Anderson 1936

Download as pdf or txt
Download as pdf or txt
You are on page 1of 30

AN DE-EM 35 - 0001

1 28 Cone-sheets, Ring-dykes, and Caldron-subs£dences.

IX.-The Dynamics of the Formation of Cone-sheets, Ring-


dykes, and Caldron-subsidences. By E. M. Anderson,
M.A., D.Sc., F.G.S. (With Nine Text-figures.)

(MS. received April 7, 1936. Read June 1, 1936.)

CONTENTS.
PAGE PACE
1. I ntroduction 128 5. Origin of the Magma Reservoirs 154
2. Mathematical Solutions 137 6. Summary 156
3. Application to Cone-sheets 142 7. References to Literature 157
4. Application to Ring-dykes, Stocks,
and Subsidences 150

I. INTRODUCTION.
A SYSTEM of intrusions of the type now known as cone-sheets was first
definitely recognized and mapped by Harker in the Cuillin district of
Skye, and was described by him in the Geological Survey memoir issued
in 1904. The members of the complex were simply designated" inclined
sheets," but they incline inwards to\\"ards a common centre from north,
west, \ south, and south-east. 'Vhen the Scottish Survey undertook the
mapping of ?vI ull , in the years preceding the war, two further series of
centrally inclined intrusions were soon recognized . Owing to the number
of these sheets, and consequent liability to confusion, individual members
can seldom be traced very far, or shown to have arcuate curvature. But
the systems as a whole sweep round centres in continuous curves. In one
case the cun·e was originally closed, though it has been somewhat inter-
rupted by later intrusions. In the other case the main body of sheets
only forms a horse-shoe. Nevertheless the term cone-sheets, which was
here applied to these intrusions by .Profes~or Bailey, is extremely
appropriate.
The intrusions forming the horse-shoe are the earlier in date. They
dip towards their centre at angles of about 45°, and average maybe 30 ft .
in thickness. They include two main petrological types, but the petrology
of these and the other igneous rocks here dealt with has already been
fully described, and will only be incidentally referred to in this paper.
The width of the horse-shoe is about ten miles. The cone-sheets which
belong ~o it are intersected by those of the other system, which dip some-
what more steeply towards a centre lying two or three miles farther to
,,

Idrolt-subsz'deilces. Cone-sheets, Rz'ng-dykes, and Cald?'on -subsz'dwces, 12 9

the north-west. The earlier sheets are very numerous, while the later
are so abundant as to be often contiguous, lea\'ing only. remnants of the
ianeous
b
rocks into which they were originally intruded (Bailey, 1924).
Well-developed cone-sheet systems also occur in Ardnamurchan (see
n of Cone-sheets, Ring- fig. 1), where they have been described by Richey ( 1930), and a few sheets
;. By E. M. Anderson, have been found in Rum, There is also a series round the Carlingford
xt-figures.) plutonic centre in Northern Ireland, where they had been mapped by
d June 1,1936.) Trail!, but their identity as a cone-sheet system was first recognized by
Richey. This does not complete the list, as, except for Arran, cone-sheets
seem to occl.tr round all the important British Tertiary centres.
PAGE
n of the Magma Reservoirs 154
Outside of Britain the cone-sheet principle of intrusion has hardly
lary 156 as yet been recognized. Recently, however, what appears to be a series
cnces to Literature 157 of very thick cone-sheets has been found by Bain in the Kudaru Hills of

