbrite A and Pyroclastic Fall 2

The pumice samples from these two eruptions are identical and will

be described together. They are composed of a vesiculated and colour-

less glass enclosing phenocrysts of sanidine, clinopyroxene, biotite,

plagioclase, opaque oxides and accessory apatite.

Sanidine is the dominant phenocryst phase and forms sub-euhedral

laths (--- 5.0mr4 which may carry inclusions of plagioclase and biotite.

Clinopyroxene forms stumpy sub-euhedral phenocrysts (—> 0.5mm) .

Fresh biotite forms stumpy and elongated laths 0.5mm) while sub-
■
euhedral plagioclase forms stumpy laths (----) 0.75mm) which may be

zoned or mantled by alkali-feldspar. Opaque oxides form an-subhedral

phenocrysts (- 4 0.5mm) while sparse apatite forms euhedral laths

(-4 0.1mm). Sanidine, plagioclase, clinopyroxene and opaque oxides

may form glomeroporphyritic clumps.

The groundmass glass is generally colourless. However it may be

turbid in patches due to incipient devitrification (3801 and 2902) . Micro-

lites of prismatic alkali-feldspar are common in the glass of some samples

(e.g. 3801).

Ignimbrite B

Ignimbrite B pumices are composed of colourless glass containing

 196

phenocrysts of sanidine, clinopyroxene, plagioclase, leucite--?analcime,

sphene, opaque oxides and accessory apatite.

Sanidine is the dominant phenocryst phase and forms sub-euhedral

laths (—) 5.0mm). Inclusions of plagioclase, clinopyroxene, apatite

and glass may be present in these laths. Some pumice samples are noted

for the fragmented or disrupted nature of the sanidine phenocrysts.

Plagioclase forms stumpy sub-euhedral phenocrysts

Green clinopyroxene forms an-subhedral phenocrysts (--~ 1, 5mm) which

may carry inclusions of plagioclase. The clinopyroxene phenocrysts may

be zoned. Leucite altered to analcime is present as rounded to euhedral

phenocrysts (--)5.0mm). They may carry inclusions of plagioclase, clino-

pyroxene, opaque oxides and brown glass . Sphene forms sub-euhedral pheno-

crysts 0.3mm) , opaque oxides form rounded to euhedral phenocrysts

I
(—p 0.4mm) and apatite forms accessory euhedral phenocryst laths (—›

0.4mm).

Sanidine, plagioclase, clinopyroxene, sphene and opaque oxides may

form glomeroporphyritic clots.

The groundmass glass may be incipiently devitrified in patches in

certain samples (5501, 3713, 4008) . 3901 displays extensive incipient

devitrification. Microlites of alkali-feldspar and clinopyroxene are com-

mon as well as sub-spherical microphenocrysts (--0.1mm) of leucite---)

analcime (the latter are very similar in outline to the leucite micropheno-

crysts seen in the lava samples) .

 197

Ignimbrite C

The pumices from ignimbrite C are composed of vesiculated brown

to dark brown glass enclosing phenocrysts of sanidine, clinopyroxene,

plagioclase, opaque oxides as well as phenocrysts of leucite-4 anal-

cime, biotite and accessory apatite. The phenocrysts are more common

in the more "basic" samples (i.e. lower DI samples).

Sanidine is the dominant phenocryst phase and forms unzoned sub-

euhedral elongated laths (-4 5.0mm). Pale green clinopyroxene forms sub-

euhedral phenocrysts (—> 1.5mm) . Plagioclase forms stumpy sub-euhedral

laths (---) 1.0mm) which may be zoned and/or mantled by sanidine. Bio-

tite forms elongated laths up to 1.25mm long. Leucite -) analcime is

present as sub-euhedral phenocrysts (--->1.0cr4 usually with inclusions of

glass and clinopyroxene.

All the phenocrysts show an increase in size in the more "basic"

pumice samples. Glomēroporphyritic clots formed by the phenocrysts are

also more common in these latter samples.

The glass varies from clear to incipiently devitrified with spherulites

of alkali-feldspar and turbid glass (--j 0.5mm) . The colour varies from

brown in the less "basic" to darker brown in the more "basic" pumice

samples. Leucite--analcime may be present as "quench" micropheno-

crysts (cf. lavas) in the groundmass glass (1405, 1101, 1102).

Ignimbrite D

The samples from this ignimbrite are noted for phenocrysts of sanidine,

* DI = Thornton and Tuttle (1960) differentiation index

 198

plagioclase, clinopyroxene, biotite, sphene, apatite and opaque oxides

set in a colourless to pale yellow vesiculated glass.

Sanidine is the dominant phenocryst phase and forms unzoned sub-

euhedral laths (-. 4.0mm). Plagioclase forms stumpy laths (— 0.5mm)

which may be zoned. Pale green clinopyroxene forms sub-euhedral phe-

nocrysts (—, 1.0mm) . Sphene is conspicuous (4301, 3006, 4305) form-

ing an-euhedral phenocrysts (--->0.3mm) , while biotite forms fresh laths

(-40.4 mm) and opaque oxides (-+ 0.5mm) form an- euhedral crystals.

With the exception of biotite the phenocrysts may form glomero-

porphyritic clots (especially clinopyroxene, plagioclase and opaque oxides).

The colourless to pale yellow glass may carry patches of turbid

insipient devitrification (e.g. 4301, 4305) . 805 is noted for the pale

brown colour of the pumice glass and for the presence of microlites of

alkali-feldspar in the glass.

Ignimbrite E

Ignimbrite E is a mixed-magma eruption (Sparks, 1974) and the

light and dark pumice samples fall into two distinct groups based on pet-

rographic and chemical characteristics. Three of the analysed pumice

samples (901, 2501 and 4201) from this ignimbrite are identical to the

black pumices collected from ignimbrite F. These three samples (group

label = E-2) will be described together with the samples from this latter

ignimbrite in the next section.

The other group of pumice samples (2804, 2806, 2805; group

 199

label = E-1) are light grey in hand specimen and are composed of pheno-

crysts of sanidine, clinopyroxene, plagioclase, biotite, opaque oxides and

apatite. These phenocrysts are set in a colourless to pale yellow glass.

The dominant phenocryst is sanidine, forming rounded angular to

euhedral laths (-45.0mm). Some pumices (e.g. 2804) are noted for the

disrupted state of the sanidine phenocrysts and these pumices are rich in

angular fragments (down to <0.1mm) of this mineral. Plagioclase forms

stumpy phenocrysts (—H1.5mm) which may be zoned. Green clino-

pyroxene forms stump an-euhedral phenocrysts (—p 1 .0mm) , often with

inclusions of fresh brown glass. Fresh biotite (-* 2.0mn) and opaque

oxides (—p0.5mm) are conspicuous.

The phenocrysts may form glomeroporphyritic clots (e.g. plagio-

clase; clinopyroxene, opaque oxides and biotite).

2804 carries occasional anhedral phenocrysts f----) 0.75mm) of

melilite (?). These colourless .phenocrysts are noted for their anomo-

Tous "Berlin blue" birefringence, relatively high RI (< CB )*, extinction

parallel to an imperfect cleavage and the presence of a very fine cross

hatching normal to the cleavage. It was not possible to obtain an inter-

ference figure. The presence of a dark irregular corona around these

crystals suggests that they may have been in disequilibrium with the sur-

rounding glass (i.e. the liquid) at the time of eruption.

Patchy incipient devitrification is present in sample 2804.

* RI = refractive index, CB = Canada balsam.

 200

Ignimbrite F

The pumice. samples from this ignimbrite together with 901, 2501

and 4201 (E-2 pumices) are noted for their dark devitrified groundmass

glass and for phenocrysts of sanidine, clinopyroxene, leucite -+ anal-

cime, plagioclase, biotite, apatite and opaque oxides.

Pale green clinopyroxene forms an-euhedral phenocrysts (-1.00mm)

which may be zoned. Plagioclase forms stumpy an-subhedral phenocrysts

(- 40 .4mm) which may carry inclusions of brown glass. Sanidine forms

anhedral phenocrysts (-0 0.75mm) . In some pumices phenocrysts of

plagioclase (3016) and lathy sanidine (—*2.0mm,2501) are surrounded by

an aureole of vesicle rich pale yellow-brown undevitrified glass. Some

of these phenocrysts have a corroded aspect and this suggests that they

may have been in disequilibrium with the groundmass glass (i.e. liquid)

at the time of eruption.

Biotite forms fresh unaltered laths (-12.0mm) . Rounded pheno-

crysts (-05.0mm) of leucite-+analcime are present in hand specimen

and occasionally in thin section (4201). Fresh leucite phenocrysts

(_-p0.5mm) are present in 2501. A small (0.3mm) phenocryst of fresh

leucite mantled by analcime is present in 3012 (unanalysed ignimbrite F

pumice) . Leucite .. analcime microphenocrysts and quench "stars" are

common in the groundmass (e.g. 901).

Glomeroporphyritic clots of plagioclase, clinopyroxene and opaque

oxides occur occasionally in 3002 and 4201.

The groundmass glass is noted for its brown-black colour and turbid
 201

and devitrified character. 101 is however an exception and the ground-

mass of this pumice contains undevitrified brown-black glass as well as

devitrified areas.

A brown pumice xenolith (#.. 2.0mm) is present in 3002. It is com-

posed of phenocrysts of plagioclase, sanidine, pale green clinopyroxene

and biotite set in a pale brown glass. This xenolith glass is incipiently

devitrified and carries microlites of alkali-feldspar.

Ignimbrite X

These pumices are composed of phenocrysts of sanidine, clino-

pyroxene, plagioclase, biotite, opaque oxides and accessory apatite set

in a vesiculated brown glass. Microlites of alkali-feldspar are.present •

in the groundmass glass.

Sanidine is the dominant phenocryst phase and forms subhedral

laths (— 5.0mm) which may contain glass inclusions. The phenocrysts

may be disrupted to form angular fragments. Clinopyroxene forms stumpy

an-subhedral phenocrysts (—p0.75mm) . Plagioclase forms stumpy an-

subhedral phenocrysts (---) 1.0mm) which may be zoned. Biotite forms

fresh laths (—* 1.25mm) while opaque oxides form an - subhedral pheno-
crysts (—p 0.4mm) often in association with clinopyroxene phenocrysts.

Accessory apatite may form elongated laths (—+0.3mm).

Ignimbrite Y

Pumice samples from this ignimbrite are composed of phenocrysts

 202

of sanidine (—r 4.0mm) , clinopyroxene (-0.5mm), fresh biotite (-4

0.5mm) , plagioclase (-4 0.5mm) and opaque oxides 0.3mm) set in a

colourless to very pale yellow vesiculated glass. Accessory crystals of

sphene and apatite (both <0.2mm) are present. The phenocrysts may form

glomeroporphyritic clots. A colourless phenocryst (—f 0.5mm) with RI <

CB and birefringence = nil, is present in minor amounts. It has been

provisionally identified as sodalite.

Pyroclastic Fall 4

The pumice samples are noted for phenocrysts of sanidine, clino-

pyroxene, plagioclase, fresh biotite, opaque oxides and accessory apa-

tite. Sanidine (-4 2.0mm) is the dominant phenocryst phase. The vesi-

culated groundmass glass is colourless to very pale yellow.

 203

APPENDIX C

PETROGRAPHY OF THE LAVA SAMPLES ANALYSED IN THIS.WORK

In this appendix the main petrographic features of the lavas collected

and analysed by the author are described. In these descriptions the leucitic

lavas have been provisionally identified according to the petrographic scheme

proposed by Appleton (1970, p.14), i.e.:

Leucitite 3 leucite forms > 90% of the "leucite plus . feldspar"

component.

Tephritic leucitite +leucite > plagioclase; san < 10%.

Leucite tephrite -s plag > leucite; san < 10%.

Phonolitic leucite tephrite +leucite + plag + san; plag > san.

Tephritic leucite phonolite +leucite + plag + san; san > plag.

Leucite phonolite+ san > leucite; plag < 10%.

This scheme is similar to that proposed by Streckeisen (1966, 1967). It

should be noted that accurate petrographic classification of the Vulsini

lavas using the above scheme is difficult owing to the problem of accurately

estimating the relative proportions of the minerals in the fine grained lava

groundmass . In view of this the grouping of the lavas (e.g. LV-L,

LV-TEL etc.) adopted in this work has been based on the various chemical

and normative grouping criteria discussed in Chapter 6.3.2.

LV-L group (leucitites)

These lavas (204, 3201) are composed of phenocrysts of clinopyroxene

and occasional leucite set in a groundmass of these minerals as well as

 204

alkali feldspar, opaque oxides and occasional biotite and plagioclase.

Clinopyroxene forms eu-subhedral phenocrysts. (--►1.5mm) . Some

phenocrysts are pale green in colour while others are zoned with darker

green cores and show yellow green - green pleochroism. Leucite forms

eu-subhedral phenocrysts (—>1.5mm) .

The groundmass is dominated by sub-spherical grains (---> 0.2mm)

of leucite and granular to prismatic clinopyroxene. The groundmass also

carries disseminated opaque oxides and poikilitic alkali-feldspar noted for

its inclusions. and irregular interstitial occurrence. Biotite, plagioclase

and amphibole (pleochroic brown olive green brown) form accessory phases

in the groundmass.

LV-TEL group (tephritic leucitites)

These lavas (2301, 1202, 5401, 2302, 1301, 6001, 6107) are charac-

terised by clinopyroxene and leucite phenocrysts set in a groundmass of the

above minerals plus plagioclase. Olivine phenocrysts are present in the

more basic (low DI *) samples.

Leucite forms eu-subhedral macrophenocrysts (—p 2.5mm) , usually

free of inclusions. All the lavas in this group contain ubiquitous sub-

hedral to sub-spherical leucite microphenociysts (--).0.4mm) which are

often noted for their radially distributed inclusion patterns. Clinopyroxene

forms eu-subhedral phenocrysts (-- 3.0mm) . They are faintly pleochroic

(pale green - pale yellow green) and may be zoned with darker green cores.

* DI = Thornton and Tuttle (1960) differentiation index.

 205

The groundmass is composed of fine (< 0.1mm) granular to prismatic

clinopyroxene, orientated plagioclase laths, blebs and grains of leucite

and disseminated opaque oxides. Accessory phases include variable

amounts of interstitial K-feldspar, biotite, apatite and amphibole.

Petrographic variations: lava 2301 carries euhedral plagioclase

(An 91-92) phenocrysts (—>1.5mm) and heavily resorbed biotite laths

(-41.0mm) . The groundmass of 1202 is turbid and the leucite micro-

phenocrysts form "quench" star patterns (Appleton 1970) , suggesting

rapid cooling of this lava. Occasional phenocrysts of anhedral . olivine

(-->0.75mm) occur in 2302 and 1301 while in 6001 and 6107 sub-anhedral

phenocrysts (--41.0mm) are more common. The olivine displays marginal

red-brown alteration to iddingsitē .

LV-TRP group (tephritic leucite phonolites)

Four lavas. (205, 4001, 4004, 3401) belong to this rock group. The

main phenocryst phases are leucite, clinopyroxene, plagioclase and opaque

oxides set in a groundmass composed of the above minerals plus alkali-

feldspar.

Leucite is present as subhedral macrophenocrysts (- 46.0mm).

Irregularly distributed inclusions of clinopyroxene, plagioclase and opaque

oxides are present in some of the leucite macrophenocrysts. A second

population of leucite microphenocrysts (—+0.35mm) is present. These

crystals are subspherical and may carry radially distributed turbid inclu-

sions. An-euhedral phenocrysts (—+1.0mm) of clinopyroxene show green-

yellow pleochroism and zoning is present in certain phenocrysts.Plagioclase

 206

forms sub-euhedral phenocrysts (—+1. 0mm) . They may be zoned and/or

mantled by alkali-feldspar. Anhedral opaque oxides (—+0.5mm) are

present and apatite forms accessory microphenocrysts. Glomeroporphy-

ritic clots of clinopyroxene and/or plagioclase and/or leucite and/or opa-

que oxides are present.

The groundmass (<0.1mm) is composed of poorly formed alkali-

feldspar laths along with lesser amounts of turbid devitried glass, granu-

lar to prismatic clinopyroxene, plagioclase laths, sub-spherical leucite

and disseminated opaque oxides.

Petrographic variations: lava 205 is more porphyritic than 4001,

4004 and 3401. Glomeroporphyritic clots are also well developed in 205

and sanidine laths (—>0. 75mm) are present in this lava. 3401 is noted

for increased concentrations of plagioclase, groundmass alkali-feldspar

and accessory apatite, as well as lower concentrations of leucite. 3401

and 4004 are distinguished by the presence of biotite laths (-41.0mm).

(heavily to completely altered to opaque oxides) as well as fresh inter-

stitial mica. Occasional brown amphibole. crystals (< 0.15mm) are also

present in 3401.

LV-TRAB group (trachybasalt)

This lava (6701) is composed of phenocrysts of clinopyroxene

2.0mm, pale green - pale yellow green pleochroism) , sub-anhedral oli-

vine (- p1.0mm, marginal iddingsite absent) , and corroded plagioclase

(— ) 0.75mm) (An 88) . The phenocrysts may form glomeroporphyritic

clots. The groundmass (<0.1mm) is composed of laths of plagioclase

 207

and granular clinopyroxene as well as disseminated opaque oxides.

LV-QTR group (trachyte)

This lava (202) is composed of phenocrysts of sanidine, plagioclase,

clinopyroxene, biotite and opaque oxides set in a trachytic groundmass.

Sub-euhedral sanidine is the dominant phenocryst (—p3.0mm)*. Plagio-

clase (An 74) forms sub-euhedral phenocrysts (---+ 3.0mm) which may be

zoned and/or mantled by alkali-feldspar. Spongy core textures (often

infilled by brown glass) are common in the plagioclase phenocrysts.

Clinopyroxene forms subhedral laths (—+0.75 mm) exhibiting pale green -

yellow green pleochroism. Euhedral biotite laths 2.0mm) show fresh

margins unaffected by the alteration seen in biotites in lava 401 (below) .

Opaque oxides form sub-anhedral phenocrysts (-40.75mm) . Apatite

forms occasional microphenocryst laths (—+0.2mm) . Glomeroporphyritic

clots composed of phenocrysts of plagioclase, clinopyroxene, biotite and

opaque oxides are present.

The trachytic groundmass is composed of orientated alkali-feldspar

laths as well as lesser amounts of prismatic pyroxene and sparse laths of

plagioclase (mantled by alkali-feldspar). Devitrified glass fills inter- -

stitial areas between the groundmass laths.

Samples collected from the same locality as those analysed by

Schneider (1965) and Trigila (1969b) (LV-QTR analyses listed in
Chapter 6, Table 6.8B, Nos. 8101, 8102, 8095) carry ubiquitous
sanidine macro-phenocrysts (-+10.0mm) and this gives these
LV-QTR samples a coarse grained texture.
 208

LV-MTR group (trachyte)

This lava (401) is composed of phenocrysts of plagioclase, sanidine,

clinopyroxene, biotite and opaque oxides set in a trachytic groundmass.

Euhedral plagioclase (An 82; -+3.0mm) is the dominant phenocryst phase.

They may be zoned and/or mantled by alkali-feldspar. Sanidine forms

occasional laths (--+1.0mm) noted for rounded and apparently resorbed

margins. Subhedral clinopyroxene ( -)0.75mm) exhibit pale green -

pale yellow green pleochroism. Biotite laths (- +2.0mm) show corroded

margins altered to opaque oxides. In the smaller biotite laths the entire

mineral may be altered to a network of granular (< 0.1 mm) opaque oxides.

Sub-anhedral phenocrysts of opaque oxides (-30.5mm) are present.

With the exception of sanidine the phenocrysts may form glomeroporphyritic

clots.

The groundmass is dominated by orientated alkali-feldspar laths.

Prismatic pyroxene, disseminated opaque oxides and sparse laths of plagio-

clase (mantled by alkali-feldspar) are also present.

 209

APPENDIX D

VULSINI PUMICE AND LAVA ANALYSES

This appendix lists geochemical data for the Vulsini pumice and

lava samples analysed by the author (Table D.1 and D.2). Vulsini

lava analyses taken from the literature are also listed in Table D.2.

The X-ray fluorescence techniques used by the author in the analysis of

major and trace elements are described in Chapters 2 and 3. Wet

chemical techniques were used in the analysis of Na2O and FeO (see

Appendix E). The samples analysed in this work have been corrected

for leucite to analcime alteration and for excess H20+ due to hydration

of pumice glass (see Appendix A and discussions in Chapter 6.1.3).

* analyst: Mr. P. Watkins (Dept. of Geology, Imperial College)

 210
Table D.1 Vulsini pumice analyses

PF-2 A B
3701 2911 3707 4103 3709 3801 3708 3901 3712 4101

SI02 61.14 61.67 61.79 61.79 61.60 61.33 61.30 58.51 57.98 57.94
1IO2 .32 .33 .33 .32 .31 .32 .33 .44 .46 .45
AL203 18.73 18.34 18.44 18.34 18.51 18.34 18.56 19.14 19.50 19.63
FE203 1.05 1.28 .82 1.11 1.49 1.07 1.17 1.93 1.83 1.79
FEO 1.34 1.15 1.54 1,42 1.13 1.37 1.36 .76 1.01 1.08.
MNO .15 .15 .15 .15 .14 .15 .15 .12 .14 .14
MOD .20 .27 .28 .52 .30 .30 .48 .26 .35 .40
CAO 2.05 2.19 2.08 2.13 2.17 2.19 2.20 2.64 2.94 2.95
NA20 4.36 4.29 4.12 4.08 4.30 4.22 4.14 3.06 3.51 3.68
K20 8.32 8.64 8.50 8.64 8.66 8.72 8.51 10.08 10.32 10.12
P205 .07 .08 .09 .08 .08 .08 .09 •06 .01 .07
H20- .47 .22 .26 .17• .26 .19 .46 1.08 .29 .33
LOI 1.05 - 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05
TOTAL 99.25 99.66 99.45 99.80 100.00 99.33 99.80 99.13 99.39 99.63

F3/F2-1-F3 .44 .53 .35 .44 .57 .44 .46 .72 .64 .62
K20/NA20 1.91 2.01 2.06 2.12 2.01 2.07 2.06 3.29 2.94 2.75
M00/K20 .02 .03 .03 .06 .03 .03 .06 .03 .03 .04
F3-1-F2/CA 1.32 1.23 1.32 1.35 1.33 1.27 1.30 1.08 1.04 1.05
K20/CAO 4.06 3.95 4.09 4.06 3.99 3.98 3.87 3.82 3.51 3.43
CAO/MGO 10.25 8.11 7.43 4.10 7.23 7.30 4.58 10.15 8.40 7.38
F3/F2+F3 = FE203/(FE0 + FE203)
F31-F2/CA = (FE3+ + FE2+)/CA2+

• V 30 36 38 38 41 43 41 54 •63 64
CR • 5 3 9 6 9. 7 • 8 7 7 7
CO 0 1 0 0 2 1 2 1 1 1
NI 7 4 5 5 7. 6 6 5 5 8
CU 9 9 5 1- 1 6 5 5 2 5
ZN 89 91 85 85 81 86 86 73 89 88
ZR 727 732 716 674 •
664 659 665 689 725 709
RB 660 686 675 655 663 667 653 469 459 492
SR 200 250 229 282 317 280 312 405 386 395
TH ' 119 121 116 109 113 105 100 138 140 133
Y 65 68 69 69 62 62 60 52 53 57
. BA 68 97. 103 84 118 57 115 108 103 86

K/RB 105. 105. 105. 110. 108. 109. 108. 178. 187. 171.
RB/SR 3.300 2.744 2.948 2.323 2.091 2.382 2.093 1.158 1.189 1.246
TH/K .0017 •.0017 .0016 .0015 .0016 .0015 .0014 .0016 .0016 .0016

DI 85.66 86.91 85.59 85.79 86.89 86.22 85.24 82.57 83.58 83.53
OZ 0 0 .29 0 0 0 0 0 0 0
ZR .14 .14 .14 .13 .13 .13 .13 .14 .14 .14
OR 49.37 51.27 50.44 51.27 51.39 51.74 50.50 59.72 61.13 59.96
PL 42.42 40.01 41.47 40.63 39.98 38.26 41.22 27.91 20.78 21.70
(AB) 35.56 34.85 34.86 34.52 34.46 33.02 34.39 19.26 13.88 14.62
(AN) 6.86 5.16 6.61 6.11 5.52 .5.24 6.82 8.64._ -6.90 -7.08
NE .72 .79 0 0 1.04 1.46 .35 3.59 8.57 8.95
WO 0 .73. 0 0 .64 .09 0 1.01 2.23 1.85
DI 2.50 2.93 2.74 3.36 2.73 4.21 3.00 1.40 1.88 2.19
(WO) 1.22 1.47 1.34 1.69 1.39 2.08 1.51 .75 1.01 1.17
(EN) .35 .67 .38 .79 .75 .75 .72 .65 .87 1.00
(FS) .93 .79 1.02 .88 .59 1.38 .77 0 0 .02
HY 0 0 1.18 .58 0 0 0 0 0 0
(EN) 0 0 .32 .27 0 0 0 0 0 0
(FS) 0 0 .86 .30 0 0 0 0 0 0
OL .42 0 0 .36 0 0 .72 0 0 0
(FO) .11 0 0 .16 0. 0 .33 0 0 0
(FA) .32 0 0 .20 . 0 0 .38 0 0 0
MT 1.52 1.86 1.19 1.61 2.16 1.55 1.70 1.57 2.38 2.60
IL .61 .63 .63 .61 .59 .61 .63 .84 .87 .85
HM 0 0 0 0 0 0 0 .85 .19 0
AP .17 .19 .21 .19 .19 .19 .21 .14 .02 .17

3701 /AV 2 PF-2 3801 /AV 2 IG A

2911 /AV 2 PF-2 3708 /AV 2 IG A
3707 /AV 2 IG A 3901 /AV 1 IG B
4103 /AV 2 IG A 3712 /AV 2 ID B
3709 /AV 2 IG A 4101 /AV 2 IG B
 211

Table D.1 (continued)

B
3713 4007 3902 5501 • 4008 4102 5502 1401 1104 1402

SI02 58.18 58.13 58.10 56.76 57.41 57.49 56.71 59.28 59.35 59.25
1102 .46 - .44 .48 .45 .46 .48 .47 .43 .42 .42
AL203 19.40 19.66 19.19 19.22 18.48 18.69 18.57 19.24 19.30 19.13
FE203 1.83 1.87 1.64 1.74 1.92 1.75 1.83 1.73 2.01 1.93
EEO 1.06 1.02 1.50 1.15 1.45 1.66 1.72 .85 .84 1.01
MNO .14 .13 .13 .14 .11 .12 .12 .12 .12 .11
MGO .20 .52 .35 .21 .53 .84 .68 .44 .22 .26
CAO 2.95 3.10 3.17 3.29 3.65 3.80 4.18 2.33 2.47 2.53
NA20 3.75 3.61 3.69 3.26 2.77 2.79 2.89 3.85 , 3.83 3.57
K20 10.21 9.92 9.94 10.30 10.67 10.29 10.10 9.77 9.68 10.07
P205 .07 .07 .10 .07 .14 .13 .15 .05 .07 .08
H20- .27 .86 .18 .97 .12 .18 .54 .23 .17 .33
LOI 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05
TOTAL 99.57 100.38 99.52 98.61 98.76 99.27 99.01 99.37 99.53 99.74

F3/F2+F3 .63 .65 .52 .60 .57 .51 .52 .67 .71 .66
K20/NA20 2.72 2.75 .2.69 3.16 3.85 3.69 3.49 2.54 2.53 2.82
MG0/K20 .02 .05 -.04 .02 .05 .08 .07 .05 .02 .03
F3+F2/CA 1.06 1.00 1.10 .96 1.01 1.00 .95 1.19 1.22 1.25
K20/CAO 3.46 3.20 3.14 3.13 2.92 2.71 2.42 4.19 3.92 3.98
CAO/M60 14.75 5.96 9.06 15.67 6.89 4.52 6.15 5.30 ._11.23 9.73
F3/F2+F3 = FE203/(FE0 + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V 67 64 69 63 74 81 81 53 60 65
CR 5 7 6 10 6 10 8 6 5 7
CO 1 5 1 ' 3 4 .4 9 2 2 2
NI 6 9 5 8 4 5 6 4 6 5
CU 1 6 6 8 4 4 4 5 11 4
ZN 89 93 83 85 71 73 69 83 82 77
ZR 719 697 621 695 475 485 471 696 647 595
RB 492 475 553 466 430 454 454 627 589 566
SR 390 447 623 487 1013 986 1028 364 463 750
TH 128 140 110 124 75 77 79 193 179 157
Y 56 39 52 49 51 40 40 66 59 40
BA 115 186 232 93 558 564 633 113 144 229

K/RB 172. 173. 149. 183. . 206. 188. 185. 129. 136. 148.
RB/SR 1.262 1.063 .888 .957 .424 .460 .442 1.723 1.272 .755
TH/K .0015 .0017 .0013 .0015 .0008 .0009 .0009 .0024 . .0022 .0019

DI 84.71 82.57 82.99 81.40 80.76 78.89 77.85 85.37 85.28 85.27
ZR .14 .14 .12 .14 .09 : .10 .09 .14 .13 .12
OR 60.49 58.77 58.91 61.01 63.19 60.95 59.83 57.94 57.39 59.69
PL 21.21 23.88 21.99 19.18 17.03 19.24- 18.21 27.62 29.34 26.46
(AB) 15.34 15.81 15.63 11.86 10.62 11.23 10.41 21.35 22.55 20.12
(AN) 5.87 8.07 6.35 7.32 6.41 8.01 7.80 6.26 6.79 6.34
NE 8.88 7.98 8.45 8.52 6.94 6.71 7.61 6.08 5.34 5.47
WO 2.95 1.44 1.99 2.87 2.70 1.02 2.26 .87 1.53 1.74
DI 1.07 2.79 3.48 1.49 3.83 6.46 5.71 2.36 1.18 1.40
(140) -.58 1.50 1.76 -.77 1.99 3.33 2.92 1.27 .63 .75
(EN) .50 1.30 .87 .52 1.32 2.09 1.69 1.10 .55 `.65
(FS) 0 0 .85 .19 .52 1.03 1.09 0 0. 0
MT 2.54 2.44 2.38 2.52 2.78 2.54 2.65 1.88 1.88 2.40
IL .87 .84 .91 .85 .87 .91 .89 .82 .80 .80
HM .08 .19 0 0 0 0 0 .43 .71 .28
AP .17 .17 .24 .17 .33 .31 .36 .12 .17 .19

3713 /AV 2 IG B
4007 /AV 2 I0 0
3902 /AV 2 IG B
5501 /AV 2 IO B
4008 /AV 1 IG B
4102 /AV 2 IG B
5502 /AV 2 IG B
1401 /AV 2 IG C
1104 /AV 2 IG C
1402 /AV 2 IG C
 212
Table D.1 (continued)

C D
1103 7003 2915 1406 1405 1101 1102 501 3715 4301

SI02 59.25 58.85 57.79 56.38 55.62 56.09 55.92 61.93 59.33 60.33
1IO2 .43 .52 .51 .48 .48 .47 .49 .31 .48 .54
AL203 18.95 18.94 18,85 18.92 18.80 19.08 19.03 18.27 19.29 19.34
FE203 2.04 1.68 2.09 2.46 2.76 2.64 2.72 1.33 1.57 1.81
FED 1.06 1.44 1.46 1.62 1.72 1.76 1.82 .94 .99 1.06
MHO .11 .14 _ .12 .12 .13 - .- _.12 ..12 •16.___ _4_13_ .11
MGO .52 .77 .68 .82 1.14 .90 .99 .09 .23 .22
CAO 2.68 3.04 3.18 3.90 4.05 4.11 4.37 2.00 2.43 2.49
NA20 3.44 3.87 3.15 3.38 2.84 2.98 3.04 4.49 4.06 3.17
K20 9.88 9.67 10.34 9.94 9.89 9.74 9.65 8.31 9.43 9.96
P205 .07 .12 .16 .16 .18 .18 .19 .05 '.05 .07
H20- .28 .12 .14 .16 .12 .16 .10 .26 .20 .38
LOI 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1..05 1.05 1.05
TOTAL 99.76 100.21 99.52 99.39 98.78 99.28 99.49 99.19 99.24 100.53

F3/F2+F3 .66 .54 .59 460 .62 .60 .60 .59 .61 .63
K20/NA20 2.87 2.50 3.28 2.94 3.48 3.27 3.17 1.85 2.32 3.14
MGO/K20 .05 .08 .07 .08 .12 .09 .10 .01 .02 .02
F3+F2/CA_1.24 1.14 1.22 1.14 1.20 1.17 1.13 1.24 1.15 1.25
1(20/CAO 3.69 3.18 3.25 2.55 2.44 2.37 2.21 4.16 3.88 4.00
CAO/MG0 5.15 3.95 4.68 4.76 3.55 4.57 4.41 22.22 10.57 11.32
F3/F2-1-F3 = FE203/(FE0 + FE203)
F3-1-F2/CA = (FE3+ + FE2+)/CA2+

V 61 77 81 95 104 106 105 34 57 64•

CR 4 7 7 7 8' 7 6 7 8 6
CO ' 2 0 4 5 6 4 7 0 0 3
NI 5 .5 .3 6 6 6 7 5. 6 5
CU 8 10 5 7 6 5 7 27 40 22
ZN 82 89 76 82 82 81 84 100 104 75
ZR 582 532 443 507 471 462 463 773 707 551
RB 548 430' 401 461 394 380 411 681 630 570
. . SR 790 810 1253 1241 1594 1586 1673 142 222 369
TH 153 105 80 125 104 103 98 125 139 115
. Y 39 55 42 35 34 46 32 65 51 59
BA 270 543 1079 923 1440 1788 1467 106 70 ' 62

K/RB 150. 187. 214. 179. 208. 213. 195. 101. 124. 145.
RB/SR .694. .531 .320 .371 .247 .240 .246 4.796 2.838 1.545
TH/K .0019 .0013 .0009 .0015 .0013 .0013 .0012 .0018 .0018 .0014

DI 83.82 83.40 81.75 79.05 75.63 76.24 75.68 87.64 85.44 84.61
OZ 0 0 0 0 0 0 '0 .32 0 0
ZR .11 .11 .09 .10 .09 •.09 .09 .15 .14 .11
OR 58.56 57.28 61.23 58.89 58.57 57.68 57.16 49.32 55.93 59.04
PL 27.72 23.96 19.95 17.22 18.10 20.54 19.72 43.04 30.26 33.12
(AB) 20.72 18.28 13.26 10.19 8.82 10.68 10.00 37.99 23.80 24.09
(AN) 7.00 5.68 6.69 7.02 9.28 9.86 9.71 5.05 6.46 9.04
NE 4.54 7.84 7.26 9.97 8.24 7.87 8.52 0 5.72 1.48
WO 1.07 .95 1.36 2.26 .78 1.23 1.52 1.30 1.57 .61
DI 2.79 5.37 4.28 5.11 6.74 5.77 6.28 1.26 1.24 1.18
(140) 1.50 2.80 2.25 2.69 3.57 3.03 3.30 .62 .66 .63
(EN) 1.30 1.92 1.69 2.04 2.84 2.24 2.47 .22 .57 .55
(FS) 0 .66 .34 .37 .33 .50 .51 .41 0 0
MT 2.53 2.44 3.03 3.57 4.00 3.83 3.94 1.93 2.22 2.21
IL .82 - .99 .97 .91 .91 .89 .93 .59 .91 1.03
HM .30 0 0 0 0 0 0 0 .04 .29
AP .17 .28 .38 .38 .43 .43 .45 .12 .12 .17

1103 /AV 2 IG C
7003 /AV 2 IG C
2915 /AV 2 IG C
1406 /AV 2 IG C
1405 /AV 2 .IG C
1101 /AV 2 IG C
1102 /AV 2 IG C
501 /AV 2 IG D
3715 /AV 1 ID D
4301 /AV 1 I8 D
Table D.1 (continued) 213

D E-1 . E-2
3006 4305 805 804 2906 2804 2806 2805 901 2501

SI02 59.23 59.40 59.69 59.72 57.90 59.30 59.61 50.88 48.50 48.15
1IO2 .47 .50 • .45 .47 .45 .48 .50 .50 .81 .75
AL203 19.11 19.03 18.75 18.55 18.97 18.20 18.09 18.22 18.41 17.63
FE203 1.56 1.78 1.84 1.84 1.59 2.12 1.86 2.06 5.92 4.69
FEO .98 1.01 1.34 1.58 1.05. 1.12 1.87 1.89 2.26 2.66
MNO .14 .12 .10 .10 .14 .12 .09 .10 .14 .17
MGO .42 .31 .64 .71 .62 .53 .94 1.09 3.42 2.88
CAO 2.57 2.62 2.80 2.83 3.07 3.12 3.39 3.57 8.57 8.89
NA20 4.21 3.48 3.07 2.62 4.21 2.97 2.88 2.62 .99 1.72
K20 9.65 9.70 10.20 10.11 9.00 10:06 9.20 9.60 8.12 8.22
P.205 .07 .07 .12 .14 .10 .11 .20 .22 .51 .91
H20- .20 .34 .07 .06 .67 .30 .10 .11 .65 1.08
L0I 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.41 1.60
TOTAL 99.66 99.41 100.12 99.98 98.82 99.48 99.78 99.91 99.71 99.35

F3/F2+F3 .61 .64 .58 .54 .60 .65 .50 .52 .72 .64
K20/NA20 2.29 2.79 3.32 3.59 2.14 3.39 3.19 3.66 8.20 •4.78
MGO/K20 .04 .03 .06 .07 .07 .05 .10 .11 .42 .35
F3+F2/CA 1.07 1.15 1.25 1.34 .94 1.12 1.23 1.23 1.01 .89
K20/CAO 3.75 3.70 3.64 3.57 2.93 3.22 2.71 2.69 .95 .92
CAO/M0O 6.12 8.45 4.37 3.99 4.95 5.89 3.61 3.28 2.51 3.09
F3/F2-1-F3 = FE203/(FE0 + FE203)
F3-1-F2/CA = (FE3+ + FE2+)/CA2+

V 62 59 74 72 67 73 82 87 218 221
CR 5 6 6 7 7 9 11 11 16 20
CO 0 1 4 2 2 2 6 2 26 21
NI . 5 6 6 8 5 7 7 6 26 15
CU 81. 14 6 5 30 21 19 24 18 19
ZN 128 79 61 60 95 72 69 72 80 86
ZR 710 585 430 457 700 469 408 383 321 330
R8 638 598 556 550 636 545 518 503 653 611
SR 205 . 439 804 841 227 840 92A 1074 1461 1456
TH 145 110 89 89 142 85 84 65 38 46
Y 61 61 32 35 50 46 34 32 38 42
BA 55 76 260 223 181 258 564 610 904 941

K/R8 126. 135. 152. 153. 117. 153. 147. 158. 103. 112.
RB/SR 3.112 1.362 .692 .654 2.802 .649 .559 .468 .447 .420
TH/K .0018 .0014 .0011 .0011 .0019 .0010 .0011 .0008 .0006 .0007
•

DI 86.19 84.00 83.51 82.34 82.31 82.68 79.03 78.31 50.29 53.69
OZ 0 0 0 0 0 0 .13 0 0 0
ZR .14 .12 .09 .09 .14 .09 .08 .08 .06 .07
OR 57.23 57.51 60.45 59.92 53.38 59.62 54.53 56.89 36.85 35.04
PL 25.73 30.55 26.78 28.72 27.20 27.14 33.55 30.06 21.70 16.01
(AB) 21.08 22.99 19.61 20.71 21.01 20.60 24.37 20.54 0 0
(AN) 4.65 7.56 7.17 8.01 6.18 6.53 9.18 9.52 21.70 16.01
LC 0 0 0 0 0 0 0 0 8.89 10.77
NE 7.88 3.50 3.45 1.71 7.91 2.45 0 .88 4.54 7.88
WO 2.01 1.25 .43 0 1.65 2.04 0 0 0 1.14
DI 2.26 1.67 4.16 4.38 3.58 2.85 5.46 5.81 14.11 15.64
(WO) 1.21 .89 2.18 2.26 1.90 1.53 2.81 3.01 7.57 8.38
(EN) 1.05 .77 1.59 1.46 1.54 1.32 1.73 1.98 6.54 7.17
(FS) 0 0 .38 .65 .13 0 .92 .82 0 .09
HY 0 0 0 0 0 0 .93 0 0 0
(EN) 0 0 0 0 0 0 .61 0 0 0
(FS) 0 0 0 0 0 0 .32 0 0 0
OL 0 0 0 .32 0 0 0 .75 1.38 0
(FO) 0 0 0 .22 0 0 0 .52 1.38 0
(FA) 0 0 0 .11 0 0 0 .24 0 0
MT 2.25 2.20 2.67 2.67 2.31 2.61 2.70 2.99 5.39 6.80
IL .89 .95 .85 .89 .85 .91 .95 .95 1.54 1.42
HM .01 .26 0 0 0 .32 0 0 2.20 0
AP .17 .17 .28 .33 .24 .26- .47 .52 1.21 2.16
3006 /AV 2 IG 0
4305 /AV 2 IG D
805 /AV 2 IG D
804 /AV 2 IO D
2906 /AV 2 I0 D
2804 /AV 2 IG E-1
2806 /AV 2 IG E-1
2805 /AV 2 IG E-1
901 /AV 2 IG E-2
2501 /AV 2 IG E-2
 214
Table D.1 (continued)

E-2 F X
4201 101 .6101 305 303 3003 3002 3016 6502 6003

SI02 48.97 49.83 50.12 49.99 49.26 48.48 48.56 48.77 60.74 60.02
TI02 .82 .78 .77 • .78 .78 .82 .82 .82 .36 .37
AL203 17.62 18.44 18.48 18.47 18.12 16.83 16.93 17.06 18.21 18.33
FE203 4.88 4.79 4.94 4.50 4.19 4.65 4.78 4.36 1.73 1.72
FEU 3.07 .2.77 2.66' 2.91 3.25 3.47 3.36 3.72 1.21 1.43
MNO .14 .15 .15 .15 .15 .14 .14 .15 .14 .10
MOO 3.58 3.14 3.16 3.11 3.12 4.61 4.62 4.80 .52 .68
CAO ' 9.06 7.81 8.20 8.31 8.39 10.04 10.06 10.10 2.43 2.77
NA20 1.17 2.00 2.03 1.84 1.84 •1.15 1.44 1.55 3.49 2.81
K20 7.83 7.87 8.58 8.36 8.27 7.25 7.03 7.44 9.38 10.10
P205 .55 .57 .58 .56 .60 .51 .52 .53 .13 .18
H20- .57 .35 .25 _.38 .37 .58 .59 .41 .11 .21
LOI 1.18 1.39 .06 .37 .78 1.31 1.29 .75 1.05 1.05
TOTAL 99.44 99.89 99.98 99.73 99.12 99.84 100.14 100.46 99.50 99.77

F3/F24-F3 .61 .63 .65 .61 .56 .57 .59 .54 .59 .55
K20/NA20 6.69 3.94 4.23 , 4.54 4.49 6.30 4.88 4.80 2.69 3.59
MGO/K20 .46 .40 .37 .37 .38 .64 .66 .65 .06 .07
FM-F2/CA .95 1.05 1.00 .97 .98 .89 .89 .89 1.32 1.26
K20/CAO .86 1.01 1.05 1.01 .99 .72 .70 .74 3:86 3.65
CAD/MGO 2.53 2.49 2.59 2.67 2.69 2.18 2.18 2.10 4.67 4.07
F3/F2+F3 = FE203/(FE0 + FE203)
FM-F2/CA = (FE3+ + FE2+)/CA2+

V 210 212 213 226 224 228 219 . 232 50 . 63

CR 17 11 11 14 17 42 38 41 5 10
CO - 20 27 25 25 37 37 34 3 6
NI 22 14 13 13 14 40 37 38 7 8
CU 37 41 9 .27 13 54 55 50 4 9
ZN 85 98 78 88 84 76 77 80 77 67
ZR 316 333 335 315 328 288 294 305 579 422
RB 704 402 224 728 796 553' 555 558 624 580
SR 1535 1415 1496 1512 1490 1313 1338 1428 487 886
TH. 39 48 41 44 39 30 31 30 89 67
Y 36 40 36 42 38 39 29 31. 58 48
BA 849 1020 1116 1072 868 715 671 928 276 600

K/RB 92. 163. 318. 95. 86. 109. 105. 111. 125. 145.
RB/SR .459 .284 ' .150 .481 .534 .421 .415 .391 1.281 .655
TH/K .0006 .0007 .0006 .0006 .0006 .0005 .0005 .0005 .0011 .0008

DI 49.92 55.66 57.11 55.60 54.53 45.58 46.06 47.13 84.74 82.83
ZR .06 .07 .07 .06 .07 .06 .06 .06 .11 .08
OR 37.50 45.97 37.00 38.19 35.08 30.46 31.26 25.04 55.63 59.87
PL 19.59 18.03 15.94 17.33 16.63 .19.26 18.88 17.53 34.84 29.47
(AB) 0 0 0 0 0 0 0 0 28.62 21.99
(AN) 19.59 18.03 15.94 17.33 16.63 19.26 18.88 17.53 6.22 7.48
LC 7.06 .53 10.80 8.97 11.01 9.85 8.21 14.98 0 0
NE 5.36 9.17 9.31 8.43 8.43 5.27 6.60 7.11 .49 .97
WO 0 0 0 0 0 0 0 0 .27 0
DI 17.56 13.76 16.86 16.43 17.22 21.90 22.17 23.57 3.66 4.43
(W0) 9.36 7.37 9.04 8.74 9.07 11.60 11.77 12.42 1.90 2.29
(EN) 7.75 6.29 7.81 7.12 6.84 9.15 9.47 9.44 1.30 1.47
(FS) .45 .10 0 .57 1.32 1.15 .93 1.70 .46 ...68
OL .87 1.09 .04 .47 ..79 1.86 1.58 2.11 0 .24
(FO) .81 1.07 .04 .44 .65 1.63 1.43 1.76 0 .16
(FA) .05 .02 0 .04 .14 .22 .15 .35 0 .08
MT 7.08 6.95 6.83 6.52 6.08 6.74 6.93 6.32 2.51 2.49
IL 1.56 1.48 1.46 1.48 1.48 1.56 1.56 1.56 .68 .70
HM 0 0 .23 0 0 0 0 0 0 0
AP 1.30 1.35 1.37 1.33 1.42 1.21 1.23 1.26 .31 .43

4201 /AV 2 IO E-2

101 /AV 1 IG F
6101 /AV 2 IG F
305 /AV 2 IG F
303 /AV 1 IG F
3003 /AV 2 IG F
3002 /AV 2 IO F
3016 /AV 2 IG F
6502 /AV 1 IG X
6803 /AV 2 IO X
 Table D.1 (continued) 215

X Y. PF-4
6801 6809 6805 2910 6904 6903 6908 2402 . 2403 2401

8102 60.13 59.84 60.00 59.42 60.25 59.18 59.10 58.65 58.41 58.69
1IO2 .37 .39 .37 .40 .36 .46 .45 .49 .50 .50
AL203 18.34 18.52 18.23 18.03 19.73 19.21 19.39 19.19 19.39 19.47
FE203 2.05 3.30 1.52 2.35 1.85 3.02 1.50 1.53 1.46 1.52
FE0 1.23 - 1.55 1.26 .62 - 1.27 1.22 1.25 1.30
MO .11 .1 1 .10 .11 .19 .13 .16 .14 .14 .14
MGO .82 .92 .69 .89 .32 .36 .45 .42 .33 .30
CAO 2.77 2.87 2.93 3.05 2.24 2.59 2.62 2.58 2.58 2.64
NA20 2.91 2.97 2.84 2.54 4.66 4.07 4.26 3.99 4.00 3.58
K20 10.01 9.71 10.06 10.13 8.51 8.93 9.01 9.88 9.84 9.91
P205 .18 .19 .19 .20 .06 .12 .05 .08 .06 .08
H20- .11 .12 .10 .10 .28 .42 .16 .17 .15 .35
LOI 1.05 1.05 1.05 1.05_ 1.05 1.05 1.05 1.05 1.05 1.05
TOTAL 100.08 99.99 99.63 99.53 100.12 99.54 99.47 99.39 99.16 99.53

F3/F2+F3 .63 - .50 .65 .75 - .54 .56 .54 .54

K20/NA20 3.44 3.27 3.54 3.99 1.83 2.19 2.12 2.48 2.46 2.77
MG0/K20 .08 .09 .07 .09 .04 .04 .05 .04 .03 .03
F3+F2/CA 1.28 - 1.17 1.27 1.16 - 1.17 1.18 1.16 1.18
K20/CAO 3.61 3.38 3.43 3.32 3.80 3.45 3.44 3.83 3.81 3.75
CAO/MGO 3.38 3.12 4.25 3.43 7.00 7.19 5.82 6.14 7.82 8.80
F3/F2+F3 = FE203/(FE0 + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V 59 •
64 64 65 52 64 60 75 74 78
'CR 9 9 9 7 3 5 7 . 4 6 4
CO 0 2 2 3 0 2 5 3 0 0 •
NI 6 8 6 7 6 6 7 6 4 7
CU '9 28 26 13 17 35 19 42 21 6
ZN 71 79 76 73. 115 113 103 109 97. 86
ZR 437 461 424 411 895 722 717 600 589 600
RB. 565 580 567 562 617 512 500 452 452 457
SR 884 809 925 981 275 476 495 617 620 603
TH 65 69 66 68 201 • 157 146 122 117 121
Y 48 54 46 55 71 66 64 51 56 58
BA 666 492 655 700 226 298 361 41 17 141

K/RB 147. 139. 147. 150. 115. 145. 150. 181. 181. 180.
RB/SR .639 .717 .613 .573 2.244 1.076 1.010 .733 .729 .758
TH/K .0008 .0009 .0008 .0008 .0028 .0021 .0020 .0015 .0014 .0015

DI 83.06 82.33 82.50 81.18 85.83 83.94 83.86 85.15 84.78 83.67
ZR .09 .09 .08 .08 .17 .14 _ .14 .12 .12 .12
OR 59.33 57.57 59.63 60.04 50.48 52.93 53.40 58.53 58.29 58.71
PL 30.00 32.76 28.69 28.50 38.20 34.64 30.94 23.39 23.62 26.39
(AB) 22.68 24.33 21.49 20.72 30.52 26.95 23.85 18.19 17.80 18.68
(AN) 7.33 8.43 7.19 7.79 7.69 7.69 7.10 5.20 5.82 7.72
NE 1.06 .43 1.38 .42 4.83 4.06 6.61 8.44 8.69 6.29
WO 0 0 0 .14 .40 .42 ' .55 1.45 1.42 .75
DI 4.43 3.11 5.33 4.78 1.72 1.93 3.63 3.06 2.76 2.68
(140) 2.36 1.67 2.72 2.56 .92 1.04 1.87 1.59 1.41 1.36
(EN) 1.93. 1.44 1.55 2.22 .80 .90 1.12 1.05 .82 .75
(FS) .15 0 1.05 0 0 0 .65 .43 .52 .57
OL .09 .60 .20 0 0 0 0 0 0 0
(FO) .08 .60 .12 0 0 0 0 0 0 • 0.
(FA) .01 0 .09 0 0 0 0 0 0 0
MT 2.97 0 2.20 3.26 1.57 0 2.17 2.22 2.12 2.20
IL .70 .24 .70 .76 .68 .28 .85 .93 .95 .95
HM 0 3.30 0 .10 .76 3.02 0 0 0 0
PF 0 .45 0 0 0 .53 0 0 0 0
AP .43 .45 .45 .47 .14 .28 .12 .19 .14 .19

6801 /AV 2 IG X
6809 /AV 2 IG X
6805 /AV 2 IG X
2910 /AV 3 IG X
6904 /AV 2 IG Y
6903 /AV 3 IG Y
6908 /AV 2 IG Y
2402 /AV 2 PF-4
2403 /AV 2 PF-4
2401 /AV 1 PF-4
 216
Table D.2 Vulsini lava analyses

QTR VTR TRP

8101 8102 8095 202 8121 8105 B207 8111 8206 8020

SI02 62.10 62.40 62.24 60.42 56.40 58.00 58.00 57.70 55.07 55.42
1IO2 .46 .48 .48 .65 .57 .52, .82 .52 .82 .49
AL203 17.30 16.80 17.18 17.54 20.10 18.50 19.11 18.70 20.83 20.17
FE203 2.65 1.60 2.76 2.13 3.30 3.25 3.55 2.70 2.12 2.21
FED 2.20 3.20 2.34 2.23 1.50 1.20 1.00 1.75 1.99 1.61
MNO .13 .12 .14 .10 .14 .12 - .12 .12 .14
MGO 1.00 .95 .91 1.17 1.40 1.80 1.05 1.60 1.00 .71
CAO 3.00 3.10 3.21 3.33 3.60 3.70 3.76 4.20 3.37 3.57
NA20 2.80 2.70 2.79 2.96 2.70 2.90 2.84 3.00 4.00 3.84
K20 7.20 7.40 7.31 8.05 9.30 8.50 8.86 8.40 8.65 9.29
P205 .20 .22 .24 .23 .27 .25 .20 .23 .19 .12
H20- - - - .10 - - .11 - .59 -
L0I/H2O+ .70 1.00 .81 .48 1.10 1.0.0 .54 .70 .77 .89
TOTAL 99.74 99.97 100.41 99.39 100.38 99.74 99.92 99.62 99.52 98.46

F3/F2+F3 .55 .33 .54 .49 .69 .73 .78 .61 .52 .58
K20/NA20 2.57 2.74 2.62 2.72 3.44 2.93 3.12 2.80 2.16 2.42
MGO/K20 .14 .13 .12 .15 .15 .21 . .12 .19 .12 .08
F3+F2/CA 1.79 1.80 1.76 1.47 1.42 1.27 1.26 1.15 _1.36 1.17
K20/CAO 2.40 2.39 2.28 2.42 2.58 2.30 2.36 2.00 2.57 2.60
CAO/MGO 3.00 3.26 3.53 2.85 2.57 2.06 3.58 2.62 3.37 5.03
F3/F2+F3 = FE203/(FE0 + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V 73
CR 5
CO 4
Ni
CU 11
ZN 75
ZR 457
RB 469 388
SR 681 1218 1610
TH 90 74
Y -~ - 47
BA - - 895 1 -
2240

•
.K /RB - - 142. 199.
RB/SR - .689 .319
TH/K - - .0013 .0010

DI 76.58 76.15 76.95 76.99 74.93 74.77 76.21 74.22 77.75 77.50
OZ 10.34 9.57 10.15 4.22 0 0 0 0 0 0
CO 0 0 0 0 0 0- 0 0 .72 0
ZR 0 0 0 .09 0 0 0 0 0 0
OR 42.55 43.73 43.20 47.72 55.08 50.23 52.36 49.64 51.12 54.90
PL 37.07 34.71 36.37 35.77 31.49 36.90 36.88 36.37 29.47 21.27
(AB) 23.69 22.85 23.61 25.05 16.30 24.54 23.65 23.62 18.10 10.90
(AN) 13.37 11.87 12.76 10.72 15.20 12.36 13.23 12.75 11.36 10.36
NE 0 0 0 0 3.55 0 •.21 .95 8.53 11.70
WO 0 0 0 0 0 0 0 0 0 .59
DI .17 1.75 1.30 3.80 1.00 3.40 3.21 5.17 0 4.91
(WO) .09 .87 .66 1.96 .54 1.82 1.72 2.75 0 2.56
(EN) .05 .33 .38 1.23 .46 1.57 1.49 2.23 0 1.77
(FS) .03 .55 .25 .61 0 0 0 .19 0 .58
HY 3.74 5.47 3.12 2.52 0 1.70 0 0 0 0
(EN) 2.44 2.04 1.88 1.69 0 1.70 0 0 0 0
(FS) 1.30 3.43 1.23 .84 0 0 0 0 0 0
DL 0 0 0 0 2.12 .84 .79 1.35 2.63 0
(FO) 0 0 0 0 2.12 .84 .79 1.23 1.75 0
(FA) 0 0 0 0 0 0 0 .12 .89 0
MT 3.84 2.32 4.00 3.09 3.64 2.75 .85 3.91 3.07 3.20
IL .87 .91 .91 1.23 1.08 .99 1.56 .99 1.12 .93
HM 0 0 0 0 .79 1.35 2.97 0 0 0
AP .47 .52 .57 .54 .64 .59 .47 .54 1.94 .28
8101 /AV 1 LV-DTR SCHNEIDER(1965) TABLE 20/A
8102 /AV 1 LV-QTR SHNEIDER(1965) TABLE 20/B
8095 /AV 1 LV-QTR TRIGILA(1969B) TABLE II/LM9
202 /AV 2 LV-QTR THIS WORK
8121 /AV 1 LV-VTR VGLLMER(1975) TABLE 2/VLS-2
8105 /AV 1 LV-VTR SCHNEIDER(1965) TABLE 20/E
8207 /AV 1 LV-VTR WASHINQTON(1965) P.146/6
8111 /AV 1 LV-VTR SCHNEIDER(1965) TABLE 20/L
8206 /AV 1 LV-TRP WASHINGTON(1965) P.146/12
8020 /AV 1 LV-TRP APPLETON(1970) TABLE E3/•AQUA'
 217
Table D.2 (continued)

TR") TEL
8125 8054 8217 4004 4001 205 3401 0057 8055 8058

SI02 55.70 55.00 55.85 54.85 54.30 54.23 56.54 53.20 53.80 52.00
1102 .53 .40 .59 .51 .53 .50 .63 .55 .55 .55
AL203 18.90 20.60 19.34 19.82 19.90 20.32 18.67 19.50 18.90 18.20
FE203 2.40 3.40 3.77 4.14 4.23 3.23 3.65 5.00 4.40 4.00
FEO 1.80 6 0 '1.88
. .27 .34 1.10 1.65 1.10 1.30 2.40
HNO .17 .15 - .15 .15 .13 .13 .14 .14 .12
MGO 1.20 1.20 1.73 .84 1.03 .95 1.18 2.00 2.00 2.80
CAO 3.70 3.80 3.84 4.08 4.12 4.16 4.55 6.30 6.70 .7.60
NA20 4.10 2.70 3.39 3.23 2.80 3.02 2.75 2.00 2.40 2.30
K20 9.80 9.20 8.77 9.21 9.23 9.91 8.72 7.50 7.70 8.40
P205 .16 .13 '.38 .14 .15 .17 .31 .44 .36 .43
H20- - - - .95 .87 .42 .17
L0I/H204 1.40 2.80 1.14 1.56 1.43 1.20 .62 1.90 1.50 .90
TOTAL 99.86 99.98 99.68 99.75 99.08 99.34 99.57 99.63 99.75 99.70

F3/F2+F3 .57 .85 .67 .94 .93 .75 .69 .82 .77 .63
K20/NA20 2.39 3.41 2.59 2.85 3.30 3.28 3.17 3.75 3.21 3.65
MG0/K20 .12 .13 .20 .09 .11 .10 .14 .27 .26 .33
F3+F2/CA 1.25 1.07 1.58 1.08 1.11 1.09 1.24 1.00 .89 .91
K20/CAO 2.65 2.42 2.28 2.26 2.24 2.38 1.92 1.19 1.15 1.11
CAO/1160 3.08 3.17 2.22 4.86 4.00 4.38 3.86 3.15 3.35 2.71
F3/F2+F3 = FE203/(FEO + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V - - - 109 108 103 101

CR - - 7 6. 6 7
CO - .6 5' 7 2
NI - 5 5 8 6
CU - 7 1 13 18 33
ZN - - - 89 102 78 75
ZR - 587 592 555 434
RB 360 - 389 314 491 494
SR 1636 - 1864 1883 1812 1160
TH 142 - - 162 157 143 84
Y - - 53 54 53 34
BA - - 1820 1750 1672 1176

K/RB 226. - - 197. 244. 168. 147.

RB/SR .220 - - .209 .167 .271 . .426
TH/K .0017 - - .0021 .0020 .0017 .0012

DI 80.11 72.75 75.46 74.31 71.99 74.33 72.92 61.04 62.74 60.54

ZR 0 0 0 .12 .12 .11' .09 0 0 0

OR 58.03 54.37 51.82 54.55 54.64 58.72 51.69 44.32 45.50 49.64
PL 11.33 30.02 26.59 23.14 24.25 16.41 31.58 38.56 31.66 15.31
(AB) 7.17 13.10 17.66 10.82 9.83 3.86 18.81 16.48 13.60 .78
(AN) 4.17 16.92 8.92 12.32 14.42 12.54 • 12.77 22.08 18.06 14.53
NE 14.91 5.28 5.97 8.95 7.51 11.75 2.42 .24 3.63 10.12
WO 1.58 0 0 .85 0 .56 .10 0 0 .07
DI 7.88 .84 5.95 4.51 4.66 5.10 6.34 4.90 9.99 15.83
(WO) 4.13 .45 3.19 2.42 2.50 2.74 3.40 2.63 5.36 8.44
(EN) 2.99 .39 2.76 2.09 2.16 2.37 2.94 2.27 4.63 6.97
(FS) .76 0 0 0 0 0 0 0 0 .42
OL 0 1.82 1.09 0 .28 0 0 1.90 .25
(F0) 0 1.82 1.09 0 .28 0 0 1.90 .25 0
•MT 3.48 1.26 4.35 0 .05 2.52 3.92 2.41 3.05 5.80
IL 1.01 .76 1.12 .89 1.01 .95 1.20 1.04 1.04 1.04
HM 0 2.53 .77 4.14 4.20 1.49 .95 3.34 2.29 0
PF 0 0 0 .07 0 0 0 0 0 0
_ _ AP___ ....38 - .31 _ ..._.90 ..33 . _ .36 ___ __ .40_.._ .73 1.04_ _ ._,.85 _. _ 1.02

8125 /AV 1 LV-TRP VOLLMER(1975) TABLE 2/VLS-11

8054 /AV 1 LV-TRP SCHNEIDER(1965) P.394/4
8217 /AV 1 LV-TRP WASHINGTON(1965) P.146/19
4004 /AV 2 LV-TRP THIS WORK
4001 /AV 2 LV-TRP THIS WORK
205 /AV 2 LV-TRP THIS WORK
3401 /AV 2 LV-TRP THIS WORK
8057 /AV 1 LV-TEL SCHNEIDER(1965) P.394/7
8055 /AV 1 LV-TEL SCHNEIDER(1965) P.394/5
8058 /AV 1 LV-TEL SCHNEIDER(1965) P.394/8
 218
Table D.2 (continued)

. TEL
8122 8203 2301 8204 8080 8053 1202 5401 8081 8017

6102 51.10 50.24 49.32 51.21 51.07 49.90 48.77 48.39 48.64 48.77
TI02 .96 1.19 .77 1.43 1.43 .60 .75 .75 1.08 .94
AL203 18.30 18.43 19.41 18.28 19.73 10.90 19.02 19.52 17.79 17.90
FE203 3.40 2.54 3.11 3.07 3.82 4.05 6.72 3.99 3.36 5.36
FED 4.00 5.56 3.99 4.19 2.16 3.40 1.26 3.86 4.60 3.34
MNO .15 - .15 - .08 .15 .16 .18 .11 .15
MGO 3.20 3.65 3.31 3.47 2.76 3.10 3.08 2.60 3.16 4.49
CAO 7.70 7.83 7.84 7.86 7.95 8.10 8.14 8.48 9.50 9.67
NA20 2.40 2.45 2.99 2.49 2.03 1.80 2.42 2.81 1.36 1.90
K20 8.40 7.45 7.11 6.60 6.57 7.60 7.90 7.56 6.80 5.74
P205 .50 .47 .50 .35 .42 .48 .48 .52 .35 .45
H20- - - .18 .16 .14 - .93 .42 $38 0
L0I/H204- .40 . .36 .45 .56 1.25 1.70 .50 .81 2.47 .57
TOTAL 100.51 100.17 99.13 99.67 99.41 99.78 100.13 99.89 99.60 99.28

F3/F2+F3 .46 .31 .44 .42 .64 .54 .84 .51 .42 .62
K20/NA20 3.50 3.04 2.38 2.65 3.24 4.22 3.26 2.69 5.00 3.02
MG0/K20 .38 .49 .47 .53 .42 .41 .39 .34 .46 .78
F3+F2/CA 1.09 1.21 1.03 1.05 .81 1.02 1.00 1.03 .96 .98
K20/CAO 1.09 .95 .91 .84 .83 .94 .97 .89 .72 .59
CAO/MGO 2.41 2.15 2.37 2.27 2.88 2.61 2.64 3.26 3.01 2.15
F3/F2+F3 = FE203/(FE0 + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V - 233 - 198 211 -

CR - 26 - - - 9 4 - -
CO - 20 - - 20 24 - -
NI - 27 - - - 12 9 - -
CU - - 72 - 22 25 - -
ZN - 82 - - 89 98 - 77
ZR - 295 - - - 368 377 - 265
RB 464 - 396 - - 410 400 - 453
SR 1806 - 1464 - - 1591 2008 - 1411
TH 72 62 - - 75 96 - 27
Y - 42 - - 47 51 - 42
BA - 1220 - - 7 1398 1529 - 1126

K/RB 150. - 149. - - - 160. 157. - 105.

RB/SR .257 - - .270 - - - .258 .199 - .321
TH/K .0010 - .0011 - - - .0011 .0015 - .0006 •

DI 58.83 54.66 54.97 54.35 53.44 53.40 55.12 54.18 46.17 45.02
ZR 0 0 .06 0 0 0 .07 .08 0 .05
OR 40.70 41.26 38.06 39.00 38.82 44.91 33.91 28.59 39.04 34.07
PL 14.28 17.29 18.48 27.80 36.91 21.58 17.64 18.26 22.35 28.17
(AB) 0 0 0 8.59 11.59 .53 0 0 0 4.89
(AN) 14.28 17.29 18.48 19.21 25.32 21.05 17.64 18.26 22.35 23.29
LC 7.13 2.17 3.20 0 0 0 10.12 12.72 •.90 0
NE 11.00 11.23 13.71 6.76 3.03 7.96 11.09 12.88 6.23 6.06
DI 17.09 15.10 14.43 14.01 8.86 12.78 15.88 17.24 18.23 17.53
(WO) 8.86 7.72 7.46 7.31 4.75 6.68 8.52 8.92 9.39 9.36
(EN) 5.86 4.40 4.78 5.07 4.11 4.75 7.36 5.76 5.81 7.86
(FS) 2.37 2.98 2.18 1.64 0 1.35 0 2.57 3.02 .30
OL 2.14 5.73 3.65 3.40 1.94 2.73 .22 .75 2.27 2.43.
(F0) 1.48 3.28 2.42 2.51 1.94 2.08 .22 .50 1.44 2.33
(FA) .66 2.45 1.22 .89 0 .65 0 .25 .83 .10
MT . 4.93 3.68 4.51 4.45 3.08 5.87 2.41 5.79 4.87 7.77
IL 1.82 2.26 1.46 2.72 2.72 1.14 1.42 1.42 2.05 1.79
HM 0 0 0 0 1.70 0 5.06 0 0 0
AP 1.18 1.11 1.18 .83 .99 1.14 1.14 1.23 .83 1.07

8122 /AV 1 LV-TEL VOLLMER(1975) TABLE 2/VLS-3

8203 /AV 1 LV-TEL WASHINGTON(1906) P.146/33
2301 /AV 2 LV-TEL THIS WORK
8204 /AV 1 LV-TEL WASHINGTON(1965) 0.10.6/27
8080 /AV 1 LV-TEL MATTIAS(1965) TABLE 5/10
8053 /AV 1 LV-TEL SCHNEIDER(1965) P.394/3
1202 /AV 2 LV-TEL THIS WORK
5401 /AV 3 LV-TEL THIS WORK
8081 /AV 1 LV-TEL MATTIAS(1965) TABLE 5/11 0.68 CO2 INCL IN H20+/l.OZ)
8017 /AV 1 LV-TEL APPLETON(1970) TABLE E2/17
 Table D.2 (continued) 219

TEL
2302 1301 8113 8114- 8059 8015 8109 8108 8005 6001

SI02 48.30 48.35 46.90 48.40 48.60 48.46 46.70 47.00 40.16 48.56
TI02 • .80 .79 .81 .80 '.70 .90 .88 .90 .77 .78
AL203 17.84 16.68 17.40 17.50 16.40 15.63. 17.20 16.30 14.39 13.69
FE203 3.58 5.31 6.10 • 5.50 3.30 3.04 6.85 4.15 5.54 4.15
FED 3.72 2.68 2.22 2.70 4.90 5.18 2.00 4.30 2.31 3.68
MND .15 .15 .15 .16 .14 .15 .17 - .13 .13
MOO 4.84 5.45 6.00 5.00 5.80 6.63 5.80 6.50 7.16 7.40
CAO 10.39 11.08 11.20 11.20 11.60 11.71 12.30 12.50 12.54 12.68
NA20 2.43 1.79 2.30 2.30 1.60 1.49 2.10 1.75 1.21 1.08
K20 6.19 6.48 5.10 5.00 5.10 5.17 4.40 4.80 5.63 4.86
P205 .38 .44 .36 .36 .42 .38 .37 .39 .42 .47
H20- .28 .14 - - - - - - - .60
L0I/H20+ .94 .63 .80 .80_ 1.00 .58 .80 .90 .46 1.26
TOTAL 99.84 99.97 99.34 99.72 99.56 99.32 99.57 99.49 98.72 99.34

F3/F21-F3 .49 .66 .73 .67 .40 .37 .77 .49 .71 .53
.K20/NA20 2.55 3.62 .. 2.22 2.17 __. 3.19 _ -3.47 . _ 2.10 . _ 2.74._-9.65 4.50
MGO/K20 .78 .04 1.18 1.00 1.14 1.28 1.32 1.35 1.27 1.52
F3+F2/CA .79 .77 .78 .78 .81 .81 .75 .76 .66 .69
K20/CAO .60 .58 .46 .45 .44 .44 .36 .38 .45 .38
CAO/MGO 2.15 2.03 1.87. 2.24 2.OQ 1.77 2.12 1.92 .1.75 1.71
F3/F2+F3 = FE203/(FE0 + FE203)
F3-1-F2/CA = (FE3+ + FE2+)/CA2+

V 173 194 - - - - - 194

CR 23 42 - - - 279
CO • 24 25 .. - - - - - - 26
NI 48 40 - - - - - 92
CU 129 56 - - - - - - 83
ZN 67 75 - - • 61 - - 62 60
ZR 304 298 - - - 225 - - 196 197
RB 431 489 - - 430 - - 513 905
SR 1586 1431 - - 1097 - 1194 993
TH 59 54 - - 29 - - 30 29
Y 31 40 - - - 34 - - 38 36
BA 1158 1216 - - - 973 - 1089 951
r.
K/RB 119. 110. - - - 100. - - 91. 45.
RB/SR .272 .342 - - - .392 - - .430 .911
TH/K .0011 .0010 _ - - .0007 - - .0006 .0007

DI 45.16 43.48 39.51 41.48 37.72 36.52 35.46 34.41 36.49 34.10
ZR .06 .06 0 0 0 .05 0 0 .04 .04
OR 24.24 23.73 24.71 29.55 30.14 26.06 25.24 19.23 21.87 29.01
PL 19.42 18.26 22.09 25.70 23.05 20.62 24.51 22.44 17.12 18.30
(AB) 0 0 0 3.04 .55 . 0 0 0 0 .30
(AN) 19.42 18.26 22.09 22.66 22.50 20.62 24.51 22.44 17.12 18.01
LC 9.78 11.54 4.26 0 0 3.63 .60 7.17 9.07 0
NE 11.14 8.21 10.54 8.89 7.04 6.83 9.63 8.02 5.55 4.79
DI 24.17 26.88 24.22 23.78 25.95 28.51 26.54 29.33 33.43 33.37
(140) 12.69 14.42 12.99 12.76 13.49 14.84 14.23 15.46 17.93 17.68
(EN) 9.30 12.46 11.23 11.03 9.08 10.12 12.30 11.72 15.50 13.96
(FS) 2.18 0 0 0 3.38 3.55 0 2.16 0 1.73
OL 2.43 .78 2.60 1.00 5.30 6.21 1.50 3.77 1.64 3.56
(F0) 1.93 .78 2.60 1.00 3.76 4.48 1.50 3.13 1.64 3.13
(FA) .50 0 0 0 1.54 1.73 0 .64 0 .43
MT 5.19 6.84 5.30 6.91 4.78 4.41 4.45 6.02 5.64 6.02
IL 1.52 1.50 1.54 1.52 1.33 1.71 1.67 1.71 1.46 1.48
HM 0 .59 2.45 .74 0 0 3.78 0 1.65 0
AP .90 1.04 .85 .85 .99 .90 .88 .92 .99 1.11

2302 /AV 2 LV-TEL THIS WORK

1301 /AV 2 LV-TEL THIS WORK
8113 /AV 1 LV-TEL SCHNEIUER(1965) TABLE 20/N
8114 /AV 1 LV-TEL SCHNEIDER(1965) TABLE 20/0
8059 /AV 1 LV-TEL SCHNEIDER(1965) P.394/9
8015 /AV 1 LV-TEL APPLETON(1970) TABLE E2/15
8109 /AV 1 LV-TEL SCHNEIDER(1965) TABLE 20/I
8108 /AV 1 LV-TEL SCHNEIDER(1965) TABLE 20/H
8005 /AV 1 LV-TEL APPLETON(1970) TABLE E2/5
6001 /AV 2 LV-TEL THIS WORK


Table D.2 (continued) 220

TEL L
8003 8006 6107 8t20 8118 8131 8130 8215 8072 8079

SI02 49.01 49.34 48.52 48.10 47.10 47.55 48.07 47.39 45.80 47.92
1IO2 .77 .72 .81 .73 • .66 .92 .95 .45 1.11 .78
AL203 13.98 13.32 14.19 18.70 18.60 16.12 15.60 14.79 14.56 15.59
FE203 4.55 4.08 4.43 3.70 5.80 3.02 3.92 3.10 2.91 2.61
FE0 2.92 3.65 3.41 3.90 1.95 4.52 3.31 5.08 5.46 A.19
MMO .14 :13 .14 .13 •.20 .21 .18 - .12 '•12
MGO 7.47 8.06 • 6.66 3.20 3.90 4.36 5.24 6.77 6.78 6.02
CAO 13.27 13.31 13.36 9.10 9.60 10.12 10.68 11.61 12.04 12.04
NA20 1.11 1.11 1.37 2.20 2.20 2.00 1.86 1.49 1.59 1.02
K20 4.80 5.08 5.25 8.60 8.00 8.01 7.89 6.93' 5.84 7.17
P205 .38 .40 .47 .48 .42 .58 .56 .15 .32 .36
H20- - .23 .73 .45 .28 .46 .29
LOI/H20+ .41 .59 .93 1.30 1.30 1.68 1.76 .77 2.20 1.32
TOTAL 98.81 99.79 99.77 100.14 99.73 99.82 100.47 98.81 99.19 99.43

F3/F2-1-F3 .61 .53 .57 .49 .75 .40 .54 .38 .35 .38
K20/NA20 4.32 4.58 3.83 3.91 3.64 4.00 4.24 4.65 3.67 7.03
MGO/K20 1.56 1.59 1.27 .37 .49 .54 •66 .98 1.16 .84
F3+F2/CA .61 .65 .65 ;94 .85 .85 .75 .81 .81 .65
.60
K20/CAO .36 .38 .39 .95 .83 .79 .74 .60 .49
CAO/MGO 1.78 1.65 2.01 2.84 2.46 2.32 2.04 1.71 1.78 2.00
F3/F2+F3 = FE203/(FE0 + FE203)
F31-F2/CA = (FE3+ + FE2+)/CA2+

V 199
CR 208
CO 36
NI 79
CU 100
ZN 59 52 60
ZR 193 196 221
RB 389 485 538 407
SR 1101 1020 1143 2418
TH 21 29 . 32 95 .
Y 35 33 34
BA . 1045 979 917
•
K /RB 102. 87. 81. 175.
RB/SR .353 .475 .471 .168
TH/K .0005 .0007 .0007 .0013

DI 33.52. 33.88 35.33 53.22 50.35 48.94 47.99 39.58' 35.42 39.64'

ZR .04 .04 .04 0 0 O 0 O 0 0
OR 28.23 23.74 21.22 14.75 14.79 12.32 13.46 2.95 4.94 8.08
PL 18.92 16.28 •16.98 15.68 17.25 11.35 10.92 13.20 15.34 16.79
(AN) 18.92 16.28 16.98 15.68 17.25 11.35 10.92 13.20 15.34 16.79
LC .20 5.05 7.82 28.39 25.47 27.45 26.00
29.80 23.19 26.89
NE 5.09 5.09 6.28 10.08 10.08 9.17 8.53 6.83 7.29 4.68
WO 0 0 0 0 .30 O 0 O 0 0
DI 35.14 37.55 36.76 21.80 20.95 28.28 30.27 34.85 33.95 32.45
_ (WO) .18.78 19.90 _ __11.31 __11.24 14.64_ _ 1a.13_._12..66 _16.95
(EN) 15.85 15.77 15.78 7.52 9.71 9..53 12.67 12.32 12.01 11.94
(FS) .50 1.87 1.45 2.96 0 4.10 1.56 4.40 4.28 3.56
OL 2.00 3.41 .62 •.45 0 1.37 .30 4.44 4.76 2.84
(F0) 1.93 3.01 .57 .31 O .93 .26 3.18 3.42 2.14
(FA) .07 .39 .06 .14 O .44 .04 1.25 1.34 .70
MT 6.60 5.92 6.42 5.36 5.02 4.38 5.68 4.49 4.22 3.78
IL 1.46 1.37 1.54 1.39 1.25 1.75 1.80 .85 2.11 1.48
HM 0 0 0 0 2.33 0 0 0 0 .0
AP .90 .95 1.11 1.14 .99 1.37 1.33 .36 .76 .85

8003 /AV 1 LV-TEL APPLETON(1970) TABLE E2/3

8006 /AV 1 LV-TEL 'APPLETON(1970) TABLE E2/6
6107 /AV 2 LV-TEL THIS WORK
8120 /AV 1 LV-L VOLLMER(1975) TABLE 2/VLS-1
8118 /AV 1 LV-L SCHNEIDER(1965) TABLE 20/S
8131 /AV 1 LV-L TRIGILA(1969A) OUADRO 7/L2
8130 /AV 1 LV-L TRIGILA(1969A) RUADRO 7/M2
8215 /AV 1 LV-L WASHINGTON(1906) P.146/39
8072 /AV 1 LV-L MATTIAS(1965) TABLE 5/2 0.56 CO2 INCL IN H20+/L04)
8079 /AV 1 LV-L MATTIAS(1965) TABLE 5/9


Table D.2 (Continued)

221

L BSN TRAB
8002 0074 204 0218 3201 8071 0075 0073 0216 .8063

8IO2 47.69 47.39 47.10 46.24 47.26 46.59 46.43 46.54 44.09 55.30
1IO2 .75 1.32 .81 1.17 .80 1.11 1.24 1.04 .95 .50
AL203 14.64 14.17 15.56 14.42 14.14 12.19 13.57 12.95 12.73 17.60
FE203 5.17 4.12 3.16 4.06 3.53 3.30 2.13 2.37 3.31 3.40
FE0 2.66 4.15 4.58 4.36 4.15 4.81 4.75 5.90 4.35 1.90
MNO .12 .16 .15 - .15 .11 .09 .12 .12'
MGO 6.48 5.63 6.04 6.99 7.21 8.37 8.25 7.38 13.71 4.20
CAO 12.49 12.60 12.67 13.24 13.62 15.70 15.97 16.07 12.95 6.50
NA20 1.00 1.58 1.50 1.65 1.23 1.24 1.80 .92 1.02 2.80
K20 6.92 6.57 6.54 6.37 5.73 4.74 4.90 4.69 3.66 5.10
P205 .44 .45 .45 .41 .44 .31 .26 .38 .23 .22
H20- - .12 .21 .57 .48 - .18 .16 .27
LOI/H20+ .59 1.06 1.01 .78 1.31 .55 .74 .68 1.59 2.00
TOTAL 98.95 99.32 99.78 100.26- 100.05 99.02 100.31 99.20 99.66 99.64

F3/F2+F3 .66 .50 .41 .48 .46 .41 .31 .29 .43 .64
K20/NA20 6.92 4.16' 4.36 3.86 4.66 3.82 2.72 5.10 3.59 1.82
MGO/K20 .94 .86 .92 1.10 1.26 1.77 1.68 1.57 3.75 .82
F31-F2/CA .67 .73 .70 .71 .64 .59 .50 .61 .67 .88
K20/CAO .55 .52 .52 .48 .42 .30 .31 .29 .28 .78
CAO/MGO 1.93 2.24 2.10 1.89 1.89 1.88 1.94 2.18 .94 1.55
F3/F2+F3 = FE203/(FE0 + FE203)
F3+F2/CA = (FE3+ + FE2+)/CA2+

V 198 - ' 191

CR - - 22 - 62
CO - - 29 - 30
NI - 64 - 75
CU - 93 - 121
ZN 72 - 66 - 68
ZR 224 - • 308 280
RB 653 - 558 - 533
SR 1105 - 1278 - 1113
TH 39 - 47 - 42
Y 36 - 42 41
BA 1273 - 1161 - 1130

. K /RB 88. - 97. - 89.

RB/SR .591 - . .437 .479
TH/K .0007 - .0009 - .0009

DI 39.03 40.30 37.77 37.08 33.40 27.65 30.96 25.95 21.74 55.97
OZ 0 0 0 0 0 0 0 0 0 2.14
ZR .05 0 .06 0 .06 0 0 0 0 0
OR 10.27 12.12 2.07 0 4.98 0 0 0 .48 30.14
PL 14.91 12.17 46.32 13.13 16.05 13.70 14.48 17.35 19.35 44.09
(AB) 0 0 0 0 0 0 0 0 0 23.69
(AN) 14.91 12.17 16.32 13.13 16.05 13.70 14.48 17.35 19.35 20.39
LC 24.18 20.94 28.82 29.52 22.78 21.96 22.70 21.73 16.58 0
NE 4.58 7.24 6.88 7.56 5.64 5.68 8.25 4.22 4.68 0
WO .03 1.51 0 0 0 0 0 0 0 0
DI 34.81 34.64 35.41 36.25 39.07 44.89 37.14 43.41 34.15 8.11
(WO) 18.67 18.28 18.47 19.15 20.55 23.60 19.47 22.52 18.12 4.35
(EN) 16.14 14.02 12.88 14.79 15.38 17.53 14.16 14.91 14.47 3.74
(FS) 0 2.33 4.06 2.31 3.14 3.77 3.50 5.99 1.56 .03
HY 0 0 0 0 0 0 0 0 0 6.77
(EN) 0 0 0 0 0 0 0 0 0 6.72
(FS) 0 0 0 0. 0 0 0 0 0 .05
OL 0 0 2.05 2.15 2.21 2.87 5.69 3.51 15.43 0
(FO) 0 0 1.52 1.83 1.81 2.32 4.47 2.43 13.79 0
(FA) 0 0 .53 .32 .41 .55 1.22 1.08 1.64 0
CS 0 0 0 1.24 0 1.75 5.08 1.84 0 0
MT 6.79 5.97 4.58 5.89 5.12 4.78 3.09 3.44 4.80 4.93
IL 1.42 2.51 1.54 2.22 1.52 2.11 2.36 1.98 1.80 .95
HM .49 0 0 0 0 0 0 0 0 0
AP 1.04 1.07 1.07 .97 1.04 .73 .62 .90 .54 .52
8002 /AV 1 LV-L APPLETON(1970) TABLE E2/2
8074 /AV 1 LV-L MATTIAS(1965) TABLE 5/3A
204 /AV 2 LV-L THIS WORK
8218 /AV 1 LV-L WASHINGTON(1906) P.146/41
3201 /AV 2 LV-L THIS WORK
8071 /AV 1 LV-L MATTIAS(1965) TABLE 5/1
8075 /AV 1 LV-L MATTIAS(1965) TABLE 5/4
8073 /AV 1 LV-L MATTIAS(1965) TABLE 5/3
8216 /AV 1 LV-BSN WASHINGTON(1965) P.146/40
8063 /AV 1 LV-TRAB SCHNEIDER(1965) P.394/13
 222
Table D.2 (continued)

TRAB MTR
6701 8092 8124 8093 8110 8123 8060 8014 401

SI02 53.37 52.89 54.25 53.70 52.50 52.20 52.20 52.17 56.65
TI02 .75 .69 .90 .70 .70 .76 .65 .75 .86
AL203 15.85 15.76 15.00 15.74 16.10 16.90 15.70 15.98 17.55
FE203 2.12._ 2.35 -.. _ 1.90 2.03.__.2.20 .___3.30 _._2.3Q____3.D5. _ 2.02
FED 3.60- 4.49 3.90 4.15 3.70 3.00 3.80 3.25 4.33
MNO .12 .12 .12 .12 .13 .13 .12 .11 .12
MGO 7.03 6.95 7.00 6.79 7.50 7.30 7.50 7.46 2.54
CAD 8.87 8.93 8.95 8.96' 9.20 9.40 9,80 10.09 `5.55
NA20 2.52 2.90 2.80 2.73 2.75 2.60 2.45 - 2.45 2.38
K20 4.33 4.34 4.15 4.45 4.20 4.10 4.00 3.78 6.79
P205 .23 .23 .20 .22 .21 .23 .24 .22 • .34
H20- .20 - - - - - - - .07
L0I/H20+ ..65 .56 .20 .58 ..40 .20 .50 .44 .43

TOTAL 99.64 100.21 100.17 100.17 99.59 100.12 99.26 99.75 99.63

F3/F2+F3 .37 .34 .33 .33 .37 .52 .38 .48 .32
K20/NA20 1.72 1.50 1.48 1.63 1.53 1.58 1.63 1.54 2.85
MG0/K20 1.62 1.60 1.69 1.53 1.79 1.78 1.88 1.97 .37
F3+F2/CA .74 .89 .76 .80 .74 .75 .72 .70 1.34
K20/CAD .49 .49 .46 .50 .46 .44 .41 .37 1.22
CAO/MGO 1.26 '1.28 1.28 1.32 1.23 1.29 1.31 1.35 2.19

F3/F2+F3 = FE203/(FE0 + FE203)

F3+F2/CA = (FE3+ + FE2+)/CA2+. •

V 156 132.
CR 233 8
CO 31 8
NI 111 9
CU 49 72
ZN 62 - - - 60 122
ZR 247 - - - 226 346
RB 319 322 - 289 - 390 391
SR 563 • 543 - 662 - 640 791
TH 69 51 - - 49 64 •53
Y 37 - -. - 33 39
BA 543 - - - 640 1039

K/RB 113. - .107. - - 118. - 80. 144.

RB/SR .567 - .593 - - .437 - .609 .494
TH/K .0019 - .0015 - - .0014 . - .0020 .0009

DI 47.01 47.47 48.15 48.16 45.59 44.71 43.05 42.28 61.75

DZ 0 0 - 0 0 0 0 0 0 1.36
ZR .05 0 0 0 0 0 0 .05 .07
OR 25.69 25.65 24.63 26.30 24.82 24.32 23.64 22.46 40.25
PL 40.42 35.79 41.55 37.95 37.01 40.77 37.89 40.10 37.23
(AB) 21.32 18.62 23.31 20.39 17.82 18.48 17.86 18.72 20.14
(AN) 19.10 17.17 18.23 17.55 19.18 22.29 20.03 21.38 17.09
NE 0 3.21 .21 1.47 2.95 1.91 1.55 1.09 0
DI 18.83 20.51 19.94 20.35 19.93 18.13 21.48 21.74 7.14
(WO) 9.89 10.70 10.45 10.63 10.47 9.63 11.28 11.51 3.63
(EN) 7.33 7.48 7.60 7.50 7.77 7.75 8.32 9.05 1.94
(FS) 1.61 2.33 1.89 2.22 1.69 .75 1.88 1.18 1.56
HY .07 0 0 0 0 0 0 0 7.90
(EN) .06 0 0 0 0 0 0 0 4.38
(FS) .01 0 0 0 0 0 0 0 3.52
OL 8.81 9.25 8.78 8.75 9.48 8.10 9.07 7.63 0
(FO) 7.10 6.89 6.89 6.60 7.65 7.31 7.26 6.67 0
(FA) 1.72 2.36 1.89 2.15 1.83 .78 1.81 .96 0
MT 3.07 3.41 2.75 2.94 3.19 4.78 3.33 4.42 2.93
IL 1.42 1.31 1.71 1.33 1.33 1.44 1.23 1.42 1.63
AP .54 .54 .47 .52 .50 .54 .57 .52 .81

6701 /AV 2 LV-TRAB THIS WORK

8092 /AV 1 LV-TRAB TRIGILA(19698) TABLE II/126
8124 /AV 1 LV-TRAB VOLLMER(1975) TABLE 2/VLSLA
8093 /AV 1 LV-TRAB TRIGILA(1969B) TABLE II/S2 .
8110 /AV 1 LV-TRAB SCHNEIDER(1965) TABLE 20/K
8123 /AV 1 LV-TRAB VOLLMER(1975) TABLE 2/VLS-5
8060 /AV 1 LV-TRAB SCHNEIDER(1965) P.394/10
8014 /AV 1 LV-TRAB APPLETON(1970) TABLE E2/14
401 /AV 2 LV -MTR THIS WORK
 223

APPENDIX E

Na2O AND FeO WET CHEMICAL ANALYSIS

The Vulsini pumice and lava Na 2 O and FeO analyses were carried

out by Mr. P. Watkins (Dept. of Geology, Imperial College). Na 2O was

analysed by decomposing the sample with. a HF, HCIO4 and HNO3 solu-

tion followed by determination using an EEL Mk 2 flame photometer and

synthetic standard solutions. A radiation buffer was added to both

sample and standard solutions. The relative standard deviation of

the method is 0.8 - 2.4% while the relative accuracy is estimated as

being a 2.0% (based on the results for standard rock samples) .

FeO was determined, on a separate portion of the sample powder,

by decomposition with a HF/metavanadate solution followed by titration

with K2Cr2O7 . The relative standard deviation of the method is 0.6 -

0.9% and the average relative accuracy for the analysis of chromite free

standards is 2•.1%.
 224

REFERENCES

Abbey, S. (1973) . Studies in "Standard Samples" of silicate rocks and

minerals. Geol. Surv. Canada , Paper 73-36, 25p.,

Abbey, S. (1975a). Studies in "Standard Samples" of silicate rocks and

minerals. Part 4 : 1974 edition of usable values. Geol. Surv.
Canada, Paper 74-41.

Abbey, S. (1975b). 1972 values for geochemical reference samples.

Geochim. et Cosmochim. Acta, 39, 535-537.

Abbey, S. (1977). "Standard Samples" : how "standard" are they?

Geostandards Newsletter, 1, 39-45.
•

Ahrens, L.H. (1957) . A survey of the. quality of some of the principle

abundance data of geochemistry. in Physics and Chemistry of
the earth, Vol.2. Eds: Ahrens, L .H. , Press, F., Rankama, K. ,
Runcorn, S.K. Pergamon Press, London, 259 p.

Alvarez, W. (1972). Rotation of the Corsica - Sardinia microplate.

Nature Phys. Sci ., 235, 103-105.

Appleton, J.D. (1970) . The petrology of the potassium-rich lavas of

the Roccamonfina Volcano, Italy. PhD. thesis, Univ. of Edin-
burgh.

Appleton, J.D. (1972). Petrogenesis of potassium-rich lavas from the

Roccamonfina Volcano, Roman Region, Italy. J•. Petrol., 13 ,
425-456. r.

Bailey, D.K. and Schairer, J.F. (1964). Feldspar-liquid equālibria in

peralkaline liquids : the orthoclase effect. Am. J. Sci., 262,
1198-1206.

Baker, B.H. ,: Goles, G.G.', Leeman, W.P. and Lindstrom, M.M. (1977).
Geochemistry and petrogenesis of a basalt-benmoreite-trachyte
suite from the southern part of the Gregory Rift, Kenya . • Contrib.
Mineral. Petrol . , 64, 303-332.

Baker, I. (1969) . Petrology of the volcanic rocks of Saint Helena

Island, South Atlantic. Bull. Geol. Soc. Am., 80, 1283-1310.
 225

Baker, P.E., Gass, I.G., Harris, P.G. and Le Maitre, R.W. (1964).
The volcanological report of the Royal Society Expedition to Tristan
da Cunha, 1962. Phil.. Trans. Roy. Soc. Lond., ser. A, 256,
439-578.

Barberi, F., Ferrara, G., Santacroce, R., Treuil, M. and Varet J. (1975) .
A transitional basalt-pantellerite sequence of fractional crystal-
lization, the Boina Centre (Afar rift, Ethiopia) . J. Petrol . , 16,
22-56.

Barberi, F., Gasparini, P., Innocenti, F. and Villari, L. (1973) .

Volcanism of the southern Tyrrhenian sea and its geodynamic impli-
cations. J. Geophys. Res., 78, 5221-5232.

Barbieri, M., Penta, A. and Turi, B. (1975) .. Oxygen and strontium.

isotope ratios in some ejecta from the Alban Hills volcanic area,
Roman comagmatic region. Contrib. Mineral. Petrol., 51,
127-133.

Berlin, R. and Henderson, C.M.B. (1968). A reinterpretation of Sr

and Ca fractionation trends in plagioclases from basic rocks.
Earth and Plan. Sci. Let., 4, 79-83.

Berlin, R. and Henderson, C;M.B. (1969). The distribution of Sr and

Ba between the alkali feldspar, plagioclase and groundmass phases
of porphyritic trachytes and phonolites. Geochim. et Cosmochim.
Acta, 33, . 247-255.

Borley, G.D. (1977). Problems of silicate analysis by X-ray fluores-

cence using a fusion technique and a single reference standard.
Technical Report No. XRF-1, Dept. of Geology, Imperial College,
Univ. of London, 7p.

Bowen, N.L. (1937) . Recent high-temperature research on silicates

and its significance in igneous geology. Am. J. Sci., 33, 1-21.

Bryan, W.B., Finger, L.W. and Chayes, F. (1969). Estimating pro-

portions in petrographic mixing equations by least squares approxi-
mation. Science, 163, 926-927.

Buie, B.F. (1941). Igneous rocks of Highwood Mountains, Montana.

Pt. III, Dikes and related intrusives. Bull. Geol. Soc. Am.,
52, 1753-1808.
 226

Burns, R.G. (1970). Mineralogical applications of crystal field theory.

Cambridge Univ. Press, London, 224p.

Burri, C. (1968). Potentielle Mineralbestande kalitrachy-basaltisch-

karbonatischer Syntektika und die Zukunft des Besuvmagmas . •
Geol. Rundsch., 57, 744-765.

Carmichael, I.S.E., Turner, F.J. and Verhoogen, J. (1974). Igneous

Petrology. McGraw-Hill, New York, 739 p.

Chayes, F. (1969). A last look at G-1 - W-1. Carnegie Inst.

Washington Year Book 67, 1967-68, 239-241.

Chayes, F. (1970). Another last look at G-1 - W-1. Internat. Assoc.

Math. Geol. J., .2, 207-209. .

Christie, O.H.J. and Alfsen, K.H. (1977). Data transformation as a

means to obtain reliable consensus values for reference materials.
Geostandards Newsletter, 1, 47-49.

Compston, W., Chappell, B.W., Arriens, P.A. and Vernon, M.J. (1970) .
The chemistry and age of Apollo 11 lunar material. Proc. Second
Lunar Sci. Conf., Geochim. et Cosmochim. Acta, Suppl e 2,
Vol. 2, 1123-1138.

Cox, K. G., Hawkesworth, C.J., O'Nions, R.K. and Appleton, J.D. (1976).
Isotopic evidence for the derivation of some Roman Region volcanics
from anomalously enriched mantle. Contrib. Mineral. Petrol., 56,
173-180.

Cundari, A. (1975). Mineral chemistry and petrogenetic aspects of the

Vico lavas, Roman volcanic region, Italy. Contrib. Mineral.
Petrol., 53, 129-144.

Cundari, A. and Le Maitre, R.W. (1970) . On the petrogeny of the leucite-

bearing rocks of the Roman and Birunga Volcanic Regions. J. Petrol . ,
11, 33-47.

Cundari, A. and Mattias, P.P. (1974). Evolution of Vico Lavas, Roman

Volcanic Region, Italy. Bull. Volc . , 38, 98-114.

Deer, W.A., Howie, R.A. and Zussman, J. (1966) . An introduction to

the rock-forming minerals. Longman, London, 528p.
 227

De Fino, M. and Mattias, P.P. (1965). La lava di Ferento presso -

Viterbo. Period. Mineral., 34, 219-233.

Dickinson, W.R. (1975). Potash-depth (K-h) relations in continental.

margin and intra-oceanic magmatic arcs. Geology, 3, 53-56.

Drake, M.J. and Weill, D.F. (1975). Partition of Sr; Ba, Ca, Y,
En2+, En3+ and other REE between plagioclase feldspar and mag-
matic liquid : " an experimental study. Geochim. et Cosmochim.
Acta, 39, 689-712.

Draper, N.R. and Smith, H. (1966). Applied regression analysis.

Wiley, New York, 407 p.

Dupuy, C. (1968). Rubidium et caesium dans biotite sanidine et verre

des ignimbrites de Toscane (Italie). Chem. Geol., 3, 281-291.

Earle, S. (in prep.). Geological interpretation of the geochemistry of

stream-sediments, waters, and soils in the Bristol district, with
particular reference to the. Mendip Hills, Somerset. Ph.D.
Thesis, Dept. of Geology, Imperial College, Univ. of London.

Ellis, P.J., Copelowitz, I. and Steele, T.W. (1977). Estimation of

the mode by the dominant cluster method. Geostandards News-
letter, 1, 123-130.

Engels, J.C. and Ingamells, C.O. (1977). Geostandards - A new

approach to their production and use. Geostandards Newsletter,
1, 51-60.

Ewart, A., Bryan, W.B. and Gill, j.B. (1973). Mineralogy and geo-
chemistry of the younger volcanic islands of Tonga, S.W. Pacific.
J. Petrol., 14, 429-465.

Fabbi, B.P. (1972). A refined fusion X-ray fluorescence technique,

and determination of major and minor elements in silicate stan-
dards. Am. Mineral., 57, 237-245.

Fairbairn, H.W., Schlecht, W.G., Stevens, R.E., Dennen, W.H.,

Ahrens, L.H. and Chayes, F. (1951). A co-operative investi-
gation of precision and accuracy in chemical, spectrochemical
and modal analysis of silicate rocks. U.S. Geol. Surv. Bull.
980 , 71p.
 228

Flanagan, F.J. (1969). U.S. Geological Survey Standards - II.

First compilation of data for the new USGS rocks. Geochim. et
Cosmochim. Acta, 33, 81-120.

Flanagan, F.J. (1973). 1972 values for international geochemical

reference samples. Geochim. et Cosmochim. Acta, 37, 1189-
1200.

Flanagan, F.j. (1975). Author's reply. Geochim. et Cosmochim.

Acta, 39, 537-540.

Flanagan, F.J. (1976a). 1972 compilation of data on USGS standards.

U.S. Geol. Surv., Prof. Paper 840, 131-183.

Flanagan, F.J. (1976b). G-1 et W-1 : Requiescant in pace: U.S.

Geol. Surv., Prof. Paper 840, 189-192.

Fleischer, M. (1969). U.S. Geol. Survey Standards - I. Additional

data on rocks G-1 and W-1, 1965-1967. Geochim. et Cosmo-
chim. Acta, 33, 65-79.

Frisch, M.J. and Liddiard, L. (1976) . MNF (Mi Nnesota FORTRAN)

reference manual for CDC 6000/7000/Cyber series computers.
Published by University Computer Center, Univ. of Minnesota,
Minneapolis, U.S.A.

Ghiara, M.R. and Lirer, L. (1976). Mineralogy and geochemistry of

the "low potassium" series of the Roccamonfina volcanic suite
(Campania, South Italy). Bull. Volc., 40, 39-56.

Govindarajn, K. and Roelandts, I. (1977). Neutron activation analysis .

of two USGS rock reference samples (Granite G-2, Andesite AGV-1)
and discussion in the light of compiled data. Geostandards
Newsletter, 1, 163-179.

Gunn, B.M. (1967a). Matrix corrections for X-ray fluorescence spec-

trometry by digital computer. Can. Spectrosc., 12, 41-46 and
64.

Gunn, B.M. (1967b). The incident attenuation factor in X-ray emission

analysis and the determination of some heavy elements in standard
rocks. Can. Spectrosc., 12, 163-168.
 229

Gupta, A.K. and Fyfe, W.S. (1975). Leucite survival : the alteration
to analcime. Can. Mineral., 13, 361-363.

Hamilton, D.L. and MacKenzie, W.S. (1965). Phase equilibria

studies in the system NaA1SiO4 (nepheline) - KA1SiO4 (kalsilite)
Si02 - H2O. Mineral. Mag., 34, 214-231.

Harris, P.G. (1967). Segregation processes in the upper mantle.

in Mantles of the Earth and Terrestrial planets (ed: Runcorn, S.K.).
Interscience, New York.

Harvey, P.K., Taylor, D.M., Hendry, R.D. and Bancraft, F. (1973).

Accurate fusion method for the analysis of rocks and chemically
related materials by X-ray spectrometry. X-Ray Spectrom., 2,
33-44.

Hatherton, T. and Dickinson W.R. (1969). The relationship between

andesitic volcanism and seismicity in Indonesia, the Lesser
Antilles, and other island arcs. J. Geophys. Res., 74, 5301-
5310.

Haukka, M.T. and Thomas, I.L. (1977). Total X-ray fluorescence

analysis of geological samples using a low-dilution lithium meta-
borate fusion method,. Matrix corrections for major elements.
X-Ray Spectrom., 6, 204-211.

Heinrich, K.F.J. (1966). p. 1035 in The Electron Microprobe. eds:

Mckinly, T.D. , Heinrich, K.F.J. and Wittry, D.B. Wiley,
New York.

Henderson, C.M.B. (1965) . . Minor element chemistry of leucite and

pseudoleucite. Mineral. Mag. 35, 596-603.

Ingamells, C.O. (1970). Lithium metaborate flux in silicate analysis.

Anal. Chim. Acta, 52, 323-334.

Irving, A.J. (1978). A review of experimental studies of crystal/liquid

trace element partitioning. Geochim. et Cosmochim. Acta, 42,
743-770.

Jenkins, R. and de Vries, J.L. (1967). Practical X-ray spectrometry

(First Edition) . Philips Technical Library, Eindhoven, 184p.
 230

Keith, T.E.C. and Muffler, L.J.P. (1978). Minerals produced during

cooling and hydrothermal alteration of ash flow tuff from Yellow-
stone drill hole Y-5. J. Volc. Geoth. Res., 3, 373-402.

Kerr, P.F. (1959). Optical Mineralogy. McGraw Hill, New York,

442p.

Klug, H.P. and Alexander, L.E. (1973). X-ray diffraction procedures

for polycrystalline and amorphous materials. (2nd Edn.) . Wiley,
New York, 966p.

de Laeter, j.R. and Hosie, D.J. (1977) . Mass spectrometric isotope

dilution analyses of barium in geochemical reference samples.
Geostandards Newsletter, 1, 143-145.