I
Nigeria (1934) .
A point of some importance 'with regard to the mechanism of these
intrusions was noticed by Harker in Skye, 'where the inner members of
the complex were found to be dipping at steeper angles than the outer.
wn as cone-sheets was first This is also the case, at least to a limited extent, with regard to the later
'r in the Cuillin district of sheets in Mull. The rule has, in addition, been found by Richey to apply
gical Survey memoir issued to the cone-sheets connected with" Centre 2" in Ardnamurchan, where,
simply designated" inclined however, the inner and steeper sheets are later in date than the outer.
common centre from north, In sharp contrast to these centrally inclined intrusions are the more
ttish Survey undertook the or less vertical ring-dykes. Individual ring-dykes were already known,
, war, two further series of for instance, that partly encircling the Glencoe caldron-subsidence. The
lcd. Owing to the number occurrence of a numerous system of concentric ring-dykes was, however,
fusion, individual members a possibility first realized in Mull, and it was during the survey of Mull
\'e arcuate curvature. But that the term ring-dyke was introduced by Bailey and first applied
continuous curves. In one In Mull there are two centres, which are certainly connected with the same
It has been somewhat inter- underground reservoirs which gave rise to the two sets of cone-sheets.
'c the main body of sheets Round the earlier or Beinn Chaisgidle centre there is a complex system of
'rm cone-sheets, which \\'as
'ssor Bailey, is extremely

the earlier in date, They


,
,
ring-dykes, although it must be noted that in usi.ng this word one refers
more often than not to cases in which the ring is incomplete. The arcs
of a circle round which individual ring-dykes can be traced are, however,
much larger than in the case of individual cOll'e-sheets. The complex
, and average maybe 30 ft. as a whole is about four miles in diameter. Fully half of the area which
ical types, but the petrology ! it occupies is made up of ring-dykes, although there are" screens" of pre-
'alt with has already been j
existing rock between. In the case of one of the innermost members the
-(erred to in this paper. l ring is only a mile in diameter, and nearly complete.
cs. The cone-sheets \vhich The Beinn Chaisgidle centre was the satne as that of the early cone-
er system, which dip some- •t sheets. The ring-dykes were later than these, but mainly earlier than the
or three miles farther to t cone-sheets of Centre 2. The latter centre is surrounded by one nearly
t
P.R.S.E.-VOL. LVI, 1935-36, PART II. 9

~ ::;s *. u: CF
. '" .
13 0 E. M . Anderson, The Dynamics of the Formation of

complete ring-dyke, the Loch Ba


Felsite. The curve is, however, not
strictly circular, although it has
bilateral symmetry about a north-
western aXIs. This dyke IS later
than the cone-sheets of its own centre,
and in contrast to them has an out-
ward inclination which varies from
45° to 70° or 80°, as judged from
its columnar structure. It followed
a pre-existing fissure which was itself
a complete ring. This fissure was a
fault with central do\ynthrow, and
the amount of throw is estimated at
one point to be not less than 3000 ft.
The Glen More ring-dyke is
classed as the outermost member of
the suite connected with the earlier
or Beinn Chaisgidle centre. Its
form, however, suggests that it came
under the influence of both centres,
and it is of special interest also in
another respect . . At one point this
nearly vertical intrusion splits, send-
ing off a horizontal apophysis, which,
however, soon abruptly turns up-
wards, and itself becomes vertical.
"It seems fair to claim these ex-
posures as furnishing an ocular
demonstration of the simultaneous
employment of two parallel fissures
by the ascending magma of a ring-
dyke" (Bailey, 1924, p. 314).
In Ardnamurchan Richey has
shown the presence of two series of
ring-dykes (fig. 2). The earlier of
these was developed round Centre
2, which has already been noticed,
partly during the interval of time
which separated the intrusion of the
two series of cone-sheets. The later
ring-dykes are developed round
ii, •

,. tlte Formatz"oiZ of
. Cone-sheets, Rz"ng-dykes, and Caldron-subsz·dences. 13 1
ring-dyke, the Loch Ba
another centre, and are extremely perfect. They border one another
The curve is, however, not
with hardly any remnants or "screens" of the intruded rock, and the
ircular, although it has
order of intrusion was towards the centre. The rings thus became
symmetry about a north-
successively smaller in diameter, and the centre is occupied by a boss.
axis. This dyke is later
:one-sheets of Its own centre,

t
ntrast to them has an out-
lination which varies from
° or 80°, as judged from
nar structure. It followed
ting fissure v\'hich was itself
'e ring. This fissure was a
1 central down throw, and
./
t of throw is estimated at
to be not less than 3000 ft.
len More ring-dyke is
the outermost member of
connected '"ith the earlier ./
Chaisgidle centre. Its
ever, suggests that it came
influence of both centres,
of special interest also in

f' ,.,
spect. At one point this
ical intrusion splits, send-

~.-'~'.-. -~. w~d :""'~YKSSc..ntr.