Leake, B.E., Hendry, G.L., Kemp, A., Plant, A.G., Harvey, P.K.,
Wilson, J.R., Coats, J.S., Ancott, J.W., Lunel, T. and Howarth,
R.J. (1969) . The chemical analysis of rock powders by auto-
matic X-ray fluorescence. Chem. Geol., 5, 7-86.

Le Maitre, R.W. (1962). Petrology of volcanic rocks, Gough Island,.

South Atlantic. Bull. Geol. Soc. Am., 73, 1309-1340.

Leoni, L. and Saitta, M. (1976) . Determination of yttrium and niobium

on standard silicate rocks by X-ray fluorescence analyses. X-Ray
Spectrom., 5, 29-30.

Leroux, J., Lennox, D.H. and Kay, K. (1953) . Direct quantitative

X-ray analysis. Anal. Chem., 25, 740-743.

Lipman, P.W. (1965). Chemical comparison of glassy and crystalline

volcanic rocks. U.S. Geol. Surv., 1201-D, D1-D24.

Lipman, P.W. (1967). Mineral and chemical variations within an ash-

flow sheet from Aso caldera, S.W. Japan. Contrib. Mineral.
Petrol., 16, 300-327.

Lipman, P.W., Christiansen, R.L. and O'Connor, J.T. (1966) . A com-

positionally zoned ash-flow sheet in southern Nevada. U.S.
Geol. Surv. Prof. Paper 524-F, 47p.

Lipman, P.W., Christiansen, R.L. and Van Alstine, R.E. (1969).

Retention of alkalis by calc-alkaline rhyolites during crystallization
and hydration. Am. Mineral., 54, 286-291.
 231

Lister, B. and Gallagher, M.j. (1970) . An inter-laboratory survey of

the accuracy of ore analysis. Trans. Instn Min. Metall.,
Section B, 79, B213-B237.

Locardi, E. and Mittempergher, M. (1967) . Relationship between some

trace elements and magmatic processes. Geol. Rundsch., 57,
313-334.

Luth, W.C. (1967). Studies in the system KA1SiO4 Mg2SiO4 - Si02 -

H20. I. Inferred phase relations and petrologic applications.
J. Petrol., 8, 372-416.

Madansky, A. (1959) . The fitting of straight lines when both variables

are subject to error. J. Am. Statist. Assoc., 54, 173-205.

Marinelli, G. and Mittempergher, M. (1966) . On the genesis of

typical Mediterranean (Potassic) suite. Bull. Volc. , 29, 113-
140.

Mattias , P.P. (1965) . Lave dell'apparato Vulsino : M. Jugo -

Commenda - M. Razzano. Period. Mineral., 34, 137-199.

Mittempergher, M. (1965). Volcanism and petrogenesis in the

S. Venanzo area (Italy) . Bull. Volc . , 28, 85-94 .

Nappi, G. (1969). Genesi ed evoluzione della caldera di Latera.

Boll. Serv, Geol. Ital., 90, 61-81.

Ninkovich, D. and Hays, J.D. (1972). Mediterranean Island arcs and

origin of high potash volcanoes. Earth Plan. Sci. Let., 16,
331-345.

Niskanen, E. (1964). Quantitative mineral analysis of kaolin-bearing

rocks by X-ray diffraction. Anal. Chem., 36, 1268-1271.

Noble, D.C. (1967). Sodium, potassium and ferrous iron contents of

some secondarily hydrated natural silicic glasses. Am. Mineral.,
52, 280-285.

Noble, D.C. (1970). Loss of sodium from crystallized comendite welded

tuffs of the Micence Grouse Canyon member of the Belted Range Tuff,
Nevada. Bull. Geol. Soc. Am., 81, 2677-2688.
 232

Noble, D.C. and Hedge, C.E. (1970). Distribution of rubidium

between sodic sanidine and natural silicic liquid. Contrib.
Mineral. Petrol., 29, 234-241.

Norrish, K. and Hutton, J.T. (1969). An accurate X-ray spectrographic

method for the analysis of a wide range of geological samples.
Geochim. et Cosmochim. Acta, 33, 431-453.

Norrish, K. and Taylor, R.M. (1962) . Quantitative analysis by X-ray

diffraction. Clay Minerals Bull., 5, 98-109.

Palme, C. and Jagoutz, E. (1977) . Application of the fundamental

parameter method for determination of major and minor elements on
fused geological samples with X-ray fluorescence spectrometry.
Anal. Chem., 49, 717-722..

Parker, R.J. (1975). The petrology and origin of some glanconitic and
glanco-conglomeratic phosphorites from the South African conti-
nental margin. J. Sediment. Petrol., 45, 230-242.

Parker, R.J. (1977). Factors affecting the quality of major element

rock analysis by X-ray fluorescence combined with flux-fusion
sample preparation. Tech. Rep. No. XRF-2B, Dept. of Geology,
Imperial College, London, 35p.

* Parker, R.J. (1978a) . Quantitative determination of analcime in

pumice samples by X-ray diffraction. Mineral. Mag., 42, 103-
106.

* Parker, R.j. (1978b). An iterative method for determining background

intensities used in XRF calibration lines for flux-fusion silicate
rock analysis. X-Ray Spectrom., 7, 38-43.

Parker, R.J. and Siesser, W.G. (1972). Petrology and origin of some
phosphorites from the South African continental margin. J. Sedi-
ment. Petrol., 42, 434-440.

* Parker, R.j. and Willis, J.P. (1977). Programs SORT, REORD and MW
for major element XRF data processing. Computers and Geo-
sciences, 3, 115-172.

Philpotts, J.A. and Schnetzler, C.C. (1970). Phenocryst-matrix

partition coefficients for K, Rb, Sr and Ba, with applications to
anorthosite and basalt genesis. Geochim. et Cosmochim. Acta,
34, 307-322.

reprints of these articles are enclosed in the pocket on the inside of

the rear cover
 233

Poldervaart, A. and Parker, A.B. (1964). The crystallization index as

a parameter of igneous differentiation in binary variation diagrams.
Am. J. Sci. , 262, 281-289.

Raby, R.E. (1976). SIMPLE (SIMulated PLottEr). ICCC Bulletin,

6.3/6, Imperial College Computer Centre, Imperial College,
Univ. of London, London.

Ridley, K.J.D., Turek, A. and Riddle, C. (1976). The variability of

chemical analyses as a function of sample heterogeneity, and the
implications to the analyses of rock standards. Geochim. et
Cosmochim. Acta, 40, 1375-1379.

Riley, j.P. (1958).. Simultaneous determination of water and carbon

dioxide in rocks and minerals. Analyst, 83, 42-49.

Ringwood, A.E. (1966). The chemical composition and origin of the

Earth. in Advances in earth sciences (ed: Hurley, P.M.).
M.I.T. Press, Boston.

Rose, H.J., Adler, I. and Flanagan, F.J. (1963). X-ray fluorescence

analysis of the light elements in rocks and minerals. Appl.
Spectrosc., 17, 81-85.

Ross, C.S. and Smith, R.L. (1955). Water and other volatiles in
volcanic glasses. Am. Mineral., 40, 1071-1089.

Rittman, A. (1933). Die geologisch bedingte Evolution und Differen-

tiation des Somma-Vesuvmagmas. Zeitschr. Vulk., 15, 8-94.

Rubeska _ , I. (1977) . The state of the art of trace element analysis of

geological samples as derived from results on standard rock sam-
ples. Geostandards Newsletter, 1, 15-20.

Sahama, T.G. (1960). Kalsilite in the lavas of Mt. Nyirangongo

(Belgian Congo) . J. Petrol . , 1, 146-171.

Savelli, C. (1967). The problem of rock assimilation by Somma-

Vesuvius magma. I. Composition of Somma and Vesuvius lavas.
Contrib. Mineral. Petrol., 16, 328-353.

Savelli, C. (1968). The problem of rock assimilation by Somma-

Vesuvius magma. II. Composition of sedimentary rocks and car-
bonate ejecta from the Vesuvius area. Contrib. Mineral. Petrol.,
18. , 43-64.
 234

Schneider, H. (1965) . Petrographie des Latervulkans und die Magme-

nentwicklung der Monti Volsini (Prov. Grosseto, Viterbo und
Orvieto, Italien) . Schweiz. Mineral. Petrog. Mitt . , 45, 331-
445.

Scott, R.B. (1971). Alkali exchange during devitrification and hyd-

ration of glasses in ignimbrite cooling units. J. Geol., 79,
100-109.

Smith, R.L. and Bailey, R.A. (1966) . The Bandelier Tuff, a study of
ash-flow eruption cycles from zoned magma chambers. Bull.
Volc., 29, 83-104.

Sparks, R.S.J. (1974). The products of ignimbrite eruptions.

unpublished Ph.D. Thesis, Imperial College, University of London.

Sparks, R.S.J. (1975). Stratigraphy and geology of the ignimbrites of

Vulsini volcano, central Italy. Geol. Rundsch., 64, 497-523.

Sparks, R.S.J., Self, S. and Walker, G.P.L. (1973). Products of

ignimbrite eruptions. Geology, 1, 115-118.

Strasheim, A. and Brandt, M.P. (1968) . A quantitative X-ray fluo-

rescence method for the determination of minor and trace elements
in geological samples using a correction technique for the matrix
effects. Spectrochim. Acta, 23B, 535-541.

Streckeisen, A.L. (1966). Die klassification der eruptivgesteine.

Geol. Rdsch, 55, 478-491.

Streckeisen, A.L. (1967). Classification and nomenclature of igneous

rocks. Neues Jahrb. Mineral. Abhandl . , 107, 144-214.

Thompson, R.N. (1977). Primary basalts and magma genesis. III

Alban Hills, Roman comagmatic province. Contrib. Mineral.
Petrol., 60, 90-108.

Thornton, C.P. and Tuttle, O.F. (1960). Chemistry of igneous rocks;

1 Differentiation Index. Am. J. Sci., 258, 664-684.

Till, R. (1974). Statistical methods for the earth scientist - an intro-

duction. Macmillan, London, 154 p.
 235

Till, R. (1977). The HARDROCK package, a series of FORTRAN IV

computer programs for performing and plotting petrochemical calcu-
lations. Computers and Geosciences, 3 , 185-243.

Trigila_ , R. (1966). Studio geopetrografico dell' edificio vulcanico di

M. Calvo (Farnese-Viterbo). Period. Mineral., 35, 1023-1095.

Trigila,R. (1969a). Studio geopetrografico del complesso vulcanico

di Latera (Vucani Volsini) . Nota I: Le vulcaniti leucitiche del
medio corso del Fiume Olpeta. Period. Mineral., 38, 155-223.

Trigila, R. (1969b) . Sulla genesi dei magmi a carattere mediterraneo.

Period. Mineral., 38, 625-652.

Trigila, R., Ferrini, V. and Cozzupoli, D. (1971) . Studio geopetro-

grafico del complesso vulcanico di Latera (Vulcani Volsini) . Nota II:
La serie piroclastica di Stenzano fra it F. Olpeta e la Selva del
Lamone. Period. Mineral., 40, 125-169.

Turi, B. and Taylor, H.P. (1976). Oxygen isotope studies of potassic

volcanic rocks of the Roman province, Central Italy. Contrib.
Mineral. Petrol., 55, 1-31.

Turner, F.J. and Verhoogen, J. (1960). Igneous and metamorphic

petrology (second edition). McGraw-Hill, New York, 694 p.

Verdurmen, E.A. Th. (1977). Accuracy of X-ray fluorescence spectro-

metric determination of Rb and Sr concentrations in rock samples.
X-Ray Spectrom., 6, 117-122.

Vistelius, A.B. (1970). Statistical model of silicate analysis and

results of investigation of G-1 and W-1 samples. Internat.
Assoc. Math. Geol. J., 2, 1-14.

Vistelius, A.B. (1971). Some lessons of the G-1 - W-1 investigation.

Internat. Assoc. Math. Geol. J. 3, 323-326.

Vollmer, R. (1975). Rb - Sr and U - Th - Pb systematics of alkaline

rocks : the alkaline rocks from Italy. Ph.D. thesis, Swiss
Federal Inst. of Technology, Zurich.

Vollmer, R. (1976) . Rb - Sr and U- Th - Pb systematics of alkaline

rocks : the alkaline rocks from Italy. Geochim. et Cosmochim.
Acta, 40, 283-295.
 236

Vollmer, R. (1977) . Isotopic evidence for genetic relations between

acid and alkaline rocks in Italy. Contrib. Mineral. Petrol., 60,
109-118. .

Walker, G.P.L. (1972). Crystal concentration in ignimbrites.

Contrib. Mineral. Petrol., 36, 135-146.

Washington, H.S. (1906) . The Roman comagmatic region. Carnegie

Institution of Washington, Publication No.57, 199 p.

Welday, E.E., Baird, A.K., McIntyre, D.B. and Madlem, K.W. (1964).
Silicate sample preparation for light-element analysis by X-ray
spectrography. Am. Mineral., 49, . 889-903.

Williams, P.P. (1959) . Direct quantitative diffractometric analysis.

Anal. Chem., 31, 1842-1844.

Willis, J.P., Ahrens, L.H., Danchin, R.V., Erlank, A.J., Gurney, J.J.,
Hofmeyr, P.K., McCarthy, T.S. and Orren, M.J. (1971). Some
inter-element relationships between lunar rocks and fines, and
stony meteorites. Proc. Second Lunar Sci. Conf. , Geochim. et
Cosmochim. Acta, Suppl. 2, Vol.2, 1123-1138. MIT Press.

Willis, J.P., Fesq, H.W., Kable , E.J.D. and Berg, G.W. (1969).
The determination of barium in rocks by X-ray fluorescence spectro-
metry. Can. Spectrosc., 14, 150-157.

Wolf, J.A., Wright, J.V., Smith, A.L. and Clough, B.J. (in press).
Mixed pumice deposits. Bull. Volc.

Zielinski, R.A. (1975) . Trace element evaluation of a suite of rocks

from Reunion Island, Indian Ocean. Geochim. et Cosmochim.
Acta, 39, 713-734.
 MINERALOGICAL MAGAZINE, MARCH 1978, VOL. 42, PP. I03-6

Quantitative determination of analcime in

pumice samples by X-ray diffraction
R. J. PARKER
Department of Geology, Imperial College, London

SUMMARY. A quantitative X-ray diffraction method has leucite --> analcime alteration, the concentration of
been successfully applied to the determination of anal- analcime in the samples was required. Apart from
cime in pumice rock samples. A calibration line was variable amounts of analcime, the pumice samples
constructed from spiked pumice standards (range: o to are composed essentially of a volcanic glass matrix
42%) and the mean relative error of the standards on the
and minor amounts of feldspar and pyroxene
calibration line was < o'6 %, The major-element compo-
sitions of the samples and standards were available, and phenocrysts. The friable nature of the analcime in
this allowed the total mass absorption coefficients to be these samples results in a significant loss of anal-
calculated. The latter were then used to correct the sample cime on thin-sectioning the sample. This precluded
and standard intensities for absorption effects resulting the use of quantitative point-counting methods
from compositional variations. When compositional data using optical microscopy.
already exist, the calculation of the total absorption Table I lists three leucite and three analcime
coefficient provides a rapid and accurate alternative to analyses. The analcime samples were purified using
direct measurement, or to the use of an internal standard. a combination of hand picking, sieving through
VARIOUS techniques for quantitative X-ray —40o mesh (the analcime was very friable and
diffractometric (XRD) determination of mineral could be gently brushed through the mesh, leaving
concentrations have been discussed by Klug and the other mineral components behind) and electro-
Alexander (1973). If an internal standard is not magnetic separation. Leucite samples were purified
used then a knowledge of the mass absorption by hand picking and electromagnetic separation.
coefficients of the samples and standards is XRD scans were used to check the purity of the
required. These coefficients may be determined mineral separations.
directly on the powders (Leroux et al., 1 953; The XRD determinations were based on a calib-
Williams, 1959; Norrish and Taylor, 1962; ration line constructed from seven standards pre-
Niskanen, 1964). As an alternative to direct pared by spiking a pumice sample with varying
measurement, the absorption coefficients may be weight fractions of purified analcime. The required
calculated, provided that major-element concen- range of the calibration line was o to 4o % analcime.
tration data is available for the powders. While The XRD intensities from the standards were
Klug and Alexander (1973) discuss in detail a large adjusted for absorption effects due to the change in
number of quantitative XRD techniques, they do the total mass absorption coefficient resulting from
not include the method of calculating the absorp- the addition of the spiked analcime. The major
tion correction, and the purpose of this paper is to element compositions of the standard powders
draw attention to the potential of this technique in were used to calculate total mass absorption
quantitative XRD. coefficients using the formula:
The present work is concerned with the quanti-
Pt = W
tative determination of analcime in certain pumice
samples recovered from the Vulsini volcano in where p; = total mass absorption coefficient for the
central Italy. The volcanic rocks from this area are standard at a given wavelength, p; = mass absorp-
noted for their high potassium values and the tion coefficient for pure element i in the standard at
presence of leucite. Leucite readily alters to anal- the given wavelength, and W = weight fraction of
cime even at relatively low surface (25 °C) tem- element i in the standard. The p; were calculated for
peratures (Gupta and Fyfe, 1975). The author is Cu-K0 from an algorithm supplied by K. Norrish to
engaged in a geochemical study on these Vulsini M. T. Frost (pers. comm.). Alternative sources for
pumice samples and in order to correct the these coefficients abound (Heinrich, 1966). The
determined whole-pumice geochemistry for the weight fraction of oxygen in the powder was
© Copyright the Mineralogical Society

104 R. J. PARKER
TABLE I. Leucite and analcime analyses*-all data on moisture (H 2O-)free basis

Leucite Analcime

Sample 2503 4601 4701 Ave. 2502 2802 6506 Ave.

SiO2 55'24 55'41 55'41 55'35 54'78 5505 5505 54'96

TiO2 0-05 0.05 0.05 0.05 0.06 0.05 o-o6 0-06
Al2O3 22.49 22.28 2255 22.44 2185 2I.90 22.11 2I-95
Fe2O3 t 0.41 0.42 0.45 0-43 0.50 0.43 0-48 0.47
MgO 0.44 0.28 0.38 0- 37 0.25 0.21 0.18 0-21
CaO 0.05 0.06 o o6 0.06 0.25 0.50 0.48 0.41
Na2O 0-28 0.36 0.31 0-32 1 1.92 12.19 11.97 12-03
K 2O 20.17 20.32 20.49 20.33 0'97 0.27 0.71 0-65
Rb2O 0,23 0.15 0.16 0.18 0'31 0.20 0.29 0.27
Ign.$ 1.60 1.02 0.62 I-o8 9'32 9.20 9'54 9'35

Total 100.96 100-35 100.48 100.61 100-21 I00'00 100-87 100.36

* Samples analysed using XRF techniques described in Norrish and Hutton (1969), Parker and
Willis (1977), and Parker (1978).
t Total iron.
$ Loss on ignition, 85o C for 3o minutes, calculated as H2O.

The ionic ratios per six oxygens are: leucite Si 2.014, Ti 0.001, Al 0.963, Fe 0.012, Mg 0.020, Ca 0-002,
Na 0.023, K 0.944, Rb 0.004, $ H2O 0.131, Si+Al+Fe = 2.99, K+Na+2Ca+2Mg+Rb = 1.02;
analcime Si 2.032, Ti 0.002, Al 0-957, Fe 0.013, Mg 0-012, Ca 0.016, Na 0.862, K 0.031, Rb croo6, $H20
1.155, Si+Al+Fe = 3.00, K+Na+2Ca+2Mg+Rb = 0.95.

determined by difference and the absorption due to computer program will expedite the calculation of
this element was included in the calculation of pt. the sample pt from the major element composition
The correction for the variation of pt in the of each sample.
standards was as follows. The observed XRD Experimental. A Philips X-ray diffractometer
intensity (lobs) for a given weight fraction of was used with a nickel-filtered Cu tube operating at
mineral x in a standard of coefficient µt, is given by 4o kv and 20 ma. The divergence and scatter slits
Klug and Alexander (1973): lobs. = W . K/px p1, were i° and the receiving slit was 0'2 mm. The X-ray
where K = instrumental constant, W„ = weight intensities were measured on a scintillation counter
fraction of mineral x, and p.= density of mineral x. having a linear response over the observed intensity
Now if we postulate a second, hypothetical stan- range (max. observed intensity < 700 c.p.s.).
dard containing the same weight fraction of All the standard and sample powders (0.5 g
mineral x, but having a different total mass absorp- aliquots) were hand ground in an agate mortar to
tion coefficient pt= then the theoretical intensity pass a-goo mesh nylon sieve. A test powder,
(ltheor) that would be observed in this standard ground to -400 mesh, was found to give the
will be /them = W,0K/pxpt2 and hence Itheor. _ strongest diffraction intensity for analcime when
lobs.11t,/pt 2. Thus standards of dissimilar pit may be compared with coarser grindings (-300 and-170
compared by normalizing the observed inten- mesh). The standards were homogenized using a
sities of the standards with respect to a selected high-speed shaker. By far the greatest variation in
pt. These normalized intensities will then produce the diffracted intensity in repeated analyses of the
a linear relationship when compared to the weight same powder was found to be associated with the
fractions of mineral x in the standards. Note loading of the powder in the diffractometer cavity-
that the pt normalization may be carried out with mount. In order to assist the production of repro-
respect to the p, of any one of the standards. The ducible loadings, the cavity-mount was clamped
unknown samples may then be compared against upside down on a glass plate using a small metal
the normalized calibration, provided the sample Pt clamp attached to a wooden base. The powder was
has been calculated and used to normalize the then pressed into the cavity from the back and a
sample intensity with respect to the calibration small glass slide applied as a backing. The glass
line p,. For large batches of samples a suitable slide was held in place on the back of the cavity-
 DETERMINATION OF ANALCIME IN PUMICE SAMPLES io5

mount by `sticky' tape. The reproducibility of this relative standard deviation (as a percentage of
method of loading will be further discussed in the mean counts). These data were not computed for
results section below. standard 7 because it was only loaded twice. With
XRD scans (7 A-1.5 A) on the pumice powders regard to the other standards, the number of
produced sharp analytical peaks and a generally loadings for each standard are not enough to allow
low and flat background reflecting the dominantly completely reliable relative standard deviation
glassy matrix. Intensity measurements were made (RSD) data to be computed for each individual set
on the analcime 5.6o A line. The leucite lines 5'54 A of standard loadings. However, the narrow range
and 5'39 A are possible sources of interference, but of the computed RSD's, and the absence of large
careful scanning in this region did not detect the changes in RSD with concentration, indicates that
presence of these lines in any of the pumice the mean RSD should be significant. This mean
powders. Initially integrated intensities were mea- value (2•o7 %) is similar to the best mean RSD of
sured by counting the diffracted X-rays while 2•o % reported by Niskanen (1964) for pure milled
scanning over the peak. It was found, however, that quartz loaded ten times in a rotating sample
by carefully setting the goniometer on the 5.7o A holder.
line and counting for a fixed time, good calibration The factors that affect the variability of quanti-
data were produced. All the powders were counted tative XRD data have been discussed by Klug and
in this manner for two consecutive periods of 40 sec Alexander (1973), and the reproducibility of the
and the results averaged. The background inten- standard loadings indicated in Table II is con-
sities were measured for 4o sec either side of the sidered to reflect three factors: fine grain size
5.6o A line (normally ±0.5 20), and the averaged achieved by hand grinding all powders to pass
results subtracted from the peak intensities. Each -40o mesh; the very poor cleavage exhibited by
standard powder was loaded and counted as above analcime meant that preferred orientation effects
at least in duplicate. The net peak intensities from resulting from loading were negligible or absent;
the loadings of a given standard were finally the ease of producing uniform loadings using the
averaged. mounting clamp described above. It may be noted
A reference sample was permanently mounted in that the reproducibility of the loadings would have
a separate cavity-mount and this sample was been further improved by the use of a rotating
diffracted at - 2 hour intervals to provide data to sample holder and a larger primary beam (2-40).
correct the peak-minus-background measurements Furthermore, any preferred orientation effect
for any machine drift. The stability of the X-ray (often present with other minerals) could have
generator and counting circuits were such that been mitigated by using a `rough' pressing sur-
negligible drift occurred within this period. face (ground glass or filter paper, see Norrish
Results and discussion. The XRD calibration data and Taylor, 1962, p. 107).
for the seven standards are presented in Table II. In Table II lists the calculated total mass absorp-
order to evaluate the reproducibility of the sample tion coefficient and the absorption correction
loading technique, the standard deviation for the factor (ACF) for each standard. This last factor is
loadings of each analcime standard were computed simply the pt of the individual standard ratioed
and are listed in Table II with the corresponding against the pt for the first standard. The ACF's were

TABLE II. Calibration data

Std. Load- Mean S.D. R.S.D. E4 A.C.F. Corr. % Total Calc. Abs. Rel.
ings P-B C/4o S P-B Spike anal- anal- error error %*
C/4o S C/4o S cime cime

1 4 2 458 47'0 1.91 57.77 1-000 2458 0.00 5-67 5.62 0.05 0.88
2 4 4762 137.2 2-88 56.53 0'979 4662 5.00 10•67 10.66 0•01 0-09
3 6 7241 85.9 1'19 55'36 0.958 6937 10.00 15.67 1 5'87 0.20 P28
4 6 9626 193.8 2.01 54.04 0'935 9000 15.00 20.67 20.59 o o8 0.39
5 4 13563 296.2 2.18 52.10 0.902 12 234 22.50 28.17 27.98 0-19 0-67
6 7 1 7 992 402.9 2.24 50'25 0.870 1 5 653 30.00 35.67 35'80 0•13 0.36
7 2 21 887 - - 48.63 0.842 18 429 3650 42.17 42'15 0.02 0'05

S.D. = Standard deviation. P-B = Peak minus background. R.S.D. % = Relative standard deviation as a percentage
of mean loading intensity; mean for t to 6, 2.07 %. C/4o S = Counts per 4o seconds. A.C.F. = Absorption correction
factor. * Mean for I to 7, 0.S3 %.
 to6 R. J. PARKER

used to compute a least-squares regression line care, the accuracy of this technique of mineral
tensities for the variations in lc. analysis can be brought within the t % relative-
The corrected intensities, together with the per- error band.
centage spiked analcime in each standard, were The advent of rapid and accurate whole rock
then used to compute a least-squares regression line major element analysis by X-ray fluorescence
through the data. The slope (337.2 counts per t spectrometry combined with flux-fusion sample
analcime) and intercept (2478 counts) of this re- preparation (e.g. the Norrish method: Norrish and
gression allowed the analcime concentration in the Hutton, 1969; Harvey et al., 1973; Parker and
unspiked; standard to be calculated (i.e. 5.67 Willis, 1977; Parker, 1978), has greatly facili-
analcime). From this the total analcime concentra- tated the acquisition of major-element geochemical
tions in the spiked standards were calculated and data. In petrological studies involving major-
the results are listed in Table II. Fig. t shows the element analyses as well as the quantitative deter-
standard concentrations plotted against intensities. mination of sample mineral concentrations, the
application of calculated mass absorption correc-
tions to quantitative XRD intensities is recom-
mended. This approach provides an accurate and
20 rapid alternative to direct measurement of the mass
m absorption correction, or to the use of internal,
v standard techniques.
c
0 15
U Acknowledgements. My thanks are extended to Dr. I. L.
e) Gibson of Bedford College, University of London, who
N kindly made available the X-ray diffractometer used in
0 this work. Fruitful discussions were held with M. T. Frost
10 and P. Suddaby both of Imperial College, University of
London. The chemical analyses of leucite and analcime
m were carried out at Imperial College on a Philips 1212 X-
0
ray spectrometer purchased with the aid of a N.E.R.C.
x
ul 5
research grant.

c o measured XRD intensity

REFERENCES
7 • corrected n
0 Gupta (A. K.) and Fyfe (W. S.), 1975. Can. Mineral. 13,
U
361.
10 20 30 40 Harvey (P. K.), Taylor (D. M.), Hendry (R. D.), and
w t . % analcirrie Bancroft (F.), 1973. X - Ray Spectrom. 2, 33.
Heinrich (K. F. J.), 1966. In The Electron Microprobe,
FIG. 1. Analcime XRD calibration line. p. 1035; eds. T. D. Mckinly, K. F. J. Heinrich, and
D. B. Wittry. Wiley, New York.
The slope of the calibration line and the corrected Klug (H. P.) and Alexander (L. E.), 1973. X - Ray
XRD intensities were then used to back-calculate Diffraction Procedures for Polycrystalline and
Amorphous Materials, 2nd edn., p. 966. Wiley, New
the analcime concentrations in the standards. This
York.
allowed the calculation of the individual standard Leroux (J.), Lennox (D. H.), and Kay (K.), 1953. Anal.
absolute) errors, and hence the standard relative Chem. 25, 74o.
errors. Finally, the mean relative error (0.53 %) was Niskanen (E.), 1964. Ibid. 36, 1268.
calculated. In order to check the above calibration Norrish (K.) and Hutton (J. T.), 1969. Geochim.
technique, a new pumice sample (p, = 53-5) was Cosmochim. Acta, 33, 431.
spiked to produce four standards (4.77, 9.77, 14.77, and Taylor (R. M.), 1962. Clay Minerals Bull. 5, 98.
and 19.77 wt % analcime). These standards were Parker (R. J.), 1978. X - Ray Spectrom. 7, 38.
mounted only in duplicate and the resulting cali- and Willis (J. P.), 1977. Computers and Geosciences,
3, 115.
bration data computed as above. This produced a
Williams (P. P.), 1959. Anal. Chem. 31, 1842.
mean relative error of 0.52 %, with a relative error
range of ,o- to % to 0.92 %, for the four standards.
The mean relative errors (0.53% and 0.52%) for [Manuscript received 14 January 1977;
these calibration lines indicate that with adequate revised 1 September 1977]
 An Iterative Method for Determining Background
Intensities used in XRF Calibration Lines for
Flux-Fusion Silicate Rock Analysis

R.J. Parker
Dept. of Geology, Imperial College, London, England

In major element rock analysis using XRE and fused samples and standards, the construction of standard
calibration lines is facilitated by the accurate knowledge of the background intensities for the elements
analysed. As an alternative to measuring the background `off peak' or on blanks, an iterative technique is
proposed. This technique is based on repeatedly incrementing the background value until the best calibra-
tion line has been found. The quality of the calibration line is evaluated in terms of the average relative and
average absolute errors. Comparative data, presented for calibration lines constructed from (1) backgrounds
measured from blanks and (2) backgrounds determined by iteration, show that the iterative technique
produces superior calibrations. The analytical techniques have been tested by analysing international rock
standards as unknown samples.

INTRODUCTION method raised by Ingamells' and Fabbi.6 Fabbi (Ref. 6,

p. 237) considers that preparing ground glass briquettes is
In the X-ray fluorescence analysis of fused silicate rock and quicker than casting glass discs. However, experience in our
mineral samples the construction of multi-standard calibra- laboratory has shown that the plunger apparatus enables
tion lines is a common technique in many laboratories. A satisfactory glass discs to be cast on a routine basis at about
brief outline of the methods used at Imperial College is twice the speed at which ground glass briquettes can be
given below. prepared.
(2) Loss on ignition (LOI) values are determined by
(1) Standards, blanks and samples are pre-ignited and
igniting standard and sample powders at 850 °C for 30 min.
then fused with a flux. Two different fusion techniques
Higher temperatures were found to cause certain glassy
are available. Pure lithium tetraborate flux is used in the
pumice samples and volatile rich granites to become
first method of fusion, and the flux to sample ratio is
partially fused. The recommended concentrations for the
normally 7 :1. The fusion melt is quenched to a glass and
calibration standards are corrected for the effect of loss (or
then ground and briquetted. Successful analysis using this
gain) on ignition.
fusion technique requires careful attention to the briquette
(3) All count rate data are corrected for machine drift
preparation process. Factors such as the briquette pressing
and dead time. The count rate data from briquettes prepared
time and pressure should be standardized, and the surface
by the first method of fusion are corrected for matrix
of the pressing die should be maintained in a mirror smooth
effects using the methods of Gunn.' The matrix correction
condition. The selected grinding technique should ensure
procedures described by Norrish and Hutton3 are used to
that the ground glass grain size distribution is kept below
correct the count rate data from discs prepared by the
10pm for all sample fusions (see Ingamells,t p. 327). These
second method of fusion.
and other factors affecting this method are discussed in
(4) In constructing the calibration line for a given
some detail by Parker.'
element a number of standards (usually from 6 to 14),
The flux for the second method of fusion is the Norrish
each prepared in duplicate or triplicate, are analysed.
mixture of lithium tetraborate, lithium carbonate and
Each preparation is normally counted at least twice. Matrix
lanthanum oxide (47.0: 36.6 :16.3). The flux (Spectroflux
corrections are applied to the XRF intensities, rather than
105, Johnson-Matthey Ltd, London) and sample powder
to the standard concentrations as done by Norrish and
weights are 1.5 g and 0.28 g respectively, and 0.02 g of
Hutton3 and Harvey et al.5 The background count rate
NaNO3 is also added to the fusion which is cast as a glass (determined on blanks or by iteration) is subtracted from
disc. This method has been described in detail by Norrish
the peak count rate to give the net peak intensity for each
and Hutton.3 NaNO3 is added as recommended by these standard count. Mg backgrounds for standards and samples
workers to ensure oxidizing conditions. NaNO3 also assists
are corrected for the effect of crystal (RbAP) fluorescence
in forming a homogeneous glass.' Harvey et aL 5 discuss the
using techniques described in Parker and Willis.8
Norrish method in detail and the plunger apparatus described
(5) The matrix corrected net peak intensities for each of
by these latter workers greatly facilitates the ease with
the standards used in the calibration are then summed and
which the glass discs may be cast. Using this plunger appara-
the standard concentrations are also summed. The slope
tus means that no special expertise is required in preparing
factor (SF) of the calibration line is then given by:
the glass discs and this removes the objections to this
Sum of standard concentrations
SF —
© Heyden & Son Ltd, 1978 Sum of net standard counts s
0049-8246/78/0007-0038 $03.00
38 X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978
 ITERATIVE METHOD FOR DETERMINING BACKGROUND INTENSITIES USED IN XRF CALIBRATION LINES

(6) The count rate data for unknown samples are samples and standards is significantly increased; (ii) for
converted into background corrected concentrations using certain elements accurate background values are difficult to
the calibration slope factors and the appropriate Gunn or obtain due to the sloping nature of the background. An
Norrish matrix corrections. The matrix correction pro- example is Ti Ka which is on the shoulder of the La L11 line
cedures are iterated. (La is the heavy absorber in the Norrish fusion flux);
(7) Na2O and FeO values for the unknown samples are (iii) major element flux impurities will increase the effec-
determined by wet chemical techniques. These data plus tive backgrounds for the elements concerned. Backgrounds
H2O (110 °C) and LOI (850 °C) are incorporated in the measured off peak will therefore be too low. The same
final sample analysis, and the XRF oxide concentrations effect will occur where there is primary X-ray beam spectral
are adjusted accordingly. contamination (e.g. Fe Ka from Fe contamination in the
(8) Variations in the weight of powder taken for stand- target of tungsten or chromium X-ray tubes, see Norrish
ard or sample fusions are allowed, and the actual weight and Hutton,3 p. 437).
taken is ratioed against the recommended weight in order Norrish and Hutton (Ref. 3, p. 443) recommend that
to correct the concentrations assigned to the standards or background values should be determined on blanks. As an
samples (see Parker and Willis,8 p. 143). This facility allows alternative to using blanks, an iterative technique for deter-
considerable variation in powder weight and this can be mining background values is described in this paper. The
useful when reduced quantities of material are available, `blank' and `iterative' methods for determining backgrounds
as can be the case in mineral analysis. Norrish and Hutton do not suffer from the problems outlined above for the `off
(Ref. 3, p. 446) describe these procedures for their method peak' method.
and they note that they have successfully analysed samples
varying in weight by 0.25 to 2.0 times the recommended
weight.
(9) The analyses are carried out on a Philips PW 1212 QUALITY OF THE CALIBRATION LINE
XRF spectrometer. The instrumental conditions are listed
in Table 1. The count rate data are punched out on paper- In constructing calibration lines it is useful to have criteria
tape and the data tape is then read into the college com- for evaluating the quality of any given calibration. The
puter system for processing. criteria adopted are the average relative and average absolute
The computational procedures outlined in points (3) to errors of the standards with respect to the calibration line
(8) above are described in more detail in Parker and Willis', (see also Fabbi6). They are calculated as follows. The net
while various factors contributing to errors in the above peak intensity for each individual standard is multiplied by
analytical techniques are discussed in Parker? the SF to give the calculated (as opposed to recommended)
concentration for that standard on the calibration line. The
difference between the recommended and calculated con-
BACKGROUND INTENSITIES centrations is then the absolute error for that standard. The
relative error (as a percentage) is similarly calculated. The
A feature of the method of calibration described above is absolute and relative errors for all the standards used in the
that the calibration line will automatically go through the calibration are then averaged to produce average relative
origin. Because of this any errors in the background count and average absolute errors for that calibration. A diagram
rate used to calculate the net peak intensities will directly of the calibration line construction is given in Fig. 1.
affect the quality of the calibration line. In this method of constructing the calibration line the
Harvey et a1.5 recommend that in their method of high concentration and therefore high intensity standards
analysis the background count rates for Norrish fusion discs contribute relatively more to the resulting slope factor as
should be determined off peak at convenient 20 settings for compared to the lower concentration standards. This is
each element. For several reasons backgrounds measured in desirable because of the better counting statistics at the
this way have not been used in constructing calibration higher concentrations. Furthermore, the quality of the
lines as described above: (i) the total counting time for analytical data for the lower concentration standards may

Table 1. Instrumental settings

M gd Ti Ca K Si Al P Fe Mn Cr Ni

X-ray tube Cr Cr Cr Cr Cr Cr Cr W W W W
kV 60 60 40 40 60 60 60 40 60 60 60
mA 32 24 16 24 32 32 32 8 32 32 32
Crystals RbAP LiF LiF LiF PET PET GE LiF LiF LiF LiF
Collimator') C F F F C C C F F F F
Counter Fl Fl Fl Fl Fl Fl FI Fl FI FI+Sc Fl+Sc
Counting time (s) 100 40 40 40 F.0 .c 100 100 20 100 100 100

8 LiF = LiF200.
b C = coarse (480pm), F = fine (160µm) collimator.
c Fl = flow, Sc = scintillation, vacuum path used for all elements.
d Asymmetric window set on pulse height analyser for Mg.
e F.C. = fixed count (300000).

X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978 39

 R.J. PARKER

This method of regression is identical to the slope factor

(SF) calibration method discussed in the introduction,
except that the SF is in fact = 1/b due to the way in which
the nominal sample concentrations are calculated, i.e. in
it program MW,8 the sample XRF count rate data are multi-
' plied by the SF, and not divided by the SF which would be
' std
the case if the SF = b. With regard to size of the corrections
U
0 required for matrix effects, dead time, machine drift and
backgrounds, the following points may be noted. For
normal silicate rocks the matrix corrections are on average
± 3% (relative) for Norrish fusions and ± 6% (relative) for
C2 C,
the pure lithium tetraborate fusions (7 : 1 dilution) (see
Concentration (wt.%) —~ Norrish and Hutton3 and Parker2). Dead time and machine
drift corrections are normally less than 1-5% (relative),
c, = recommended concentration for standard
c2 = calculated concentration for standard while the background correction as a relative proportion
n = number of standards of the total count is on average less than 2-3% for elements
d = I c, — c, ( = absolute error such as Si, Al, Fe, Ca and K.
Ed A further point to be considered when using classical
— = average absolute error
n least squares regression methods is the underlying assump-
100 •d/c, = relative error (%)
tion that while the dependent variable (i.e. the XRF count
E (100 • did ) o
' average relative error (%) rates) may contain errors, the independent variable (i.e.
E nc,
the standard concentrations) should not. This may not be
— slope factor the case for XRF calibration lines based on fused rock
E std. c.p.s.
standards. The adopted standard concentrations for the disc
Figure 1. Construction of calibration line or briquette will contain errors both in terms of the 'recom-
mended' values for these rock standards, and in terms of
not be as good as that for the higher concentration stand- the errors associated with the actual preparation of the
ards. In their seminal study on the variability of inter- standard (i.e. the errors associated with the ignition,
laboratory analytical data for international rock standards, weighing, fusing, casting or pressing of the disc or briquette)
Fairbum et a1.9 show that for the majority of oxides there (see Till, tt p. 99 and Madansky12 for further discussions
is an increase in the precision of the reported data with an with regard to regressions involving errors in both variables).
increase in the concentration level of the oxide being
determined (see also Ahrens24 and Lister and Gallagher25). BACKGROUNDS FROM BLANKS
Many workers use classical least squares regression tech-
niques in constructing X-ray fluorescence calibration lines. Figure 2 shows a calibration line constructed for TiO2. The
In these techniques all the standards are equally weighted standards and blanks in this calibration were prepared using
in the calculation of the regression line. While procedures the method of Norrish and Hutton.3 Note that the relative
are available to weight the data points in a least squares errors are higher for the lower concentration standards.
regression1', it appears that such procedures are seldom This is considered to reflect the poorer counting statistics,
used as no references to weighted regressions were found in and may also reflect the quality of the recommended data,
the literature on XRF calibration techniques applied to at these lower concentrations. The recommended standard
fused rock standards. data used in the calibrations are taken from Abbey13 and
Draper and Smith (Ref. 10, p. 80) discuss weighted least Flanagan.14 The differences in the actual concentration data
squares regression and in particular the case where the assigned to the standards in Fig. 1 as compared to the data
variance of the dependent variable y is proportional to the recommended by Abbey and Flanagan reflect factors dis-
size of the corresponding independent variable x (i.e. cussed in points (2) and (8) in the introduction.
a2 = V(y) = kx). The standard deviation of XRF counts is The matrix corrected background count rate value used
a= Vcount, and hence the variance of the count is a2 = count. in constructing the TiO2 calibration line in Fig. 2 was
In the calibration procedures described in the introduction determined from the average of two SiO2 and four CaO
the original XRF counts (or intensities) measured on the blanks. Matrix corrected background count rates deter-
fusion discs or briquettes require correction for dead time, mined on blanks prepared from different oxides may vary
machine drift, backgrounds and matrix effects, i.e. the by up to 10%. As a result, Norrish and Hutton3 recommend
variance of the original count is not exactly proportional that blanks should be prepared from several different oxide
to the concentration x in the standard. This lack of propor- compositions and the results averaged. Furthermore, in
tionality will be dependent on the size of these corrections. order to guard against contamination, each blank should be
For those elements for which the corrections are relatively prepared at least in duplicate.
small the variance V (y) of the original XRF count will be Further examples of calibration lines constructed using
proportional, to the first approximation, to the concen- backgrounds determined from blanks are given in Parker
tration x of the element in the standard. Draper and Smith and Willis (Ref. 8, pp. 154-157).
show that for the case where the variance of y is propor-
tional to x, the slope b of the weighted regression line is: BACKGROUNDS BY ITERATION
Ey Table 2 shows the determination of an iterated background
b=
for a TiO2 calibration line. The iteration procedure for a

40 X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978

 ITERATIVE METHOD FOR DETERMINING BACKGROUND INTENSITIES USED IN XRF CALIBRATION LINES

100.00
Table 2. Determination of iterated background for Ti02°
80.00 (iteration delta = 5 counts s t)
0 Background
60.00 (counts s') a.a.e. a.r.e.
iN
105 0.013 3.74
40.00
110 0.012 3.46
115 0.011 3.18
20.00 120 0.010 2.92
125 0.009 2.65
0 00 4,4 130 0.008 2.38
0.00 0.50 1.00 1.50 2.00 2.50
Percent Ti 02
135 0.008 2.12
S 0.00039 AAE 0.02 ARE 4.52 140 0.007 1.86
145 0.006 1.64
STD RC CC AE RE CCPS 150 0.006 1.42
BCR/B 2.24 2,26 -0.02 -0.75 5788 155 0.006 1.22
8CR/B 2.24 2.24 0.00 0.09 5740 160 0.005 1.06
BCR/D 2.24 2.26 -0.02 -1.09 5800 165 0.005 0.94
BCR/D 2.24 2.25 -0.01 -0.35 5758 170 0.006 0.87 Selected by iteration
BCR/E 2.23 2.25 -0.02 -0.74 5772 175 0.006 0.95
BCR/E 2.23 2.23 0.00 0.05 5727 0.007
180 1.18
W-1/E 1.08 1.06 0.02 1.89 2711 185 0.008 1.47
W-1/D 1.08 1.05 0.03 2.70 2689
190 0.009 1.75
W-1/A 1.08 1.07 0.01 1.12 2733
195 0.010 2.04
AGV/C 1.05 1.04 0.01 0.93 2672 200 0.010 2.33
AGV/B 1.05 1.05 -0.00 -0.06 2696
205 0.011 2.63
AGV/A 1.05 1.05 -0.00 -0.33 2701
210 0.012 2.93
GSP/C 0.66 0.66 0.01 1.25 1684
215 0.013 3.23
GSP/B 0.66 0.64 0.02 3.52 1642
GSP/A 0.66 0.64 0.02 3.29 1646 220 0.015 3.54
225 0.016 3.85 Determined from blanks
G-1 /C 0.26 0.23 0.03 10.90 598
G-1 /B 0.26 0.24 0.02 7.41 622 230 0.017 4.16
G-1/A 0.26 0.24 0.02 9.27 607 235 0.018 4.47
NIMP/C 0.20 0.18 0.02 11.01 459 240 0.019 4.79
NIMP/B 0.20 0.18 0.02 10.41 462 245 0.020 5.10
N I M P/A 0.20 0.18 0.02 10.75 460 250 0.021 5.42
NIMG/C 0.09 0.09 0.00 5.03 222 255 0.022 5.75
NIMG/B 0.09 0.09 0.01 5.52 220 260 0.023 6.07
NIMG/A 0.09 0.11 -0.02 -19.95 280
a The TiO2 iteration data in Tables R and 3 are from different
SI02/B 0.00 0.09 0.00 0.00 242
calibrations.
S102/A 0.00 0.10 0.00 0.00 244
CAO/C 0.00 0.10 0.00 0.00 254
CAO/C 0.00 0.10 0.00 0.00 258
CAO/B 0.00 0.10 0.00 0.00 254 be speeded up by applying a coarse (e.g. 5 counts s t)
CAO/A 0.00 0.10 0.00 0.00 258 increment to begin with, and then a fine increment (e.g.
Figure 2. TiO2 calibration line 0.25 counts s) in the region of the a.r.e. minimum.
STD = standard. RC = recommended concentration, CC = calculated Table 3 lists major element calibration background
concentration, AE = absolute error, RE = relative error, CCPS = count rates, a.r.e.'s and a.a.e.'s, determined first by itera-
corrected counts per second, S = slope factor, AAE = average tion and then from blanks. Table 4 lists the standards used
absolute error, ARE = average relative error in these calibrations. The iteration method produces lower
a.r.e.'s for all calibrations. With regard to the a.a.e.'s, Mn0
and K20 give the same values for both methods, while the
iteration method gives lower a.a.e.'s for all the remaining
given calibration line is carried out by initially setting the calibrations.
background count rate to zero counts cl (c.p.s.) followed Further examples of calibration lines constructed using
by the calculation of the average relative error (a.r.e.) and iterated backgrounds are given in Parker and Willis (Ref. 8,
the average absolute error (a.a.e.). The background is then pp. 165-167).
incremented by delta c.p.s. and the calibration line recalcu-
lated to produce a new a.r.e. Successive increments of the
background count rate lead to a progressively decreasing ACCURACY OF THE NORRISH ANALYTICAL METHOD
a.r.e. until a minimum is reached, whereupon further
increases in the background c.p.s. will lead to an increasing In order to provide an estimate of the accuracy of the
a.r.e. The optimum background count rate is therefore at analytical techniques described above for Norrish fusions,
the minimum a.r.e. Note that the changes in a.r.e. are six standards were analysed as unknown samples. This
smooth. As XRF data processing is normally carried out analytical run included the ultrabasic standards NIM-P,
with the aid of a digital computer, the iteration process NIM-D and DTS in the Si02 calibration. The data for each
may conveniently be built into the calibration section of of the standards was processed with that standard removed
the data processing program.8 The iteration process may from all of the calibration lines (the K-feldspar standard,

X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978 41

 R.J. PARKER

Table 3. Comparison of calibration data

Si02 TiO2 A1203 Fe203 MnO MgO CaO K20 P20, Cr20, NiO
Background iteration
(counts s ') 49 172 16 61 144 44.0 81 27 59 665 146
Background blank
(counts s') 27 201 22 49 144 47.5 96 29 51 654 141
a.r.e. iteration 0.52 1.98 0.68 1.49 9.14 5.39 1.73 0.90 6.06 13.75 9.28
a.r.e. blank 0.53 4.50 3.66 1.60 9.17 8.77 1.93 0.92 8.66 32.48 11.04
a.a.e. iteration 0.339 0.008 0.080 0.098 0.011 0.175 0.049 0.027 0.013 0.014 0.005
a.a.e. blank 0.343 0.011 0.090 0.102 0.011 0.190 0.054 0.027 0.018 0.019 0.006
No. of stds. 12 12 8 14 6 12 12 12 6 4 6
Range of stds. 39 -76 0.05- 0.26- 1.3- 0.03- 0.38- 0.68- 0.25- 0.14- 0.01- 0.01-
(wt.%) 2.7 17.4 17.0 0.78 49.7 14,2 15.4 0.50 3.6 0.31

British Chemical Standards No. 376, was not used in the

Table 4. Standards used in calibrations
calibrations). Duplicate preparations of each standard were
Si Ti Al Fe Mn Mg Ca K P Cr Ni analysed, and the averaged results are presented in Table 5.
There is good agreement between the XRF values listed
GH + + + + +
in Table 5 and the recommended (or preferred) values for
G-1 + + + + + + + +
the analysed standards. Note that the range of these igneous
G-2 + + + + + + + +
NIMG + + + + + + standards covers acid, basic, ultrabasic and syenitic compo-
GSP + + + + + + + + + sitions. Norrish and Hutton3 and Harvey et al.5 provide
AGV + + + + + + + + further data indicating that the technique is capable of
NIMS + + + + + + + analysing a considerable range of compositions. The author
BCR + + + + + + + + has successfully used the technique to analyse P, Ca and Fe
NIML + + + + + + rich rocks from the sea floor.15,16 The Norrish technique
W-1 + + + + + + + + + has been used extensively in the analysis of lunar
NIMN + + + + + + materials.17-23
BR + + + + + +
The `recommended LOI' values listed in Table 5 are the
NIMP + + +
+ +
sum of the recommended H2O'' plus CO2 values for each of
DTS
NIMD + + + + the standards. The determined LOI values were corrected
for the effect of the oxidation of FeO to Fe203. Tests have

Table 5. Accuracy test on Norrish analyses using iterated backgrounds for the calibration lines

G-1 VV-1 BR NIMS NIM-D K-FELDf

Det. Rec.' Det. Rec .b Det. Rec .b Det. Rec .b Det. Rec .b Det. Rec .b
S102 72.26 72.68 52.84 52.72 38.33 38.39 63.39 63.54 39.02 38.97 67.18 67.10
TiO2 0.26 0.26 1.08 1.07 2.65 2.61 0.05 0.05 0.02 0.02 0.01 0.01
AI20, 14.18 14.05 15.10 14.87 10.14 10.25 17.15 17.16 0.35 0.26 17.75 17.70
Fe2O3 0.89 0.87 1.37 1.40 5.57 5.61 1.02 1.08 0.79 0.70 0.08 0.10
FeOC 0.96 0.96 8.73 8.73 6.60 6.60 0.29 0.29 14.67 14.67 0 0
Cr20, 0 0 0 0 0 0 0 0 0.44 0.41 0 0
MnO 0.03 0.03 0.17 0.17 0.20 0.20 0.02 0.01 0.22 0.21 0.01 0
MgO 0.37 0.38 6.50 6.63 13.53 13.35 0.43 0.48 43.43 43.68 0.01 0.03
CaO 1.30 1.39 11.02 10.98 13.78 13.87 0.68 0.68 0.29 0.26 0.48 0.54
Nape 3.32 3.32 2.15 2.15 3.07 3.07 0.41 0.41 0.06 0.06 2.83 2.83
K20 5.55 5.48 0.67 0.64 1.42 1.41 15.60 15.40 0.01 0.02 11.14 11.20
P20, 0.09 0.09 0.16 0.14 1.00 1.05 0.13 0.14 0 0.03 0.03 0
NiO 0 0 0 0 0 0 0 0 0.27 0.29 0 0
LOId 0.35 0.41 0.66 0.59 3.23 3.17 0.25 0.33 0.72 0.68 0.30 0.35g
Total 99.53 99.92 100.42 100.09 99.50 99.58 99.42 99.57 100.29 100.26 99.82 99.86
Fe203e 1.96 1.94 11.07 11.11 12.90 12.98 1.34 1.44 17.09 16.99 0.08 0.10
a Ref. 14, recommended values on moisture (H20-1 free basis.
b Ref. 13, recommended values on moisture (H20-1 free basis.
c FeO and Na20 = recommended values.
d LOI (Det.) = ignition loss at 850°C (corrected for FeO Fe203).
LOI (Rec.) = sum of recommended H20.plus CO2 .
e Total iron.
f K-FELD = British Chemical Standards No. 376.
g Recommended LOI as per British Chemical Standards certificate.

42 X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978

 ITERATIVE METHOD FOR DETERMINING BACKGROUND INTENSITIES USED IN XRF CALIBRATION LINES

shown that ignition at 850 °C (and at 950 °C) does not `off peak'. This latter method suffers from a number of
oxidize all of the Fe0 to Fe203 (e.g. Fe0 in NIMD = 5.47% inherent errors and significantly increases the analytical
after ignition at 850 °C for 30 min) and the LOI oxidation time. The `Slope Factor' calibration method (backgrounds
correction was adjusted for this effect. Apart from H2O+ by iteration or from blanks) weights the calibration line
and CO2, other volatiles lost on ignition may include towards the higher quality data available from the higher
elements such as F, Cl and S; and this will contribute to the concentration standards. This calibration method is pre-
variation in the determined LOI with respect to the sum of ferred to the classical least squares technique in which all
the recommended H2O+ plus CO2. There is nevertheless the standards are equally weighted.
reasonable agreement between these data for the standards The accuracy test on the Norrish method of sample
analysed in Table 5. fusion and matrix correction combined with weighted
calibration lines (using iterated backgrounds) has produced
satisfactory results.
CONCLUSIONS

The iterated background method produces superior calibra- Acknowledgements

•
tions as compared to calibrations using backgrounds from
Fruitful discussions were held with G.D. Borley, R. Howarth, P.
blanks, and furthermore the overall analysis time is reduced. Suddaby and M. Thompson of the Dept. of Geology, Imperial
Both the iteration and the blank methods of background College. The XRF data was produced on a Philips 1212 X-ray
• determination are preferred to determining backgrounds spectrometer purchased with the aid of a N.E.R.C. research grant.

REFERENCES

1. C.O. Ingamells, Anal. Chim. Acta 52, 323 (1970). P.K. Hofmeyr, T.S. McCarthy and M.J. Orren, Proc. Second
2. R.J. Parker, Factors affecting the quality of major element Lunar Sci. Conf., Geochim. Cosmochim. Acta Suppl. 2, Vol. 2,
rock analysis by X-ray fluorescence spectrometry combined MIT Press, 1971, pp. 1123-1138.
with flux-fusion sample preparation. Technical Report No. 18. LSPET (Lunar Sample Preliminary Examination Team),
XRF-2B, Dept. Geology, Imperial College, Univ. London, Science, 175, 363 (1972).
1977, p. 35. (Copies may be obtained from the author). 19. J.P. Willis, A.J. Erlank,J.J. Gurney, R.H. Theil and L.H. Ahrens,
3. K. Norrish and J.T. Hutton, Geochim. Cosmochim. Acta 33, Proc. Third Lunar Sci. Conf., Geochim. Cosmochim. Acta
431 (1969). Suppl. 3, Vol. 2, MIT Press, 1972, pp. 1269-1273.
4. C. Palme and E. Jagoutz,Ana/. Chem. 49, 717 (1977). 20. LSPET (Lunar Sample Preliminary Examination Team),
5. P.K. Harvey, D.M. Taylor, R.D. Hendry and F. Bancroft, Science 179, 22 (1973).
X-Ray Spectrom. 2, 33 (1973). 21. LSPET (Lunar Sample Preliminary Examination Team),
6. B.P. Fabbi,Am. Mineral. 57, 237 (1972). Science 182, 659 (1973).
7. B.M. Gunn, Can. Spectros. 12, 41, 64, 163 (1967). 22. J.M. Rhodes, K.V. Rodges, C. Shih, B.M. Bansal, L.E. Nyquist,
8. R.J. Parker and J.P. Willis, Comput. Geosci. 3, 115 (1977). H. Wiesmann and N.J. Hubbard, Proc. Fifth Lunar Sci. Conf.,
9. H.W. Fairburn et al., U.S. GeoL Surv. Bull. 980, 71 pp. (1951). Geochim. Cosmochim. Acta, Suppl. 5, Vol. 2, Pergamon,
10. N.R. Draper and H. Smith, Applied Regression Analysis. Wiley, 1974, pp. 1097-1117.
New York, 1966. 23. A.R. Duncan, A.J. Erlank, J.P. Willis, M.K. Sher and L.H.
11. R. Till, Statistical Methods for the Earth Scientist — An Ahrens, Proc. Fifth Lunar Sci. Conf., Geochim. Cosmochim.
Introduction, Macmillan, London, 1974. Acta, Suppl. 5, Vol. 2, Pergamon, 1974, pp. 1147-1157.
12. A. Madansky, J. Am. Stat. Assoc. 54, 173 (1959). 24. L.H. Ahrens, Physics and Chemistry of the Earth, Vol. 2,
13. S. Abbey, Geol. Surv. Can. Pap. 73-36, pp. 25 (1973). Pergamon, 1957, pp. 30-45.
14. F.J. Flanagan, Geochim. Cosmochim. Acta 37, 1189 (1973). 25. B. Lister and M.J. Gallagher, Trans. lnstn Min. Metall. 79,
15. R.J. Parker and W.G. Siesser, J. Sediment. Petrol. 42, 434 B213 (1970).
(1972).
Received 25 October 1976; accepted (revised) 21 September 1977
16. R.J. Parker, J. Sediment. Petrol. 45, 230 (1975).
17. J.P. Willis, L.H. Ahrens, R.V. Danchin, A.J. Erlank, J.J. Gurney, n Heyden & Son Ltd, 1978

X-RAY SPECTROMETRY, VOL. 7, NO. 1, 1978 43

Computers el Geosciences, Vol. 3, pp. 115-171. Pergamon Press, 1977. Printed in Great Britain

COMPUTER PROGRAMS SORT, REORD, AND MW FOR

MAJOR-ELEMENT XRF DATA PROCESSING
R. J. PARKER
Department of Geology, Imperial College, Prince Consort Road, University of London, London SW7 2BP, England

and

J. P. WILLIS
Department of Geochemistry, University of Cape Town, Cape Town, South Africa

(Received 7 September 1976)

Abstract—Three computer programs, SORT, REORD, and MW, have been written in ASA FORTRAN IV for
processing major-element analytical data produced by X-ray fluorescence of rock samples. The programs have been
written for the analytical technique of determining sample concentrations from calibration lines constructed from
known standards. SORT and REORD are preprocessing programs that transform the initial raw data into a form
suitable for processing by program MW. This program constructs calibration lines for each of the elements analyzed,
and then converts the sample data into matrix corrected oxide concentrations.
Program MW will process data according to either the Norrish and Hutton or the Gunn methods of sample fusion
and matrix correction. Background intensities for the calibration lines may be determined by two different methods,
and provision is made for correcting the Mg0 calibration line data for interference due to crystal fluorescence.
Wet-chemical data on Fe0 or Na20 for the samples also may be entered and these data will be incorporated in the final
output analyses.

Key Words: Data processing, Norrish and Hutton method, Gunn method, Spectrometer, X-ray fluorescence,
FORTRAN, Geochemistry.

INTRODUCTION For any given laboratory this generalized data file is

Three interrelated computer programs, SORT, REORD, best produced by a program specifically written to accept
and MW, have been written, in ASA FORTRAN IV, to the raw data from the particular spectrometer in question.
process major-element analytical data produced by X-ray For most XRF systems this should not prove to be too
fluoresence (XRF). The various techniques used in XRF arduous a programming task. A common spectrometer is
analysis of rock samples have as their underlying the Philips 121211220 range and program SORT has been
principle that of comparing the XRF intensity produced designed to accept raw data from this machine and to
by a given element in an unknown sample against the XRF produce a generalized data file suitable for input to
intensity observed for the same element in a standard of program REORD.
known composition. The comparison may be made by Program SORT has been constructed so as to allow
using a single standard, or a multistandard calibration line. maximum flexibility in the input data, that is, the raw data
The programs have been written with this latter technique from the analysis of samples, blanks, and standards may
in mind, and with reference to firstly the Norrish and be input in any order, and for any sample, blank or
Hutton (1969) method of preparing fusion discs for standard, the order of element analysis may be changed.
analysis and of making matrix corrections, and secondly Up to fifteen major elements may be analyzed. The
to the Gunn (1967a, 1967b) methods of making matrix "Norrish" method will correct Si, Ti, Al, Fe, Cr, Mn, Mg,
corrections and preparing fusion discs. Ca, K, P, and Ni for matrix effects, whereas the "Gunn"
It is not feasible to write a general purpose program to method will matrix correct the preceding elements as well
take the output data from the many different XRF as Na, Sr, and Ba.
spectrometers found in geological laboratories. Pro- The output from program SORT is written to a
gram REORD therefore has been written to accept an temporary disc file which is rewound in preparation for
input data file in a generalized form having the following reading in by program REORD. This latter program
characteristics (see Tape 11-N, Tape 11-G). builds up standard/blank "calibration" data sets for the
(1) The first six lines of the file must consist of header elements analyzed and produces "sample" data sets.
information. These data sets are written out to a second temporary disc
(2) The samples, blanks, and standard must be num- file which is rewound to serve as input to program
bered as described. MW.
(3) The samples, blanks, and standards should be Program MW reads the standard/blank data sets from
sorted into ascending numbers, that is the unknown sample the disc file and constructs a calibration line for each of
data first, followed by blank, and then standard data. the elements analyzed. The calibration line information is
(4) The data for each of the elements analyzed must be used to convert the sample data (read from the disc file)
corrected for dead time and machine drift. into nominal concentrations which are matrix corrected to
(5) The format of the file must follow that of 1012 produce the final analyses. A schematic flowchart for the
format in program REORD. programs is given in Figure 1.
115
116 R. J. PARKER and J. P. WILLIS

Tape 51= Pgm SORT Tape 60=

XRF raw data file lineprinter file

Tape 61=
lineprinter file Tape 62-
lineprinter file

Tape 11= Tape 12= Tape 7=

Pgm REORD
temporary disc file temporary disc file punch card file

Tape 53=
Tape 52= sample name &
standard/blank data file weighing file

Note: All the 'tapes' are disc files

Figure I. Schematic flowchart for programs SORT, REORD and MW.

SAMPLE, BLANK, AND STANDARD NUMBERING PROGRAM SORT

The programs require that each sample, blank, or This program is designed to accept, as input data, the
standard should be assigned, for identification, a four digit punched paper-tape output produced by a Philips
"prime" analysis number (a-number) which must end in 1212/1220 XRF spectrometer connected to an Addo
zero. Sample a-numbers must be between 0010 and 7990, paper-tape punch or teletype terminal. These data must be
whereas blank a-numbers must be between 8000 and 8990, proceded by header information (Tape 51-N) which must
and standard a-numbers between 9000 and 9899. Standard consist of on the first line the method option followed by
a-numbers 9900 and above are reserved for Mg0 two lines of descriptive information for the analyses, the
interference standards (see further discussion later). dead-time constant on line 4, the option key on line 5 (see
Three fusion discs for each sample, blank, or standard program MW documentation), and the sample holder
may be prepared for analysis and "disc" a-numbers correction factors on line 6. The latter corrections allow
should be assigned as follows: first disc =a-number+ 41 for any variations in the XRF response of the sample
(eg 1150), second disc = a-number + 3 (eg 1153) and third holder positions 2-3-4 in the spectrometer with the first
disc = a-number + 6 (eg 1156). Each disc may be analyzed position taken as unity. Line 7 must carry the machine
(or counted) up to three times and the "disc" a-number element numbers which identify the elements being
should be incremented by 1 or 2 for the second or third analyzed in the spectrometer. Numbers 88 and 99 are not
analysis. Thus if the "prime" a-number (for a sample) is allowed as machine element numbers because these
1150, the three disc numbers are 1150, 1153 and 1156, and if numbers are required for other purposes in the data file
each disc is analyzed three times then the a-numbers (Tape 51-N). The sample holder of the spectrometer can
become: accomodate up to four analysis discs, and the first holder
position should contain the reference disc. Before
analyzing or counting the discs in the spectrometer, the

first disc second disc third disc operator should punch out on the paper tape the number

1150, 1151, 1152 1153, 1154, 1155 1156, 1157, 1158 88 (see Tape 51-N) followed on a new line by the first
reference a-number (99), the 1-3 a-numbers correspond-
ing to the 1-3 standards, blanks, or samples in the
This configuration allows, in a single a-number, the remaining holders, and if it is to be recounted, then finally
identification of the sample, up to three analysis discs for the second reference a-number (9999). The reference disc
that sample and up to three XRF analyses on any should be analyzed or counted (for a particular element)
particular disc. This flexible system can accomodate the followed by 1-3 standard, blank, or sample discs to
numbering requirements for duplicate and triplicate produce an analysis cycle. The reference disc may be
sample preparation and analysis. recounted at the end of the cycle. Any combination of
In order to monitor and correct for machine drift in the samples, blanks, and standards may be present in
Philips 1212/1220 spectrometer, a reference disc is analysis cycle (Tape 51-N).
analyzed (or counted at the start of each analytical cycle. Program SORT makes dead-time corrections and sorts
These counts are identified by the number 99 (Tape 51-N). the data into arrays linking a-numbers with element count
The reference disc also may be recounted at the end of the rates for sample, blank, standard, and reference discs
cycle and it then must be identified by the number 9999. (Tape 60-N). The reference disk data is used to correct the
These numbers must not be used for sample or standard element count rates for machine drift, and the resulting
a-numbers. (Note: The reference disc numbers, 99 and data are written out to the first temporary disc file (Tape
9999, are used only by program SORT, and do not occur in 11). This generalized data file is rewound in preparation for
the generalized data file read by program REORD.) reading in by program REORD.
 Computer programs SORT, REORD, and MW for major-element XRF data processing 117

00100 C
00110 C PROGRAM SORT — BY R J PARKER (DEPT, OF GEOLOGY,
00120 C
00130 C UNIV. OF LONDON)
00140 C
00150 C PROGRAM SORT PROCESSES X—RAY FLUORESENCE DATA PRODUCED BY A PHILIPS
00160 C 1212 SPECTROMETER IN THE ANALYSIS OF MAJOR ELEMENTS IN ROCK SAMPLES.
00170 C THE PROGRAM SORTS THE ANALYSES INTO ASCENDING ORDER OF
00180 C A — NUMBERS AND ASSIGNS THE ANALYTICAL (XRF) INTENSITIES
00190 C TO THE APPROP
REATE ELEMENTS OF THE ANALYSIS ARRAYS.
00200 C THE PROGRAM MAKES DEAD TIME AND MACHINE DRIFT CORRECTIONS AND
00210 C NRITES THE DATA OUT TO A TEMPORARY DISC FILE WHICH THEN SER
VES AS
00220 C THE INPUT FOR PROGRAM REORD.
00230 C
00240 C PROGRAM SORT HAS BEEN WRITTEN IN ASA FORTRAN IV. EXCEPT
00250 C FOR THE PROGRAM HEADER STATEMENT.
00260 C
00270 C ACKNOWLEDGEMENTS SOME OF THE PROCEDURES USED IN PROGRAM
00280 C SORT OWE THEIR INSPIRATION TO 8. GUNN, DEPT. OF GEOLOGY.
00290 C UNIV. OF MONTREAL.
00300 C
00310 C
00320 C I/O UNIT NUMBERS
00330 C
00340 C IRAD = 51 — ANALYTICAL DATA INPUT FILE
00350 C IRSB = 52 — STANDARDS/BLANKS INPUT FILE
00360 C IWLP = 60 — LINE PRINTER OUTPUT FILE
00370 C ITEMPI = 11 — FIRST TEMPORARY DISC FILE
00380 C
00390 C
00400 C NOTE THE FIRST PROGRAM STATEMENT 'PROGRAM SORT(TAPES)'
00410 C IS A CDC FORTRAN STATEMENT FOR DECLARING DATA FILES
00420 C
00430 C
00440 C
00450 C
00460 C ARRAYS
00470 C
00480 C METHOC(8) — NORRISH OR GUNN OPTION FOR MATRIX CORRECTIONS
00490 C DSCRIP(80) — 80 CHARACTERS FOR DESCRIPTION OF RUN ETC
00500 C NAME ( 80) — BO CHARACTERS FOR ANALYSTS NAME AND DA,
TE ETC
00510 C CHOIS(20) — OPTIONS — SEE DOCUMENTATION IN PROGRAM MW
00520 C
00530 C NNOX(151) — A — NUMBERS FOR ALL SAMPLES/BLANKS/STANDARDS
pp REF(151 15,2) — REFERENCE INTENSITIES " "
0 550 C X(151,15) — ELEMENT INTENSITIES FOR " " It
00560 C CS(5) — COUNTS FOR 1 TO 5 SAMPLES/BLANKS/STANDARDS/REF-
00570 C ERENCES IN AN ANALYSIS CYCLE
T(5) — COUNTING TIMES IN ANALYSIS CYCLE
00590 C C(5) — INTENSITIES IN ANALYSIS CYCLE
00600 C EL(5) — MACHINE ELEMENT NUMBERS FOR ELEMENTS IN AN ANALYSIS CYCLE
00610 C NNO(5) • A - NUMBERS IN AN ANALYSIS CYCLE
00620 C NNOXY(5) — TEMPORARY ARRAY FOR NUMERICAL ORDER EQUI4ALENCE
00630 C OF ANALYSIS CYCLE A — NUMBERS VS ALL A — NUMBERS
00640 C IN NNOX(151)
00650 C RFIRST(15) — INITIAL REFERENCE INTENSITIES, USED IN CONJUNCTION
00660 C WITH REF(151,15 2) FOR MAKING DRIFT CORRECTIONS
00670 C REF0 ( 15) — REFERENCE INTENSITIES FOR CALCULATING TRANSIENT DRIFT
00680 C
00690 C PARAMETERS
u0700 C
00710 C OTC — DEAD TIME COEFICIENT
00720 C LINE — DATA LINE COUNTER
00730 C
00740 C NSI
00750 C — ) XRF MACHINE NUMBERS
G0760 C ' — ) FOR ELEMENTS
00770 C NELZ — )
00780 C
00790 C JK — NUMBER OF REF'S/BLANKS/STANDARDS/SAMPLES IN AN ANALYSIS CYCLE
00 800 C JKX — FLAG FOR PRESENCE OF SECOND REFERENCE COUNT IN
00810 C JKK — IS EQUAL TO JK, OR IS EQUAL TO 'JK — 1' IF SECOND REF'NCE
00820 C COUNT IS PRESENT IN THE ANALYSIS CYCLE
00830 C JZ — IS THE INDEX OF BLANK/STANDARD/SAMPLE A — NUMBERS IN
00840 C ARRAY NNO (READ IN FROM AN ANALYSIS CYCLE)
00850 C JZZ — IS THE INDEX OF BLANK/STANDARD/SAMPLE A — NUMBERS IN
00 ARRAY NNOX (HOLDS ALL A — NUMBERS READ IN)
6 70 C COUNT — TOTAL NUMBER OF BLANKS/STANDARDS/SAMPLES READ IN
00880 C FROM THE INPUT DATA FILE
u0890 C MAXSPL — UPPER LIMIT FOR TOTAL NUMBER OF BLANKS/STANDARDS/SAM-
00900 C PLES THAT CAN BE READ IN FROM THE INPUT DATA
00 910 C A — DEAD TIME COEFICIENT
00920 C LIMIT — COUNT RATE LIMIT
00930 G
00940 C A°RIM — ) POSITION CORRECTION FACTORS FOR SECOND THIRD
00950 C °PRIM — ) AND FOURTH SAMPLE HOLDER POSITIONS IN T HE XRF ANALYSING
00960 C CPRIM — ) MACHINE. FIRST POSITION TAKEN AS UNITY
00970 C
00980 C LINEX — JATA LINE COUNTER FOR ANALYSIS CYCLE
00990 C
01000 C IY — ) INDECES FOR ARRAY ELA(25) CONTAINING ELEMENT SYMBOL
01010 C IE — ) EQUIVALENTS OF 'MACHINE ELEMENT NUMBERS — 50'
01020 C
01030 C KNO — A — NUMBER FOR REFERENCE/BLANK/STANDARD/SAMPLE
01040 C IN AN ANALYSIS CYCLE
01050 C JXX — NUMBER OF BLANKS/STANDARDS/SAMPLES IN AN ANALYSIS CYCLE
01060 C NX — INDEX FOR ASSIGNING REFERENCE COUNTS TO BLANKS/STAND-
01970 G ARDS/SAMPLES IN AN ANALYSIS CYCLE
01080 C DRIFT — COUNT RATE DIFFERENCE BETWEEN TWO CONSEQUTIVE
'01090 C OBSERVATIONS FOR THE SAME ELEMENT ON THE
01100 C REFERENCE DISC.
O111C C XORIFT — DALGULATED DRIFT
01120 C LORIFT — DRIFT LIMIT
01130 C IS — COUNTER FOR BLANKS/STANDARDS/SAMPLES IN ANALYSIS CYCLE
01140 C AND IS USED FOR EXTRACTING NZ FROM ARRAY NNOXY
01150 C NZ — INDEX FOR ASSIGNING ANALYSIS CYCLE INTENSITIES FOR
118 R. J. PARKER and J. P. WILLIS

01160 C BLANKS/STANDARDS/SAMPLES TO THE MAIN INTENSITY

01170 C ARRAY(X(NZ,J))
G1i8C C
01190 C KKX - )
01200 C MKR - ) PARAMETERS FOR SORTING
41210 C NNX - I A - NUMBERS, AND ASSOCIATED
41220 G XA - l INTENSITIES INTO
01230 C REFA •• ) ASCENDING ORDER
01240 C RE FB - )
01250 C
01260 C
41270 C
01280 C
01290 PRO;,RAM SCRT(TAPE60 TAPE51 TAPE11,OUTPUT)
01300 COMMON NNOX(i51) REF(151 , 15,21 X(151,15) C(5)
0131C +T(5), CS(5), EL(5), NNO(51, NNOXY(5), NNOXL(5), RFIRST(15),
01320 +NAM=(10),CSC.RIP(80),CHOIS(2O),METHOD(8),£LA(25),REFD(15)
01330 INT E GER COUNT,CTR EL,DSCRIP,CHOIS
01340 REAL LIMIT, LORIFT
01350 C
01360 C SET I/O UNIT NUMBERS
U1370 C
01380 IRAQ = 51
01390 IWLP = 60
01400 ITEHPI = 11
01410 C
41420 C INITIALIZE VARIABLES AND ARRAYS
01430 G
01440 NX = 0
01450 COU>1T = 0
01460 LIN= = 0
01470 LIMIT = 50000.0
0148C LDRIFT = 3.0
01490 MAXSPL = 150
01500 00 399 J = 1 15
u1510 REFJ(J) = 0.3
01520 RFIRST(J) = 0.0
01530 999 CONTINUE
41540 DO 1011 I = 1,151
01550 NNOK(I) = 0
01560 00 1090 J = 1,15
01570 X (I J) = 0.0
u158C REF(I,J,1) = 0.0
01590 REF(I,J,2) = 3.0
01600 1000 CONTINUE
01610 1001 CONTINUE
01620 C
01630 C READ IN HEADER INFORMATION AND WRITE IT OUT TO
01640 C TEMPORARY DISC FILE
01650 C
01660 REA7(IRAD,7081) METHOD
61670 REA3(IRAD,7081) 0SCRIP
01680 REA](IRAD,7081) NAME
41690 READ (IRAD,7083) DIC
0170) READ(IIRA0,7085) CHOIS
01710 READ(IRAO,7088) APRIM BPRIM CPRIM
01720 READ(IRAO,70691 NSI,N+ I,NAL, IFE,NCR,NMN,NMG,NCA,NNA,NK,
01730
+WRITE( I TEMP1,7081)EMETHOD
01750 WRITE(ITEMP1,7081) DSCRIP
01760 WRITE(ITEMP1,7081) NAME
01770 WRITE(ITEMP1,7083) DTC
41780 WRITE(ITEMP1,7085) CHOIS
01790 WRITE(ITEMFI 7088) APRIM, BPRIM, CPRIM
01800 WRITE(IWLP,7390) METHOD
01810 WRITE(IWLP,7091) OSCRIP
01620 WRITE(IWLP,7091) NAME
01830 WRITE(IWLP,7U92) OTC
01640 WRITE CIWLP,7093) CHOPS
01850 WRITE(IWLP,709 g4) APRIM BPRIM CPRIN
p g T I,NAL,NFE,NCR,NMN,NMG,NCA,NNA,NK,
01860 +WRITE(IWLP 7NE5) NSIIN
01880 7081 FORMAT(80A1)
01890 7083 FORMAT(F3.1)
01900 7085 FORMAT(20I1)
01910 7088 FORMAT(3(F6.4,1X))
01920 7089 FORNAT(15(I2,X))
01930 7090 FORMAT(IHl //X,8A1)
11
01940 7091 FORMAT(X,80A
01950 7092 FORMAT(X,F3.1)
01960 7093 FORMAT(X,20I1)
01970 7094 FORMAT(X,3(F6.4,X))
01980 7095 FORMAT(/X 23HMACCHINE ELEMENT NUMBERS,/,
01990 +X,45HSI T 1 AL FE CR MN MG CA NA K P NI EX EY EZ ,/,
02000 +X,15(I2,X)/)
02010 C
02020 C STATEMENTS 1100 CONTINUE TO 5018 CONTINUE = DATA INPUT SEQUENCE
02030 C
02040 1100 CONTINUE
02050 00 398 J = 1,5
02360 EL(J) = 0
02070 C(J) = 0.0
02080 T(J) = 0.0
02090 CS(J) = 0.0
02100 C (J) = 0.0
02110 998 CONTINUE
02120 C
02130 C READ IN ANALYSIS NUMBERS AND COUNT RATE DATA FOR EACH
02140 C ANALYTICAL CYCLE
02150 C
42160 DO 1210 I = 1,5
02170 1160 CONTINUE
02180 READ(IRAD,1170) SL(I),T(I),CS(I)
02190 1170 FOR'(AT(I2,X,F7.2,X,F8.3)
02200 LINE = LINE + 1
 Computer programs SORT, REORD, and MW for major-element XRF data processing 119

0221C IF(EL(I).EC.-9) GO TO 5018

02220 IF(°L(I).NE.88) GO TO 1042
02230 DO 1057 JC = 1,5
02240 1057 NNO(J0)=0
02250 READ(IRA0,10441(NNO(JO),JQ = 1,5)
02260 1044 FORAA7(I2,1X,4(I4,1X))
02270 LINE = LINE + 1
02280 JK = 0
G229G JKX = 0
02300 DO 1039 N = 1t5
02310 IF('(NO(N).GT.0) JK = JK + 1
0232C IF(KNO(N).EQ.9999) JKX = 1
02330 1039 CONTINUE
02340 JKK = JK
02350 IF(JKX.£0.1) JKK = JK — 1
02360 C
0237C C IF NO. READ INTO NNO NOT ALREADY PRESENT IN NNOX,
82380 C THEN TRANSFER COPY OF NNO INTO NNOX
02390 C
02400 C PLACE NUMERICAL ORDER EQUIVALENCE(JZZ) OF ANALYSIS CYCLE
02410 C A — NUMBERS (HELD IN NNO) VS ALL SAMPLE NUMBERS (HELD
02420 C IN NNOX) INTO NNOXY.
02430 C
02440 DO 9392 J = 1,3
02450 NNOXY(J) = 0
02460 9392 CONTINUE
02470 DO 9302 JZ = 2 JKK
02480 00 9304 JZZ = i,COUNT
02490 IF(NNO(JZ).EQ.NNUX(JZZ)) GO TO 9306
0250C 9304• CONTINUE
02510 COUNT = COUNT r 1
02520 IR•OUNT•07.MAXSPL1 HRI7E(IKLP,9305) MAXSPL
02530 9305 FOR4AT(/X,32HNB — NUMEER OF INPUT SAMPLES GT .I4/)
u254C IF(0OUNT.GT.MAXSPL) GO TO 5018
02550 NNOX(COUNT) = NNO(JZ)
02560 NNOXY(JZ-1) = COUNT
02570 GO TO 9302
u2580 9306 CONTINUE
02590 NNOXY(JZ-1) = JZZ
02600 9302 CONTINUE
02610 GO TO 1160
02620 1042 CONTINUE
02630 IE = EL(I) — 50
02640 45 CONTINUE
02650 IF(EL(I).NE.O.AND•7(I)•NE.O.O.AND.GS(I).NE.O.0) GO TO 1190
02660 NRITE(IWLP 1187) NNO(I), LINE EL(I), T(I), CS(I)
02670 1187 FORMAT(/X,24MINCOMPLETE DATA FOR SPL ,I4,6H DL ,I4,6H EL ,
02680 2I2,5H I ,F6.2,7H CTS ,F8.0/)
02690 EL(I) = 0
02700 T(I) = 0 0
02710 CS(I) = 0.1
02720 C(I) _ —0.0
02730 IF(I.EQ.JK) GO TO 1220
0274C GO TO 1210
02750 1190 CONTINUE
02760 C
02770 C REDUCE TIMES AND COUNTS TO INTENSITIES AND MAKE
02780 C DEAD TIME CORRECTIONS
02790 C
12800 A = DTC/1.E6
02810 C(I) = CS(I) / T(I)
02820 C(I) = C(I) / (1.0 — C(I) + A)
02830 IF(NNO(I).GE.8700.AND.NNO(I).LT•90001 GO TO 1221
02840 IF(r,(I).LT.LIMIT) GO TO 1221
02850 HRITE(INLP 1223) LIMIT NN0(I).LZNE,EL( II,C(I)
02860 1223 FORMAT(/X,i1HCT RATE GT ,F7.0,8H FOR ,I4,8H DL ,I4,6H EL
02870 +I2 11H CT RATE ,F7.0/)
02880 1221 CONT
INUE
02890 C
02900 C HAKE SAMPLE HOLDER POSITION CORRECTIONS
02910 C
02920 IF(')NO(I).EQ.9999) GO TO 1220
0293C IF(I.°_0.2) C(I) = C(I) * APRIM
02940 IF(I.EQ.3) C(I1 = C(I) • APRIM
G2950 IF(I.EQ.4) C(I) = C(I) " CPRIM
02960 IF(I.EQ.JK) GO TO 1220
02970 1210 CONTINUE
02980 C
02990 C STATEMENT 1210 CONTINUE IS THE END OF THE INPUT READING LOOP
03000 C (00 1210 I = 1,5) FOR ANY GROUP OF UP TO 5 SPL/STO/BLK/REF DISCS
03710 C
03420 1220 CONTINUE
03030 IS = 0
03040 LIN_X = LINE — JK
03050 C
0396C C STATEMENT DO 5021 I = 1,JK STARTS THE LOOP ASSIGNING
03070 C COUNT RATES TO ELEMENTS F OR A GIVEN SAMPLE
03080 C
6309E DO 5020 I = 1,JK
03100 LIN_X = LINEx + 1
03110 IY = EL(I) — 50
03120 KNO = NNO(I)
03130 C
03140 C SORT ACCORDING TO ELEMENT
03150
C J n
03170 IF(EL(I).EG.NSI) J = 1
03180 IF(=L(I).EO.NTI) J = 2
03190 IF(.L(I).£O.bAL) J = 3
03200 IF(':L(I).EQ.NFE) J = 4
03210 IF(ELtII.EC.NCR) J = 5
03220 IF(EL(I).EQ.NMN) J = 6
63230 IF(EL(I).EO.NMG) J = 7
03240 IF(E L(I).EC.NCA) J = 8
03250 IF(EL(I).EC.NNA) J = 9
120 R. J. PARKER and J. P. WILLIS

03260 IF(EL(I).EQ.NK1 J = 10
03270 IF(FL(I).EC.NP) J = 11
u328C jF(_L(I).EC.NNI) J = 12
03290 IF(EL(I).EG.NELXI J = 13
03300 IF(:L(I).EQ.NELY) J = 14
03310 IF(°_L(I-).EO.NELZ) J = 15
4332E IF(J.NE.0) GO TO 112
03330 WRITE(IWLP 1230) KNO LINEX, EL(I)
I 4HUNIDENTI FIED EL FOR SPL ,I4,6H DL ,I4,6H
03340 1230 FOR4AT(/X L EL ,I2/)
03350 GO TO 5021
03360 112 CONTINUE
03370 IF(KNO.EQ.99) GO TO 2001
03380 IF(KNO.E0.9999) GO TO 2302
03390 GO TO 3000
03400 C
03410 C ASSIGN REFERENCE INTENSITIES TO APPROPREATE ELEMENTS
03420 C OF APPROPREATE REFERENCE ARRAYS
03430 C
03440 2001 CONTINUE
03450 JXX — JKK — 1
03460 DO 9314 JR = 1,JXX
03470 NX = NNOXY(JR)
03480 REF(NX,J,1) = C(I)
03490 9314 CONTINUE
4350C C
03510 C ASSIGN INITIAL REFERENCE INTENSITY FOR ELEMENT J
03520 C
03530 IF(RFIRST(J).E4.0.0) RFIRST(J) = C(I)
03550 2002 CONTĪNUE21
03560 JXX = JKK — 1
03570 DO 9315 JR = 1,JXX
03580 NX = NNOXY(JR)
03590 REF(NX,J,2) = C(I)
03600 9315 CONTINUE
03610 C
03620 C CALC. TRANSIENT DRIFT ON REFERENCE INTENSITIES
03630 C
03640 5021 CONTINUE
03650 IF(REFD(J).E0.0.0) GO TO 5022
03660 IF(C(II.E4.0.0) GO TO 5020
03670 DRIFT = ABStREFO(J) —C(I))
03680 XORIFT = (DRIFT/REFD(J)) * 100.0
03690 IF(XDRIFT.GT.LDRIFT) WRITE(IWLP,3920) KNO,LINEX,EL(I),
+XDRIFT,LDRIFT (NNO(IABC),IABC=2,JKK)
03700 5022
03720 GO TO 5020
03730 3920 FORHAT(/X,16H••- DRIFT ON REF ,I4,6H DL ,I4,6H EL ,I2,
03740 +10H DRIFT F4.1,9H LDF ,F4.1/
03750 + X, 24H AFFEC TS STO/BLK/SPL ,5(I4,2X)/)
03760 C
03770 C ASSIGN SAMPLE/BLANK/STANDARD INTENSITIES TO APPROPREATE
03780 C ELEMENTS OF APPROPREATE ARRAYS
03790 C
03800 3000 CONTINUE
03810 IS = IS + 1
03820 NZ = NNOXYtIS)
03830 X(NZ Jl = C(I)
03840 5020 CONT INUE
03850 C
03860 C GO TO 1130
03870 C
03880 GO TO 1100
03 890 C
03900 C SORT A — NUMBERS INTO ASCENDING ORDER AND
03910 C MAKE INTENSITIES FOLLOW SUITE
03920 C
03930 5018 CONTINUE
03940 00 4040 M=1,COUNT
03950 KKX = COUNT — M
03960 NKR = 0
03970 DO 4042 L = 1,KKX
03980 IF(1NOX(L).LE.NNOX(L+1)) GO TO 4042
03990 MKR = NKR + 1
04000 NNX = NNOX(L)
04010 NNOX(L) = NNOX(L+1)
04020 NNOX(L+1) = NNX
04030 DO 4048 J = 1,15
04040 XA = X(L,J)
04050 X(L,JI = X(L+1,J)
04060 X(L+1,J) = XA
04070 REFA = REF(L,J 1)
04080 REF(L,J,1) = R~F(L+1,J,1)
04090 REF(L+1,J,1) = REFA
04100 REFS = REF(L,J22)
04110 REF(L,J,2) = RtF(L+1,J,2)
04120 REF(L+1 J,2) = REFS
04130 4048 CONTINUE
04140 4042 CONTINUE
04150 IF (MKR.EC.0) GO TO 4046
04160 4040 CONTINUE
04170 4046 CONTINUE
04180 C
04190 C IF SECOND REF = 0.0, THEN ASSIGN FIRST REF TO
04200 C SECOND REF
04210 C
04220 DO 9706 I = 1,COUNT
04230 DO 9708 J = 1,15
04240 IF(REFlI1J,2l.EQ•0.0) REF(I,J,2) = REF(I,J,1)
04250 9708 CONTINUE
L4260 9706 CONTINUE
04270 HRITE(IWLP,5050)
04280 WRITE(IWLP,5027)
04290 5050 FOR4AT(58X,11HSORTED DATA//)
 Computer programs SORT, REORD, and MW for major-element XRF data processing 121

04300 5027 FORMAT(11X,118H SI TI AL MG

+ CA NA /) K P NI FE ELK CR ELY MN ELZ
04320
04330 COUNT = COUNT + 1
04340 NNOX(COUNT) = -99
04350 DO 5019 I=1 COUNT
04360 WRITE(IWLP,5016) NNOX(I), (X(I,J),J=1,15)
04370 WRITE(IWLF,5917) (REF(I,J,1).J=1,15)
04380 WRITE(IWLP 5917) ( REF(I,J,2) J=1,15)
01.390 5016 FORMAT(//1X( I4 4X,i5(F7.D,1X;)
64400 5917 F0R'1AT(9X,T5(F3.O,iX))
04410 5019 CONTINUE
04420 C
04430 C CORRECT FOR DRIFT
04440 C
04450 WRITE(IWLP,500)
04460 WRIT°_(IWLF,502) RFIRST
04470 500 FORMAT(///,X 'INITIAL REFERENCE COUNT RATES - SI, TI, AL, ETC')
04480 502 FORMAT(/,X,15(F7.0))
64490 DO i105 L = 1,151
04500 IF(1NOX(L).E0.-99) GO TO 1110
04510 DO 1101 J = 1,15
04520 IF(REF(L,J,11.EQ.0.0.0R.REF(L,J,2).EQ.0.0) GO TO 1101
04530 XB = 2.0 • RFIRST(J)/(REF(L,J,1) + REF(L,J,2))
04540 X(L J) = X(L,J) • XB
04550 1101 CONTINUE
04560 1105 CONTINUE
64570 1110 CONTINUE
04590 c WRITE OUT SORTED DATA TO FIRST TEMPORARY DISC FILE
04600 C
04610 DO 6070 L = 1 COUNT
04620 WRITE(ITEMP1,,302) NNG((L), (X(L,J),J=1,15)
61.636 6600 CONTINUE
04640 6002 FORYAT(I4,4X 15F8.2)
04650 REWIND ITENPI
04660 RETURN
04670 END

PROGRAM REORD preignited before fusion. The small differences in the

After reading in the generalized data file (Tape 11-N, concentrations listed for the different disc preparations in
Tape 11-G), program REORD reads blank/standard the NORRISH standards in Tape 52—N reflect small
concentrations and matrix factors from Tape 52, each differences in the weight of the standard actually taken for
blank and standard being identified by the four digit fusion. The same weight of powder was taken for all of the
"disc" a-number (see blank/standard numbering). These GUNN standards (see further discussion on fusion
blank/standard data are written out to the line printer weights).
(Tape 61-N, Tape 61-G). The matrix factors in Tape 52-N Program REORD reorders the blank and standard
or Tape 52-G must be calculated using the method of count rate data to produce calibration data sets for each
Norrish and Hutton (1969) or Gunn (1967a, 1967b). The element. These data sets link the count rates for blanks
standard concentrations listed in Tape 52-N and Tape 52-G and standards to their concentrations and matrix factors.
were taken from Abbey (1973) for all the standards except They are written out to the second temporary disc file.
G-1, and for this latter standard the data were taken from Finally a-number/count rate data sets for the samples are
Flanagan (1973). The concentration data were corrected written out to this secondary temporary disc file (Tape
for loss on ignition as all standards (and samples) were 12-N, Tape 12-G).

00100 CC
00110
00120 C PROGRAM REORD - BY R. J. PARKER (DEPT. OF GEOLOGY,
00130 C IMPERIAL COLLEGE,
00140 C UNIV. OF tONDON).
00150 C
00160 C PROGRAF REORD
00170 C REORDERS THE OUTPUT FROM PROGRAM SORT.
00180 C IT DEVELOPES FOR EACH ELEMENT A STANDARD/BLANK STOATA ARRAY,
00190 C AND THESE ARRAYS ARE WRITTEN OUT TO THE SECOND TEMPORARY DISC
00200 C FILE. THE SAMPLE DATA IS ALSO WRITTEN OUT TO THIS FILE.
00210 C
00220 C MAXIMUM NO OF BLANKS IS 20, MAXIMUM NO OF
00230 C STANDARDS IS 100.
00240 C
00250 C I/O UNIT NUMBERS
00260 C
00270 C ITEMPI = - FIRST TEMPORARY DISC FILE
00280 C ITEMP2 = 12 - SECOND TEMPORARY DISC FILE
00290 C IRSB = 52 - STANDARDS/BLANKS INPUT FILE
00300 C IWLP = 61 - LINE PRINTER OUTPUT FILE
U0310 C
00320 C ARRAYS
00330 C
110340 C METHOD(8) - NORRISH OR GUNN OPTION FOR MATRIX CORRECTIONS
00350 C DSCRIF(80) - DESCRIPTION OF RUN ETC
00360 C NAME(80) - ANALYSTS NAME AND OATS ETC
00370 C CHOIS(20) - OPTIONS - SEE DOCUMENTA
TION IN PROGRAM MW
30380 C
00390 C NTITLE (80)
00400 C ABL(20,8) -▪ FĪTH8R
RS HATS
CARC E FOBNKDATAON
EROLA N AAST0 /8 K
00410 C FILE
00420 C NB (20,2) - BLANK NUMBER AND NAME(BLK/STD FILE)
00430 C BCX(20,15) - BLANK GONGS FOR 15 ELEMENTS(BLK/STO FILE)

122 R. J. PARKER and J. P. WILLIS

00440 C BAN(20,15) — BLANK ABSORPTION FACTORS(BLK/STD FILE)

00450 C B7OTAL(20) — BLANK TOTALS(BLK/STD FILE)
00460 C AST(i00,8) — FIRST 6 CHARACTERS OF STANDARD DATA ON BLK/STD
00470 C FILE
U0480 C NS(100 2) — STANDARD NUMBER AND NAME(BLK/STD FILE)
00 490 C SCX(10I,15) — STANDARD GONGS FOR 15 ELEMENTS(BLK/STD FILE)
00500 C SAN(100,15) — STANDARD ABSORPTION FACTORS(BLK/STD FILE)
00510 C 0) — STANDARD TOTALS(BLK/STD FILE)
STOTAL(11
00520 C
00530 C STDATA(332,5) — ARRAY USED FOR BUILDING UP CALIBRATION DATA
00540 C SETS FOR EACH ELEMENT
00550 C STDATX(5) — TEMPORARY ARRAY USED IN SORTING STDATA(332,5)
00560 C INTO ASCENDING ORDER OF CONCENTRATION
00570 C
60580 C QU O
00590 C RNB(20 2) — ) A N DLNS( iOŌS2) O VFACILI T A T E T RANSFER OF
00600 C RNS(100,2) — ) ALPHANUMERICS IN ARRAYS NB AND NS INTO
00610 C ARRAY STOATA
00620 C
00630 C ENOX(15) — IS USED TO TERMINATE THE SAMPLE DATA
00640 C ON THE TEMPORARY DISC FILE
00650 C
U0660 C
00670 C PARAMETERS
00680 C
00690 C DTC — DEAD TIME COEFICIENT
00700 C NBX — NUMBER OF BLANKS ON BLANK/STANDARD FILE
00710 C NSX — " STANDARDS " " "
u0720 C K — INDEX FOR BLANKS/STANDARDS IN CALIBRATION ARRAY STDATA
00730 C CTR — COUNTER (OR INDEX) FOR A — NUMBER AND INTENSITY ARRAYS
00740 C BCOUNT — NUMBER OF BLANK A — NUMBERS
00750 C SCOUNT — NUMBER OF STANDARD A — NUMBERS
00760 C
00770 C NNXYZ —
00780 C NNNX — ) PARAMETERS FOR REDUCING INPUT DATA
00790 C NXXN — ) A — NUMBERS TO DISC A — NUMBERS
00800 C NZNZ — )
00810 C
00 820 C KX — ) PARAMETERS FOR SORTING CALIBRATION ARRAY STOATA
00830 C NTX — ) INTO ASCENDING ORDER OF CONCENTRATIONS
00840 C
00851; C CTX — COUNTER FOR A — NUMBERS AND INTENSITIES, IS USED
00860 C IN WRITING OUT SAMPLE DATA TO TEMPORARY FILE
u0670 C NEND — TERMINATOR A — NUMBER FOR TEMPORARY FILE WRITE OUT
00880 C
00890 C
00900 C
00910 PROGRAM REORD(TAPE6I,TAPE52 TAPEII,TAPEI2,OUTPUT)
00 920 DIMENSION STDATA(332,5),STDATX(5)
u0930 DIMENSION IB(20,2),BCX(27,15),BAN(20,15),NS(100,2),SCX(100,15),
00940 +SAN(190,15),BTOTAL(20),STOTAL(110),AUL(20,8),AST(160,8)
00950 INTEGER ABL AST
00960 DIMENSION NAME(80),DSCRIP(8U),CHOIS(20),METHO0(8),NTITLE(80),
60970 +ENDX(15),NNOX(151),X(151,15)
00980 DIMENSION RNO(20 2) RNS(100 2)
00990 EQUĪVALENCE (RNB ~I,i) NB(1 1)) (RNS(1 1) NS(1,S) 1
01000 L
INTEGER CTX,DCOUNT,SCOUNT,CTR,SUMSP ,ISCRIP,CHOIS
01010 DATA JF /SHF/
01020 C
01030 C SET I/O UNIT NUMBERS
u1U40 C
01u50 ITEMPI = 11
01060 ITEIP2 = 12
01070 IRSB = 52
0108C IWLP = 61
01090 C
01100 C INITIALIZE ARRAYS
01110 C
01120 DO 2994 T. = 1,20
0113C NB(I,1) = 0
01140
01150 DO02994 IJ = 115
01160 BCX(I,J) = b
01170 UAN(I,J) = 0.0
01180 2994 CONTINUE
01190 I = 1,110
01200 00299)
01210 STOTAL(I) = 0.0
01220 00 2996 J = 1,15
0123C SCX(I,J)
0124C = 0.0
01250 2996 CONTINUE
u1260 G
01270 C READ HEADER DATA FROM FIRST TEMPORARY DISC FILE AND
01280 C WRITE IT OUT TO THE SECOND TEMPORARY DISC FILE
01290 C
U1300 READ(ITEMP1,7081) METHOD
01310 READ(ITEMF1,7081) NAME
61320 REA7(I7EMP1,7081) DSCRIP
01330 READ(ITEMP1,7083) DTC
0134C READ(ITEMP1,7085) CHOIS
0 1350 READ(ITEMP1,7088) APRIM, BPRIM, CPRIM
01360 WRITE(ITEMP2,7381) METHOD
01370 WRIT=(ITEHF2,7J81) NAME
01380 WRITE(ITEMP2,7381) DSCRIP
01390 WRITE(ITEMP2,7083) DTC
01400 WRITE(ITEMF2,7085) CHOIS
01410 WRITE(ITEKP2,7J88) APRIM, BPRIM, CPRIM
01420 7081 FORIAT (80A1)
01430 7083 FORIAT(F3.1)
01440 7085 FDRMAT(23II)
01450 7087 FORIAT(6(F6.4,1X))
01460 7088 FORMAT(3(F6.4,1X))
01470 C
01480 C READ IN SORTED DATA FROM FIRST TEMPORARY DISC FILE
0149C C
 Computer programs SORT, REORD, and MW for major-element XRF data processing 123