rizontal apophysis, which,
. ,,:JWoa.uu; S~ti:.m..ctLf. ",,",v ...
SiimBCe1't.b-a- 1 _ I' J.
soon abruptly turns up- . ,., . Sch.t.r'..4 au.u,.." IrtJUrS~tJ. Om.J.re 2 .
~n.tr. G.
d itself becomes vertical.
fair to claim these ex-
~ Pr.
.&r~Sl.C
akrti. Se~~.
D' f' I...... .1V&lcaru.c 111.2
.. .. C!rti:ru
Vents. c.n.tre 1. * ~:::::i, 1&~n:olUl
JJ1V;~ li.rtIl _i.JJtin, R..£.rr.9 -a..y1u t:»rrcp~~tl.$
&n4ic~ Unporta.n,.t, lI1.l7'tu'lNe' IrIArYfJU.
furnishing an ocular
FIG. z.-Ring-dyke complexes and cone-sheets of Ardnamurchan.
rion of the simultaneous Reproduced from "The Geology of Ardnamurchan, etc. ," iT/em. Ceol.
1t of two parallel fissures Surv., PI. II, p . 7I.
ending magma of a ring-
The structure of this centre has influenced the ~ontours of the ground,
iley, 1924, p. 3 14).
forming a striking case of "lunar topography_" .
lna murchan Richey has
A fi'n e instance of a Tertiary ring-dyke complex has also been descnbed
presence of two series of
by Richey round Slieve Gullion in Ireland ( 1932). Here, as in t~e ~ase
(fig. 2). The earlier of
of the Loch Ba Felsite individual rings seem to have an outward 1I1chna-
developed round Centre
tion which can be as'certained at a few places to be about 70°. The
Cariingford plutonic complex consists largely of a ring of gabbr~ (eucrite),
as already been noticed,
ing the interval of time
which may be a ring-dyke, surrounding a central stock of grallite. Both
rated the intrusion of the
these intrusions are obviously developed from the centre ,\-hich has been
)f cone-sheets. The later
responsible for the formation of the cone-sheets.
are developed round

~"'i!
'., ,
'"
.. ' ':';47r 9

·I32 E. M. Anderson , The Dynamics of the Formation of

Outside of Britain the best-known ring-dyke complex is probably


that of the Pilansberg in South Africa. This has been recently described
by Shand, who, however, explains the structure as being that of a modified
laccolite (I928). Similar intrusive complexes have also been found in
New Hampshire (Kingsley, I93 I), in Utah, and elsewhere, and it seems
probable that as geological investigation proceeds their number will be
extended.
The type instance of a dissected caldron-subsidence is that of Glencoe,
though similar structures may at present be in course of formation under,
for instance, Santorin, and other examples of the wider type of volcanic
vents. In Glencoe, however, dissection has sho\yn how an annular fault
ga.ve rise to a wide crater of subsidence, and itself probably formed the
channel through which a succession of lavas were poured out. These
occupied the subsidence, and possibly overflo\\·ed . The fault dips out-
wards on the northern margin at angles of 50° to 70°, but this is balanced
by an inward dip on the southern side. It seems probable in fact that
when formed the marginal fracture was in most places nearly vertical,
and that there has been subsequent tilting.
The classical account of the Glencoe subsidence which was given by
Clough, Maufe, and Bailey ( I909) contains also another suggestion
which is probably of great importance in the tectonics of intrusion. The
Glencoe fault came to the surface, but it is possible for a ring-fracture to
form in depth without reaching the surface, but extending downwards
to some liquid substratum, and thus leaving the rock within its circuit
hanging as it were, and dependent partly on th e tensile strength of the
layers above. Under these circumstances the stresses which gave rise
to the fault-ring would be profoundly modified, and a cross-fracture might
be formed, under which the block as a whole might sink. The more or
less cylindrical space left vacant might be filled by an ascending magma,
for which, as before, the ring-fracture would act as a channel. A granite
stock might well be formed in this manner, and the hollow space it might
be supposed to fill was named in this paper a "subterranean cauldron."
The adjacent Cruachan granite, which belongs to the same period of
igneous activity as the Glencoe subsidence, was shown by the writers to
be steep-sided , while in the case of its inner core, the Starav granite, the
vertical nature of the margin had already been made out. These cases
are not, however, exceptional. The very word "stock" implies a steep-
sided intrusion. Often, however, the sides have an outward inclination,
and are not precisely vertical, as shown in the writers' diagram (reproduced
as fig. 3).
A structu re which is probably very closely allied to that of ring-dykes
tlte Formatz"on of Cone-slzeets, Rz'ng-dykes, and Caldron-subsz'dences. I33