01500 00 1010 L = 1,151

u1511 READ(ITEMF1 1 3 )12 NNOX(L),(X(L,J),J=1,15)
01520 IF(~lNOX(L).Ē0.-99) GO TO 1014
01530 101U CONTINUE
0154C 1012 FORMAT(I4,4X,15F8.2)
U1550 1014 CONTINUE
01560 C
01570 C
01580 C READ IN BLANK AND STANDARD DATA (CONCENTRATIONS AND
01590 C MATRIX FACTORS). NO OF BLANKS IS NBX, NO OF STANDARDS
0160C C IS NSX.
01610 C
01620 READ(IRSB 4) NTITLE
4 FORMAAT(80 A 1)
3 0
016 I = 0
01651 NBX = 0
01660 3000 I = I + 1
01670 NBX = NBX + 1
61680 READ(IRSB,3006) (ABL(I,J),J=1,8),NB(I,1),NB(I,21,(BCX(I,J),J=1,15)
01690 IF(ABL(I 1 E1I.JF)
). GO TO 3015
01700 REA7(IRSR,3011) (BAN(I,J) J=1,15)
01710 IF(NBX.£Q.20) WRITE(IWLP,~J012)
01720 IF(NBX.E0.20) GO TO 3015
01731 GO TO 3000
01740 3015 I = 0
01750 NSX = 0
01761 3020 I = I + 1
01770 NSX = NSX + 1
01780 READ(IRSB 3006) (AST(I J) J=1,8),NS(I,1),NS(Is2),lSCX(I,J),J=1,151
01790 IF(AST(I,1).EQ.JF) GO TO 3025
01800 READ(IRS8,3011) (5AN(I,J1,J=1,15)
01810 IF(NSX.EQ.100) WRITE(IHLP,3013)
01820 IF("1SX.E0.300) GO TO 3025
01830 GO TO 3020
01840 3006 FORHAT(8A1•,2X I4,2X,A6,2X 8F7.3/24X,7F7,3)
01850 3011 FORHAT(24X,8F7)4/24X 7F7.~,)
01860 3012 FORHAT(/,X,27HNB - MAX NO OF BLKS READ IN)
01870 3013 FORMAT(/,X,27HNB - MAX NO OF STDS READ IN)
01880 3025 CONTINUE
01890 no 1001 I = 1,NBX
01900 DO 1001 J = 1,15
31910 BTOTAL(I) = 8TOTAL(I) * BCX(I,J1
01920 1001' CONTINUE
01930 DO 1002 I = 1,NSX
01940 00 1002 J = 1,15
01950 STOTAL(I) = STOTAL(I) + SCX(I,J)
01960 1002 CONTINUE
01970 C
01980 C WRITE CUT BLANK AND STANDARD DATA
01990 C
02000 WRITE(IWLF,5) NTITLE
02010 5 FOR4AT(1H1///115X 80A1///14X
02020 +113H SI02 TIO2 AL203 FE 203 CR203 MNO MGO CAD NA20
02030 + K20 P205 NIO ELX ELY ELZ TOTAL /)
02040 00 6 I = 1,NBX
02050 WRITE(IWLP,71 NB(I 1) NB( I,2) (BCX(I,JI,J=1,15),BTOTAL(I)
02060 WRITE(IWLP,8)(BAN(),J),J =1,15)
0200 6 CONTINUE
02080 7 FORMAT(/1x,I4 1X,A6 1X,15(F7.3),F9.3)
02090 8 FORMAT(13X,15(F7.4) )
02100 00 9 I = 1,NSX
02110 WRITE(IWLP,7) NS(I 1) NS(Is2) (SCX(I,J),J=1,15),STOTAL(I )
02120 WRITE(IWLP,8)(SAN(I,J1,15 ),J= )
02130 9 CONTINUE
02140 C
02150 C DO 40 J = 1,15
02160 C BUILD UP CALIBRATION ARRAY(STDATA(332,5)) FOR EACH ELEMENT
02170 C
02180 00 40 J = 1,15
02190 DO 1005 I = 1,332
02200 DO 1005 K = 1,5
02210 1005 STDAT A(I,K) = 0.0
02230 . CTR = 0
02240 BCOUNT = 0
02250 SCOUNT = 0
02260 STOAT/M(11) = J
02270 C
02280 C REDUCE A - NUMBER TO PRIME A - NUMBER, THEN SUBTRACT THIS PRIME
EAU-
02300 C TIONN(NXXN) FTHENTĪNDĪCATESANETHER NTHEEORI GĪN A L NUMBER REFERS
02310 C THE FIRST SECOND, OR THIRD DISC PREPARATION. THE A - NUMBER IS THEN
62320 C SET (IK NNXYZ) TO THE THE PRIME A - NUMBER PLUS NZNZ (IE PLUS 0s
02330 C 3 OR 6) DEPENDING ON NETHER IT IS THE FIRST, SECOND, OR THIRD DISC
02340 C PPARATION.
RE
02350 C
02360 10 CONTINUE
02370 K = K + 1
u2380 11 CTR = CTR + 11
02390 NNXYZ = ((NNOX(CTR)/10) a 10)
02400 NNNX = NNOX(CTR)
02410 NXXN = NNNX - NNXYZ
02420 IF(4)(XN.LE.2) NZNZ = 0
02430 IF(NXXN.GE.3.AND.NXXN.LE.5) NZNZ = 3
U2440 IF(NXXN.GE.6) NZNZ = 6
02450 NNXYZ = NNXYZ + NZNZ
02460 C
02470 C TEST FOR A - NUMBERS LESS THAN 8000 (IE SAMPLES), IF FOUND THEN
02480 C IGNORE. ALSO CHECK THAT COUNTS IN X(CTR,J) ARE PRE SENT, IF NOT
02490 C THEN IGNORE.
02500 C
02510 IF(NNOX(CTR).LT.8000) GO TO 11
IF(X(CTR,J).GT.C.0) GO TO 20
02530
20
124 R. J. PARKER and J. P. WILLIS

02540 IF(INOX(CTR+1).E0.-99) GO TO 30
02550 GO TO 10
02560 20 CONTINUE
0257C C
02580 C INCREMENT BLANK OR STANDARD COUNTER DEPENDING ON A - NUMBER
0259C C
02600 IF(TN0X(GTR).L7.9000) 3COUNT = BCOUNT + 1
0261C IF(ANOX(CTR).GE.9000) .COUNT = SCOUNT + 1
02620 C
0263C C TEST FOR BLANK OR STANDARD A - NUMBER. IF BLANK THEN LOOK FOR
0264C C EQUIVALENCE BETWEEN NNXYZ AND BLANK A - NUMBERS STORED IN NB(I,1).
U2650 C WHEN EQUIVALENCE FOUND THEN TRANSFER BLANK DATA INTO STDATA ARRAY
02660 C
02670
U 2680
02690
IF(NNOX
STDAT.4(K 1) = NNOX(CTR)
UTR).GE.9U00) GO TO 401
DO 22 I=1,NBX
02700 IF(9NXYZ.NE.NB(I,1)) GO TO 22
02710 IF(3CX(I,J).E0.0.01 GO TO 21
02720 dCOUNT = BCOUNT - 1
02730 K = K - 1
02740 GO TO 10
02750 21 CONTINUE
02760 STDATA(K,2) = RNB(I,2)
02770 GO TO 24
0278U 22 CONTINUE
02790 WRITE(IWLP,28) NNOX(CTR)
02800 28 FORMAT(//1X,30HWARNING - NO DATA FOR STD/BLK ,I4//)
02810 °COUNT = BCOUNT - 1
02820 K = K - 1
02830 GO TO 405
0284C 24 STDATA(K,3) = BCX(I,J)
02850 STDATA(K,4) = BAN(I,J)
02860 STDATA(K 5) = X(CTR,J)
02870 GO TO 40 5
02880
0289C
02901
C401
C
CONTINUE
LOOK FCR EQUIVALENCE BETWEEN NNXYZ AND STANDARD A - NUMBER STORED
02910
02920
02930
CC IN ARRAY NS(I,1). WHEN FOUND TRANSFER STANDARD DATA INTO STRATA ARRAY
DO 402 I = 1,NSX
02940 IF('INXYZ.NE.NS(I,1)) GO TO 402
02950 IF(SCX(I J).NE.0.0) GO TO 1401
C2960 IF(J.EQ.7) GO TO 1401
02970 SCOJNT = SCOUNT - 1
02980
02990
03000
IF
K = K - 1
(INOX(CTR+1).EQ.-99) GO TO 30
GO TO 10
03010 1401 CONTINUE
03020 STDATA(K12) = RNS(I,2)
03030 GO TO 403
0374E 402 CONTINUE
03050 WRITE(IWLP 28) NNOX(CTR)
03060 SCOJNT = SCOUNT - 1
03')70 K = K -
03080 GO TO 405
03'J9L 403 STDATA(K,3) = SCX(I,J)
03100 STDATA(K,4) = SAN(I,J)
03110 STDATA(K15) = X(CTR,J)
03120 405 CONTINUE
03130 IF (1NOX (CTR+1). EQ.-99) GO TO 30
03140 GO TO 10
03150 C
03160 C INSERT INTO STRATA ARRAY THE NUMBER OF BLANKS AND STANDARDS FOUND
03170 C FOR ELEMENT U, ANC THEN ADD TERMINATING DATA TO STRATA ARRAY
03180 C
03190 30 CONTINUE
03200 IF(K.IT.2) GO TO 35
03 210 STDATA(1,4) = DCOUNT
03220 S1DATA(1,5) = SCOUNT
03230
03240 STDATA(K 1) _ -1
03250 IF(")NOX(G -
( TR+1).E0. 99.AND.J.EQ.15)STDATA(K,1) = -99
03260 35 CONTINUE
03270 IF(K.LT.2ISTOATA(K+1 1) _ -1
03280 IF(K.LT.2.AND.J.EQ.15)3TOATA(K+1,1) = -99
03290 IF(K.LT.2) K = K + 1
03300 C
03310 C SORT BLK/STD ARRAY (STDATA) INTO ASCENDING ORDER OF CONCS
03320 C
03330 DO 5075 N = 1,K
03340 KX = K - N - 1
0335C NTX = 0
03360 DO 5015 I = 1,KX
03370 IF(STOATA(I,3).LE.STDATA(I+1,3)) GO TO 5015
03380 NTX = 1
03390 DO
03400 STDATX(JX)
X(JX) = ST
DATA(I+i,JX)
03410 STDATA(I+1,JX) = STDATA(I,JX)
03420 STOATA(I,JX) = STDATX(JX)
03430 5020 CONTINUE
03440 5015 CONTINUE
03450 IF(JTX.EQ.01 GO TO 5025
03466 5005 CONTINUE
03470 5025 CONTINUE
03480 C
03490 C WRITE CALIBRATION ARRAY FOR ELEMENT J TO
03500 C SECOND TEMPORARY DISC FILE
03510 C
03520 WRITE(ITEMPZ 9601) ((STDATA(KX,L),L=1,5),KX=1,K)
03530 9601 FORMAT(F6.0,22X,A6,2X,F7.3,2X,F7.4,2X,F8.2)
03540 40 CONTINUE
U3550 C
03560 C WRITE SAMPLE DATA (ELEMENTS 1 TO 15) TO
03570 C SECOND TEMPORARY DISC FILE
03580 C
03590 CTX = 0
03600 150 CONTINUE
 Computer programs SORT, REORD, and MW for major-element XRF data processing 125

03610 CTX = CTX rr 1

03620 IF111110 MT X).EQ.-99) GO TO 210
03630 IF(lNOX(CTXI.GE.8000) GO TO 210
03640 WRITEtITEMP2,9602) Mal( (CTX), (X(CTX, hi) ,M=1,15)
03651. 9602 FORMAT(I4,4X,15F8.21
03660
210 CONTĪNUEO
03680 NEND = —999
03690 DO 211 11 =1,15
03700 ENOX(9) = 0.0
03710 211 CONTINUE
u3720 WRITE(ITENP2,9602) NEND, (ENDX(M),M=1,15)
03730 REWIND ITENP2
03740 URN
END

PROGRAM MW These standards should be made up in the same manner as

Subroutine SLOPE reads the header data from the the other standards and the data on their concentrations
second temporary disc file and then reads the calibration and calculated matrix factors should be entered in the
data sets for each of the elements analyzed. The standardstblanks file (Tape 52-N).
calibration data sets are used to calculate the "blank" and The first time the Mg0 calibration line is calculated, no
"slope" of the calibration lines for the elements, as well as interference correction is applied. Subroutine MGIS is
the absolute and relative errors with respect to the called and the Mg0 interference correction factors are
calibration line for each standard counted (Tape 62- N, calculated for the six interfering oxides, as well as the
Tape 62-G). total c/s Mg0 interference to be expected for each
The slope of the calibration line is calculated by standard. Control returns to subroutine SLOPE which
dividing the sum of the standard concentrations by the recalculates the Mg0 calibration line, subtracting the
sum of standard count rates. This method of calculating appropriate Os Mg0 interference correction from each
the slope means that the high concentration (and therefore standard count rate. The absolute and relative errors for
high count rate) standards contribute relatively more to the corrected Mg0 calibration line are improved con-
the resulting slope factor. This is desirable as the higher siderably (Tape 62-N). Note that because the Mg0
XRF intensities give better counting statistics, and the interference standards have 0.0 Mg0 they do not
standard concentration data at these higher levels may be contribute to the standard summations used for deter-
of superior quality as compared to data for low- mining the slope of the Mg0 calibration line. If the
concentration standards. A linear regression through the background count rates are being determined by iteration,
data would weight the data equally, and the resulting slope then the Mg0 calibration is printed only once after the
factor would not reflect the better quality data at the high iteration has ended. Test data from subroutine MGIS will
concentrations. be written out if the chois (9) option is switched on.
Two methods are available for determining the "blank" If desired (see option documentation in program MW)
or background intensity for the standards and samples. the calculated calibration data for each element are
This value may be determined directly from the blanks written out (Tape 62-N. Tape 62-G), and the calibration
(see Norrish and Hutton, 1969, p. 443). This method also lines may be plotted (Fig. 2).
can be used for blanks prepared using the Gunn Subroutine CONC reads the first sample data set from
technique. Alternatively the background count rate may the temporary disc file. It reads in from Tape 53 the
be determined by iterating (from zero background c.p.s.) corresponding a-number, name, and weighings for the
the calibration line calculation so as to select the sample, as well as sodium and FeO values if the latter has
background intensity that gives the lowest average been determined from wet chemistry (see Tape 53-N,
relative error (see option documentation in program MW). Tape 53-G). The weighings are used to calculate the
For most calibrations this iterative technique produces
lower average relative and average absolute errors as
compared to measuring the backgrounds directly on 100.00
blanks (see Parker, in press).
The count rate observed for Mg0 is affected by crystal
80.00
fluorescence which is sample/standard composition
ō

dependent, as well as being dependent on the pulse height

settings on the spectrometer. In order to correct for this,
Counts /secon d

6000
six Mg0 interference standards should be prepared
containing (say) 70% Si02 and 30% of the interfering oxide
(that is, Ti02, Fe203, Cr203, MnO, CaO, and K20). These 4000
interference standards must be assigned "prime" a-
numbers as follows:
2000

Ti02 interference standard = 9900

Fe203 interference standard = 9910 I I I I I
20.00 4000 60.00 60.00 100.00
Cr203 interference standard = 9920
WT. PC. SI02
Mn0 interference standard = 9930
Ca0 interference standard = 9940 Figure 2. Calibration data for each element may be written out on
K20 interference standard = 9950 plotter.
126 R. J. PARKER and J. P. WILLIS

00100 C
MW - BY J WILLIS (DEPT. GEOCHEMISTRY,
UNIV. OF
00110 C
00120 C PROGRAM. P
00130 C OF CAPE TOWN)
00140
00150 C
C
AND R J PARKER (DEPT. OF GEOLOGY
IMPERIAL COLLEGE,
00160 C
00170 C
UNIV. OF LONDON(
00180 C PROGRAM. MW CONVERTS MAJOR ELEMENT X-RAY FLUORESCENCE (XRF)
00190
00200 C
C
INTENSITIES MEASURED IN THE ANALYSIS OF ROCK SAMPLES,
INTO OXIDE CON
CENTRATIONS. THE
CONVERSION PROCESS, FOR
00210
00220 C
C
UP
00230 C FROM XRF INTENSIT IES MEASURED ON KNOWN ROCK STANDARDS.
00240
00250 C
C
MATRIX CORRECTIONS ARE APPLIED TO THE XRF INTENSITIES USING
EITHER THE METHODS OF NORRISH AND HUTTON(1969 GEOCHIM. ET
00260
00270 C
C
COSMDCHIM. ACTH V33,PP431-453) (SEE SUBR NORFAC) OR OF
GUNN( 1968,CANAD IAN JOUR. OF SPECTROSCOPY,V12,PP41-46,PP163-168)
00280
00290 C
C
(SEE SUBR GUNFAC).
00300
00310 C
C
PROGRA1 HW HAS BEEN WRITTEN IN ASA FORTRAN IV, EXCEPT
FOR THE 'PROGRAM HEADER STATEMENT AND FOR
00320
00330 C
C
DATA STATEMENTS(SEE COMMENT NOTES)
00340
00350 GC NOTE PROGRAM HW IS DESIGNEDTO RUN IN CONJUNCTION WITH
PROGRAM REORD. THIS LATTER PROGRAM WRITES TO A TEMPORARY
00360
00370 C
C DISC FILE THE INPUT DATA FOR PROGRAM MW.
00380
00390 C
C NOTE THE OUTPUT FROM PROGRAM MW REQUIRES A LINE PRINTER
WITH 132 PRINT POSITIONS.
60400 C
00410 CC OPTIONS
60420
00430 CC IF METHOD = NORRISH THEN NORRISH + HUTTON MATRIX CORRECTIONS
00440
00450 C
IF METHOD = GUNN THEN GUNN MATRIX CORRECTIONS
0C466
00470 C
C
IF CHOIS(1) = 1 THE OXIDE CONCS. ARE PUNCHED ON CARDS
IF CHOIS(2) = 1 THE ELEMENT CONCS. ARE PUNCHED ON CARDS
00480
00490 C
C
IF CHOIS(3) = 1
IF CHOIS(4) = 1
THE PROGRAM EXITS AFTER CALC. CALIBRATION LINES
THE WORKING CURVES ARE PLOTTED VIA SUER. GRAPH
00500
00510 C
C IF CHOIS(5) = 1
IF CHOIS(6) = 1
THE PLOTS ARE DRAWN HALF SIZE
THE WORKING CURVE DATA IS NOT PRINTED
00520
00530 CCC IF CHOIS(7) = 1
IF CHCIS(8) = 1
ANY NEGATIVE GONGS. ARE SET TO ZERO
THE IRON CONC. IS PRINTED OUT AS FEO
00540
00550 CC IF CHOIS(9) = 1
IF CHCIS(IC) = 1
S
L A NKO VA LUES FOR
T HE PGM SETSUSLOPETAND•BL
0056C
06570 C IF CHOIS(11) = 1
NORRISH TEST DATA
THE PGH SETS SLOPE AND BLANK VALUES FOR
00580
06590 C
C IF CHOIS(12) = 1
GUNN TEST DATA
BKGDS ARE DETERMINED BY ITERATION OF THE CALIBRA-
00600
00610 C
C IF CHOIS(13) = 1
TION LINES (BKGDS NORMALLY DET. FROM BLANKS)
THE PGM WRITES OUT 'N-3' (NOMINAL MINUS BLANK)
00620
00630 C OXIDE CONCENTRATIONS
IF CHOIS(14) = 1 THE PGA WRITES OUT ELEMENT CONCENTRATIONS
00646 CCC
06650
00660 CC
60670 I/O UNIT NUMBERS
00680
00690 CC IRNW = 53 - SAMPLE NAME/WEIGHING INPUT FILE
LINE PRINTER OUTPUT FILE
6070C C ĪPNCH = 67 PUNCH CARO OUTPUT FILE
00710 ITEMP2 = 12 - TEMPORARY INPUT FILE
PRODUCED BY PROGRAM
INPUTREORD
C
00720
06730 C
C IRSTD = 52 - STANDARDS/BLANKS FILE
00740
00750 C
C I WT =6 - O UTPVT 7o TNī7_K 4cr/ v, 'TEAM /AIAl-
00760 CC
00770 NOTE THE FIRST PROGRAM STATEMENT 'PROGRAM MW(TAPES)'
IS A CCC FORTRAN STATEMENT FOR DECLARING DATA FILES
00790 C
0080C
00810 PROGRAM MW(TAPE7,TAPE12,TAPE53,TAPE62,TAPE52,TAPE27,
+OUTPUT,TAPE6=OUTPUT)
00820
6083E COMMON DSGRIP(80),NANE(80) CHDIS(2O) ,METHOD(8),0TC,
0084C
00 850
+COMMON /AA/IU(15). BLANK(151,W AVISP(6)
INTEGER CHOIS
00860
00870 ITEMP = 12
IRSTO = 52
0088C IRN4 = 53
0089C IWLP = 62
00900 IPNCH = 7
00910 I WT = 6
0C92C CALL SLOPE
00930 IF(:HOIS(3).EQ.1) GO TO 50
0094C CALL CONC
00950 50 CONTINUE
U0960
00970 STOP
END
00980 C
00990 C
01000 C
01010 C SUBROUTINE SLOPE READS THE CALIBRATION DATA FROM THE
01020 C TEMPORARY DISC FILE AND CALCULATES THE 'SLOPE' OF THE CALIBRATION
01030 C LINE AND THE 'BLANK' CONCENTRATION FOR EACH OF THE ELEMENTS
01040 C ANALYSED.
01050 C
01060 C
01070
01080 C
C
ARRAYS
01090 C
01100 C
METHOD(8) - NORRISH OR GUNN OPTION FOR MATRIX
DSCRIP(80) - 80 CHARACTERS FOR DESCRIPTION OF
CORRECTIONS
RUN ETC.
01110 C NAME(80) - 80 CHARACTERS FOR ANALYSTS NAME AND DA TE. ETC.
01120 C CHOIS(12) - OPTIONS, SEE DOCUMENTATION ABOVE
 Computer programs SORT, REORD, and MW for major-element XRF data processing 127

01130 C
01140 C U(15) - SLOPES OF CALIBRATION LINES
01150 C BLANK(15) - BLANK CONCENTTRATIONS
01160 C OXIDE(15) - OXIDE NAMES
ū117C C STDATA(100,5) - CALIBRATION DATA FOR UP TO 100 STOS/ELKS
0118E C SBNUM(100) - STD/BLK A - NUMBERS TRANSFERED FROM STDATA(100,5)
01190 C SBNAM(100) - STD/BLK NAMES TRANSFERED FROM STOATA(100 5)
L1206 C C(100) - STO/BLK CONCENTRATIONS TRANS. FROM STDATA(100.5)
01210 C B(100) - INTENSITIES FOR STO/ELKS
01220 C 0(100) - STO/BLK CONES. CALC. FROM CALIB. SLOPE • STD/BLK INTEN.
01230 C G(100) - ABSOLUTE ERROR C(I) - D(I) IE DIFFERENCE BETWEEN
01240 C RECOMENDED STD/BLK CONC. AND CALC. STO/BLK CONC.
01250 C H(100) - RELATIVE(OR PERCENTAGE) ERROR G(I)/C(I) * 100
01260 G XXF(15) - AVERAGE ABSOLUTE ERROR OF STD/BLK CONCS. N.R.T.
01270 C CALIBRATION LINES FOR THE 15 ELEMENTS
01280 C XXH(15) - AVERAGE RELATIVE(OR PERCENTAGE) ERROR OF STO/BLK CONES.
61290 G N.R.T. CALIBRATION LINES FOR THE 15 ELEMENTS
01300 C XFON(15) - FIGURE OF MERIT ARRAY FOR THE 15 CALIBRATION LINES
01310 C NUM(50) --- SEE SUBR MGIS
01320 C SUMCS(50) -
01330 C SUMCSX(100) - MG C/S INTERFERENCE CORRECTION FOR STANDARDS
61340 C
0135C C PARAMETERS
0i360 C
01370 C DIG - DEAD TIME COcFICIENT
01380 C NPLOT - NUMBER OF PLOTS
01390 C SUMSB - NUMBER OF BLANKS + NUMBER OF STANDARDS + 2
01460 C M - ELEMENT NUMBER
01410 C N9 - NUMBER OF BLANKS
01420 C NS - NUMBER OF STANDARDS
01430 C RNS — NUMBER OF 5T05 USED IN CALIBRATION LINE
01440 C ISX - SEE SUBR MGIS
01450 C MGCTR - COUNTER FOR NUMBER OF TIMES SUBR MGIS IS CALLED
G1460 C HUMP - PRIME STD A - NUMBER
01470 C ZNB FLOATING POINT EQUIVALENT OF NB
01480 C L - INDEX FOR ARRAYS S3NUM(100) SBNAM ( 100) AND C(100)
01490 C XG - MATRIX ABSORPTION FACTOR F OR CORRECTING
u1500 C OBSERVED STO INTENSITY
01510 C E - OBSERVED STD INTENSITY
01520 C XC - BLANK INTENSITY
01530 C NF - FIRST INDEX OF STD IN ARRAY S C(1 30) AND B(1a0)
01540 C NL - LAST ,. .. '. .~ .. .. ..
01550 C XO - SUM OF C(I) FOR ALL STANDARDS COUNTED FOR CALIBRATION 'I'
01560 C XE - SUM OF B(I) FOR ALL STANDARDS COUNTED FOR CALIBRATION 'M'
01570 C XF - AVERAGE ABSOLUTE ERROR FOR STOS FOR A PARTICULAR
01580 C CALIBRATION LINE
01590 C XM - AVERAGE RELATIVE PERCENTAGE ERROR FOR STDS FOR
61600 C A PARTICULAR CALIBRATION LINE
01610 C XHX - XH FROM PREVIOUS BKGD ITERATION
131620 C ESCAPE - FLAG USED TO ESCAPE FROM BKGD ITERATION
01630 C FOM - FIGURE OF MERIT ((SUM OF ABSOLUTE ERRORS DIVIDED BY SUM
01640 C OF STANDARD CONCENTRATIONS) • 100.0) IE A PERCENTAGE
01650 C A - OTC CONVERTED FROM E NOTATION
01660 C NUMX - OLD 'A - NUMBER'(LAST DIGIT SET TO 0)
01670 C NUMY - NEW 'A - NUMBER'( " '' " )
61680 C
01690 C
01700 SUBROUTINE SLOPE
01710 COMMON OSCRIP(30) NAME(80) CHOIS(20),METH00(8),DTC,
01720 +ITEMP,IWLP,IPNCH IRNW,IRST Ō IWT
01730 COMMON /AA/ U(15). BLANK(15), AVISP(6)
01740 COMMON /BB/ 0(100) BX(100), C(100) OXIDE(15)
01750 COMMON /CC/ NA4(50),Sl.ONCX(15),SCONOZ(50,15).AVIS(6),
0176C +NOS(6),IS(6),CIS(6),NUM(50),SUMCS(50),SUMCSX(100),
01770 +SBNUM(100)
01780 DIMENSION STDATA(100 5) SBNAM(100),D(100),G(1001,
0179C +H(1)0) XXF(15) XXH(15), XF04(15)
01800 INTEGER CHOIS,SONUM,SUMSB
01810 C
01820 C INITIALIZE VARIABLES AND ARRAYS
01830 C
01840 NPLDT = 0
01850 ISX = 0
01860 DO 12 I = 1 6
01870 AVISP(I) = 0.0
01880 12 CONTINUE
01890 DO 904 J =1,15
01900 U(J) = 0.0
01910 BLANK(J) = 0.0
01920 904 CONTINUE
01930 C
01940 C READ IN HEADER DATA FROM TEMPORARY DISC FILE
01950 C
01960 R£AD(ITEMP,5010) METHOD
01970 READ(ITEMP,5010) DSCRIP
01980 REAO(ITEMP,5010) NAME
01990 REAO(ITEHP,5040) OTC
02000 READ(ITEMP,5060) CHOIS
088) APRIM, BPRIH, CPR IM
5010 FRAT1
20 OM(80Aj
02030 5040 FORMAT(F3.1)
02050 5080 FORHAT(6(F6.4,X))
5088 FORHAT(3(F6.4,IX)) 02.5g0
020 60 [
02080 C THE SEOENCE TO STATEMENT NUMBER4249- READS FROM THE TEMPORARY
02090 C DISC FILE THE CALIBRATION DATA FOR THE 'NTH" ELEMENT.
02100 C THIS DATA IS THEN ASSIGNED TO APPROPREATE VARIABLES
02110 C AND ARRAYS.
02120 C
02130 C H IS ELEMENT NUMBER
02140 C NB IS NUMBER OF BLANKS
02150 C NS IS NUMBER OF STANDARDS
02160 C •E' IS BLK/STD/SPL INTENSITY
02170 C SBNUM(L) IS ELTH STD/BLK NUMBER
02180 C SBNAM(L) IS ELTH STO/BLK NAME
02190 C C(L) IS ELTH STD/BLK CONC
128 R. J. PARKER and J. P. WILLIS

02200 C XG IS ABSORB FACTOR

02210 C
G2220 1001 CONTINUE
02230 1084 CONTINUE
02240 00 200 I = 10.00
02250 SUMCSX(I) = 0.0
02260 DO 20q 0 IX = 1,5
02270 STDĀTA(I,IX) = 0.0
02280 200 CONTINUE
02290 SUMS3 = 0
02300 NBLK = 0
02310 NSTD = 0
02320 REA3(ITEMP,199) (STDATA(i,L),L=1,5)
02330 N8 = IFIX(3TDATA(1,4))
02340 NS = IFIX(STDATA(1 51)
02350 SUMSB = NB + NS + 2
02360 READ(ITEMP 199) ((STDATA(KX,L) L=1.5),KX=2,SUMSB)
0237C 199 FORMAT(F6.00,2X A6,2X F7.3 2X F??.4,2X,F8.2)
02380 IF(SUHS0.EC.2..S7OATA
A ND G0UH B,1).EQ.-99.0) GO TO 1020
02390 IF(SUMSB.E0.2) GO TO 1084
02400 M = STOATA(1,1)
02410 L = 0
02420 00 209 K = 2 SUMSB
02430 IF(K.EQ.SUMSB) GO TO 210
02440 L = L + 1
C2450 SBNUM(L) = STDATA(K,1)
02460 SBNAM(L) = STDATA(K,2)
02470 C(L) = STDATA(K 3)
02480 XG = STOATA(K,4j
02490 E = STDATA(K,5)
02500 C
02510 C NORMALIZE INTENSITY E FOR. AFACTOR XG ANO
02520 C ASSIGN RESULT TO ARRAY 8(L)
0253C C
02540 BX(L) = E * XG
02550 209 CONTINUE
02560 216 CONTINUE 02140
82570 C
02580 C THE SECUENCE TO STATEMENT NUMBER) CALCULATES THE
02590 C AVERAGE BLANK INTENSITY (XC)
02600 C
02610 XC = 9.0
02620 IF(NB.EQ.0) GO TO 4
32630 00 1919 I — 1 N8
02640 XC = XC + BX(I)
u2650 1010 CONTINUE
02660 ZNB = FLOAT(NB)
02571 XC = XC/ZN9
02680 4 CONTINUE
02690 NF = NB + 1
02700 NL = NB + NS
02710 C
82720 C NF ANC NL ARE NOW THE FIRST ANO LAST INOICES OF THE
02730 C STANOARDS(AS OPPOSED TO BLANKS) IN THE ARRAYS C(I) AND B(I)
02740 C
02750 C
02760 C TRANSFER BLANKS FROM 8X(I) TO 8(I)
02770 C
02780 00 3912 I = 108
02790 8(I) = BX(I)
u2800 3912 CONTINUE
02810 C
02820 C BACKGROUND ITERATION OPTION.
02830 C
02840 2999 CONTINUE
02850 MGCTR = 0
02660 IF(CHOIS(12).EQ.0) GO TO 3010
02870 ESCAPE = 0.0
02880 XC = —1.3
02890 XHX = 0.0
u2900 3005 CONTINUE
02910 XC =' XC + 1.0
02920 IF(XC.GT.1000) GO TO 3999
62930 3019 CONTINUE
02940 C
02950 C SUBTRACT BLANK INTENSITY FROM EACH STANDARD
02960 C INTENSITY.
02970 C
02980 DO 1011 I = NF,NL
u2990 B(I) = BX(I) - XC
03000 1011 CONTINUE
03010 C
03(120 C CALL SUBR. MGIS
03030 C
03040 IF(4.NL.7) GO TO 1199
03050 IF(CHOIS(12).E(1.0.AND.MGCTR.EQ.0) GO TO 1199
03060 6001 CONTINUE
03070 MGCTR = MGCTR + 1
03080 CALL MGIS(NL,ISX,MGCTR)
03990 1199 CONTINUE
03100 C
03110 C
03120 C IF ELEMENT IS MG THEN APPLY
03130 C INTERFERENCE CORRECTIONS TO EACH STANDARD
03140 C
03150 C -
03160 6002 CONTINUE
03170 IF(1.NE.7) GO TO 6310
03180 IF(HGCTR.EC.0) GO TO 6019
03190 DO 6008 I = NF NL
03200 NUMD _ (SBNUM(1)/10) * 10
0321C DO 6094 J = 1 ISX
03220 IF (NUMP.GE.9900
) GO TO 6004
03230 IF(')UMP.EQ.NUM(J)) 8(I) = B(I) — SUMCS(J)
0324C IF(NUMP.EO.NUM(J)) SUMCSX(I) = SUMCS(J)
03250 6004 CONTINUE
 Computer programs SORT, REORD, and MW for major-element XRF data processing 129

03260 6008 CONTINUE

03270 6010 CONTINUE
03280 C
03290 C
03300 C
03310 C
u3320 C CALC SLOPE U(H) OF THE CALIBRATION LINE THROUGH ALL STANDARDS
03330 C AND THE ORIGIN IGNORING ANY STANDARD WHOSE OXIDE CONCENTRATION
03340 C C(I) IS 0.0 . THEN BACK CALC FROM THE SLOPE THE CONCS O(I)
u3350 C OF THE BLANKS ANO STANDARDS. USING THE BACK CALCULATED CONCENTRATIONS
03360 C CALC THE ABSOLUTE AND PERCtNTAGE RELATIVE ERRORS G ( I) ANO M(I), FOR
03370 C EACH STANDARD W. R. T. THEIR RECOMMENDED CONCENTRATIONS.
03380 C RAG E ATIVE ERRORS,
03390 C XFSA ND A XH, ASEWELL FĪGUREA OF
THESF F MERIT,EF OM,
03400 C
03410 C
03420 XD = 0
03430 XE = 0
03440 SUM::ON = 0.0
03450 C
03460 C
03470 -DO 1012 I = NF NL
03480 IF(C(I).E0.0.01 GO TO 1012
03490 SUMCON = SUMCON + C(I)
03500 XD = X0 + C(I)
03510 XE = XE + 8(I)
03520 1012 CONTINUE
03530 U(M) = XD/XE
03540 C
u3550 C
03560 BLANK(M) = XC • UtM)
03570 IF(NH.EQ.0) GO TO 1073
03580 DO £072 I = 103
03590 D(I) = U(M) a D(I)
03600 G(I) = 0.0
03610 H(I) = 0.0
03620 1072 CONTINUE
03b30 1073 CONTINUE
03640 C
03650 C
03660 DO 1013 I = NF,NL
03670 D(I) = U(M) " B(I)
03680 G(I) = 0.0
03690 H(I) = 0.0
p .OD ( I) TO 1013
03710 G(Ī)(11C(I)O
03720 H(I) = 100.0 s G(I)/C(I)
1013 CONTINUE
03730 0
03750 C
03760 RNS = 0.0
u3770 XF = 0.0
03780 XH = 0.0
03790 DO 1014 I = NF,NL
03800 IF(C(I).E0.0.0) GO TO 1014
03810 XH = XH + ABS(H(I))
03820 XF = XF + ABS(G(I))
03830 RNS = RNS + 1.0
03840 1014 CONTINUE
03850 FOM = XF/SUMCON + 100.0
03860 XH = XH/RNS
03870 XF = XF/RNS
03880 XXH(M) = XH
03890 XXF(M) = XF
03900 XFOM(M) = FOM
u3910 C
03920 C
03930 C
03940 C BKGD ITERATION OPTION — TEST FOR XH MINIMUM
0395C C
03960 IF(CHOIS(12).EQ.0) GO TO 3999
03970 IF(ESCAPE.EQ.1.01 GO TO 3999
03980 IF(XC.EQ.0..0) GO TO 3997
03990 IF(XH.GT.XHX) GO TO 3998
04000 3997 CONTINUE
04010 XHX = XH
04920 GO TO 3000
04030 3998 CONTINUE
04040 XC = XC — 1.0
04050 ESCAPE = 1.0
04060 GO TO 3010
04070 3999 CONTINUE
04080 C
04090 C IF CHOIS(6) = 0 THEN WRITE OUT CALIBRATION DATA FOR
04180 C ELEMENT 'M'
04110 C
04120 IF(CHOIS(6).EQ.1) GO TO 1080
04130 WRITE(IWLP,5011)
04140 WRITE(IWLP.5012) METHOD
04150 WRITE(I)&P,5012) DSCRIP
04160 WRITE(IWLP,5012) NAME
04170 WRITE(IWLP,5013) OXIDE(M)
04180 5011 F0RMAT(1H1)
04190 5012 FORIAT(X,80A1)
04200 5013 FORMAT(//X.2IHCALIBRATION LINE FOR ,A61)
04213 IF(1.£Q.7) WRITE(IWLP,1015) (AVISP(I),I=1,E)
04220 1015 FORMAT(/ X 'MG INTERFcRENCE CORRECTION IN C/S PER',
04230 4' P_RCENT OXIDES',/ X ' TI FE CR MN ',
04240 +' CA K',/,X,6F~7.2)
04250 A = JTC/1.E6
04260 WRITE(IWlFs1016) A
0427C 1016 FORMAT(//1h ,3X,14HDEAD TIME IS ,F9.7,2X,
04280 * 7.HSECONCS // 1H . 3X, 91HSTANDARD STD. CONC. CALC. CONC.
04290 + ABS ERROR REL ERROR PERCENT CORRECTED C.P.S.,
04300 317H MG CORRECTION 11
04310 NUMX = 0

CAGEO Vol. 3, No. 1—I

130 R. J. PARKER and J. P. WILLIS

04320 DO 1076 I = 1,NL

0433C NUMY = ((SBNUM(I)/10)•10)
6434C IF(NUMY.NE.NUMX) WRITE(IHLP,1077)
04350 NUMX = NUMY
04360 WRITE(IHLP,1018)SBNUM(I),SBNAM(I),C(I),D(II,G(I),HtI),3(2),
04370 +SUM.1SX(I)
04380 1076 CONTINUE
04390 1077 FORIAT(X)
04400 1018 FOR'3AT(1H ,3X,1I6,2X,A6,F8.3sFl5.3,F1l.3,FS7.2,F2l.kF12. 1)
04410 WRITE(INLP,1081) FOM
04420 1081 FORMAT(//1H 3X 18HFIGURE OF MERIT IS,F8.21
04430 WRITE(IWLF,1019I XF XH,U(M),BLANK(M)
04440 1019 FORMAT(//lh ,3X,17HAV. ABS. ERROR IS,F8.3,
64450 +20X,17HAV. REL. ERROR IS F8.2
04460 +//1H ,3X 26HSLOPE OF CALIBRATION LINE IS ,E10.4,
04470 +5X,22HBLANK CONCENTRATION IS F6 3, 9H PERCENT//)
04480 IF(CHOIS(12).EQ.1) WRITE(IWL',1382) XC
04490 IF(CHO.IS(12).EQ.3) WRITE(IWLP 1083) XC
04500 1082 FORMAT(1)1 ,3X,20HITERATED BKG0 CPS IS,F7.1)
0451C 1083 FORMAT(1H ,3X,20HAVERAGE BLANK CPS IS,F7.1)
04520 1680 CONTINUE
04530 C
64540 C IF CALIBRATION LINES TO BE PLOTTED THEN CALL SUBR GRAPh
04550 C
04560 IF(DHOIS(4).EQ.1) CALL GRAPH(U,M,NF,NL,NPLOT)
04570 C
04580 C IF TEST FOR MG CALIBRATION RECALCULATION (NON—ITERATIVE BKGD)
04590 C
04600 IF(M.NE.7) GO TO 2299
04610 IFCHOIS(12).E(1.0.AND.MGCTR.EQ.0) GO TO 6001
04620 2299 CONTINUE
u4630 C
04640 C TEST FOR END OF CALIBRATION DATA
04650 C
04660 IF(STOATA(SUMSB,1).NE.-99.0) GO TO 1001
04670 1020 CONTINUE
04680 IF(CHOIS(4).NE.1) GO TO 1022
04690 CALL ENPLOT(20.0)
04700 1022 CONTINUE
04710 C
04720 C WRITE OUT SUMMARY OF CALIBRATION DATA
04730 C
04740 WRITE(IWLP,100)
04750 100 FORMAT(/H1//X,27HSUMMARY OF CALIBRATION DATA)
04760 WRITE(IWLP,105)
04770 105 FORMAT(/X,48HELEMENT SLOPE BLANK AAE ARE,
04780 F10H FOMI
04790 DO 111 I = 1,15
64800 IF(U(I).EQ.0.0) GO TO 110
04810 WRITE(IWLP 115) OXIDE(I), U(I), BLANK(I), XXF(I), XXH(I) XFOM(I)
04820 WRITF(IWT,115 ) OXIDE(I), U(I),BLANK(I),XXF(I),XXH(I),XFON(I)
04830 110 CONTINUE
64840 115 FORMAT(/X,A6,3X,E10.4,5X,F6.3,4X,F6.3,4X,F6.3,4X,F6.3)
04850 116 CONTINUE
04860 RETURN
04870 END
04880 C
04890 C BLOCK LATA SUBPROGRAM FOR ARRAY OXIDE
04900 C
04910 BLOCK DATA
04920 COMMON /BB/ B(100), BX(1001s C(100), OXID£(15)
04930 C
u4940 C NOTE LATA STATEMENT IS NON STANDARD
04950 C
04960 DATA OXIDE
u4970 +/6HSI02 6HTIO2 } 6HAL2036HFE203 6HCR203
04980 +`6HMN0 , iHMGO .6HCAO , tiiHNA20 ,IHK20 ,6HP205 ,6)1NIO ,
u4990 •6HELX ,6HELY ,6HELZ /
05000 END
05,310 C
05020 C
05030 C
65040 C
05050 C SUBROLTINE MGIS CALCULATES FROM INTERFERENCE SIDS THE
05060 C CORRECTION FACTORS REQUIRED FOR THE MG CALIBRATION LINE
05070 C
05080 C ARRAYS
05090 C NUM(50) — STD 'PRIME' A — NUMBERS
05100 C NAM(50) STD NAMES
05110 C SCONCX(15) — STD GONGS
05120 C SCONCZ(50,15) — STD CONICS FOR ALL STOS READ IN
05130 C CHOIS(12) — SEE OPTION DOCUMENTATION AT START OF PGM
05140 C NOS(6) — 'PRIME' A — NUMBERS FOR MG INTERFERENCE STDS
05150 C CIS(6) — CONE OF INTERFEARING OXIDE IN INTERFERENCE STOS
05160 C SBNUM(100) — SEE SUBR SLOPE
05170 C B(100) — SEE SUBR SLOPE
65180 C AVIS(6) — AVERAGE C/S MG INTERFERENCE FOR OXIDES TI — K
05190 C IN THE INTERFERENCE STDS
05200 C AVISP(6) — AVERAGE C/S MG INTERFERENCE PER PERCENT
05210 C INTERFEARING OXIDES TI - K
05220 C IS(6) — OXIDE NAMES TI — K
05230 C SUMCS(50) — TOTAL C/S MG INTERFERENCE FOR STANDARDS COR-
05240 C RESPONDING TO NUM(50)
05250 C
05260 C
05270 G PARAMETERS
05280 C ISX — STDS
05290 C
05300 C
05310 C NTĪTLE ■ STĪTLEEC A RDO OF BLANK/STANDARD FILE
05320 C NFIN — FLAG TO TEST FOR END OF BLANKS
05330 C NF — '• '' •' STANDARDS
05340 C NAMX — NEW STD NAM
05350 C NUMX — NEW PRIME S1'O NUMBER
 Computer programs SORT, REORD, and MW for major-element XRF data processing 131

05360 C NUMT — OLD PRIME STD NUMBER

05370 C NUMTX — PRIME STD NUMBER FROM SBNUM(L)
05380 C TE STANDARD B6NUN(E)
05390 C DĪV IS— NUMBERLOF/A NALYSES OUSEDRT OAACCUMULA
05400 C CSTI — )
05410 C CSFE
05420 C — I C/S MG INTERFERENCE FROM OXIDES IN STD SCONCZ(ISX,J)
05430 C CSMN
05440 C - )
05450 C CSK — )
05460 C
05470 SU82OU7INE MGIStLU,ISX,MGCTRI
05480 COMMON DSCRIP(80),NAME(80),CHOIS(20),METH00(8),DTC,
05490 2ITE•'IP,IWLP,IPNCH,IRNW IRSTD,IWT
05500 COMMON /AA/ 0(15) BLANK(15), AIISP(6)
05510 COMMON /88/ 9(100 1X(100)
B C(1001 OXIDE(15)
05520 COMMON /CC/ NAM(501 ,SCONCX(I5) ,SCONCZ(50,
1 5),AVIS(6),
05530 +NOS(6) IS(6),CIS(6),NUM(5O)sSUHCS(50),SUMCSX ( l00),
05540 +SBNJM(10
0)
05550 INTEGER SBNUM, CHOIS
05560 C
05570 C
05580 C PRESET ARRAYS
05590 C
05600 IF(MGCTR.GT.1) GO TO 200
05610 DO 100 I = 1,50
05620 NUM(I) = 0
05630 NAM(I) = 6H
05640 DO 110 J = 1,15
05650 5G0NCZ(I,J) = 0.0
05660 110 CONTINUE
05670 100 CONTINUE
05680 C
05690 C
05700 C REWIND BLANK/STANDARD FILE AND READ TO ENO OF BLANKS
0571C C
05720 REWIND IRSTD
05730 READ(IRSTO,120) NTITLE
05740 120 FORMAT(A1)
05750 130 READ(IRSTD,140) NFIN
05760 140 FORMAT(A1 I)
05770 IF(4FIN.NE.1HF) GO TO 130
05780 C
05790 C
05800 C READ Its STANDARD DATA
05 810 C
05520 NUMT = 0
05830 ISX = 0
05840 150 READ(IRST0,160) NF,NUNX,NAMX,(SGONCX(J),J=1,15)
05850 160 FORMAT(A1 9X,I4 2X A6 2X,8F7.3/2'.X,7F,7.3)
05860 IF(NF.EQ.1FF) GO TO 200
05870 NUMX = (NUMX/10) • 10
05880 IF(MUMX.EO.NUMT) GO TO 150
05890 NUMT = NUMX
05900 ISX = ISX + 1
05910 IF(ISX.E(1.51) WRITE(INLP,165)
05920 IF(ISX.E0.51) GO TO 200
05930 165 FORMAT(/,X,'WARNING — MORE THAN 50 STOS FOR MGFAGS')
05940 NUM(ISX) = NUMX
05950 NAM(ISX) = NAMX
05960 DO 179 J = 1 15
05970 SCOACZ(ISX,J) = SCONCX(J)
05980 170 CONTINUE
05990 GO TO 150
06000 200 CONTINUE
06010 IF(ISX.EQ.51) ISX = 5G
06020 C
06030 C
06040 C IF TESTING WRITE OUT STANDARD DATA
06050 C
06060 IF(CHOIS(9).EQ 0) GO TO 220
06070 NRITF(IWLP,l163)
06080 1160 FORMAT(1H1,/,X 42HTESTING — MGO INTERFERENCE CORRECTION DATA ,/)
06090 DO 210 K = 1 ISX
06100 WRITE(IWLP,161) NUM(K) NAM(K),(SCONCZ(K,J),J=1,15)
06110 161 FORMAT(X,I4,2X,A6,2X,1SF7.3)
06120 210 CONTINUE
06130 220 CONTINUE
06140 C
06150 C
06160 C PRESET PRIME A — NUMBER AND CONCENTRATION DATA FOR THE
06170 C MG INTERFERENCE STANDARDS
36180 C
06190 C NOS(1) = PRIME A — NUMBER FOR 1IO2 INTERF. STD.
06200 C NOS(2) IS FE203
06210 C NOS(3) " CR203 ill
06220 C NOS(3) •' MNO 411 "
66230 C NOS(5) " CAO IS
06240 C NOS(6) " K20
06250 C
06260 C CIS(1) = CONC OF INTERFEARING OXIDE IN STD = NOS(1)
06270 C CIS(21 ETC
06280 C
06290 NOS(1) = 9900
0630 NOS(2) = 9917
0631C NOS(3) _ 9920
66320 NOS(4) = 9930
06330 NOS(5) = 9940
06340 NOS(6) = 9950
06350 CIS(1) = 30.0
0636C CIS(2) = 30.0
06370 CIS(3) = 30.0
06380 CIS(4) = 30.0
06390 CIS(5) = 30.0
06400 CIS(6) = 30.0
06410 C
132 R. J. PARKER and J. P. \Visus

06430 C
06430 C CALC MG INTERFERENCE AS C/S PER PECENT INTERFEARING OXIDE
06440 C
06450 DO 24' K = 1,6
06460 SUMIS = C.0
06470 DIV = 0.0
06480 DO 230 L = 1,LU
06490 IF(SBNUM(L).LT.9909) GO TO 230
06500 NUMTX = (SBNUM(L)/10) } 10
06510 IF(`IOS(K).EQ.NJMTX) SUMIS = SUMIS + 3(L)
06520 IF(NDS(K).EQ.NUMTX) DIV = DIV + 1.0
06531 230 CONTINUE
06540 IF(3UMIS.E0.0.0) GO TO 240
06550 AYIS(K) = SUMIS/DIU
06560 AVISP(K) = AVIS(K)/CIS(K)
06570 240 CONTINUE
06580 C
06590 C
06600 C IF TESTING WRITE OUT CALL INTERFERENCE DATA
06610 C
06620 IF(CHO2S(9).0Q.0) GO TO 260
06630 IS(1) = 2HTI
06640 IS(2) = 2HFE
U6650 I5(3) = 2HCR
06660 IS(4) = 2HMN
66670 IS(5) = 2HCA
06680 IS(6) = 2HK
06690 DO 250 K = 1,6
06700 WRITE(IWLP ,21.5) IS(K),NOS(K) AVIS(K),AVISP(K)
06710 245 FORMAT(/,X,A2,2X,I4,2X,F6.2,) 2X,F6.2)
06720 250 CONTINUE
66730 260 CONTINUE
06740 C
06750 C
06760 C CALC TOTAL C/S MG INTERFERENCE CORRECTION FOR STANDARDS,
06770 C USING ALL INTERFEARING OXIDES
06780 C
36790 DO 270 I = 1,ISX
06800 CSTI = SCONCZZ(I,2) " AVISP(1)
06810 CSFE = SCONCZ(I,4) +` AVISP(2)
06820 CSCR = SCONCZ(I,5) * AJISP(3)
06830 CSMN = SCONCZ(I,6) * AVISP(4)
06840 CSCA = SCONCZ(I,8) * AVISP(51
06850 CSK - SCONCZ(I,10) ' AVISP(6)
06860 SUMOS(I) = CSTI + CSFE + CSCR + CSMN + CSCA + CSK
06870 270 CONTINUE
06880 C
06890 C
06900 C IF TESTING WRITE OUT TOTAL C/S MG INTERFERENCE FOR STANDARDS
06910 C
06920 IF(OHOIS(9).EQ.0) GO TO 300
06930 DO 290 I - 1,ISX
U6940 WRITE(IWLP,280) NUM(I), HAF1(I), SUMCS(I)
06950 280 FORMAT(/,X,I4,2X,A6,2X,F6.1)
06960 2911 CONTINUE
86970 300 CONTINUE
06980 RETURN
06990 ENO
07000 C
07010 C
07020 C
07030 C
07040 C
07050 C SUBROUTINE GRAPH PLOTS THE CALIBRATION LINE FOR EACH
07060 C ELEMENT ANALYSED. IT IS WRITTEN FOR A CALCOMP 760 PLOTTER
07070 C
07080 C
07090 C ARRAYS
07100 C
07110 C ME'TH00(8)0SCRIP(80),NAHE( 8O),GHOIS( 20) — SEE SUBR CALIB

07120 C U(15),6(100),C(100),OXIOE(15) - SEE SUBR SLOPE

07130 C
07140 C DEP(95) - CONTAINS STANDARD CONCENTRATION DATA FROM C(I)
07150 C UNDEP(55) " " INTENSITY " " B(1)
07160 C XP(6) — [ SCALING AND LINE
07170 C YP(6) I PLOTTING DATA ARRAYS
07180 C XLIST(3) - CONTAINS ALPHANUMERICS FOR X-AXIS
07190 C YLIST(10) -• " " " Y —AXIS
07200 C XPAGEX(15) — I CO— ORDINATES OF PLOT ORIGINS
07210 C YPAGEY(15) — [ FOR UP TO 15 GRAPHS
07220 C
07230 C PARAMETERS
07240 C
07250 C AXLEN — AXIAL LENGTH OF GRAPH
07260 C XPAGE — [ PLOT
ORIGIN
07270 C YPAGE — I FOR GRAPH
07280 C NPLOT — COUNTER FOR NUMBER OF TIMES SUBR. GRAPH IS CALLED
0729C C XMAX — [ CO— ORDINATES OF UPER LIMIT
07300 C YMAX I OF PLOTTED LINE
07310 C M — ELEMENT NUM9°_R(1 TO 15)
07320 C NF — INDEX OF FIRST STANDARD IN ARRAYS C(I),B(I)
07330 C NL — " LAST "
07340 C NI — [
07350 C NJ — I INDEXING VARIABLES FOR
07360 C IJ — I PLOTTING PARAMETERS IN ARRAYS OE:P
07370 C NN — I AND UNDEP
07380 C NM — (
07390 C
07400 C
07410 C
07420 SUBROUTINE GRAPH(U MX NFX NLX NPLOTX)
07430 COMMON DSCRIP(80) FAME(80 j GHŌIS(2O),METH0O(8),OTG,
07440 2ITENPIHLP,IPNCH IRNW,IRSTD,IHT
07450 COMMON/BB/ B(100) BX(100) C(10G), OXIDE(15)
07460 DIMāNSION 0(155),DE ~'(105) UND
(105),XP(6),YP(6),XLIST(3),YLIST(10
EP )
07470 DIH_NSION XPAGEX(15), YPAGEY(15)
 Computer programs SORT, REORD, and MW for major-element XRF data processing 133

07480 INTEGER
07490 CC NOTECHOIS
07500
07510 C LATA STATEMENTS ARE NON STANDARD
07520
07530 DATA XPAGEX
+0.0.0.0t0.0.9.0.0.0.0.0/
+10.00.010.00.0110.0t0.0.9.0.0.0,0.0.
07540
07550 DAT YP
AGEY
07560 +/3.0.12.0.21.0
07570 +3.0,12.0,221.0,3.0,1 2.0,21.0/
3.0 12.0 21.0,3.3,12.0,21.0,
Q
07590 NF = MNFX
07600
07610 NL = NLX
XMAX = 0.0
07620
07630 CC SET PLOT ORIGIN AND AXIAL LENGTH FOR THE GRAPH
07640
07650 CC XPAGE YPAGE ARE THE ORIGIN(BEARING MIND THAT
07660
07670 C C XPAGE
CO—ORDINATES IS AN OF THE
INCREMENT W.R.T. THE ORIGIN OF THE
07680 C PREVIOUS GRAPH BUT THAT
07690 C M.R.T. THE BOTMTO EDGE OF PAPER EACH TIME).
YPAGE MUST BE RESET
07700
07710 CC TO YPAGE AVOID
MUSTOVERWRITING
BEMUST
GREATER THE
THANNON—VARIABLE
3.0 TITLE,
.29.0(FOR
07720
07730 C YPAGE+AXLEN
CC AXLEN IS THE ALWAYS BE .LE. CALCOMP 760).
07740
07750 CC LENGTH OF THE AXES IN INCHES (BOTH AUTOMATICALLY EQUAL)
07760
07770 C
07780 NPLOT
07790 NPLOT == NPLOTX
07800
07810 NPLOTX ==NPLOT
AXLEN NPLOT+ 1
7.0
07820 XPAGE
YPAGE == YPAGEY(NPLOT)
07830 IF(NPLOT.EC.1) XPAGEX(NPLOT)
07840
07850 CC IF CHOIS(5)CALL START
07860
07870 C = 1 , THEN PLOT IS HALF SIZE
07880 IF(CHOIS(5).EQ.1)
07890 CC IF FIRST CALL TO SUER CALL FACTOR(0.5)
07900
07910 CC NAME AND OSCRIP GRAPH THEN PLOT TITLE,
07920
07930 CALL
IF(NPLOT.GT.11 GO TO —3)90
07940
07950 CALL PLOT(0.0,-30.0
PLOT(0.0,1.0. 3)
07960 +N
07970 CALL BY SYMBOL(0.0,1.0,0.28,59HCALIBRATION
XRF SFECTROHETRYs0.0159) LINES FOR MAJOR ELEMENT RU
07980
07990 XPOS
DO = 0.0
I = 1.80
08000
08410 CALL88SYMBOL(XPOS
CALL SYMBOL(XPOS,0.5,0.280SCRIP(I)10.0,1)
0.0,3.28,NAME(I),0.U,1)
08020
08030 XPOS
88 = XPOS + 0.128
CONTINUE
08040
08050 90 CONTINUE
C CALCULATE
08060
08070 00C
C PLOTTING PARAMETERS
08080
08090 100 I = NF,NL
IF(C(I).LE.XMAX) GO TO 100
u B3UO XMAX = C(I)
08110
08120 100
XMAX CONTINUE
08130 YMAX
u8140 XP(1) ===XMAX/U(M)
1.05+XMAX
0.0
08150
08160 YP(1)
XP(2) == XMAX
a.0
08170
Ōb180 YP(2)
08190 NI
NJ NL= —YMAX
= NI+1
= NF * 1
08200 DEP(NJ)
08210 IJ
UNOE°(NJ) = 0.2
u8220
08230 DO =510 I = =NF,NL
0.0
08240 IJ
68250 DEP(IJ) = IJ+1= C(I)
08260
08270 UNDEP(IJ)
61 CONTINUE = 9(I)
08280
08290 C
CC PLOT THE GRAPH
0830C
08310 CALL _ _
08320 CALL PLOT(X P AGE3 YPA E3 ,-3)
08330 CALL SCALE(OEP,AXLEN,NJ 1)
08340 CALL
08350 NM
NN == NI+3 SCALE(UNDeP,AXLEN,NJ,1)
NI+2
08360
0837C
08380 XP(3)
u8390 XP(44) === DEP(NN)
YP(3) CEP(NM)
UNCEP(NN)
08400
08410 YP(4) = UNDEP(NM)
CALL AXIS(0.0,2.0 OXIDE(M) 6 AXLEN,c.3,0EP(NN) ,DEP(NH))
08420 CALL SYMOOL(1.9, .4,0.14 7HWT. PC. 0.0 7)
08430
08440 CALL 'EP(NH))AXIS ( 0.0,0.0,17HCOUNTS PER SECOND XLEN
,90.0,UNDEP(NN),UND
,17,A
08450 CALL
08460 CALL LINc(CEP,UNDEP,NJ,1s -1,3)

08480
RETURNLINE(XP,YPs2s1s0,0)
08490
08500 CC
08510
08520 CC
08530 C SUBROUTINE CONC

134 R. J. PARKER and J. P. WILLIS

08540 C CONVERTS SAMPLE COUNT RATES INTO CONCENTRATIONS - USES SLOPES

08550 C OF CALIBRATION LINES To CALC NOMINAL GONGS
08560 C
0857C C
08580 C ARRAYS
08590 C
08600 C DSCRIP(80),NAME(80),CHOIS(12) — SEE SUBR SLOPE
08610 C U(15),BLANK(15) * SEE SUBR SLOPE
08620 C
08630 C RSPL(15) — SAMPLE ELEMENT INTENSITIES
08640 C NSPL(15) — 'NOMINAL + BLANK' OXIDE CONCS. FOR SAMPLE
08650 C NSPX(15) — 'NOMINL — BLANK' OXIDE CONCS. FOR SMPLE
08660 C C(15) — CLEMENT CONCS FOR SAMPLE
08670 C V(15) — MATRIX FACTORS FOR SAMPLE
08680 C H(15) — MATRIX CORRECTED OXIDE CONCS. FOR SAMPLE
08690 C WT(5) — SAMPLE/CRUCIBLE WEIGHINGS
08700 C SPLNAM(76) — 76 CHARACTERS FOR SAMPLE NAME
08710 C WTITLE(79) — 80 CHARACTERS FOR TITLE OF NAME/WEIGHING FILE
08720 C W0(15) — OXIDE TO ELEMENT CONVERSION FACTORS
08730 C
08740 C PARAMETERS
08750 C
08760 C NX — NUMBER OF SAMPLES PROCESSED
0877C C NY — NUMBER OF SAMPLES WRITTEN OUT(RESET TO 0 AFTER EVERY 10TH)
08780 C SUMCOR — TOTAL OF MATRIX CORRECTED OXIDE CONCS FOR SAMPLE
08790 C SUMNOM — TOTAL OF 'NOMINAL — BLANK' OXIDE LONGS FOR SAMPLE
08800 C SPLNUM — SAMPLE A — NUMBER(INTENSITIES)
u8810 C SPNUMA — SAMPLE A — NUMBER(NAME)
08820 C SPNUMH SAMPLE A — NUMBER(WEIGHINGS)
08830 C XX — PERCENTAGE LOSS OF SAMPLE AT 110 C
08840 C YY .• " " " BETWEEN 110 C AND 1000 C
0885C C ZZ — WEIGHT LOSS BETWEEN 110 C AND 1000 C
08860 C N — WEIGHT OF SAMPLE TAKEN FOR FUSION
08870 C WF — WEIGHT OF FLUX USED IN FUSION (GUNN METHOD)
.08880 C HR — REFERENCE HEEIGHT OF SAMPLE USED IN FUSION (GUNN METHOD)
08890 C NSOD — OPTION FLAG FOR READ IN OF SODIUM DATA DETERMINED
08900 C FROM WET CHEMISTRY. IF SET TO S THEN SODIUM DATA MUST
08910 C BE PRESENT.
08920 C NFEO — OPTION FLAG FOR READ IN OF WET CHEM FED. IF SET
08930 C TO F THEN FED DATA MUST BE PRESENT.
08940 C RSOD - WET CHEM SODIUM DATA FOR SAMPLE SPLNUM
08950 C FED — WET CHEM FEO DATA.
08960 C FE — FED CONVERTED TO ELEMENT.
08970 C
08980 C
08990 SUBROUTINE CONC
09000 COMMON DSCRIP(80) NAME(80),CHOIS(20),METHOD(8),OTC,
09016 +ITEMP IWLP,IPNCH IRNW,IRSTD,IWT
09020 CGM9O1 /AA/ U(15), BLANK(15), AVISP(6)
09030 INTEGER ABC, SPLNXX SPLNZZ, SNOX
09040 DIMENSION RSPL(15), NSPL(15) WT(5), H(15),NSPX(15),
09056 +V(15), C(15) SPLNAM(76) WQ(15), WTITLE(79)
09060 REAL NSPL, NCF, LLIM, NSPX
09976 INTEGER CHOIS, SPLNAM, SPNAMA, SPLNUM, SPNUMA, SPNUMW, BB, USUM
09980 C
09090 C DATA WO(I) CONTAINS OXIDE TO ELEMENT CONVERSION FACTORS
09100 C
09110 C NOTE DATA STATEMENT IS NON STANDARD
09120 C
09130 DATA NO
09140 +10.4675,0.599510.5292,0.699410.6843,0.7745,0.6032,0.7147,
09150 +0.7419,0.8302,0.436410.7858,0.0,0.0,0.0/
09166 DATA LRONX /6H FE203/
09170 DATA IRONY /6H FED /
09180 C
09190 C IF TESTING NORRISH DATA THEN PRE—SET DTC,U, AND BLANK
69'200 C
09210 IF(CHOIS(1C).EQ.D) GO TO 6321
09220 OTC = 1.7
09230 U(1) = 0.8595E-02
09246 U(2) = 0.3796E-03
09250 0(3) = 0.9476E-02
09260 U(4) = 0.2814E-02
09270 U(6) = 0.5252E-03
09280 U(7) = 0.6307E-01
09290 U(8) = 0.1349E-02
09306 U(1:1) = 0.9725E-03
09310 U(11) = 0.3219E-02
09320 BLANK(1) = 0.255
09330 BLANK(2) = 0.098
09340 BLANK(3) = 0.222
09350 BLANK(4) = 0.138
09360 BLANK(6) = 0.078
09370 BLANK(7) = 2.997
09380 BLANK(8) = 0.079
09390 BLANK(10) = 3.016
09400 BLANK(11) = 0.187
09410 AVISP(1) = 1.53
09420 AVISP(2) = 0.23
09430 AVISP(3) = 0.09
09440 AVISP(4) = 0.11
09450 AVISP(5) = 1.20
09460 AVISP(6) = 0.62
09470 6321 CONTINUE
09480 C
09490 C IF TESTING GUNN DATA THEN PRE—SET OTC, U, BLANK AND AVISP
09560 C
09510 IF(CHOIS(11).EQ.0) GO TO 6322
09526 OTC = 1.7
09530 0(1) = 0.7577E-02
09540 0(2) = 0.2977E-03
u9550 0(3) = 0.9043E-02
09560 U(4) = 0.7341E-03
09570 U(6) = 0.1450E-03
09580 U(7) = 0.6159E-01
09590 U(8) = 0.1036E-02
 Computer programs SORT, REORD, and MW for major-element XRF data processing 135

09600 U(10) = 0.7728E-03

09610 U(11) = 0.2778E-32
09620 BLANK(1) = 1.15
09630 BLANK(2) = 0.01
09640 BLANK(3) = 0.22
09650 BLA4K(4) = 0.11
09660 BLANK(61 = 0.01
09670 BLA1K(7) = 1.66
09680 BLA1K(8) = 0.05
09690 BLANK(10) = 0.06
09700 BLANK(111 = 0.60
09710 AVISP(1) = 2.05
09720 AVISP(2) = 0.04
09730 AVI3P(3) = 0.04
u9740 AVISP(4) = 0.04
09750 AVISP(5) = 1.20
09760 AVISP(6) = 0.93
09770 6322 CONTINUE
09780 NX =
0979C = 10
09800 C
09810 C READ FROM THE TEMPORARY DISC FILE THE A—NUMBER
09820 C AND COUNT RATES FOR ELEMENTS ANALYSED IN THE
09830 C SAMPLE
u9840 C
09850 42 CONTINUE
09860 NX = NX + 1
09970 READ(ITEMF,4444) SPLNUM, (RSPL(I),I=1,151
09880 4444 FORIAT(I4,4)(,i5F8.2)
39890 IF(SPLNUM.E0.-999) GO TO 1059
09900 DO 43 JK = 1,15
09910 NSPL(JK) = 0.0
09920 NSPX(JK) = 0.0
09930 43 CONTINUE
09940 SUMCOR = C.0
09950 SUMNOM = 0.0
09960 FED = 0.0
09970 FE = 9.0
09980 C
09990 C IF TESTING THEN WRITE OUT SAMPLE NUMBER AND COUNT RATES
10000 C
10910 IF(CHOIS(9).EQ.0) GO TO 7378
10020 WRITE(IWLP,3333)
10030 WRITE(IWLP 2222)
10440 2222 FORMAT(X 'ZEST WRITE OUT OF SAMPLE DATA',/)
10050 WRITE(IWLP,4445) SPLNUM, (RSPL(I),I=1,15)
10060 3333 FOR`1AT(//J)
10070 4445 FORMAT(1X,I4,4X,15F8.2)
10080 7378 CONTINUE
10090 C
10100 C READ IN SODIUM(IF WET CHEM DETERMINED) FOLLOWED BY
10110 C SAMPLE A—NUMBER NAME AND WEIGHINGS FRCM
10120 C SEPARATE 'NAME AND WEIGHINGS FILE' (UNIT NUMBER = IRNW)
10130 C
10140 IF(NX.EQ.1) READ(IRNW,1031) WTITLE
10150 IF(1X.EQ.1) READ(IRNW,1032) NSOD,NFEO,WR,WF
10160 1031 FOR9AT(80A1)
10170 1032 IF(NSOD
.E .1 HS) READ(1R~IH,10331 RSOD
10190 IF(1FEO.EC.1HF) READ(IRNW,1033) FEO
10200 1033 FORHAT(F5.2)
10210 READ(IRNW,1034) SPNUMA, (SPLNAM(I), I = 1,76)
10220 REA3(IRNW,1035) SPNUMW, (WT(I),I = 1,5)
10230 1034 FORMAT(I4,76A11
10241. 1035 FORMAT(I4,1X,5F9.5)
10250 IF(SPNUMA.NE.SPNUMW.OR.SPNUMA.NE.SPLNUM) WRITE(IWLP,3331)
10260 +SPNUM4,SPNUMW SPLNUM
10270 3331 FORMAT(/)(,28HERROR IN SAMPLE NUMBERS FOR ,
10280 +23HNAME/WEIGHING/COUNTS — ,3(I4,2X))
10290 C
10300 C IF TESTING THEN WRITE OUT SAMPLE NUMBER, NAME AND WEIGHINGS
10310 C
10320 IF(CHOIS(9).EQ.01 GO TO 7903
10330 IF(NSOD.EQ.1HS) WRITE(IWLP,7900) RSOD
10340 7900 FORMAT(X,F5.2)
10350 WRITE(IWLP,7901) SPNUHA,(SPLNAM(I), I = 1,761
10360 WRITE(INLP,79021 SPNUMW, (WT(J),J=1.5)
10370 7901 FORMAT(X,I4,76A1)
10380 7902 CON I4,10X,5F9.5)
T ĪNUE.
10400 C
10410 C USE FIRST 4 SAMPLE/CRUCIBLE WEIGHINGS FOR LOSS
10420 C ON HEATING CALCULATION 5TH WEIGHT (WT(5)) IS IGNITED SAMPLE
10430 C WI TAKEN FOR FUSION. CALC PERCENTAGE LOSS AT 120 C (XX) AND
10440 C PERCENTAGE LOSS ON IGNITION AT 1000 C (YY) .
10450 C
10460 XX = WT(2) — WT(3)
10470 YY = WT(2) — WT(1)
1048g C ZZ = WT(3) — WT(4)
g
1050 0 ĪF(X X.EQ.0.0.AND.YY.EQ.0.0.AND.ZZ.EQ.0.0) GO TO 45
10510 IF(XX.EQ.0.0) GO TO 44
10520 XX = 100.0 + XX/YY
10530 44 YY = 100.0 + ZZ/YY
10540 C
10550 C CALC 'NOMINAL+BLANK' CON[ (NSPL(J)) FRCM COUNT RATES (RSPL(J)
10560 C AND CALIBRATION LINE SLOPES ( U(J) )
10570 C
10580 45 CONTINUE
10590 DD 1029 J = 1,15
10600 NSPL(J) = RSPL(J) t U(J)
10610 1029 CONTINUE
10620 C
IF SODIUM HAS BEEN WET CHEM DETERMINED,THEN ASSIGN RSOD TO
10640 C NSPLI9)
10 650 C
10660 IF(NSOD.EO.IHS1 NSPL(9) = RSOD
136 R. J. PARKER and J. P. WILLIS

10670 C
10680 C CALL SUER NORFAC OR SUBR GUNFAC
10690 C
10700 IF(NEETHOO(1).EQ.iHN) CALL NORFAC(NSPL,MSPX,W,V H NX,XX,YY)
10710 IF(`1ETHOD(l).EQ.1HG) CALL GUNFAC(NSPL,NSPX,W•Wf'",WR,V.H,NX,
10720 +XX,YY0S001
10730 C
10740 C IF CORRECTED CON[ IS NEGATIVE THEN SET IT TO 0.0,
10750 C OEPENOING ON CHOIS(7)
10760 C
10770 IF(0HOIS(7).E(1.0) GO TO 1085
10780 00 1050 I = 1115
10790 IF(1(I).LT.0.0) H(I) = 0.0
10800 1050 CONTINUE
10810 1085 CONTINUE
10820 C
10830 C IF CHOIS(8) = 1, THEN CONVERT FE203 TO FEO
10840 C
10850 IF(0HOIS(8).EQ.1) FEO = H(4) • 0.89981
10860 IF(HOIS(8).EQ.i) H(4) = 0.0
10 870 C
10880 C IF FEC VALUES PRESENT THEN CALC. TRUE FE203 AND LOI
10890 C
10900 IFINFEO.NE.1HF) GO TO 1051
10910 OXFEO = FEC * 1.111353
10920 H(4) = H(4) — OXFEO
10930 ACLOI = OXFEO — FE0
10940 YY = YY + AOLOI
10950 1051 CONTINUE
10960 C
10970 C SUM H(I) AND NSPX(I) AND ADD IN HEATING LOSES,
10980 C ANO ALSO ADO FEO TO SUMCOR.
10990 C
11000 DO 1052 I = 1,15
11010 SUM0OR = SUMCOR + H(I)
11020 SUMIOM = SLMhOM + NSPX(I)
11030 1052 CONTINUE
11040 SUMCOR = SUMCOR + XX + YY + FEO
11050 SUMVOM = SUMNOM + XX + YY
11060 C
11070 C CALC. ELEMENT CONC. FROM OXIDE CONC. USING FACTORS IN WQ(I).
11080 C IF CHOIS(8) = 1, THEN CONVERT FED TO FE
11090 C
11100 00 1053 I = 1,15
11110 C(I) = H(I) * WQ(I)
11120 1053 CONTINUE
11130 IF(3HOIS(8).EQ•1) FE = FED * 0.7773
11140 C
11150 C WRITE. RESULTS TO LINE PRINTER FILE
11160 C
11170 IF(NX.GT.1) GO TO 9500
11180 WRITE(IWLF,9501)
11190 9501 FORMAT(//,X 'ABREVIATIONS' // j
11200 +X,'OXO = WEIGHT PERCENT OXIDE'''.
11210 +X,'N-9 = NOMINAL CON[ MINUS CLANK CONC (WI PERCENT)',/,
11220 +X,'_IAT = MATRIX FACTOR',/,
11230 +X,'-LM = ELEMENT CONC (WT PERCENT)'./,
11240 +X,"I+B = NOMINAL CONC PLUS BLANK LONG (WT PERCENT)',/,
11250 +X '3LK — BLANK CON[ (WI PERCENT)',/)
11260 9500 C ON
TINUE
11270 NY = NY + 1
11280 IF(NY.LT.11) GO TO 90
11290 WRIT£(IWLF,821
11300 alRITE(IWLF,83) METHOD
11310 WRITE(IWLP,83) DSCRIP
11320 WRITE(IWLP,83) NAME
1133E WRITE(IWLP 86)
11340 82 FOR4AT(1H1)
11350 83 FORMAT(/X 80A11
11360 86 FGR'1AT(//7X,45HSIO2 1IO2 AL203 FE203 FEO CR203 MNO ,
11370 +53H MGO CAO NA20 K20 P205 NIO ELX ELY .
11380 +26H ELZ H2O— LOI TOTAL //)
11390 NY =
11400 90 CONTINUE
11410 WRITE(IWLP,9060) SPLNUM,(SPLNAM(I),I=1 76)
11420 WRITE( IWLP.9001) ( H(I),I=1,4),FED,(H( I),Ir5,15),XX,YY,SUMCOR
11430 WRITE(IWLP,9002) (NSPX(I),I=1,35)sXX,YY,SUMNOH
11440 WRITE(IWLP,9004) (V(I),I=1,15)
11450 IF(CHOIS(13).EQ.1)WRITE(IWLP,9005) (NSPL(I),I=1,15)
11460 IF("HOIS(14).£4.1)WRITE(IW1P,9003) (C(I),I=1,4),FE,(C(I),I=5,15)
11470 9009 FORNATlX,I4,76A1,/)
11480 9001 FOR4AT(1)(,4HOX0 ,5F7.2,2F6.2,5F7.2,4Fo.2,2F5.1.F8.2)
11490 9002 FORMAT(1X,4HN— D ,4F7.2,7X,2F6.2 5F7.2,4F6.2,2F5.1,F8.2)
11500 9003 FORIAT(1)(,4HEL4 ,5F7.2,2F5.2 5F7.2 4Fo.2)
11510 9004 FORIAT(1X,4HMAT ,4F7.3,7X,2F6.3,5F7,3,4F6.3)
11520 9005 FDR9AT(1X 4HN+3 ,4F7.2,7X,2F5.2,5F7.2,4F6.2)
11530 WRITE(IWLP,9036)
11540 9006 FORMAT(/)
11550 WRITE(IWT,5U00) SPLNUM,(SPLNAM(I),I=1,70)
11560 WRITE(IWT,9097)(H(I),I=1,12).XX.YY,SUMGOP,
11570 9007 FOR9AT(14F5.1,F7.1)
11586 C
11590 C IF CHOIS(1) = 1 THEN PUNCH OUT OXIDE CONCS
11600 C IF CHOIS(2) = 1 THEN PUNCH OUT ELEMENT CONES
11610 C
11620 IF(CHOIS(1).EQ.1) WRITE(IPNCH,91) SPLNUH (SPLNAH(I) I=1 76)
11530 IF ( CHOIS( 1).EQ.1 ) WRITE ( IPNCH,92) (H(I ),I=1,41,FEO,(H(I
),I=5,15),
11640 +XX,YY,SUMCGR
11650 LF(CHOIS(2).E4.1) WRITE(IPNCH,92) (C(I),I=1,4),FE,(C(I),I=5,15)
11660 91 FORNAT(I4 76A1)
11670 92 FORMAT(9F7.3/9F7.3,2X,1F7.3)
11680 GO TO 42
11690 1059 CONTINUE
11700 RETURN
11710 END
11720 C
11730 C
 Computer programs SORT, REORD, and MW for major-element XRF data processing 137

11740 C
11750 C SUBROUTINE NORFAC
11760 C CONVERTS NOMINAL CONCENTRATIONS INTO MATRIX CORRECTED
11770 C CONCENTRATIONS VIA THE METHOD OF NORRISH AND HUT7ON(1969,
11780 C GEOCHIM. ET COSMOCHIM. ACTA, V33, PP431-4531.
iS79C C
11800 C ARRAYS
11810 C IHD(15) — ELEMENT SYMBOLS
1182C C NCF(15 17 ) — NORRISH ABSORPTION CORRECTION FACTORS
11830 C NSPL(15) — 'NOMINAL + BLANK' OXIDE GONGS FOR SAMPLE
11840 C H(15) — ABSORPTION CORRECTED OXIDE CONES FOR SAMPLE
11850 C V(15) — MATRIX ABSORPTION CORRECTION FACTORS CALCULATED FOR
11860 C ELEMENTS IN THE SAMPLE
11870 C NSPX(15) — 'NOMINAL — BLANK' OXIDE CONES FOR SAMPLE
11880 C OXIOE(15) — OXIDE NAMES
11890 C AVISP(6) — SEE SUBR MGIS
11900 C HXXX(151 — SAMPLE GONGS FROM PREVIOUS ITERATION
NCFSI( 15) — NORRISH CORRECTION FACTORS
1191C C
11930 C
11940 C
11950 C NCFNI(15) — '' ''
11960 C
11970 C
11980 C PARAMETERS
11990 C
12000 C NX — NUMBER OF TIMES SUBR NORFAC IS CALLED
12010 C XX — PERCENTAGE LOSS OF SAMPLE AT 110 C
12020 C YY — PERCENTAGE LOSS OF SAMPLE BETWEEN 110 C AND 10000
1203C C WS — WEIGHT OF SAMPLE TAKEN FOR FUSION
12040 C PWO — PERCENTAGE WEIGHT DIFFERENCE OF THE WEIGHT TAKEN FOR
12050 C FUSION AS COMPARED TO 0.28GRAM
12060 C EE — LOSS ON IGNITION CORRECTION FACTOR
12070 C CC — NORRISH MATRIX ABSORPTION FACTOR CALCULATED FOR
12080 C ELEMENT J IN THE SAMPLE
12090 C HX — TEMPORARY STORE FOR H(7)
12100 C SUMSPL — TOTAL C/S MG CORRECTION FOR SAMPLE
12110 C CMG — PERCENTAGE MG CORRECTION FOR SAMPLE
12120 C XXX — ABSOLUTE DIFFERENCE BETWEEN CURRENT ITERATION AND PREVIOUS
12130 C ITERATION FOR THE ITH ELEMENT IN THE SAMPLE
12140 C IFLAG — ITERATION FLAG
12150 C NFLAG —
12160 C CSTI — )
12170 C CSFE — )
12180 C CSCR — ) C/S MG INTERFERENCE FROM OXIDES IN SAMPLE
12190 C CSMN — )
12200 C LSCA — )
12210 C CSK — )
12220 C
12230 C
12240 SUBROUTINE NORFAC(NSPL,NSPXDWSTV,H)NX,XX,YY)
12250 ĪSt2JlsHETNOD( 8),DTC,
1 6 TM+Ī EPIWLIPNCH
N PR NE ,IRNIRSTD
N
12270 COMMON / AA/ U(15), BLANK(/5), AVISP(6)
12280 DIMENSION IHO(15) NCF(15,17), NSPL ( 15), H(15), V(15),NSPX(15),
12290 +OXIDE(15), HXXX(1 0)
12300 REAL NSPL NCF, NSPX
12310 INTEGER C DCIS
12320 DIMENSION
12330 +NCFFE(15), NCFMN(15), NCFTI(15), NCFCA(15), NCFK(15), NCFP(15),
1234C +NCFSI(15) NCFAL(15) NCFMG(15) NCFNA(15) NCFCR(15) NCFNI(15)
12350 REAL NCFFĒ NCFMN,MCFTI,NCFCA,KCFK,MCFP,NCFsIsNGFAL s MCFMG,NGF
N As
12360 +NCFCR,NCFNI
12370 EQUIVALENCE (NCF(1,1),NGFSI(1)), (NCF(1,2),NCFTI(1)).
12380 +(NGF(1,31,NCFAL(111, (NGF(1,4),NCFFE(1)), (NGF(1,5),NCFCR(1))s
12390 +(N F(1s 6 ),NCFMN(1)fr (NGFCii 7) NCFMG(1)), (NCF(11 8) NCF ApIt1)
12400 +(NCF(1,9) NCFNA(1)I (NCF ( 1,10),NCFK( 1)), (NCF (i,1I),NCFP(1JĪ,
12410 +(NCF(1s1Z i,NGFNI(1),*
12420 C
12430 C NOTE DATA STATEMENT IS NON STANDARD
12440 C
12450 DATA IHD/2HSI,2HTI,2HAL,2HFE,2HCRT2HMN,2HMG,2HCA,2HNA,IHK,1HP,
12460 +2HNI/
12470 C
12480 C NORRISH CORRECTION FACTORS FOR MATRIX ABSORPTION
12490 C
12500 C NOTE DATA STATEMENTS ARE NON STANDARD
12510 C
12520
C DATA NCFSI
——
12540 +/-•41.061,0.034,0.12210.082,0.050,0.086,00093,0.042,
+0,0053,'0.05510.061,0.182,0.158,1.014110.0/
12550
12560 C
12570 DATA NCFTI
12580 +/0.11110.179,0.078,0.081,0.033,0.077,0.069,0.647,0.051,
+0.644,0.18110.14111-0.132,0.851,0.0/
12590 C
12610 DATA NCFAL
-.048,—.067,0.192,—.164,1.056,.0/ —
12620 +/•0.088,•0.032,...0.072,3.112,0.054,0.116,0.116,0.0371
12630 C +0
.058,
12650 DATA NCFFE
12660 +/-0,065 0.146,-3.074,-0.327,0.244,-0.331 —0.090,0.134,
+-0.110, 0.126,—x.060,-0.070,-3.163,1.046,0.0/
12670 C
12690 DATA NCFCR
12700 +/0.10210.69810.084,0.36910.028,0.j60,0.061,0.631,0.040.
,-0.139,0.853,0.0/
+0.647,0.136,.133
12710 C MN
12730 DATA NCFMN
12740 +/-0.763 0.146,-0.074,-0.044,-0.092,-0.044,-0.078,0.135,
+•0.10920.130,•.0.063,-.0.0472•0.163.1.045,0.0/
12760 C M
12770 DATA NCFMG
-. —
12780 +/•.0.070,0.0101-0.078,0.136,0.073,0.126,0.0841-0.021,
12790 +0.080, 043,—0.016, .221,-0.163,1.050,0.0/
138 R. J. PARKER and J. P. WILLIS

12800 C
D AT
A NCFCA
12020 +/0.129,0.065 0.105,0.090,0.036,0.092,0.068,0.130,0.051,
+0.723,0.182,0.14610.134.0.865,0.0/
128300 C NA
12850 DATA NCFNA
12860 +/0.30.010.010.010.00.90.00.0,0.020.010.00.010.01
+0.010.0/
1 28 80 C K
12890 DATA NCFK
12900 +/0.119,0.017$0.101,0.098,0.028.0.036,0.080.0.0.0.057,
1291( +C.C59, .179,0.136,- 0.139,0.897,0.0/
12920
C DATA NCFP
12940 +/0.127,-0.02U,3.i10,0.iD8,0.043,0.u94,O.094,-0.037,0.046,
12950 +-0.047,-0.063,0.158,-0.139,0.896,0.0/
12960
C DATA NCFNI
12980 +/..0.08220.124,0.089,0.337,0.238,0.334,0.096.0.100.
12990 +-[.103,Li 05, -J.071,-0.070,-0.162,1.038,0.0/
13000 C
13010 DATA OXIDE
13020 +/6H5102 ,6HTIO2 ,6HAL203 ,6HFE203 ,6HCR203 ,
13030 +6HM•1O ,6HMGO ,6HCAO ,6HNA20 ,6HK20 ,6HP205 ,6HNI0
13040 +6HELX ,6HELY ,6HELZ /
13050 C
13060 C WRITE CUT CORRECTION FACTORS
13970 C
13080 IF(NX.GT.1) GO TO 42
13090 WRITE(IWLP,1021)
13100 WRITE(IWLP 1022)
13110 1021 FORMAT(///1H ,22HNORRISH MATRIX FACTORS/)
13120 1022 FOR1AT(1H ,6)(,109HSI02 TI02 AL203 FE203 CR203 MNO
13130 +MGO CAO NA20 K20 P205 NIO LOSS FLUX /)
13140 DO 1023 J=1 12
13150 1023 WRITE(IHLP,10
24) IHO(J) (NCF(I,J),I=1,14)
1316U 1024 FORMAT(1H ,A2,14(1X,F7.3))
13170 42 CONTINUE
13180 C
13190 C WS = WEIGHT OF SAMPLE TAKEN FOR FUSION.
13200 C PWO = PERCENTAGE 'WEIGHT DIFFERENCE' ( OR 'LOSS') OF
13210 C THE WEIGHT TAKEN FOR FUSION AS COMPARED TO 0.28GRAM.
13220 C NOTE THAT 0.28GRAM IS THE 'REFERENCE WEIGHT' FOR THE SAMPLE
13230 C WEIGHTS TAKEN.

PHD IS USED IN MAKING A 'LOSS' MATRIX CORRECTION.

1324( C
13260 PWO = (0.28 - WS) * 170.0/0.28
13270 C
13280 C CALC. LOI (LOSS ON IGNITION) CORRECTION FACTOR (EE) AND
13290 C CORRECT NA (NSPL(9)) UPWARDS FOR LOI, THEN CORRECT NA FOR WI
13300 C TAKEN FOR FUSION RELATIVE TO 1.28GRAM, BECAUSE NA IS NOT DET.
13310 C ON FUSION DISC. IGN ASS NSPL(I) TO H(I).
13320 C
13330 EE = 1U0.0/(172.0 - (XX + YY))
13340 NSPL(9) = NSPL(9) * EE
13350 NSPL(9) = NSPL(9) * WS/0.28
13360 DO 1045 I = 1,15
1337E H(I) = NSPL(I)
13380 1045 CONTINUE
13390 C
13400 C IF TESTING WRITE OUT NOMINAL AND BLANK CONCENTRATIONS
13410 C
13420 IF(CHOIS(9).EQ.0) GO TO 9006
13430 SUMO = 0.0
13440 DO 5998 I = 1,15
13450 SUM)3 = SUMNB + NSPL(I)
13460 8998 CONTINUE
13470 WRITE(IWLP,9004) (NSPL(I),I=1 15) SUMNB
13480 WRIT-(IWLP,9005) (BLANK(I),I=~ 15)
13490 9004 FORM A T(//,X 'N+9',X,15F7.3,2X,F7.3)
1 50C 9005 FOR`7AT(/,X,;8LK',X,15F7.3,//) 3
..3510 9006 CONTINUE
13520 C
13530 C CALC. MATRIX ABSORPTION FACTORS, DET. CORR'TED CONCS AND MAKE
13540 C BLANK CORR'S. ITERATE_ PROCESS FOR UP TO 10 CYCLES. THE MATRIX CORR'S
13550 C ARE ONLY APPLIED TO THE 11 ELEMENTS SI,TI,AL,FE CR,MN MG CA
)~ IS NOT
13570 C dETERM INE DI ON NA FUSĪONTDISC( ĪEAPIF(J.E0.9)JA G OETOU91 0
13580 C NCF(13,J) = 'LOSS' CORRECTION FACTOR(SEE NORRISH AND
13590 C HUTTON(1963,P444 AND P446))
13600 C
13610 C PRESET ARRAYS
13620 C
13630 NFLAG = 0
13640 DO 2222 I = 1,15
13650 HXXX(I) = 0.0
13660 V(I) = 0.0
1367(. 2222 CONTINUE
13680 51 CONTINUE
13690 DO 1047 J = 1,12
13700 CC = 7
13710 IF(J.EQ.9) GOTO 910
13720 DO 1046 I = 1,12
13730 CC = CC + 0.01 * H(I) * NCF(I,J)
13740 1046 CONTINUE
13750 CC = CC + 0.01 * PHD * NCF(13,J)
13760 CC = CC + NCF(14 J)
13770 H(J) = CC * NSPL (J ) - BLANK(J)
13780 910 CONTINUE
13790 .V(J) = CC
13800 1047 CONTINUE
13810 C
13820 C CALC MG CORRECTION FOR ELEMENTS TI - K
 Computer programs SORT, REORD, and MW for major-element XRF data processing 139

13830 C
13840 CSTI = H(2) • AVISP(1)
13850 CSFE = H(4) • AVISP(2)
13860 CSC = H(5) • AVISP(3)
13870 CSM4 = H(6) • AVISP(4)
13880 CSCA = H(8) • AVISP(5)
13890 CSK = H(10) + AVISP(6)
13900 SUMSPL = CSTI + CSFE * CSCR + CSMN + CSCA + CSK
13910 CMG = U(7) • SUMSPL
13920 H(7) = H(7) - CMG
13930 C
13940 C IF TESTING WRITE OUT ITERATION DATA
13950 C
l0IS( 90).EQ.0) GO TO 9003
13960 SFC
13980 DO 8999 I = 1,15
13990 SUMI = SUMH + H(I)
14000 8999 CONTINUE
14010 WRITE(INLP,9001) H, SUMH
14020 WRITE(IWLP,9002) V
14030 NRITE(INLP,90001 CMG
14046 9000 FORIAT(X,'MGO INTERFERENCE CORRECTION',F7.3,/)
14050 91 FORMAT (X,' OXO', X,15F . 3 X, F7. 3)
14060 9002 FORHAT(X,'MAT',X,15F7.4)
14470 9003 CONTINUE
14080 C
14090 C TEST FOR ITERATION CONVERGENCE
14100 C
14110 IFLAG = 0
14120 00 1048 I = 1.12
14130 XXX = ABS(H( I ) — HXXX(I))
14140 IF(XXX.GE _.0.001) IFLAG = 1
14150 HXXX(I) — M(I)
1048
14170 IF(IFLA E0.0) GO TO 2226
14180 NFLAG = NFL AG + 1
14190 IF(NFLAG.E0.10) GO TO 2223
14200 IF(IFLAG.EQ.1) GO TO 51
14210 2223 CONTINUE
14220 WRITE(INLP,2225)
14230 2225 FORMAT(/,X,'WARNING - NO CONVERGENCE AFTER 10 ITERATIONS')
14240 2226 CONTINUE
14250 C
14260 C
14270 C CORRECT CONC FOR WT DIFFERENCE FROM 0.28 GRAN
14280 C AND FOR LOSS ON IGNITION.
1429C C
14300 C
14310 DO 1049 I = 1,15
14320 H(I) = H(I) • 0.28/NS
14330 (I)/EE
1049 CONT INUE
14350 C
14360 C SUBTRACT BLANK(LI FROM NSPL(L) AND ASSIGN TO NSPX(L)
14370 C AND THEN CORRECT NSPX(MG) CONC FOR INTERFERENCE
14380 C
14390 DO 2000 L = 1,15
14400 NSPX(L) NSPL(L) - BLANK(L)
2000
14420 CSTI = NSPX(2) • AVISP(1)
14430 CSFE = NSPX(4) • AVISP(2)
14440 CSCR = NSPx(5) • AVISP(3)
14450 CSMN = NSPX(6) • AVISP(4)
14460 CSCA = NSPX(8) • AVISP(5)
14470 CSK = NSPX(10) • AVISP(6)
14480 SUMSPL = CSTI + CSFE * CSCR + CSMN + CSCA + CSK
14490 CMG = U(7) • SUMSPL
14500 NSPK(7) = NSPX(7) - CMG
14510 C
14520 C CORRECT NOM-BK CONC FOR HT DIF FROM 0.28GRAM AND
14530 C FOR LCSS ON IGNITION
14540 C
14550 DO 2001 L = 1,i5
14560 NSPX(L) = NSPX(L) • 0.28/NS
14570 NSPX(L) = NSPX(L)/EE
14580 2001 CONTINUE
14590 URN
E NŌ
14610 C
14620 C
14630 C
14640 C
14650 C SUER GUNFAG CORRECTS NOMINAL CONCS FOR MATRIX EFFECTS
1466L C AFTER THE METHOD OF 0.M.GUNN(1968 CANADIAN JOURNAL OF
14670 C SPEGTROSGOPY,V12,PP41-46 AND PP163 - 168.)
14680 C
14690 C
14700 C ARRAYS
14710 C
14720 C IHOX(16) ELEMENT SYMBOLS
14730 C GCF(15,16) - GUNN CORRECTION FACTORS
14740 C
14750 RAF(15) - TOTAL INCIDENT
C F ACTORSAFORXELEMENT SISĪ02L T o AVERAGEDNFROMON)
14770 C THE TOTAL MATRX FACTORS CALC. FOR A SERIES OF
14780 C SILICATE ROCKS RANGING IN COMPOSITION FROM
14790 C ULTRABASIC TO ACID.
14800 C
14810 C SI(15) - GUNN CORRECTION FACTORS
14820 C
14830 C
14840 C GE(15) - '' '' "
14850 C
14860 C NSPL(15) - - - - NOMINAL+BLANK OXIDE GONGS FOR SAMPLE
14870 C NSPLD(15) - FLUX OIL. " " " " 118
14880 C BLANK(15) - - - - BLANK OXIDE CONCS FROM CALIBR. LINES
140 R. J. PARKER and J. P. WILLIS

14890 C 9 L A N KC (i 5) - 44 44 44 14 SA 14 14 SO
14900 C H(15) MATRIX CORRECTED OXIDE CONES FOR SAMPLE
14910 G H0(15) - - " " " " " " "
1492C C
14930 C HXXX(15) - SAMPLE CONES FROM PREVIOUS ITERATION
14940 C
14950 C ASPL(16) - TOTAL MATRIX CORRECTIONS FOR SAMPLE FOR 16
14960 C OXIDES SI - GE. (NOTE THE MATRIX CORRECTIONS
14970 C FOR SC AND GE ARE NEEDED TO MAKE INCIDENT
14980 C ABSORPTION CORRECTIONS, SEE CODE BELOW)
14990 C
15000 C V(15) TOTAL MATRIX CORRECTIONS AFTER NORMALIZATION H.R.T. RAF(15)
15010 C NSPX(15) - NOMINAL-9LANK OXIDE GONGS FOR SAMPLE
15120 C AVISP(6) - SEE SU3R. ((GIS
15J30 C
15740 C PARAMETERS
15050 C
15060 C WS - ACTUAL WEIGHT OF ,AMPLE TAKEN FOR FUSION
15070 C WR - REFERENCE WEIGHT OF SAMPLE FOR FUSION
15980 C WF - WEIGHT OF FLUX IN FUSION
15090 C OIL - PERCENTAGE FLUX USED IN FUSION
15100 C DILX - DILUTION FACTOR
15110 C ROILX - RLVERSE DILUTION FACTOR
15120 C
15130 C SUMN- )
15140 C SWIM - ) SUMMATIONS USED IN 'TEST'
15150 C SUMIT - ) WRITE OUTS
15160 C SUMFIN - )
15170 C
15180 C NX - )
15190 C =F
152J00 C XX - )
15210 C YY - I
15220 C NFLAG -)
15230 C IFLAG -)
15240' C CSTI - ) SEE SUER. NORFAC
15250 C CSFE - 1
15260 C CSCR - I
15270 C CSMN - )
15280 C CSCA - )
15290 C CSK - )
15300 C SUMSPL )
15310 C CMG - I
15320 C
15330 SUBROUTINE GUNFAC(NSPL,NSPX,WS,WF,WR,V,H,NX,XX,YY,NSOD)
15340 COMMON DSCRIP(80),NAME(80),CH0IS(20),HETH00(8),OTC,
1535C +ITE'IP IWLP,IPNCH IRNW IRSTO,IWT
15360 COMMON /AA/ U(15), BL K(15),
AN AVISP(6)
15370 DIMENSION IHOX(16), ASPL(16), RAF(15), NSPL(15), NSPX(15),
15380 +H(15)2 V(15)( HXXX(15), dLANK0(15), NSPLD(15), HD(15)
15390 DIMENSION SI(15) TI(15), AL(15) FE(15), CR(15), MN(15),
1540C +MG( 1 5), CA(15), ~JA(15), K(15), P(15), NI(15), SR(15), BA(i5),
15410 +SC(15), GE(15), GCF(15,16)
15420 REAL NSPL, NSPLD, NSPX, MN, MG, NA, K, NI, LI
15430 INT:GER CHCIS BASIS
15440 EQUĪVALENCE (6CF(1,1),SI(1)1 (GCF(1,2) TI(11),
15450 +(GGF ( 1,3),AL(1 ) 1, (GCF(1,4),FE(1)), (GC F(
1,5),CR(1)),
1546C +(GcF(1,6),MN(1)1, (GGF(1,7),NG(1)), (GCF(1,8),CA(1)),
15470 +(GCF(1,9) NA(1)) (GCF(1,16) K(1)) (GGF(1,11) P(1)),
15480 +(GCF(1,12 1 ,NI(1)), (GCF(1,13j,SR(l3 1, (GCF(1,1 14),BA( )),
15490 +(GCF(1,15),SC(1)), tGCF(1,16),GE(1))
15500 C
15510 C GUNN ABSORPTION FACTORS FOR 1 PERCENT OXIDES
15520 C NOTE LATA STATEMENTS ARE NON-STANOARD
15530 C
15540 DATA SI / 6.394, 6.676, 12.803, 23.034, 20.392, 19.412, 16.403,
15550 + 20.867, 10.517 18.577 9.260, 7.264, 26.833 10.725 34.120 /
15560 DATA TI / 0.471 , 1.772, 0.926, 1.607, 1.49 6 42
, 1. 7, 1.198,
1557C + 1.455, 5.706, 1.29b 5.799, 1.985, 1.978, 4.976, 2.849/
15580 DATA AL / 10.802, 10.36 0, 19.738, 9.118, 31.380, 29.844, 25.235,
15590 + 32.362, 16.203, 28.803, 14.249, 11.269, 41.233, 16.209 51.368/
1560E DAT 0A FE / .175, 0.669, 2.361, 0.609, 0.573, 0.547, 3.313,
15610 + 0.547, 2.187 0.487 2.215 0.751 0.759, 1.914, 5.807/
15620 DATA CR / 0.28) [, 1.0673 9, 3. 7 , 0. 4, 0.908, 0.867, 0.727,
97
15630 + 0.874, 3.465 0.780, 3.516, 1.198, 1.202, 3.028, 5.613/
1564C DATA MN / 0.222, 0.843, 2.957, 0.766, 0.718, 0.686, 0.575,
15650 + 0.689, 2.740 0.619 2.377, 0.944 0.951 2.396 6.195/
15660 DATA MG / 17.488, 16.706, 3 31.60 , 14.70 6, 50.120, 47.647, 40.352,
15670 + 12.451, 25.928 46.409 22.770, 18.167 65.792 25.404, 69.389/
15680 DAT A CA / 0.831, 3.093, 1.611, 2.836, 0, 2.59
2.079, 2.472,
15690 + 2.549, 1.326 2.272 10.075, 3.469 3.426 8.612, 4.806/
15700 DATA NA / 29.559, 28.123, 52.406, 24.74 6, 83.457, 79.294, 67.241,
15710 + 20.953, 43.261 10.636 37.936, 30.559 109.428, 41.443, 94.144/
15720 DATA K / 1.126, 4.181, 2.168, 3.8233, 3.487, 2.796, 3.323,
15730 + 3.447, 1.785, 3.071, 1.579, 4.683 4.603, 11.567 6.367/
15740 DATA P / 4.587 8, 16.6 89, 8.541, 15.3 22, 13.653, 12.995, 12.045,
157511 + 13.856, 7.027, 12.338, 6.193, 4.822, 17.971, 38.649, 23.291/
15760 DATA NI / 0.113, 0.435, 1.546, 0.395, 2.702, 2.697, 2.169,
15770 + 0.355, 1.429 0.317, 1.447, 0.48b 0.497 1.253, 3.826/
15780 DATA SR / 0.91 0, 0.074, 0.269, 0.063, 0.476, 0.378, 0.474,
15790 + 0.059, 0.248 0.053, 0.249 0.083, 0.642, 0.218 0.686/
15800 DATA BA / 0.38 9, 14.466, 0.76.3 7', 1 37, 1.241, 1.1813, 1.060,
1581C + 1.292, 4.734 1.072 4.807 1.642 1.641, 4.132, 2.386/
1582C DATA SC / 0.6212, 2.329, 1.21 12 4, 2. 7, 1.958, 1.868, 1.570,
15830 + 1.913, 7.473, 1.736, 7.600, 2.609, 2.5890 6.511, 3.680/
15840 DATA GE / 0.052, 0.200, 0.721, 0.182, 1.265, 1.261, 1.012,
16 666, •460.671, 0.225, 701 0.585. .807/
15860 DA01 RAF/770.0,67.0,1239.0,30.0,40.0,S1.0,1785.0,136.9,2980.0,
15670 +145.0,588.0117.0,3.0,55.10.0/
15880 DATA HOX/2HSI 2HTI 2HAL 2HFE,2HCR,2HMN,2HMG,2HCA,2HNA,1HK,
15890 +1HP,2HNI,2HSR,EHBA,hHSC,1lHGE/
15900 C
15910 C
15920 C WRITE OUT GUNN MATRIX FACTORS
15930 C
15940 IF(NX.GT.1) GO TO 30
 Computer programs SORT, REORD, and MW for major-element XRF data processing 141

15950 WRITE(IWLP,10)
15960 10 FORMAT(///X,39HGUNN MATRIX FACTORS FOR 1 PERCENT OXIDE,//,
15970 +3X,59HLI20407 SI02 TI02 AL203 FE203 CR203 MNO ,
15980 +59HMG0 CAO NA20 K20 P205 NIO SRO 0A0,/)
15990 DO 12 J = 1,16
16000 12 WRITE(IWLP,14) IHOX(J) , (GCF(I,J),I=1,15)
16010 14 FORMAT(1H ,A2,15(1X,F7.3))
16020 30 CONTINUE
16030 C
16040 C CORRECT SODIUM UPWARDS FOR LOSS ON IGNITION(WET CHEM DATA
16050 C INPUT ON DRY BASIS AND NOT ON IGNITED BASIS)
16060 C
16070 EE = 100.0/(100.0 — (XXFYY))
16080 NSPL(9) = NSPL(9) • EE
16090 C
16100 C TESTING
16110 C
16120 IF(CHOIS(9).EQ.01 GO TO 3000
16130 SUMN = 0.0
16140 DO 8092 I = 1,15
16150 SUMN = SUMN + NSPL(I)
16160 8002 CONTINUE
16170 WRITE(IWLP,8004)
16180 8004 FORMAT(// X 'NOMINAL GONG')_
16190
8006 FORMAT(/ 8X 115F7. 3N2XLF7.33j_1,15),SUMN
16210 WRITE(IWLP,8007) (Ī1LANK(I),I=1 15)
16220 8007 FORMAT(/,X,'BLANK CONC',//,8X, 15F7.3)
16230 8000 CONTINUE
16240 C
16250 C CALC. FLUX. PERCENTAGE (DIL) AND THEN
16260 C RECALC. NSPL(I) AND BLANK(I) FOR DILUTION DUE TO FUSION.
16270 C
16280 IF(WS.EQ.0.0) WS = WR
16290 DIL = WF/(WF+WS) • 100.0
16300 DILX = ((100.0 — DIL)/100.0)
16310 RDILX = (100.0/(100.0 — DIL))
16320 00 998 I = 1 15
16330 BLANKO(I) = B LANK(I) • DILX
16340 NSPLD(I) = NSPL(I) • DILX
16350 HO(I) = NSFLO(I)
16360 998 CONTINUE
16370 C
16380 C TESTING
16390 C
16400 IF(CHOIS(9).EQ.0) GO TO 170
16410 SUMO = 0.0
16420 SUMH = SUMH + DIL
16430 00 1990 I=1,15
16440 SUMH = SUMM + NSPLO(I)
16450 1990 CONTINUE
16460 WRITE(IWLP,159)
16470 159 FORM T(/,X,'DIL NOMINAL CONC')
16480 WRITE(IWLP,160) DIL,(NSPLD(I),I=1,15),SUMH
16490 160 FORMAT(/ X,16F7.3 2X F7.3)
16500 WRITE ( IWLP, 161) (BLA ~1KD(I),I=1,15)
16510 161 FORMAT(/,X,'DIL BLANK CONC ,//,8X,15F7.3)
16520 170 CONTINUE
16530 C
16540 C CALC. TOTAL ABS FACTOR FOR THE 14 OXIDES SI02 TO BAD.
16550 C NB ■ GCF(1 I ) C INS ONTA ADS. FACTOR FOR FLUX AND OIL
16560 C IS THE PERCENTA GE FLUX.
16570 C
16580 NFLAG = 0
16590 DO 72 I = 1,15
16600 HXXX(I) = 0.0
16610 72 CONTINUE
16620 9999 CONTINUE
16630 DO 40 I = 1,16
16640 ASPL(I) = 0.0
16650 ASPL(I) = GCF(1,I) • OIL
16660 JX 4O JJ +1
= 1,14

16680 ASPL(I) = ASPL(I) + GCF(JX,I) • HO(J)
16690 40 CONTINUE
16700 C
16710 C
16720 C ADD INCIDENT ATTENUATION CORRECTIONS
16730 C SEE B.M.GUNN(1967) CANADIAN JORNAL OF SPECTROSCOPY,
16740 C V12,PP163-168
16750 C
16760 ASPL(3) = ASPL(3) + ASPL(8) • 0.638
16770 ASPL(1) = ASPL(1) + ASPL(15) • 0.638
16780 ASPL(8) = ASPL(8) + ASPL(5) • 0.638
16790 ASPL(4) = ASPL(4) + ASPL(16) • 0.638
16800 C
16810 C CALC. NORMALIZED MATRIX ABSORBTION FACTORS (V(J))
16 820 C
16830 DO 5094 J = 1,15
16840 V(J) = 1.0
16850 5004 CONTINUE
16860 DO 5005 J = 1,14
16870 IF(ASPL(J).EO.0.0) GO TO 5005
16880 V(J) = ASPL(J)/RAF(J)
16890 5005 CONTINUE
16900 C
16910 C IF SODIUM DETERMINED FROM WET CHEMISTRY, THEN
16920 C MATRIX CORRECTIONS NOT APPLIED TO NSPL(9) = SODIUM,
16930 C IE V(9) = 1.0
16940 C
16950 IF(NSOD.EQ.1HS1 V(9) = 1.0
16 960 C
16970 C HAKE MATRIX CORR. ON FLUX OIL. NOM. GONGS., AND SUBTRACT
16980 C FLUX DILUTED BLANK GONGS.
16990 C
17000 DO 1047 J = 1,14
142 R. J. PARKER and J. P. WILLIS

17010 HD(J) = (NSPLO(J) * V(J)) - DLANKO(J)

17020 1047 CONTINUE
17030 C
17040 C MAKE MG INTERFERENCE CORRECTIONS
17050 C
17060 CSTI = RDILX * 110(2) * AVISP(1)
17070 CSFE = RDILX + H0(4) * AVISP(2)
17080 CSCR = RDILX * H0(5) * AVISP(3)
17090 CSMN = RDILX * H0(6) * AVISP(4)
17100 CSCA = RDILX * H0(8) * AVISP(5)
17110 CSK = RDILX * H0(10) * AVISP(6)
17120 SUMSPL = CSTI + CSFE + CSCR + CSMN + CSCA + CSK
17130 CMG = U(7) * SUMSPL
17140 CMG = CMG * OILX
17150 HD(7) = HD(7) - CMG
17160 C
17170 C TESTING
17180 C
17190 IF(.1HOIS(9).EQ.0) GO TO 180
17200 5UMIT = 0.0
17210 SOMIT = 5UMIT + OIL
17220 DO 172 I = 1115
17230 SOMIT = SOMIT + HD(I)
17240 172 CONTINUE
17250 WRITE(IMLF,1160) OIL, (H0(I),I=1,15), SOMIT
17260 WRITE(IHLP,1192) V
17270 WRITE(IHLP,1193) CMG
17280 1160 FORMAT(/,X,'OXD' X 16F7.3,2X,F7.3)
17290 1192 FORIAT(X,'MAT' 8X~ 15F7.4)
,'MGO INTERFERENCE CORREGTiOM-,F8.3)
17300 1193 FORNATŪE
17320 C
17330 C ITERATE ABS. COR. PROCESS
17340 C
17350 IFLAG = 0
17360 00 1048 I = 1.14
17370 XXX = ABS(HD(1) - RXXX(I))
17380 IF(XXX.GE.0.001) IFLAG = 1
17390 HXXX(I) = HD(I)
17400 1048 CONTINUE
17410 NFLAG = NFLAG + 1
17420 IF(NFLAG.EQ.10) GO TO 2223
17430 IF(IFLAG.EG.1) GO TO 9999
17440 2223 CONTINUE
17450 IF(4FLAG.EQ.i0) WRIT£(IWLP 22251
17460 2225 FORMAT (/,X,'HARNING - NO CON
VERGENCE AFTER 10 ITERATIONS')
17470 C
17480 C REVERSE DILUTION CALCULATION
17490 C
17500 DO 51 I = 1.15
17510 H ( I) = HO (Ij * RDILX
17520 61 CONTINUE
17530 C
17540 C SUBTRACT BLANK FROM NOMINAL CONC.
17550 C
17560 00 62 I = 1,15
= NSPL( I) - BLANK(I)
580 62 NSPX(I)
175
17590 C
17600 C APPLY MG INTERFERENCE CORRECTION TO NSPX(HG)
17610 C
17620 CSTI = NSPX(2) * AVISP(1)
17630 CSFE = NSPX(4) * AVISP(2)
17640 CSCR = NSPX(5) * AVISP(3)
17650 CSHN = NSPX(6) * AVISP(4)
17660 CSCA = NSPX(8) * AVISP(5)
17670 CSK = NSPX(10) * AJISP(6)
17680 SUMSPL = CSTI + CSFE + CSCR + CSMN + CSCA + CSK
17690 CMG = U(7) * SUMSPL
17700 NSPX(7) = NSPX(7) - CMG
17710 C
17720 C TESTING
17730 C
17740 IF(GHOIS(9).EQ.0) GO TO 204
17750 SUMFIN = 0.0
17760 DO 200 I = 1,15
17770 SUMFIN = SUMFIN + H(I)
17780 200 CONTINUE
17790 HRITE(IWLP,202)
17800 202 FOR:1AT(/,X,'FINAL CORR CONCS - RECALC FOR FLUX DILUTION')
17810 WRITE(IHLP,8006) (H(I),I=1,15), SUMFIN
17820 WRITE(IWLP,203)
17830 203 FOR'IAT (/)
17840 204 CONTINUE
17850 C
17860 C RECALC. H(I) AND NSPX(I) FOR WT. DIFFERENCE FROM REFERENCE
17870 C HEIGHT (WR) AND FOR LOSS ON IGNITION
17880 C
17890 DO 55 I = 1,15
17900 H(I) = H(I) * WS/WR
17910 NSPX(I) = NSPX(I) * WS/HR
17920 H(I) = H(I)/EE
17930 NSPX(I) = NSPX(I)/EE
17940 65 CONTINUE
17950 RETJRN
17960 END
 Computer programs SORT, REORD, and MW for major-element XRF data processing 143

sample loss at 110°C and at 850-950°C. The is Mg0 c/s per percent interfering oxide) are used in
Na20/Fe0/weighings data file must be headed by a title conjunction with the sample concentrations to correct the
card followed by a flag card indicating weather Na20 and Mg0 concentration. This process is iterated, the new
/or Fe0 data are present. If the method is "Gunn", then concentrations being used to redetermine the matrix
this flag card (Tape 53-G) must carry the weight of flux correction factors, until the absolute difference between
used in the fusion and the reference weight of the sample two consecutive iterations is less than 0.001 for all oxides.
powder taken for fusion. The weight of the sample actually If there is no convergence after 10 iterations a warning is
taken for each sample may differ and this must be entered printed and the iteration is stopped. In most situations
along with the loss on heating weighings. These weights are convergence occurs after 4 or 5 iterations. Note that the
needed to calculate the percentage dilution due to fusion, redetermined matrix factors are applied in each iteration
and to allow for variations in the weight of sample taken to the original nominal concentrations, and that the blank
w.r.t. the reference sample weight. The test samples, blanks concentration is resubtracted each time from the matrix
and standards analysed by the "Gunn" method were corrected oxide concentration. The progress of the matrix
prepared using a sample to flux (Li2B407) ratio of 1:7. correction iteration may be followed by switching on
Other dilutions also may be used (Borley, 1972), provided chois (9) (see option documentation in program MW).
the appropriate flux and reference sample weights are Examples of this test output from the program are given
entered on the flag card in Tape 53-G. in Tape 62-N-testing and Tape 62-G-testing.
For the "Norrish" method the flux weight and sample Corrections are made for loss on heating and for any
reference weight are not needed on the flag card as the variation in the sample weight taken for fusion. This is the
weights used should follow the values in Norrish and final corrected oxide concentration. A second "nominal-
Hutton (1969). Variations in the sample weight actually minus-blank" concentration is corrected for loss on
taken for fusion are allowed, provided the weight taken is heating and for variations in sample weight, but not for
entered along with the loss on heating weighings for the matrix effects, and this serves as a comparison for the
sample (Tape 53-N). matrix corrected result.
The subroutine calculates "nominal + blank" concen- After returning from subroutine NORFAC or GUN-
trations, for the sample, from the count rates and FAC control reverts back to subroutine CONC and if
calibration slope factor for each element. The next step is desired, any negative oxide concentrations are set to 0.0,
to apply matrix corrections via subroutines NORFAC or and Fe203 is converted to FeO. If wet-chemistry Fe0 data
GUNFAC. have been entered for the samples, the Fe203 and loss on
The first time subroutine NORFAC or GUNFAC is heating values are adjusted accordingly. After summing
called it writes out the appropriate matrix correction the oxide concentrations the sample name and results are
factors (Tape 62-N, Tape 62-G). It then applies matrix written out (Tape 62-N, Tape 62-G).
corrections using the matrix factors of Norrish and The above procedure is repeated until all the sample
Hutton (1969) or of Gunn (1967a, 1967b). The Mg0 data sets on the second temporary disc file (Tape 12-N,
interference factors determined in subroutine MGIS (that Tape 12-G) have been processed.
rAPL11 - N

NURR1SI]
TES1 DAlA FUR PARKER - 4ILLIS PAPER
PARKER. DEC. 75
1.7
UU0U0000010000
1.0000 1.0000 1.0000
10 6536.29 1580.11 2186.93 1619.62 0 434.00 72.78 3038.24 0 9481.87 110.12 0 0 0 0
13 6595.U3 1579.88 2186.98 1637.17 0 441.33 73.84 3044.43 0 9562.75 111.20 0 0 0 0
253 7099.14 1290.85 2024.97 1188.44 U 356.95 74.21 2104.70 010134.52 115.05 0 0 0 0
8000 30.32 U O 0 0 U 0 0 0 0 0 0 0 0 0
8003 31.06 U 0 0 0 U 0 0 0 0 0 0 0 0 0
8010 0 U U 0 0 0 48.23 0 0 0 0 0 0 0 0

•f Pue enevd •r •g
8013 0 0 0 0 0 U 48.50 0 0 0 0 0 0 0 0
9000 8559.58 U U 0 0 U 59.02 0 0 0 0 0 0 0 0
9003 0613.66 U U 0 0 0 0 U 0 0 0 0 0 0 0
9013 U U 0 0 0 U 67.85 0 0 0 0 0 0 0 0
9020 9038.8/ U 0 0 U U U 0 0 0 0 0 0 0 0
9023 9137.98 U 0 0 0 0 0 0 0 0 0 0 0 0 0
9030 /014.74 0 U 0 0 U 84.12 0 0 0 0 0 0 0 0
9033 7628.91 U 0 0 0 0 0 0 0 0 o 0 0 0 0
0 U 0 0 0 0 0 0 0

SI7"IIM •d
9046 U U 0 U 115.80 U
9050 6050.78 0 0 0 U u 166.14 0 0 0 0 0 0 0 0
9U53 6006.82 11 U 0 0 0 0 U 0 0 Ii 0 0 0 0
9056 6103.02 U U 0 0 U 0 0 0 0 U 0 0 0 0
9070 0 J 0 0 0 U 181.07 0 0 0 0 0 0 0 0
9071 0 u U 0 U 0 182.20 0 0 0 0 0 0 0 0
9UhU 4532.80 U U (1 U 0 838.77 0 0 O 0 0 0 0 0
9U83 4543.7. U (1 0 0 U 852.34 U 0 0 U 0 0 0 0
9090 U U 0 0 0 0 441.83 U 0 0 0 0 0 0 0
9096 U U U 0 0 U 455.63 0 0 0 0 0 0 0 0
5153 4284.5/ U U U U U 0 U 0 0 0 0 0 0 0
9156 4340.85 U J U U 0 0 U 0 0 0 0 0 0 0
990U U U U 0 U 0 93.10 0 0 U 0 0 0 0 0
9910 U 0 U U 0 0 53.85 0 0 0 0 0 0 0 0
9920 (1 0 U U 0 U 48.911 0 0 0 0 0 0 0 0
9930 U U U U 0 0 48.59 0 0 0 0 0 0 0 0
9940 U 0 U U U 0 84.47 u 0 0 0 0 0 0 0
995(1 U U U 0 0 U 67.46 0 0 0 0 0 0 0 0
-99 u 0 0 0 U 0 U 0 0 0 0 0 0 0 0
I-1'°N 'I 1°A03DVD

u 2.000U 17.UU
1uFt12 - W U .9801 48.23
8010. SI02 /A
8013. S102/13 U .9/99 48.50
WukKlSll 9900. SITZ /A U 1.0040 93.1U
991U. SilL/n U 1.0010 53.85
ILSI UAIA FIN P0KKLK - w1LL1S PAPER
9920. 5118 /A u 1.0229 48.90
Pnl<KLrt. ULC. IS

Computerprograms SORT,REORD,andMW f ormajor-elementXRFdataprocessing

9930. SININ /A U 1.0688 48.59
1.7 9940. u .9941 84.47
51CA /A
Uu0UUUUUU1UIJO11 u .9881 07.46
9950. DIK/A
1.000U 1.000U 1.u000 .682 .9905 59.02
9000. G-1 /A
1. U 2.0000 13.00
9013. 6 -2/U .710 .9932 67.85
8000. CA(/A U .9720 30.32
903U. AGV /A 1.36U 1.11u38 84.12
0003. CA0/6 U .9 / 20 31.08 1.U174 115.00
9046. r)C1?/E 3.481
9156. 51"8)/8 38.428 1.U454 4284.5/ 166.14
9050. )-1/1' 8.611 1.U1D7
9156. 811.1)/C 68.442 1.0455 4640.85
1.0063 181.07
901U. OtrnrJ /A /.531
9080. U7S /A 40.576 1.0438 4532.80 182.20
4U.bi/ 9071. aI•rl/n 1.561 I,006S
90 r 3. !)TS/Li 1.0443 4543.76 455.63
9096, d1i,P /C 45.138 1.0101
5U53.)-1/1) n2.548 1.0114 6006.82 25. 14 / 1.0102 441.83
9090. '11i'IP /A
9056. .%-1/E 52.548 1.0114 6106.02
9040. JIS /A 49.102 .9926 838.77
9050. A -1 /A 52.507 1.0115 6058.78 852.64
9086. I)Ts/B 49.826 .9928
9030. A 60/u 00.092 1.0040 7014.74
-1. (1 U U
91)63. AGV/b 00.116 1.0041 7028.91
8. U U U
9000. li-1 /A /2.969 .9872 8559.58
-1. U 0 U
9005. 1,-1 /Li /3.099 .9875 8613.66
U U 0
9020. iAimG//l /5.993 .9834 9038.87 9.
-1. U U U
9023. IJ1rih/8 /6.156 .9836 9137.98
10. U U U
-1. U U U U
-1. U U
2. U U u U U
11. u
-1. U U U U
-1. U U
3. u 0 (1
12. u 0 u
-1 . u u u U U U
U -1•
4. U U
13. U U U
-1, u u U
U 11 U
5. U U U -1•
U
14. U U
-1. U U U
-1. U U U
8. U U U
15. U U U
-1. u u U
-99. u 0 U
7.

10 6936.29 1580.11 2186.96 1619.62 0 434.00 72.78 3038.24 0 9481.87 1100.2 0 0 0 0

16 6 4 95.03 1579.88 2186.98 163/.17 0 441.33 73.84 3044.43 0 9562.75 111.20 0 0 0 0
253 7099.14 129u.85 2024.9/ 1188.44 U 356.95 74.21 2104.70 010134.52 115.05 0 0 0 0
-999 0 U U 0 0 0 0 0 0 0 0 0 0 0 0
146 R. J. PARKER and J. P. WILLIS

TAPL51 - N 6000005061 300002

6000004982 300000
88
NORRISII 99 9000 9003
TES' DATA FOR PARKER - v1ILL1S PAPER 6000004200 300001
PARKER, DEC. 75 6000003555 300000
1.7 6000003633 300000
00000000010000 80
1.000U 1.0000 1.0000 99 9080 9083
60 57 61 66 7u 68 51 58 54 59 72 6 6000004207 300001
88 3 14 15 6000006679 300000
139 8010 8013 9920 6000006663 300000
5100010000 71896 ' 88
5100010000 4825 99 8000 8003
5100010080 4850 6000004000 290788
5100018000 4690 600000400U 1255
88 6000004000 1265
99 9930 88
5100010000 71741 99 9030 9033
5100010000 4849 6000004000 286455
88 6000004000 278048
99 9950 6000004000 278603
5100010000 71256 88
5100010000 6685 99 9153 9156
88 6000004000 285190
99 9940 9910 6000004000 1698/0
5100010000 70985 6000004000 172085
5100010000 8359 88
5100010000 5316 99 253 10 13
88 5700004000 45304
99 9900 5700004000 51521
5100010000 70941 5700004000 63055
5108810000 9185 5700004000 63026
68 5800004000 72219
99 9070 5800004000 838158
5100010000 70255 5800004000 1209U5
5100010000 17663 5800004000 121150
86 5900004000 385187
99 9071 5900004000 396515
5100010000 69237 5900004000 373258
5100010000 17540 5900004000 376691
88 6000002000 145558
99 9046 9013 60000020011 142760
5100010000 697/6 6000002000 131568
5100010000 11236 6000002000 132757
5100010000 6584 6100004000 798/4
88 6100004000 80721
99 9030 6100004000 87153
5100010000 69093 6100004000 87155
5100010000 8083 88
88
99 10 13
99 9090 9096 bb0000'+000 44141
5100010000 70235 6600004000 64607
5100010000 43129 6600004000 65605
5100010000 444/5 6600010000 40746
88 6600010000 43566
99 9080 6800010000 44100
5100010000 695/9 88
5108010000 81059 99 253
88 6600004000 44266
99 9083 6800004000 475/6
5100010000 690/1 6608010000 40566
5100010000 817/4 6800010000 35481
88
88
99 9000 99 10 13 253
5100010000 69443 4259
7200004000
5100010000 5700 4404
7200004000
88 4447
7200004000
99 9050 7200004000 4601
5100010000 69194 88
5100010000 15985 99 10 13
88 5100010000 69354
99 9020 9023 5 100010000 702u
6000004201 300001 5100010000 7122
6000003370 300000 88
6000003334 3000u0 99 253
88 5100010000 69026
99 9050 9053 9056 5100010000 7124
6000004214 3000U1 -9
6000005018 300008
 Computer programs SORT, REORD, and MW for major-element XRF data processing 147

TANL52 - iJ

Nu68ISH STANDARDS AND bLANKS

1 CUNC 8(00 CAu /A U U U 0 0 0 0100.000
2 CUNC 8000 CAU /A U 0 U U
1 N AFAC AUOU CAU /A .972U 1.4980 1.019U 1.1800 1.4840 1.1800 1.0290 .9950
2 N AFAC 0000 CAU /A U .0970 .8590 1.1380
1 CONC 0003 CAU /L3 U U U U 0 0 0100.000
2 CONC 6003 COU /U U 0 U U
1 N AFAC 0003 CAU/0 .9720 1.4980 1.0190 1.1800 1.4840 1.1800 1.0290 .9950
2 N AFAC 8003 COU /19 U .8910 .8590 1.1380
1 CONC 8010 5102/A 100.143 0 0 U 0 0 0 0
2 CUNC 8010 5102/A 0 0 0 0
1 N AFAC 0010 S1U2 /1L .9531 .9613 .9681 .9811 .9553 .9821 .9801 .9934
2 N AFAC 8010 S1U2 /A U 1.0164 1.0234 .9561
1 CUIJC 6013 S102/8 99.857 0 0 U 0 U 0 0
2 CONC 6013 S102/8 U U 0 U
1 N AFAC 11013 5102/H .9529 .9607 .9679 .9809 .9547 .9819 .9799 .9926
2 N AFAC 0013 5102/8 0 1.0156 1.U226 .9559
F INISI-i3

1 CUNC 9940 SiCA /A 70.000 U U 0 0 0 0 30.000

2 CUNC 9940 SICA /A U 0 0 0
1 N AFAC 9940 S1CA /A .9587 1.1221 .9833 1.0407 1.1137 1.0414 .9947 .9936
2 N AFAC 994U SICA /A U .9803 .9738 1.0106
1 CUNC 9930 SIAN /A 70.000 U U U 0 30.000 0 0
2 CONC 993U 51MN /A U U 0 0
1 6 AFAC 9930 SIMN /A .9911 .9511 1.0292 .9912 .9424 .9877 1.0388 .9822
2 N AFAC 9930 S1MN /A U 1.0U61 1.0131 1.0808
1 CUNC 9950 51K /A 70.000 U U U U 0 0 0
2 CUNC 995u SIK /A 0 30.000 0 0
1 N AFAC 9950 S1K /A .9548 1.1212 .9800 1.0383 1.1185 1.0399 .9881 1.1715
2 N AFAC 995U SIK /A U 1.0010 .9708 1.0121
1 CONC 9920 SICK /A 70.000 U 0 U 30.000 0 0 0
2 CUNC 992U SICK/A U U 0 0
1 N AFAC 9920 SICK/JI .9863 .9379 1.0106 1.0737 .9328 .9733 1.0229 .9654
2 N AFAC 99201 SICK /A U .9887 .9978 1.0520
1 CUNC 9910 51FL /A 70.000 U U 30.000 0 0 U 0
2 CONC 99101 51F1 /A U U 0 U
1 N AFAC 9910 51FE /A .9713 .9280 .9944 1.0005 .9244 1.0009 1.0010 .9546
2 N AFAC 991U S1FE /A 0 .9803 .9849 .9806
1 CUNC 9900 5111/A 70.000 30.000 0 0 0 0 0 0
2 CUNC 9900 S1lI /A 0 0 U U
1 N AFAC 9900 S1TI /A .9611 .9817 .9848 1.0443 1.1338 1.0447 1.0040 .9741
2 N AFAC 9900 S111 /A 0 .9854 .9789 1.0178
1 CONC 9000 6-1 /A 72.969 .261 14.101 1.952 0 .030 .382 1.397
2 CONC 9000 0-1 /A 3.337 5.508 .09U 0
1 N AFAC 9(100 0-1 /A .9872 .9912 .9832 .9930 .9888 .9937 .9905 1.0191
2 6 AFAC 9000 6-1 /A U 1.0064 1.0052 .9761
1 CUNC '1 UO3 0-1/8 73.099 .262 14.126 1.955 0 .030 .383 1.400
2 CUNC 9003 0-1/8 3.343 5.518 .091 U
1 N AFAC 9003 0-1/6 .9875 .9916 .9833 .9932 .9893 .9939 .9907 1.0196
2 N AFAC 9003 0-1/13 U 1.0069 1.0056 .9763
1 CUNC 9013 6-2/6 69.695 .504 15.462 2.694 0 .040 .776 1.994
2 CONC 9013 6-2/ 43 4.090 4.553 .141 U
1 N AFAC 9015 0-2/0 .9924 .9880 .9868 .9926 .9869 .9932 .9932 1.0117
2 N AFAC 9013 0-2/8 U 1.0048 1.0041 .9786
1 CONC 9020 814E /A 75.993 .090 12.172 1.970 0 .020 .050 .77x
2 CONC 9020 8140/8 3.372 5.058 .020 U
1 N AFAC 9020 N146 /A .9834 .9854 .9821 .9909 .9619 .9916 .9906 1.0165
2 N AFAC 9020 814E /A U 1.0074 1.0072 .9747
1 CONC 9023 6140/0 76.156 .091 12.198 1.974 0 .020 .050 .774
2 CUNC 9023 N14G/8 3.379 5.068 .020 U
1 N AFAC 9023 81MG /I3 .9836 .98b0 .9823 .9911 .9625 .9919 .9908 1.0171
2 N AFAC 9023 6146/8 U 1.0079 1.0077 .9749
1 CUNC 9030 A1V /1L 60.092 1.057 17.327 6.899 0 .101 1.560 5.031
2 CONC 903U AGV /A 4.337 2.948 .503 U
1 N AFAC 9030 A6V /A 1.0040 .9939 .9989 .9977 .9953 .9975 1.0038 .9992
2 N AFAC 9030 AGV /A 0 .9997 .9986 1.0001
1 CONC 9033 A6V/B 60.113 1.057 17.333 6.901 0 .101 1.560 5.033
2 CONC 91)33 AbV/U 4.338 2.949 .503 U
1 N AFAC 9033 AOV/13 1.11041 .994U .9990 .9978 .9954 .9976 1.0039 .9993
2 N AFAC 9033 86V/ 43 0 .9998 .9987 1.0001
1 CONC 9046 8CK/E 54.704 2.214 13.643 13.621 0 .189 3.481 6.961
2 CONC 9046 (3CK /L 3.281 1.676 .329 0
1 N AFAC 9046 8LK /E 1.0090 .9914 1.0150 1.0044 1.0038 1.0030 1.0174 .9889
2 N AFAC 9046 HCN/E U .9960 ..9960 1.0316
1 CUNC 9050 w -1/A 52.567 1.067 14.827 11.186 .020 .170 6.611 10.948
148 R. J. PARKER and J. P. WILLIS

2 COJC 9050 w-1 /A 2.144 .668 .140 .010

1 N AFAC 9058 w-1 /A 1.0115 1.0124 1.0159 1.0066 1.0127 1.0058 1.0107 .9835
2 8 AFAC 9050 6-1 /A U .9960 .9939 1.0240
1 CONC 9053 w-1/1) 52.548 1.067 14.822 11.182 .020 .169 6.608 10.944
2 CONC 9053 w-1/U 2.143 .638 .14U .01U
1 N AFAC 9053 w-1/0 1.0114 1.0123 1.0159 1.0066 1.0126 1.0058 1.0107 .9834
2 N AFAC 9053 w-1/0 U .9929 .9938 1.024U
1 CONC 9056 w-1/E 52.548 1.067 14.822 11.182 .020 .169 6.608 10.944
2 CUNC 9056 w-1/E 2.143 .668 .140 .0111
1 N AFAC 9056 W-1/E 1.0114 1.0123 1.0159 1.0066 1.0126 1.0058 1.0107 .9834
2 N AFAC 9056 w-1/E U .9929 .9938 1.0240
1 CONC 9080 UIS /A 40.576 .010 .289 8.645 .638 .110 49.702 .150
2 CUNC 9(180 UIS /A .810 0 U .309 .
1 N AFAC 9080 UIS /A 1.0438 .9388 1.0884 .9739 .9325 .9761 .9926 .9598
2 N AFAC 9080 OIS/A 0• .9945 1.0047 .9875
1 CONC 9083 1)15/H 40.677 .010 .290 8.667 .640 .110 49.826 .150
2 CUWC 9083 01S/E3 .010 0 U .310
1 N AFAC 9083 UIS/0 1.0443 .9394 1.0889 .9741 .9331 .9763 .9928 .9603
2 8 AFAC 9083 UIS/0 0 .9950 1.8053 .9878
1 CONC 9070 NI V /A 52.450 .199 16.407 8.991 .010 .179 7.531 11.426
2 COAL 9070 NLMN /A 2.461 .249 .040 .020
1 11 AFAC 9070 8148 /A 1.0135 1.0115 1.0142 1.0040 1.0075 1.0037 1.0063 .9810
2 8 AFAC 9070 I11 44 / 1) U .9928 .9945 1.0134
1 CUNC 9090 1114P /A 50.7111 .198 4.236 12.826 3.591 .218 25.147 2.649
2 CONC 9098 N1MP /A .317 .089 .02U .069
1 N AFAC 9090 N11P /A 1.0232 .9576 1.0562 .9962 .9526 .9850 1.0102 .9691
2 N AFAC 9090 NiWIP /A 0 .9960 1.0035 1.0231
1 COUC 9096 11111,/C 50.752 .198 4.234 12.821 3.590 .218 25.138 2.648
2 CONC 9096 N1MP /C .617 .089 .020 .069
1 Id AFAC 9096 N1MP /C 1.0232 .9576 1.0532 .9961 .9526 .9850 1.0101 .9690
2 N AFAC 9096 N1MP /C 0 .9959 1.0035 1.0230
1 CONC 9153 '81110/8 38.428 .020 .256 17.028 .404 .207 43.073 .256
2 CUIJC 9153 N1MU/U .059 .020 .030 .286
1 N AFAC 9153 1J1N0/U 1.0454 .9393 1.0916 .9785 .9328 .9793 1.0106 .9601
2 N AFAC 9153 1J140/8 U .9946 1.0044 1.0230
1 CUNC 9156 41110 /C 38.442 .U20 .256 17.034 .404 .207 43.088 .256
2 CUNC 9156 NIMU /C .059 .020 .03U .286
1 N AFAC 9156 N1M0 /C 1.0455 .9394 1.0917 .9786 .9328 .9793 1.0107 .9602
2 N AFAC 9156 N1MO /C 0 .9947 1.0044 1.0231
FINISIIS

TAPE53 - N

SODIUM CUNCS, OESLR1PTI01'jS AND WEIGHINGS FOR NORRISH TEST SAMPLES

S
6.37
0011) BASAL LAVA
0010 24.1699 32.9001 32.8174 62.6741 .2803
3.37
0016 bASAL LAVA
0013 24.1699 32.9u01 32.8174 62.6741 .2801
2.93
0256 IGNIMURIIE
0253 24.5068 26.1942 26.1922 26.1541 .2795

1APL h 0 - IJ

mURh16,3
TLSI 0A1A FOR PARKER - wiLLIS PAPER
PAhRLR. JLC. 75
1.7
0 0000 8 I1 08 18 0 00
1.0800 1.0000 1.0080

mIIchINL LLLMLJ1 UUMiJLIS

Si II AL FF CR MN NI, LA NA Is P NI EX EY EL
60 57 61 66 10 68 51 58 54 59 72 64 13 14 15
 Computer programs SORT, REORD, and MW for major-element XRF data processing 149
000 000 000 000 000 000 000 000 00
N
J
W
000 000 000 0 00 0 00 0 0 0 0 0 0 000 00
J
W
0 0 0 000 000 000 0 00 000 000 00 0 0 0
J
W
00 0 000 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
N
000 000 000 000 000 00
0 .0 s0 I .0 0
. :n .0 .0
400 .-10 0
e4 ,1 r1 rl .•l rl .-1 .-S -'
000 000 000 000 0 C C 00
V 0 0 M0 0 1f100
00'01 .00•O• MO• O.
Y I Is- I"- Nr r .-I I.. N
0• 0• 0• a O• O. o T Cr,
.-I
O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00
o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 rI .-, 7 ...-1 LO 14 .-1
C .41 1-1 .-1 7 .-$ •-I 0 .-I .d
lJ 0 0 0 00 o0 .-1 0 CO
SORTED DA TA

.1 rI .4 .1 3- .-I N .-1 .-1

• .
• .
. • • • • 0 0 0 0 00 •• . •• . .
. • 0 0
0 7 7 v4 7 .-1 .-1 .-1 0 0 0 0 0
Os r 11 1n
I. TO. r- O• T NO•O" 7 V V S NN 0• T
:f1
ī .0 .0 .0 .0 .0 .0 f` N r r .0 .0
000 000 000 0 0 0 000 00
330 .4033 :II .0.0
M 00 70 0 In00
X 7 7 7 77 7 M7 7
0 0 0 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .-] 00
• • 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
4.
1 .0 r
0.0 CV 3,T
.. :V 7 0
0 7 T 7 7
U. .a .7, .-I s-4 4 -, 4 . ..
1,
4-4
0 0 0 0 0 0 0 0 0 0 0 0 000 00
r 7 7 7 7 11 7 7
J 0 0 0 0 0 0 V 0 0
.1 7 7 vI 0 0 000
:V V N N N N N V V
• 0 0 0 0 0 0 7 7 7 0 0 0 0 7 0 0 0
J 7 /1 7 11 11 . 4 1) 11
•
. 0•+) 41 0 ")'A TM+)
n rl .- r rt rt cV .4 -4
-. --I .-I -. .4 -I 44 I-1 .-I
0 0 0 0 0 0
0 Z` S r1 .-1 .4 'V .4 .-1 0 .d
11 33
• 0 41 0 S) .11 .0 0 4 M
In n r .-I I) In 'V 11 in '1) 4 4 .0 M 51)
NMM .) h ) r1) n N N ,0 CV
rrr
'A 40 '1) r17 £-")`I) 0N
c rr .0r- - N N N N 0 I- r
r1 0 7 4) 0 .ry 7 h
.. .4 1) 0 0 .-I .-1 0 0
:V 7 7 7 7 0 0
33 33 0 33 7
 7231. U o \) o o o o o o o o o o

9U13 o o o o o o 66. o o o o o o o
n u o o o o 699. o o o o o o o
o o o o o o 699. o o o o o o o

9020 9039. o o o o o o o o o o o o o
7229. U o o o o o o o o o o o o
7229. U o o o o o o o o o o o o

9138. u \) o o o o o o o o o o o
722~. U o o o o o o o o o o o o
722~. U \) \) o o o o o o o o o o o

9u3U 7034. U o o o o 81. o o o o o o o o

72'+9. o u o U o 692. o o n o o o o o
720,'1. U o o o o 692. o o o o o o o o

7Uo,'1. u \) o \) II o o o o o o o o o
72o,~. U U u o n n o o o o o o o o
7:i49. U o o o o o o o o o o o o o

9ll4h IJ U a u o o 112. o o o o o o lJ o
U U o o \) o 699. o o o o o o o o
\J U a o \) o 699. o o o o o o o o

9U5U f,uo,u. U () o o o o o o o o o o o
72Ub. U \) u o II o o o o o o o o
12Gh. U \l u iJ o o o o o o o o o

">98tl. u a \) u n o u o o o o o o o
72Ub. U o U o o o o o o a o o o o
7"00. U o n n o o o o o o o o o o

bUU4. u o o o (l o o o o o o o o o
72u". U u II U IJ o o o n o o o o o
72Uh. U o o o o o o o o o n o o o

'1U70 u u o o u o 177. a o n o o o o n
II u \) n o \) 103. o o n o o o o o

Computer programs SORT, REORD, and MW for major-element XRF data processing 151
O 0 0 0 O O O O 0 0 0 0 C 0 0 0 C O 0 0 0 C O O C O O O O C 0
O OC O O O O 0 0 0 C O O O O 0 O 0 0 0 0 O 0 0 0 0 0 0 0 0 0
O 0 00 0 0 C 0 C 0 000 000 000 000 000 000 000
O 0 0 0 0 0 0 0 O O C O O 0 0 0 C 0 0 O O O 0 0 0 0 0 0 0 0 0
0 0 0 C O O O O O O 0 0 O C 0 O O 7 O C 0 C. 7 C 0 O C' 0 0 0 0
O 0 0 C 0 0 0 0 0 0 C C C C C 0 0 0 0 0 C. C C 0 0 0 0 C o 0 0
0 oC0 C 0 0 0 0 0 0 0 0 0 0 0 0 0 0 000 ' 000 0 0 0 C o 0
0 C' O 0 0 0 0 0 0 0 0 0 0 O 0 0 0 0 O 0 0 0 0 C. 0 C O 0 0 0 0
. . . . . . • . . • . . . • • 0 0 0
. 7 0
0 . • . • • • . . .
`I) J)
'n M N n N 7` N N N 'n K) 41 'n '17 N O C M .4 T 0 0
O N 7`0' rl Q' 3' rt a, 0, III O0 7OO 7`.-1 1.1 r/ri 7NN
1~ -4 s.O .D 9 .D
C' ■0 .0 7 N N I r r 1, r N h N N
C C C C 0 0 0 O CG 0 O C O 0 0 O 0 0 C C C O O C C C O O
•.- 7 O C 7 7 O 0 0 0 7 0 0 0 0 O O 7 0 _: 0 7 C 0 O O O 0 C 7
C o C O O 0 0 C O O O O O 0 0 0 0 0 0 O 0 0 0 0 0 0 0 0 O 0 0
O C O '7 = 7 0 0 0 7 C 0 7 = 0 0 7 0 7 0 0 0 0 0 0 0 0 0 0 7 7
7 0 0 7 0 07 1 7 0 0 7 7 7 7 7 0 0 0 0 7 7 7 7 C 7
O 0 C C 0 0 0 0 0 0 C C CC 7 7 C .r
D 7` J' Ps 7. T X. '' N 7 N. t`
N .-4 .'4 .`l -. rl )s rl .1 :M --1 rl
VlNN n!V'V N NN 'n NN
C. N N 7 n r ?' N 7 N r-
7-1 7 ^'1 0 .7 h D 7 0
N 9 x1 7` m ,f) /) 0 -4 N
0 7 0 0 0 1 ..I T 7' 7,
7, 7, ū` 7, 7+ 7` m T s 3'
9930 o u 0 0 0 0 48. 0 0 0 0 0 0 0 0
0 U 0 0 0 0 718. 0 0 0 0 0 0 0 0
n u U 0 U 0 718. 0 0 0 0 0 0 0 0

9940 n 0 U 0 U 0 83. 0 0 0 0 0 0 0 0
0 U 0 0 U 0 711. 0 0 0 0 0 0 0 0
U U U 0 0 0 711. U 0 0 0 0 0 0 0

9950 0 U 0 0 0 0 67. U 0 0 0 0 0 0 0
0 U 0 0 U 0 713. 0 0 0 0 0 0 0 0
U U 0 0 0 0 713. 0 0 0 0 0 0 0 0

-99 0 U 0 0 0 0 0 0 0 0 0 0 0 0 0
U u 0 0 0 0 0 0 0 (1 0 0 0 0 0
o u 0 0 0 0 0 0 0 0 0 0 0 0 0

INITIAL REFERENCE COON( RATES - SI, TI. AL, ETC

7229. 1135. 2U04. 1106. 0 408. 720. 1811. 0 9790. 106. 0 0 0 0

TAPE61 - N

NORRISH STANDARDS AND BLANKS

51u2 1102 AL203 FE203 CR203 MNO MGO CAO NA20 K20 P205 NIO ELX ELY ELZ TOTAL

8000 CAD/A u u u 0 a U 0100.000 0 0 0 0 -0 -0 -0 100.000

.972U 1.498U 1.0190 1.1800 1.4840 1.1800 1.0290 .9950 0 .8970 .8590 1.1380 -0 -0 -0

8003 CAU/B 0 0 0 0 0 0 0100.000 0 0 0 0 -0 -0 -0 100.000

.974U 1.4988 1.0190 1.1800 1.4840 1.1800 1.0290 .9950 0 .8970 .8590 1.1380 -0 -0 -0

8010 SIU2/A 100.145 U U U 0 a a U 0 0 0 0 -0 -0 -0 100.143

.95.J1 .9613 .9681 .9811 .9553 .9821 .9801 .9934 0 1.0164 1.0234 .9561 -0 -0 -0

8013 SI02/(i 99.857 U U 0 U 0 0 0 0 0 0 U -0 -0 -0 99.857

.9549 .9807 .9679 .9809 .9547 .9819 .9799 .9926 0 1.0156 1.0226 .9559 -0 -0 -0

-U -u -U -U -0 -0 -0 -0 -U -0 -0 -0 -0 -0 -U -0 0
U U U 0 0 0 0 0 0 0 0 U 0 0 0

Computer programs SORT, REORD, and MW for major-element XRF data processing 153
0 0 0 0 0 0 N N 0' 0 0 U) 0 7 a) •i -1
0 0 0 0 0 O N 0 7 .1 N) U) a) 7' N 0' 0'
0 O O O o O O N 0' 111 I• 0 O C N) N N
0 0 0 0 0 0 0 0 7' 0' 7' 0' 0' 0 0 0 0
O O O O 0 0 0 0 0' 7 0' 0' 0' 0 0 0 0
.i .1 .i .i .1 .1 .1 .i .i .1 .1 .i
00 O 0 00 00 00 0 0 0 0 00 o C 00 00 00
00 00 00 00
1 1 1 11 11 1 1 11 I I 'CI 1 I I 1 1 1 1 1 1 I 1 1 1 1 1 1 1 I
0 0 0 0 0 0 00 0 0 00 0 0
0 0 O 0 O O 0 0 0 O O O O O 0 0 0 0 0 0 1
1 1 1 1 I I I I 1 1 I 1 11 1 1 I 1 1 1 1 1 1 11 1 1 I I 1 1 I I
00 00 00
1 1 1 1 1
7 0
1 1
01 10 00 0 0 00 0 0 00 00 00 00 0 0 0 0 00 00
1 1 1 1 1 1 I t I t 1 1 1 I 1 1 1 1 1 1 1 1 1 1
0 .D 0 0 0 .4 0 0 0 0 0 0 0 .4 0 M 0 .0 0 N 0 0' 0 .4 0 .1 0 '0 070 0 0 0 0
.-1 7 .i 7
O O N N 0 N 0 0 a) 7 7 0 0 -4 .4 0 0 M ON ON ON
.4 O .t I) 0 .i N r- t. N. N
0 O O O 0 0 0' 0'• 0'• • 0' 0' O O o • 0 • O • 0
• • •
.i .1 .4 .i .1 .i .• • .4 .-. .i .i
C 2 o.-■ 00 00 07' 00' ON .4 .0 .4 .-1 ON ON .G.0 MN 0'0 C0' 0.0 00
M M 0 r- a- O O'Il'1 0'C) 77 Nh N 1. 003 CO NZ 7M 7M 7M
1. .i 00 00 .-10 00 00 .n7. tDO' M0' .10' .i0' .10'
N 0' c0 N • 0'
an 0 7' 0' 7' 0' . 0 •o • 0 . 0 • o • 0' • 0' • 7 • 0' • 7'
v4 v4 v4 vi r4 v4
OM 0.1 00 ON OK) 07 037 100' MO a17 a10' ON MOD 00 coo O0' a) as
O .0 0 .• CO 0 M 00 P-10 1f) 7 ON 0T- 7 0' 7 0' r .0 M N) M N MN
CO 0 00 on O a) 1n0 U) CD Il)o 00 .0 00 7'0' Cr 0' .D0, .00' .0 0' '0O'
on O •0 7' 0. 0' •0 •0 • O .0 •0' • 0' • 7' •0' • 0' • 0'
• • 0 • • . . M • M • C M • M • N • N • .i • • • •
.1 M .i .4 .4 r♦ rl .i
00 00 00 00 00 00 N0 10 0 00 NO 0'0 N0 COO .10 70 I4)0 N10
7 7 a-
M 7 0' N r M M O
re) N) O to M M M N .1 .1 .i
M M 7 M N) 7 7 M N N N
O .0 O N O 1l) 0 7 C '0 C .1 r r1 O .0 7 t` N) Ul 7 r1 .i N M 'n .-1 0' 0 1(1 7 7 7 7
7 •n 7M
OM N .1 1n 7 7 0'0' 00' 7.1 r.0 IN. M0' rr, 0' '00 7M
07, 70 0' 0 7 O
00' 0 N 43 In is M .1 7 .-I 0' .1 N .-1 No-1 00, TO
•0 •0 •0' • 7' •0' • 7' .0' • 0'
• 0' 0' .1 0' 0' an •0 •0 •O .•1 • • . 'P • 1n • .a . O• 0 • 7 •
0 • .•. • • .i • 4-1 • .-1 r1 .-1
N) .-1 .-1 .1 .1 .1 .1
01• 00 0.i 07 00 00 N 31 M1• .0 N C.0 O 0 00 00' .-17 .N 4h cc 1-
7 .L O N .i 7 70 00 NM 00 1110 .0 M .0 M CO N .1 0 0 0 0 0
M 0` N^- 0' N a' 0 0' 0 7' tf) 0 N O 7. -I .0 •y 0 .I 0 .i
Cr` M 0 CV 0 0
• 0 • 0 • : • 0 • 0 • 0
7' 0 7 O O • 0' • 0' • 7' • a' . a' .4 • .i • >n .0 • .0 • .0 •
.-1 .4 .
1 .1 .i .i .i .1 .i .1
O 7' .i 1) --1 .0 0' 0 0 0 0' 3) 7' O
07 0 t• 0a' 0N) 00' 0N 0N- 07' 0 N 0.0
Iln 0 in 0 N
.1 ON . 7, M 0 7 MM M M a- re) 01.--1 N.i or Or 0Mo .i
O 0 M N 0 7 0 0' 0 T 0 0' 0 0' '0 T .1 0' .-i 0' .i . 0 .4 0 .i 0
0 . 0 . 0
0 • 7' 0 a' 0 0 . 0' • 0' • Q` • 7' • 0' • 0' • 7' • 0
•
• • • ., .4 .i .1
.-1 K) .-1 .-1 .i
O t` 0 7 0 I) 0 0 0 7 0 0 0 0 0 'I) 0 3' O 0' 0 1) 0 'C) 07 00 0N 00 00
M N CO 0 N 7 M 0 IT '0 .i N in I) M N N N N N N
.-1 7 .i O r) N M 0 0 0 0 7 0' 7' 0 0 .• 0 .i 0 .i
.10' .i • 0' 7' .1 7'. 0' 0' 7' 0' a' 0' 0 • _ • 0 • 0
o • • • • • .i .1 .i .i
. 1 • .1 'n .1
O r ON O M O r O <n 0 M N 0 t/) N 7 .0 C 0' 7 .i 0' N .1 0) r• 7 .'C N .a N .0
O .1 O N) 00 7 MM tl'IM 7'N NO r- ...4 0'N ON- Na- O.0 C.a 33 ,C
0 HO .10
7 0' M N- O 0 7 7 0' 0' 0' 07 7' 7' 7 7 037 •
•3' • 7 • : • = •0
7 7' .0 0
' • : .I•0 • 0
0 ' 0 0 • 0 : • ;
7 .I • .1 • N • .i • r/ • .a • .a • .4 n .+ • .-1 • .-1 •
.4 .i .1 .1 -4 .1 .4
.1 rt .1 M .i .1
0N) 0N 00 00 07 0 0 .1 N .0 r) N O 70 -1 0 i1 N- T M 0 n 0 r 7' Na' N0'
'1) 0' 0 0 7 7 0 'C) N M 0 0 N N 7' N N 0 N) 7 7 11 N 17 N 1) N I)
0 .-1 C) .i 0 N) 0' M 7' .0 .i 0 .i CO .i 0 .1
0 N '0 .t 0' 0 .i 0 rt 0 7 7' • 0' • 7' • a' • 7' . 0 • 0 • 0 • 0
7 o 7' 0 7' 0' • 7' • 7' • • 7 • 7 •
••• • • C.
7 7 • in • N • N • r • N • 'n • 7 .i .4
r+ . .4 .. .i .--1 .4 .1 .i .1 .i .i .1 .i .1 .1
0-4 7N 0N 0 7' 00 01• r1 c11 N .0 7 0 07 .i0 •-7' r 0 7 7 1. 7 t` 'I )•.ON '0
N .i .1 10 0.1 .0 .I '.0 .1 0 0 Cr, if) 7' .D O M to 7 4-1r-- ‘ON .0 N
N 1) N `0 N C 0 N 7' N 7' tn 0 0 0 0 0 -0 7` 0 7' N 3' 0 .1 0 .I 0 .1
.1 0' .4 7' 7' • 7' •7 .7 •7 •7 • 7' •D• • 0' • 7' •7 .0 .0
• • • 0 • • • . . • .i • .1 • N • .i •--1 .-4 • .4 •
.1 .1
.-1 r+ M
O N 0 .i 0 0 0 'n 0 n 0 .-1 7 N 0' I1 Il 7 `n 7 .D -D N O M .a 7 N I) 0 7 0 7
0 0 0 7 7 0 .0 0 -• 0-1 0 7 '- )' N T 11 n 0 ) 7- 7 -4 7 -0 7 .0 1 7 -I 7 -I
7 :f) 0 7' 0 'P 0 0 0. N 0 .0 7,.0 0 0 .0 7' 7' 0 -• 7• 7 0 .4 0 •• 7 1) .1 n .i I) -4
• 7' • 0' • 7' • 1' IN . 7' • 7' • 0' • 7 • 7 • • 0 • 0 • 0 • 0 • 0 • 0
0 • 0 • 0 • 0 • 0 • 0 • N • 0 • 7' • J7 • .0 • 0 • = • 7 • N • N • N •
N N N N N ~ r r. '. .0 .i .0 .i 1) -I 41 .i 'I) .i ,l) .-

9U23N1mG /R
9020NIMG /A
9920SICF/A

9056w -1/t
9053w -1/u
99440 SICA /A

9930SIMN / A

9050w- 1/l1
9033A0v /H

9046 6CR/E
o-2/o
90U3 o-1/(i

9030 Ao V/A
9000 o- 1 / A
9950 SIK/A

9013
9080 OTS/A 40.5/6 .010 .289 8.645 .638 .110 49.702 .150 .010 0 0 .309 -0 -0 -0 100.439
1.0468 .9388 1.0884 .9739 .9625 .9761 .9926 .9598 0 .9945 1.0047 .9875 -0 -0 -0
9083 UTS/8 40.6/7 .Ulu .290 8.667 .640 .110 49.826 .150 .010 0 0 .310 -0 -0 -0 100.690
1.0446 .9394 1.0889 .9141 .9331 .9763 .9928 .9603 0 .9950 1.0053 .9878 -0 -0 -0
9070 NImN /A 52.450 .199 16.407 8.991 .010 .179 7.531 11.426 2.461 .249 .040 .020 -0 -0 -0 99.963
1.0166 1.0115 1.0142 1.0040 1.0075 1.0037 1.0063 .9810 0 .9928 .9945 1.0134 -0 -0 -0
9090 NjWP / 1\ 50.710 .196 4.236 12.826 3.591 .218 25.147 2.649 .377 .089 .020 .069 -0 -0 -0 100.190
1.0262 .9576 1.0532 .9962 .9526 .9850 1.0102 .9691 0 .9960 1.0035 1.0231 -0 -U -0
9096 NIMP /C 50.752 .198 4.234 12.821 3.590 .218 25.138 2.648 .377 .089 .020 .069 -0 -0 -0 100.154
1.0262 .9576 1.0532 .9961 .9526 .9850 1.0101 .9690 0 .9959 1.0035 1.0230 -0 -0 -0
9153 6 1W0/0 38.428 .020 .256 17.028 .404 .207 43.073 .256 .059 .020 .030 .286 -0 -0 -0 100.067
1.0454 .9393 1.0916 .9785 .9328 .9793 1.0106 .9601 0 .9946 1.0044 1.0230 -0 -0 -0
9156 NIMU/C 38.442 .020 .256 17.034 .404 .207 43.088 .256 .059 .020 .030 .286 -0 -0 -0 100.102
1.0455 .9394 1.0917 .9786 .9328 .9793 1.0107 .9602 0 .9947 1.0044 1.0231 -0 -0 -0

simm •d I puu aaNavd •1 •g

-a -u -0 -u -o -o -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 o
u 0 u 0 0 0 0 0 0 0 0 0 0 0 0

TAPE62 - N

NORKISH
TEST UAIA FOR PARKER - WILLIS PAPER
PARKER, DEC, 75

CALIBRATION LINE I-ON SI02

DEAD TIME IS .0000017 SECONDS

STANDARD STU. CONC. CALC. CONC. ABS ERROR FIEL ERROR PERCENT CORRECTED C.P.S. MG CORRECTION

8000 CAD/A 0 .253 0 0 29.5 0

8003 CAU/U 0 .260 u 0 30.2

9153 NlmO/U 38.428 38.244 .184 .48 4449.3 0

9156 NIMD /C 38.442 38.754 -.312 -.81 4508.5 0

9080 UTS /A 40.576 40.413 .163 .40 4701.5 0

9083 DFS/U 40.677 40.531 .146 .36 4715.2 0


9053 w-1/U 52.546 51.965 .583 1.11 6045.5 0

9056 w-1/E 52.548 52.801 -.253 -.48 6142.8 0
9050 w-1 /A 52.567 52.422 .145 .28 6098.6 0
9030 ANV/A 60.092 60.281 -.189 -.31 7013.0 0
9033 AGV/U 60.113 60.410 -.297 -.49 7027.9 0

9000 G-i/A 72.969 72.377 .592 .81 8420.2 0

9UUS 6-1/U 73.099 72.858 .241 .33 8476.2 0

Computerprograms SORT,REORD,and MW f ormajor-element XRF dataprocessing

902U NImG /A 75.993 76.149 -.156 -.21 8859.0 0
9023 NI'°1G/13 16.156 77.003 -.647 -1.11 8958.3 0

FIGURE UF MERI I 1S .56

Av. ABS. ERROR IS .31b AV. REL. ERROR IS .55

SLOPE OF_ CALIbnAT1UN LINE IS .8596E-02 BLANK CUNCENTRATIUN 1S .256 PERCENT

AVERAGE BLANK LPS 1S 29.6

NUI3R1SH
TLSI UAIA FUR PARKER - wILLIS PAPER
PARKER. NEC. /5

CALIBRATION LINE FOR MGO

MI INIERFE HENCE LURRECT1ON IN L/S PER PERCENT OXIDES

I1 FL CR 4N CA K
UU U U U U

OLAU TIP'L IS .000OU17 SECUNUS

STIUJUARD SFU. CONC. CALL. CONC. ABS ERROR REL ERROR PERCENT CORRECTED C.P.S. MG CORRECTION

80111 SIU2/A 0 2.871 U 0 47.3 0

6013 Slu2/t3 U 2.686 0 0 47.5 0
9900 SITZ/A 0 2.798 0 0 46.1 0
9910 SIFE/A U .395 U 0 6.5 0
9920 SICR/A 0 .159 0 0 2.6 0
 9930 Sii4N /A 0 .187 0 0 3.1 0

9940 S 1 CH/A 0 2.224 0 0 36.6 0

9950 SIK /A 0 1.170 0 0 19.3 0

9000 6-1 /A .582 .672 -.290 -75.87 11.1 0

9013 6-2/0 .776 1.214 -.438 -56.46 20.0 0

9030 A1,V /A 1.560 2.250 -.690 -44.22 37.0 0

9046 6CR/E 3.481 4.277 -.796 -22.86 70.4 0

9050 W-1 /A 6.611 7.320 -.709 -10.72 120.5 0

9070 NIMN /A 7.531 8.188 -.657 -8.72 134.8 0

9071 I1IMN /A 7.531 8.257 -.726 -9.64 136.0 0

25.074 .064 .26 412.8

SITIM 'd f PuUlNIvd f 2I

9096 NIMP /C 25.138 0
9090 NIMP /A 25.147 24.230 .917 3.65 398.9 0

9080 DTS /A 49.702 47.688 2.014 4.05 785.2 0

9063 DTs/8 49.826 48.516 1.310 2.63 798.8 0

FIGURE OF MERII IS 4.85

AV. ADS. ERROR IS .785 AV. REL. ERROR IS 21.74

SLOPE OF CALIBRATION LINE IS .6074E-01 BLANK CONCENTRATION IS 2.879 PERCENT

AVERAGE BLANK LPS IS 47.4

NURRISH
TEST DATA FOR PARKER - WILLIS PAPER
PARKER. DEC. 75

CALIBRATION LINE FUR M6O

M6 1NTERFE HENCE CORRECTION IN C/S PER PERCENT OXIDES

TI FE CR MN CA K
1.54 .22 .09 .10 1.22 .64

DEAD TIME IS .0000017 SECONDS

STANDARD STU. CONC. CALC. CONC. ABS ERROR REL ERROR PERCENT CORRECTED C.P.S. MG CORRECTION

8010 S102 /A 0 2.981 0 0 47.3 0

8013 SIO2/fi U 2.997 0 0 47.5 0
9900 SITI /A 0 2.906 0 0 46.1 0

ComputerprogramsSORT,REORD,and MW f ormajor-element XRF dataprocessing

9910 SIFE /A u .410 U 0 6.5 0
9920 SICk /A 0 .165 U 0 2.6 0
9930 S1i4N /A 0 .194 U U 3.1 0
9940 S1CA /A 0 2.310 0 0 36.6 0
9950 S1K /A U 1.215 0 U 19.3 0

9000 6-1/A .682 .315 .067 17.57 5.0 6.1

9013 6-2/li .176 .837 -.061 -7.86 13.3 6.7
9030 AGV /A 1.560 1.632 -.072 -4.62 25.9 11.2
9046 UCR/L 3.481 3.435 .046 1.31 54.5 15.9
9050 w-1 /A 6.611 6.475 .136 2.06 102.7 17.9
9070 N1MN /A 7.531 7.469 .062 .82 118.4 16.4
9071 N1mN /A 7.531 7.541 -.010 -.13 119.6 16.4
9096 N1mP /C 25.138 25.614 -.476 -1.89 406.1 6.7
9(190 NINW /A 25.147 24.737 .41U 1.63 392.2 6.7
9080 UT5 /A 49./02 49.385 .317 .64 783.0 2.1
9083 UlS/U 49.826 50.245 -.419 -.84 796.7 2.1

FIGURE OF MERL I IS 1.17

AV. ABS. ERROR IS .189 AV. RLL. ERROR IS 3.58

SLOPE OF CAL/BK(010N LINE 1S .6307E-U1 BLANK CONCENTRATION IS 2.989 PERCENT

AVERAGL tLANK LPS IS 4/.4

 SurMARr oF CALILSRA11ON turn 00
00

ELEMENT SLOPE F3LANK AAE ARE )UM

S102 .8596E -02 .256 .316 .553 .559

mu0 .6307E -U1 2.989 .189 3.580 1.168

8888ISO MATRIX FALT08S

S102 TI02 AL203 FE206 CR203 /HNO MU) CAO )JA20 820 P205 NIO LOSS FLUX

SI -.061 -.034 .122 .082 .050 .086 .093 -.042 .063 -.055 -.061 .182 -.158 1.014
11 .11u .179 .078 .081 .033 .077 .069 .647 .051 .644 .181 .141 -.132 .851
AL -.088 -.032 -.072 .112 .054 .116 .116 -.037 .058 -.048 -.060 .192 -.164 1.056
FE -.065 .146 -.074 -.027 .244 -.031 -.090 .134 -.110 .126 -.060 -.070 -.163 1.046
CII .102 .690 .084 .069 .028 .060 .061 .631 .040 .647 .136 .133 -.139 .853

'f II
MN -.063 .146 -.074 -.044 -.092 -.U44 -.078 .135 -.100 .130 -.063 -.047 -.163 1.045
Mu -.070 .010 -.078 .156 .073 .126 -.084 -.021 .080 -.043 -.016 .221 -.163 1.050
.128 .U36 .092 .068 .130 .051 .723 .182 .146 -.134 .865

Y3NHVd
CA .UbS .105 .U9U
NA U U II U U 0 u 0 0 0 0 U 0 0
K .119 .011 .101 .U98 .028 .086 .080 0 .057 .U69 .179 .136 -.139 .897
I-' .127 -.02u .110 .108 .043 .094 .U94 -.037 .048 -.047 -.063 .158 -.139 .896

f Pue
81 -.082 .124 -.U89 .337 .238 .334 -.096 .100 -.103 .105 -.071 -.070 -.162 1.038

d
IBREVIATIONS

smiM
OXU = 'WEIGHT PEIICLNT OXIDE
N-B = NOMINAL CURL mINUS BLANK CUNC (WT PERCENT)
MAT = MATRIX FACTOR
ELM = ELEMENT CONE (WT PERCLNI)
8+8 = NOMINAL CURL PLUS BLANK CONC (NT PERCENT)
ULK = BLANK CONE t N f PERCENT)

NOURISH

TEST UAIA FOR PARKER - WILLIS PAPER

PARKER. DEC. 75

820- LOI TOTAL

SI(02 TIU2 AL2U3 FE2O3 FLO CR203 MN0 860 CAO NA20 K20 P205 NIO LLX ELY ELZ

10 BASAL LAVA
0 .9 1.6 99.37
UXI) 54.66 .SU 19.83 4.33 0 0 .15 .77 4.06 6.37 8.95 .16 0 0 0
 ▪•

• •
Computer programs SORT, REORD, and MW for major-element XRF data processing 159
.a Ii) o ON
r1 CD N M 0
O O. O O
O ON O O
r1 r• -.
43 VD MIh
a
•1 .• N N
Q• N 14) 0• 01
.0 .4 CV CV CV
0• 0• r1 rl 0, M a a 0
7 O a a O
a 0 0 0 0
o .-I ri . .4 r1 r1
00 000 000
0 0 00 0 0 0 0 0 0
00 000 000
a
0 0 00 00 00 0a
0a a a a O O a
N
a a o a a 00 aa 00
00. 00m 00.4 U-)
r,
m m m .a
a• m as
010134.52

0 a a .- OS 07 O7
.0 CO J0 t` N CO CO 7 4.1 .I r• rl
rl 0+ 4-1 r1 0. . . 0•
• as • • T • •.ON
CD N .-1 K) .i a r1 a r/ a
.D ON 7 7 0 7 N r1 r` 00 .D 7 7> Cr .00. 7 7
O• O• o 00 .D.D0 N) : .i 0• r/ 01 r1 01 r1 01
•ON • • o a • • 0. • 01 • 0. • 01
'L • 0• ON • T ON •
r• r1
.D JD O a .1 7 .1 7 11) 7
N a r- 1- a In .1 a Ul .i .D CV U') r1 I1 1 r1 Ul .4
M M M O 0, a 7 a 0)a 0a a 0 00 a
N • • • o • 0 • o .a
M M If) NN 7
T ON 7 O ON
r• rl rl r1
a
.-• .0 a .Da .pa .00 .0 a
r1 CO CO N 01 N a M N In 01 01 0• on
01M a0.N) xr•7 ON O, ON ON ON
•a • • a • • a rl
1 • 7 M • N N • N N CV N N CV
rf .-1 r/
7
N 0• 0. in r)7 M 4) MM
MCC M IL1 .D 7 7 7 0 n) n o7 cO MJ 0 M CO CO
N 0• m n ON 7,0) 01 Ul 33 a 9 7 c0 7 a 7 0 a
•01 • • 01 • • ON 0) • • •J •a •a •a
J7 n N CV • N • N • N •
0'. N .r r• .-• ..i
.0 • 0 N T Ir) In Q, Ii) 0• P 01
I) 0 0. 7 a .0 ..) .0 M .O .•)
.4 a r1 .4 a r1 rl 0 '4) .0 IT 0• 7' a' 0• 0) 0• 0• 7'
•7 • • a • • J •0. • 0• • 0' •0'
7 N • • • •
.-I
rl r• r1 a 7
Ul
0N 0 0O+ -I r- x a m1 0 -7 a .0 a .0
CV 44 • U) .r .4 .• 7 -I 7 rr a
0 .0 .+ a rt a .d a .1 0 44 0
0
• N • • • o • a • O • a
.4 44 r1 .-1 .4
0 7 N
7 N 0 0 0 .-• o Ir) O in a Ul
0, N 0• 7, 0•
.JO .a N -4 -4 -I
x • o a a a
.I ..D • 0 • J . a • 7
'4 'V .-1 7. .4 7• .4 0. ---I 7,
0I- 37 n .0 Ul 7 7 •D .0 .0 .0
M a .1 Mo .-1 r4 r- 0 7 V 1- •c) • a .-• • a .1 • 0 •-• •
•a • • 0 • • a on N 7 4) 0 o N 7 N 7 N 7
7 • 7 7 • ml '•) • a •) r4 N a N a N 0 N 0
N 0• •O •a •a •7

SAM PLL U A TA
44 r1 rl rt Y) 4 • ^.) • Z h ▪ M •Z
N
CD • .-1 0 .-I 0 r4 0 .-• 0
N .0 ' .-1 -I .r .y
N Cr .1.D M n UI Cl` N 71 N 0 `1 I- r1 M I- 10 7N 417 . -
7, 71, 70•0• V) '1l .n x N h') .J .D 7, C.) 47 7, U 0 7• .J
• 7 • • 11 • • 71
30 01 .4.! •- 10 J '- .D J N ]0 J
n
7, • 7. 0• Y) a 'i :
-4 .1 y
• O 7 •
.-I -t b • •7, Y •7T Y •7.Y • 7•.0
CD .0 7• 7 • Y '0 • X
7. 7 .-1 .4 0 M 0 r10 b• D .-I • 0
71 V N 11 U :J :J :.J
N -1 • 0 1) 7 ' N +) V N V V
i7 t 7 T O• a -) J a a 4.I a a J a 0 ~.I
a 7 N 7 a I' "NL) 0- NO 7Nt.J 3- NL)
.• J • • •a? • 0 Z •0.0 •0Z

1 GNI BRITE
r1 • - • ..0 • y • W

MU TEUUT
as -. .+ Y .-4 Y .'I Y .+ Y
'J
N Cr N IL) 1- a:.i .4 NLo .L: :41 i.....1 1)741 Li
13 BA SAL

o
1- £ . • .L) 10 N LL I) 9 .L .L) b L 11 00 l,-
7•
7 • • V 00 0 .0'0' : 07Y .0 0` "
UI r1
• a W •a W •0•11 •0, .il
CJ 4-1 • H • 4- 0 • I- '0 • F-
.-I .0 .0 •-1 7 .0 ? L 7 .0
h ,e) M .4 .4 M
.0 n U 30 I- - n r.0ro V)
X I Q N ,c I Q '/J .n • P .f. Jo S .: r 0 O 0 U •- 0 o al
O Z f 0 Z J V N N N + J X G J' X Q 0 X C :7 X .0 N
4- Z 0 0ī Z 0£ F 00.0 0 £ £
OAF) 59.32 .39 16.35 3.15 0 0 .11 .94 2.82 2.93 9.64 .18 0 0 0 0 .1 2.3 100.20
N-b 59.4? .3o 18.55 3.14 U .11 .98 2.70 2.93 9.62 .18 0 0 0 0 100.39 0
.1 2.3
MAT .99m 1.02u .989 1.004 1.019 1.005 .994 1.043 0 1.001 .994 ,991 0 0 0

IhPL11 - <i

1,UNN
1LS1 U6IA FOR F'ARKLR - WILLIS PAPER
P/11<KLR, JU j E. 76
1.7
0II00000UI001100
1.0000 1.0000 1.0000
50 1,362.h6 4/87.84 1636.6816506,74 U 13U3.35 231.96 9517.53 0 987.23 53.90 0 0 0 0
53 6355.04 4623.4) 1615.851644/.57 0 1318.23 223.01 9475.54 0 979.09 55.75 0 0 0 0
9240 0 2512.68 1787.58 0 U U 0 0 0 0 0 0 0 0 0
9241 0 2261.34 1707.20 U 0 0 0 U 0 0 0 0 0 0 0
925u 0 5620.53 2023.34 0 0 U U 0 0 0 0 0 0 0 0
9251 0 5695.19 2016.55 0 0 0 U 0 0 0 0 0 0 0 0
9270 0 1619.97 1559.07 0 0 0 0 0 0 0 0 0 0 0 0
0 J695.58 1570.77

SMIM •d•f puu uaNavd'f

9271 0 0 0 0 0 0 0 0 0 0 0 0
9290 U 67.44 50.01 (1 0 U U 0 0 0 0 0 0 0 0
9291 U 70.20 48.75 U 0 U 0 0 0 0 0 0 0 0 0
9410 U 708.43 484.97 U U 0 0 0 0 0 0 0 0 0 0
9411 U 723.63 484.82 0 0 U 0 (1 0 U 0 0 0 0 0
942u 0 360.97 1452.60 0 U 0 0 U 0 0 0 0 0 0 0
9421 0 359.69 1451.04 0 0 0 0 0 0 0 0 0 0 0 0
9430 0 687.43 1902.65 0 0 U 0 0 0 0 0 0 0 0 0
9431 U 673.13 1899.61J U 0 U 0 0 0 0 U 0 0 0 0
9440 U 1700.21 1566.19 0 0 U U 0 0 0 0 0 0 0 0
9441 U 1(55.85 1508.26 0 U 0 U 0 0 0 U 0 0 0 0
9450 U 118.94 2U/5.17 0 0 U U U 0 0 0 0 0 0 0
9451 0 161.81 2060.46 0 0 U U 0 0 0 0 0 0 0 0
9460 U 115.46 52.87 0 0 U 0 0 0 0 0 0 0 0 0
9461 0 112.51 52.10 0 0 0 0 0 0 0 0 0 0 0 0
-99

IAE'L12 - U

Gurii
rLS1 UAIA FOR PARKLR - wILL1S PAPER
PARKER, JUNE. 76
1.7
000000000(11101)
1.0000 1.0111)0 1.00110
1. U U U
-1. U U U
2. u 0 20.00
929U. UTS .01U .9146 67.44
9291. UTS .010 .9143 70.20
9460. NIMU .u2U .9136 113.46
9461. NIMU .020 .9136 112.31
9450. NIMS .050 1.0585 178.94
9451. NIMS .050 1.11583 181.81
9420. 41MG .09U .9802 560.97
9421. 41MG .090 .9802 359.69
9410. NIMP .199 .9406 108.43

Computerprograms SORT, REORD,and MW f ormajor-element XRF dataprocessing

9411. WIMP .199 .9406 723.65
9450. NIMN .199 1.0146 687.45
9431. N1 84 .199 1.0146 673.15
9440. NIML .5U2 .9839 1700.21
9441. NIML .5U2 .9839 1/35.85
9240. GSP .661 .9921 2612.68
9241. GSP .661 .9921 2261.34
925U. AGV 1.058 .9896 3620.53
9251. AGV 1.058 .9896 3693.19
9210. BCH 2.220 .9906 7619.97
9271. 8LH 2.220 .9906 7695.38
-1. U U U
3. 0 U 20.00
9460. IJL) .257 1.0645 52.87
9461. NIMU .257 1.0645 52.1U
929U. UTS .290 1.0623 50.01
9291. UTS .29U 1.0626 48.75
9410. NIMP 4.244 1.0229 484.9/
9411. NIMP 4.244 1.0229 484.82
9420. NIMG 12.172 .9441 1952.60
9421. WIMG 12.172 .9447 1451.04
927u. BCH 13.678 .979U 1559.07
9271. UCH 13.678 .979u 1570.77
9440. I'JJML 13.910 .9715 1586.19
9441. NIML 15.910 .9715 1588.26
9240. 6SP 15.348 .9524 1787.58
9241. GSP 15.648 .9524 1787.28
9460. NIMN 16.418 .9751 1902.65
9431. 81W 16.418 .9751 1899.6U
945U. NIMS 17.197 .9466 2U73.17
9451. N1MS 17.19/ .9466 2060.46
925U. 8(,V 11.341 .961U 2023.34
9251. AGV 11.347 .9610 2018.35
-1. U U U
4. U U U
-1. u 0 U
5. U U U
-1. U 0 U
6. U U U
-1. U U 0
7. U U U
 -1. U U U
8. U U U
-1. U U U
9. U U U
-1. U U U
10. U U U
-1. U U U
11. U U U
-1. U U U
12. U u U
-1. u u U
13. u U U
-1. U 0 U
14. U U U
-1. U U U
15. U U U
-99. U U U
50 6362,66 4787.84 1636.6816506.74 U 13U3.35 231.96 9517,53 0 987.23 53.90 0 0 0 0
53 6355,04 4823.40 1615,8516467.57 0 1318.23 223.01 9475.54 0 979.09 55.75 0 0 0 0

"MU113N1Vd • f •g
-999 U U U 0 0 U 0 0 0 0 0 0 0 0 0

sfll[m •d •f
 Computer programs SORT, REORD, and MW for major-element XRF data processing 163

TANE52 - G

GUNN STANDARDS ANU ULANKS

1 CUNC 8010 SI 100 .000 0 0 0 0 0 0 0
2 CONC 8010 SI U U U U
1 G AFAC 8010 SI 9042 .9457 .9315 .9391 .9509 .9665 .9742 .9977
2 G AFAC 8010 SI 9857 1.0393 1.0373 .9015
1 CUNC 6030 AL 0 0100.000 0 0 0 0 0
2 I CONC 8030 AL 0 0 0 U
1 G AFAC 8030 AL 1. 1676 .9149 .9172 .9093 .9212 .9355 .9602 .9676
2 G AFAC 8030 AL 9715 1.0084 1.0082 .8721
1 CONC 8040 SIAL 5U.000 0 50.000 0 0 0 0 0
2 CUNC 8040 SIAL 0 U 0 U
1 G AFAC 8040 SIAL 1. 0559 .9303 .9243 .9242 .9361 .9510 .9672 .9827
2 G AFAC 8040 SIAL 9786 1.0269 1.0228 .8868
FINISHB
SECOND CARD
1 CONC 9200 6-1 72.944 .261 14.099 1.949 0 .030 .382 1.396
2 CONC 9200 6-1 3.334 5.503 .090 0
1 G AFAC 9200 6-1 .9599 .9883 .9461 .9919 .9972 1.0137 .9872 1.0320
2 G AFAC 9200 G-1 .9891 1.0146 1.0127 .9803
1 CUNC 9210 6-2 69.677 .5u4 15.458 2.691 0 .040 .775 1.994
2 CUNC 9210 0-2 4.089 4.552 .141 0
1 G AFAC 9210 6-2 .9671 .9831 .9492 .9915 .9944 1.0108 .9902 1.0223
2 G AFAC 9210 6-2 .9900 1.0129 1.0109 .9896
1 CONC 942U N1MG 75.994. .090 12.172 1.968 0 .020 .050 .773
2 CUNC 9420 N1MG 3.372 5.U58 .020 U
1 G AFAC 9420 N1MG .9526 .9802 .9447 .982U .9872 1.0035 .9868 1.0280
2 G AFAC 942U NIMG .9886 1.0166 1..0149 .9727
1 CONC 925U AGV 60.162 1.058 17.347 6.899 0 .101 1.561 5.037
2 CONC 9250 AGV 4.342 2.952 .504 U
1 G AFAC 9250 AGV .9865 .9896 .9610 1.0172 1.0066 1.0234 1.0024 1.0060
2 G AFAC 9250 AGV 1.0014 1.0059 1.0027 1.0761
1 CONC 9270 HCR 54.841 2.220 13.678 13.621 0 .190 3.489 6.979
2 CONC 9270 HCR 3.289 1.680 .330 U
1 G AFAC 9270 HCR .9961 .9906 .9790 1.0507 1.0193 1.0364 1.0180 .9937
2 G AFAC 9270 UCR 1.0197 1.0004 .9974 1.2031
1 CONC 9260 w-1 52.657 1.069 14.852 11.178 .020 .170 6.622 10.967
2 CUNC 9260 w-1 2.147 .639 .140 .010
1 G AFAC 9260 8-1 1.0016 1.0143 .9787 1.0562 1.0320 1.0494 1.0099 .9850
2 G AFAC 9260 w-1 1.0149 .9972 .9949 1.1736
1 CONC 9240 0SP 68.011 .667 15.348 4.381 0 .040 .970 2.041
2 CONC 9240 GSP 2.829 5.588 .283 U
1 G AFAC 9240 GSP .9703 .9921 .9524 1.0073 1.0052 1.0219 .9926 1.0316
2 G AFAC 9240 OSP .9959 1.0114 1.0088 1.0301
1 CUNC 9430 NIMN 52.484 .199 16.418 8.978 .010 .179 7.536 11.434
2 CONC 9430 NIMN 2.462 .249 .040 .020
1 G AFAC 9430 NIMN 1.0025 1.0146 .9751 1.0415 1.0236 1.0408 1.0037 .9800
2 G AFAC 943U N1MN 1.0079 .9970 .9952 1.1281
1 CONC 9290 UTS 40.654 .010 .290 8.645 .640 .110 49.799 .150
2 CONC 9290 UTS .010 0 0 .310
1 G AFAC 9290 UTS 1.0398 .9143 1.0623 .9371 .9189 .9340 .9829 .9634
2 G AFAC 9290 UTS .9940 1.0066 1.0030 1.0307
1 CONC 9410 NIMP 50.862 .199 4.244 12.812 3.598 .219 25.193 2.653
2 CONC 9410 NIMP .378 .089 .020 .070
1 G AFAC 9410 NIMP 1.0122 .9406 1.0229 .9876 .9480 .9637 1.0072 .9704
2 G AFAC 9410 NIMP 1.0172 1.0044 1.0041 1.1814
1 CONC 9440 NIML 53.978 .502 13.910 10.205 0 .779 .236 3.270
2 CONC 9440 NIML 8.559 5.658 .072 0
1 G AFAC 9440 NIML .9866 .9859 .9713 1.0251 .9954 1.0121 1.0168 1.0130
2 G AFAC 9440 NIML 1.0038 .9894 .9874 1.1301
1 CONC 9450 RIMS 63.676 .050 17.197 1.443 0 .010 .481 .681
2 CONC 9450 RIMS .411 15.433 .14U U
1 G AFAC 9450 NIMS .9695 1.0583 .9466 1.0595 1.0663 1.0841 .9823 1.1103
2 G AFAC 9450 NIMs .9924 .9945 .9926 1.0380
1 CUNC 9460 N1M0 38.569 .U2U .257 17.034 .406 .208 43.230 .257
2 CONC 9460 NIMO .059 .020 .030 .287
1 G AFAC 9460 NIMO 1.0422 .9166 1.0645 .9614 .9185 .9336 1.0034 .9618
2 G AFAC 9460 NIMO 1.0143 1.0012 .9999 1.1687
1 CONC 9470 PCC 44.033 .010 .763 8.683 .460 .125 45.579 .554
2 CONC 9470 PCC .010 U 0 .334
1 G AFAC 9470 PCC 1.0316 .9194 1.0528 .9422 .9242 .9593 .9840 .9656
2 G AFAC 9470 PCC .9951 1.0051 1.0043 1.0331
FINISHS
SECOND CARD
 TAPE53 - G

SODIUM CONCENTRATIUNS. DESCRIPTIONS AND WEIGHINGS FOR GUNN TEST DATA

S F 0.40 2.80
1.91
7.78
0050 LAVA WITH PINK SPOTS
0050 27.67612 35.68384 35.65114 35.26487 0.4
1.91
7.78
0053 LAVA WITH PINK SPOTS - DUPLICATE
0053 27.67612 35.68384 35.65114 35.26487 0.4
3.85
0.0
0110 GRANITE
011U

SITHM 'd'f PUUaaxxvd f H

TAPE61 - G
GUNN STANDARDS AND BLANKS

514)2 1102 AL2O3 FE203 CR203 MNO MGO CAO NA20 K20 P2OR NIO ELX ELY ELZ TOTAL

8010 SI 100.000 0 0 0 0 0 0 0 0 0 0 0 -0 -0 -0 100.000

.9042 .9457 .9315 .9391 .9509 .9665 .9742 .9977 .9857 1.0393 1.0373 .9015 -0 -0 -0

9030 AL 0 0100.000 0 0 0 0 0 0 0 0 0 -0 -0 -0 100.000

1.16/6 .9149 .9172 .9093 .9212 .9355 .9602 .9676 .9715 1.0084 1.0082 .8721 -0 -0 -0

8040 SIAL 50.000 U 50.000 0 U 0 0 0 0 0 0 0 -0 -0 -0 100.000

1.0359 .9303 .9243 .9242 .9361 .9510 .9672 .9827 .9786 1.0239 1.0228 .8868 -0 -0 -0

-0 -U -U -U -0 -0 -U -0 -0 -0 -0 -0 -U -0 -0 -0 0
U 0 U 0 0 0 0 0 0 0 0 0 0 0 0

9200 G-1 72.944 .261 14.099 1.949 U .030 .382 1.396 3.334 5.503 .090 0 -0 -0 -0 99.988
.9599 .9883 .9461 .9919 .9972 1.0137 .9872 1.0320 .9891 1.0146 1.0127 .9803 -0 -0 -0

9210 6-2 69.6/7 .504 15.458 2.691 0 .040 .775 1.994 4.089 4.552 .141 0 -0 -0 -0 99.921
.96/1 .9831 .9492 .9913 .9944 1.0108 .9902 1.0223 .9900 1.0129 1.0109 .9896 -0 -0 -0

9420 NLJIG 75.934 .090 12.172 1.968 0 .02U .050 .773 3.372 5.058 .020 U -0 -0 -0 99.517
.9526 .9802 .9447 .9820 .9872 1.0035 .9868 1.0280 .9886 1.0166 1.0149 .9727 -0 -0 -0

9250 AGO 60.162 1.058 17.347 6.899 U .101 1.561 5.037 4.342 2.952 .504 U -0 -0 -0 99.963
.9865 .9896 .9610 1.0172 1.0066 1.0234 1.0024 1.0060 1.0014 1.0059 1.0027 1.0761 -0 -0 -0

Computer programs SORT, REORD, and MW for major-element XRF data processing 165
N .i CO a. 0 N 0 N N .-1 0
ei N In 0 ei Nl .0 N N U)
.1 3 .-I O .D N) ei In N) 1)
0 0 0
0 0 '
0 7 o 0
0 0 O 0 O 0 Q' 0' 0 O
.-1 .-1 .i ei .-1 el .4 .1
MGCORREC TION

00 0 0 00 00 00 00 00 00 00 00 00
I 1 1 1 1 1 1 1 1 I I 11 I t 1 I I t 1
0 0 0 0 0 0 0 0 O O 0 O O O 0 0 O O O 0 O C
1 I 11 1 1 II III I I I t 1 1 I I 1 I I
00 00 00 0O 00
COR RECTED C.P.S.

00 00 00 OC D O Oc
I t 1 1 1 1 1 I t I I 1 1 I I I I t
0.4 0.O C5 IA 0.i 0 1- 7 3 Cel O7 NN Sel 0 O
N) .i N1 O N O .i O N ei O O N.0CO CO N7 N1 1
7 On 'h 7N 'n.) 0O 'n N) .1 "s1
N • ei 0 • ei • 0 • ei .i O • .i • O
ei ei ei ei .4 .i .i ei .4 .i
0? C 0' N) 7 O N O O C .-I N S 0 .0 O 0' C N) C C
S 3 0 0 7 1n N)
Ma Is- N7 Ps Ps ?N N) 0' 7 1
M 0'.i0' N0 00' O 00 CO .i0' 00' 0
.0' . 0' .0 • 3' CD . 0 •0' • 0' . 0' 0
.i .i .-I .i
'0
REL ERRORPERCEN T

03 0'N CO 3 0' O O .0 0 0 O? N1 U) 0 N 0 e 00 N
CO O 'O N CO .-I ? N N) O ? U) 0' Al 3 N CD.i 1n 1
.DO .00' Ul r4 N0' O CD CD •0' 'DC S 0' CD O '0
•0 • 0' • 0 • 0' 0 •0 •0' •O o
.-I • • In • . U) • 0 • •
.-. .i ei el .i ei .i
0'O ei3 0'N) 0.-1 0 0
0'1N NO' ON O' N0' 00 ON
00' S 3 NN 1) r- .i 3 .1 Nb 11 Ill r4 I) 3
.-1 in CD 7 O' o ei 00'
U1
ei
Nel .i ei 00' 3 0 0 0'
•O • 0 . 7. •7 • Cl, • O •0 • a. • 0 • 0'
.1 ,..)• N • N • N • ••0••
ei el ei ei ei
0' r- 1` 7 ei .0 3 O CD 7 10 3 0 0 .-I N) N' 0 3 D 7 0
N N) .D U1 3.1 .1 O In N1 11 O N N) CO O '0 .-1 U) 1) .D 1
0'0' 0' 0 0)') 70 el .0 .ON NI el .0.i N .0 0
• 0' . 0' • 0 •0' • 7' • 0• • 0 • ei • O' • 0' 7
.0 • 7 • CV • .-C • • CM • '+1 • • • • C
ei ei .i .i ei •

ER R OR
0'0 Na' O.0 .ON 0' 0' •r) CU •J00 .-C N) 07 O' 7 C O
0 0 N 7' r N N) N) 7• N 0' r` 'C) '0 0 N '7 N) N 3 1
?.4 .DO 0'0' U10 NO .i 0 Nel •0' 30 N O U)0
•J • O •0. • O •0' •0 •O • 0 • 0`

A f3S
In • .0 • • r • 7' • U) • • .1 • In •
.i ei el 3 N .i ei 7 ei 3
03 O3 O7' a, 0 0 O 7'N T ei 7.-i O.0 1)`C) 0
0"O r0. 3 .-I r• O el d- el .1 N N el 4- o.1 N7' 1
.i N1 ei 3 O CO el 3 ei N) N .0 1` ei 7 O N 'C) el 'C)
•0 •3 • 0 •O .0' •7' •O •O • O' •'T
•
•
I-4 ei ei el • ei el
• 0\
_U C
0 N) 0 0 7 N C 17 0 1' 0 0 CD 3 O M .O 1) 0 CM 7 0
0' N N 11 el .7 7 0 0' O 0' U) 'C 0 0 .0 3 1 •
o N) O O N 10 .i 11 3
ei .0 3 ei 3 N U
0 •7 O • 7 •0'. • T 0' 0 • T • T.

PAPER
• N) • ••• 7
ei e1 ei .i 1.4 U
J
.-1 1.• 0 N ei M 0 11 11 ei N 1) O .i N) tn 3 7 M N 0 0 U G.

Wi LL 1S
N O 1- .0 01s Is ei 3 1- •4 Ps o11 3 0' .7 .-4 O N 1 W L)
.O1) ei U) •0 O J' 3 .0'+) 0 0 NN 3 J) CD ..JD .O3
• •7 •7 • 7 •3' •1' • 7 •7 • T •0'
N) r .i • 3 • 0 • 7 • Nel • O • ei • I~ • 0 •
ei ei .i .i

00110 017
el ei ei ei ei ei CD
0 0 N"- 0 3 'a r 0.'n 3 T 0 'C) ss D Ps f) 0 O 7 7 U ei
C- O' 0 0 3 N ei 11 O' N 3 N ei .i .i 0' .D 11 3 10 N 1
'.0 N 'D N .111 71s N'O NN 0 r• 3 N.D t`0 7 •
•7' . 7' . 0' . 7' • 0 • 0 • T . 0' • 3 • .0 U
N) • 3 • U) • ' 0 • •3• '+) • N • • •

.
el ei el ei 1 ei ei ei e/ ei
3
3JD 3'.) N ei T .0 7'+1 T 0 N7' 0 N1 0 .0 7 7 J
N 0 '1/ 7 .D N 0' 7 ei 3 7' 0 O'n 1) 0 N K1 ei 0' t 5/)
N 0' = .i D 7' .-I ei 0 .1 ei 3,
3 'f) 0 3 •1l 7 ei 7 ei :n
•T • 7 • : . : • 3, • •T • 0 • 1, • ei r
• • • '
N .i

UEAD11I' L
e4 ,y ei

S T A !JUARLi
N N 0 .0 .O 1) 3' N "1 .0 0 0 U
.-1 .i N .0 ■-I Y) 7 1) 3 _a .0
3 D n -4 -i 7 D 3 N n 7' N '- .0 '- 7, .O N n -in 1
0 7' .D 7 7 '- S JD +1 7 .l T 7 .D .0 1) 7 7 ' 0.
• ' • 3 • : • 7 • 3 • 3 • ' • : • O • 7 N
3 • N • 0 • N • O • 7 • '0 • '0 • 0 • 3 • N 7'
U) sfl -4 'C n ei 7 e+ 11 -♦ 01, .0 'n ei ;ei 0
G
a. J 1)
ei
3 'n __ =~ _ - 0 ._ 1
.
7 0 0 7 7 J 0 7 0 7 0
N 10 ? Nl T .i 7 U) .0 r`
7
1
N N N ? N 3 7 7 3
T 7, O' 1' 0' 0\ 7'
 9291 LAS .010 .0 10 -.000 -1.21 34.2 0

946)) NIMU .02U .022 -.OU2 -9.04 73.7 0

9461 0 1 411) .020 .021 -.001 -7.49 72.6 0

9450 NDMS .050 .U47 .003 5.62 159.4 0

9451 NIMS .050 .048 .002 3.83 162.4 0

942U NIMG .090 .096 -.006 -6.53 323.8 0

9421 NIMG .090 .096 -.006 -6.12 322.6 0

9410 N1M1-' .199 .188 .011 5.32 636.3 0

9411 NIMP .199 .193 .006 3.19 650.6 0

9430 0 14 0 .199 .198 .001 .69 667.5 0

9431 0 1 40 .199 .193 .006 2.85 653.0 0

944 1) NIML .n02 .486 .016 3.10 1642.8 0

9441 NIML .502 .497 .005 1.03 1677.9 0

Si111M d •f Pus xa)iavd •j •g

9240 GSP .667 .670 -.003 -.52 2264.4 0
9241 GSP .667 .655 .012 1.74 2213.5 0

9250 AGV 1.058 1.052 .006 .58 3552.7 0

9251 AGV 1.058 1.073 -.015 -1.44 3624.8 0

9270 NCR 2.220 2.226 -.006 -.27 7518.3 0

9271 BCR 2.220 2.248 -.028 -1.27 7593.0 0

FIGURE nF MER1I IS 1.35

AV. ABS. ENROrt IS .007 AV. REL. ERROR IS 3.41

SLOPE OF CALIBRATION LINE IS .2961E-03 BLANK CONCENTRATION IS .009 PERCENT

ITERATED BKGO LPS 1S 30.0

GUNN
1tS1 DATA FOR PARKER - WILLIS PAPER
PARKER. JUNE, 76

CALIBRATION LINE FOR AL203

DEAD TIME IS .0000017 SECONDS

 STANDARD STU. CONC. CALC. CUNC. ABS ERROR ,REL ERROR PERCENT CORRECTED Q.P.S. MG CORRECTION

9460 NIMU .257 .265 -.O05 -3.24 29.3 0

9461 NIMU .257 .258 -.001 -.35 28.5 0
9290 DTS .29U .237 .053 18.37 26.1 0
9291 UfS .29U .225 .065 22.55 24.8 0

Computerprograms SORT, REORD,andMW f ormajor-element XRF dataprocessing

9410 N1MP 4.244 4.251 -.007 -.15 469.1 0
9411 NI4P 4.244 4.249 -.005 -.12 468.9 0
9420 NiMG 12.172 12.19U -.018 -.15 1345.3 0
9421 NIMG 12.172 12.177 -.005 -.04 1343.8 0
9270 13CR 13.678 13.586 .092 .67 1499.3 0
9271 BCR 13.678 13.690 -.012 -.09 1510.8 0
9440 NII4L 13.910 13.716 .194 1.39 1513.7 0
9441 NIML 13.910 13.734 .176 1.26 1515.7 0
9240 6SP 15.348 15.183 .165 1.08 1675.5 0
9241 GSP 15.346 15.180 .16b 1.09 1675.2 0
9430 N1MN 16.418 16.567 -.149 -.91 1828.3 0
9431 N1414 16.418 16.540 -.122 -.74 1825.3 0
9450 N1MS 17.197 17.538 -.341 -1.98 1935.5 0
9451 8143 17.197 17.429 -.232 -1.35 1923.4 0
9150 AGV 11.347 17.375 -.026 -.16 1917.4 0
925] AGv 17.347 17.331 .016 .09 1912.6 0

FLOURE Hi mLNII IS .b4

AV. ABS. ERROR 1S .U96 AV. RLL. ERROR IS 2.79

SLOPC 01 CALIUKATIUN LINL 1S .90b2L-02 BLANK CONCENTRATION 1S .245 PERCENT

ITLRATtP BKGU LPS IS 27.0

SUMMARY uF CALIUKnI IUI`I UATA

tuLMLNT SLOPE BLANK AAE ARE FUR

1102 .2961t-u5 .009 .007 3.406 1.352

 81203 .90o21-u2 .245 .U93 2.790 .838

GUNN M.ATislx FACTORS FUR 1 PLRLENT OXIUL

L12U4U7 S102 1102 AL2U3 F1203 CR203 MOO 860 CAO NA2O K20 P205 NIO SRO BAO
S1 6,394 6.670 12.803 23.034 20.392 19.412 16.403 20.867 10.517 18.577 9.260 7.264 26.833 10.725 34.120
11 .471 1.772 .928 1.607 1.496 1.427 1.198 1.455 5.706 1.296 5.799 1.985 1.978 4.976 2.849
AL 10.802 10.360 19.73h 9.118 31.380 29.844 25.235 32.362 16.203 28.803 14.249 11.269 41.233 16.209 51.368
FL .115 ,669 2.361 .609 .576 ,547 3.313 .547 2.187 .487 2.215 .751 .759 1.914 5.807
Ch .262 1.069 3.137 .914 ,908 .867 ,727 .874 3.465 .780 3.516 1.198 1.202 3.028 5.613
RN .222 .843 2.951 .766 .718 .686 .575 .689 2.740 .619 2.777 .944 .951 2.396 6.195
hRG 17.408 16.70o 31.603 14.705 50.128 47.647 40.352 12.451 25.928 46.409 22.770 18.167 65.792 25.404 69.389
CA .831 3.09/ 1.611 2.830 2.594 2.079 2.472 2.549 1.326 2.272 10.075 3.469 3.426 8.612 4.806
No 29.556 20.123 52.406 24.746 83.457 79.294 67.241 20.953 43.261 10.606 37.936 30.559 109.428 41.443 94.144
K 1.125 4.181 2.160 3.826 3.487 2.796 3.323 3.447 1.785 3.071 1.579 4.683 4.603 11.567 6.367
P 4.5hb 16.670 8.549 15.612 13.653 12.995 12.045 13.856 7.027 12.338 6.193 4.822 17.971 38.649 23.291
41 .113 .43~ 1.546 .395 2.697 2.169 .355 1.429 .317 1.447 .488 .497 1.253 3.826

Wd 'f 2I
2.702
58 .010 .074 .269 .087 .475 .378 .474 .059 .248 .053 .249 .083 .642 .218 .686
ISA .389 1.46o .76/ 1.33/ 1.241 1.183 1.060 1.202 4.734 1.072 4.80 7 1.642 1.641 4.132 2.386
SC .622 2.329 1.214 2.127 1.958 1.868 1.570 1.913 7.473 1.706 7.60 0 2.609 2.589 6.511 3.680

SITIIM d'f P ue 11T1

GL .052 .2Uu .721 .182 1.265 1.261 1.012 .163 .666 .146 .671 .225 1.701 .585 1.807

AURLV1AT IONS

Ux(J = WEIGHT PERCLIJT 0x1UL

N-13 = N0i01Ii•1AL Cowl MINUS ULANK CUNC (wT PERCENT)
MAT = MATRIX FACIUW
ELM = ELERREUT COWL ('WT PERCFNI )
N+U = l)0IlINAL cum. PLUS ULANK CONC (81 PERCENT)
BLK = BLANK CUNC Iwl PLRCEN11

GUNN
TEST DATA FOR PARKER - WILLIS PAPER

pAkKER• JUNE, 76

S102 T102 AL203 FE203 FEU CR203 MNO MG0 CAD NA20 K20 P205 NIO ELX ELY ELL H20- LOI TOTAL

5U LAVA WITH PINK SPOTS

Ox) 45.29 1.34 13.73 3.33 7.78 U .18 11.21 9.15 1.91 .66 .14 0 0 0 0 .4 5.7 100.82
N-U 44.60 1.34 13.82 11.38 0 .17 11.04 9.30 1.91 .67 .14 0 0 0 0 .4 5.7 100.46
MAT 1.015 1.003 .994 1.052 1.023 1.041 1.012 .984 1.000 .997 .995 1.192 1.124 1.010 1.000
 53 LAVA wITH PI8K SPuTS - DUPLICATE
'°N 'E"PA 030V0

UXU 45.06 1.35 13.51 3.25 7.78 0 .18 10.67 9.07 1.91 .66 .15 0 0 0 0 .4 5.7 99.68
N-U 44.54 1.35 13.64 11.35 U .17 10.52 9.26 1.91 .66 .15 0 0 0 0 .4 5.7 99.65
MAT 1.011 .990 .991 1.048 1.019 1.036 1.010 .980 1.000 .993 .991 1.186 1.118 1.006 1.000
-1-1

TAPE.62 - u - TLSTIHt.

Computerprograms SORT,REORD,and MW f ormajor-element XRF dataprocessing

TEST MUTE' our OF SAmPLE OATH
U 1318.23 223.01 9475.54 0 979.09 55.75 0 0 0 0
53 6355.04 4823.40 1615.8516467.57
1.91
53 LAVA WITH PINK SPOTS - OuPLICATE
53 27.67612 35.68384 35.65114 35.26487 .40000

NOMINAL CONC

48.152 1.436 14.612 12.089 U .191 13.735 9.817 2.015 .757 .155 0 U 0 0 102.959

BLANK CUNC

1.150 .010 .22U .110 U .010 1.660 .050 0 .060 0 0 0 0 0

OIL NUU1HRAL CONC

87.500 6.019 .179 1.827 1.511 U .024 1.717 1.2.27 .252 .095 .019 0 0 0 0 100.370

DIE CLANK CUNC

.144 .001 .028 .014 U .001 .208 .006 0 .008 0 0 0 0 0

UXU 87.500 6.005 .180 1.798 1.586 0 .024 1.410 1.209 .252 .087 .019 0 0 0 0 100.070
MAT 1.0215 1.0102 .9993 1.0586 1.0313 1.0487 1.0125 .9905 1.0000 1.0031 1.0008 1.1899 1.1214 1.0177 1.0000
MOO 1NTEHFEAHENCE COKHLCTION .121

OAt 87.500 5.955 .179 1.785 1.576 0 .024 1.409 1.200 .252 .087 .019 0 0 0 0 99.985
MAI 1.0133 1.0020 .9921 1.0521 1.0231 1.0403 1.0118 .9830 1.0000 .9961 .9938 1.1923 1.1242 1.0094 1.0000
M 0O INTEHFF HENCE CURRLCIION .120

UxD 87.500 5.948 .178 1.783 1.571 0 .024 1.408 1.197 .252 .087 .019 0 0 0 0 99.965
MAT 1.0118 .9992 .991U 1.0490 1.0202 1.0374 1.0108 .9807 1.0000 .9939 .9917 1.1882 1.1201 1.0066 1.0000
MOO INTERFE HENCE CORRECTION .120

OXU 87.500 5.944 .170 1.782 1.570 O .024 1.408 1.196 .252 .086 .019 0 0 0 0 99.959
MAT 1.0114 .9986 .9907 1.0482 1.0195 1.0367 1.0105 .9801 1.0000 .9934 .9912 1.1868 1.1188 1.0059 1.0000
MOO INTERFE. HENCE CORRECTION .120


OxU 87.500 5.945 .178 1.782 1.570 0 .024 1.407 1.196 .252 .086 .019 0 0 0 0 99.958 J
MAI 1.I111. .9984 .9906 1.0479 1.0193 1.0365 1.0104 .9800 1.0000 .9933 .9911 1.1864 1.1184 1.0058 1.0000
MG0 INTLf1FE RLNCE CUHHICI1ON .120
FINAL CURB CONCS - HECALC FUR FLUX DILUTION

47.541 1.424 14.255 12.558 U .188 11.260 9.570 2.015 .692 .153 0 0 0 0 99.662

53 LAVA :iTH RINK SHUTS - DUPLICATE

OxU 45.06 1.35 13.51 3.25 7.78 0 .18 10.67 9.07 1.91 .66 .15 0 U 0 0 .4 5.7 99.68
N-B 44.54 1.35 13.64 11.35 U .17 10.52 9.26 1.91 .66 .15 0 0 0 0 .4 5.7 99.65
MAT 1.011 .998 .991 1.048 1.019 1.036 1.010 .980 1.000 .993 .991 1.186 1.118 1.006 1.000

SImMd'f P ue x3N Vd• f •d

•
 Computer programs SORT, REORD, and MW for major-element XRF data processing 171

DISCUSSION data were produced at Imperial College by a Philips 1212 X-ray

The three programs have been kept separate as this spectrometer purchased with the aid of a N.E.R.C. research grant.
results in reduced central memory requirements (approx- The inspirations for some of the procedures used in program
SORT owe their origins to a program written by B. M. Gunn
roximately 20k, CDC 6400), and easier debugging and (University of Montreal). Finally, the first author wishes to
testing. The programs have been tested at Imperial acknowledge that part of this work was done during the tenure of
College, University of London, on a CDC 6400 running a N.E.R.C. research assistantship.
under the Kronos operating system. The programs are REFERENCES
stored in a precompiled form on the CDC 6400 permanent Abbey, S., 1973, Studies in "standard samples" of silicate rocks
file-storage system. The data input files also are stored on and minerals: Geol. Surv. Canada Paper 73-36, 25 p.
this system. The relatively small central memory needed Borley, G. D., 1972, Problems of silicate analysis by X-ray
for the programs, as well as the relatively short central fluorescence using a fusion technique and a single reference
processor time required, means that the programs can be standard: Proc. 8th X-ray analytical conference, Pye-Unicam
Ltd., Philips Analytical Department, York Street, Cambridge,
run interactively under the Kronos operating system. England, p. 18.
Rapid turn-around is possible between the XRF analysis Flanagan, F. J. 1973, 1972 values for international geochemical
followed by the input of the paper-tape data at an reference samples: Geochim. Cosmochim. Acta, v. 37, no. 5, p.
interactive terminal and the processing of the data. 1189-1200.
Gunn, B.M., 1967a, Matrix corrections for X-ray fluorescence
A prelimary study on the relative merits of the two spectrometry by digital computer: Can. Spectrosc., v. 12, p.
methods of analysis shows that the "Norrish" method 41 -46 and p. 64.
produces better calibration lines, although those produced Gunn, B. M., 1967b, The incident attenuation factor in X-ray
by the "Gunn" method are nevertheless satisfactory for emmision analysis and the determination of some heavy
elements in standard rocks: Can. spectrosc. v. 12, p. 163-168.
most applications. A detailed comparitive study is in
Norrish, K., and Hutton, J. T., 1969, An accurate X-ray
progress. spectrographic method for the analysis of a wide range of
geological samples: Geochim. et Cosmochim. Acta, v. 33, no. 4,
Acknowledgments—Thanks are extended to the staff at the p. 431-453.
Imperial College Computer Centre for their helpful advice. Parker, R. J. (in press), An iterative method for determining
Fruitful discussions were held with G. D. Barley and P. Suddaby, background intensities used in XRF calibration lines for silicate
both of the Department of Geology, Imperial College. The test rock analysis. X-ray Spectrometry.