Irke complex is probably has been brought to light by Richey's study of the relations of the different
las been recently described granites constituting the Mourne Mountains in Ireland (Richey, I9 z8 ).
as being that of a modified
have also been found in
Id elsewhere, and it seems
~ eds their number will be

sidence·is that of Glencoe ,


:ourse of formation under
,he wider type of voIcani~
i'wn how an annular fault
tself probably formed the
were poured out. These
,·ed. The fault dips out-
o 70°, but this is balanced
. ms probable in fact that FIG.3.- Diagram illustrating subaerial and subterranean cald ron·subsidences
st places nearly vertical, accompanied by volcanic and plutonic accumu la tions of igneous rock.
Reproduced from "The Geology of Ben Nevis and Glen Coe,"
ence which was given by Mem. Geol. Surv., fig. 24, p. 126.

also another suggestion


tonics of intrusion. The o 2 3 4 5 e 7 MIr.ES
ible for a ring-fracture to
t extending downwards
Ie rock \vithin its circuit
Ie tensile strength of the
stresses which gave rise
nd a cross-fracture might
ight sink. The more or
y an ascending magma,
as a channel. A. granite
he hollow space it might
subterranean cauldron."
. to the same period of
shown by the writers to
FIG. 4.-Diagram illustrating supposed mode of intrusion of granites of
, the Starav granite, the
Eastern Mourne Mountains.
made out. These cases
Reproduced from" The Structural Relations of the Mourne Granites" (J. E. Richey),
's tock" implies a steep- Quart.J our". Geol. Soc., vol. lx.."xiii, 1928, fig. 7, p. 685.
an outward inclination
TS' diagram (reproduced There are here three successive intrusions. The first two partially sur-
round the third in a manner which is more or less arcuate. The boundary
cd to that of ring-dykes surfaces of all three are in part vertical, or somewhat outwardly inclined.
134 E. M. Anderson, The Dynamics oj tlte Formation oj

'When a certain height is reached, however, these steeply dipping surfaces


bend over abruptly to a position which is nearly horizontal (fig. 4). Thus
granite No. 2 overlies as well as surrounds granite NO.3, and No. I has
the same relation to No.2. There is a suggestion also that the outer
boundary of granite No. I behaves in the same way. Thus granite No. I ,
may be regarded as having filled a subterranean caldron-subsidence.
This, however, was only the first stage of intrusion . After No. I had con-
solidated, a second caldron-subsidence must have formed within it, with
lesser radius and a lower roof. This process was again repeated, leading
to the intrusion of NO.3. The flat roof must in each case mark different
dynamical conditions from the nearly vertical boundary, in the manner
which has been explained above. In other words, it must have been a
cross-fracture.
It is well known that granite stocks are often margined by rings of
igneous rock of somewhat different composition from the interiors. The
successive formation of caldrons with diminishing radius may therefore
have been a common process, but there is no other known case in which
the difference in roof level of these caldrons can so well be seen .
The outward inclination of a granite margin is finely illustrated in the
case of the Castle Peak stock, British Columbia, which has been figured
by Daly (1912, figs. 36 to 40). The roof of the intrusion seems, however,
in this instance, to have been formed 'in a method different from cross-
fracture. This is well shown in two of the diagrams (figs. 36, 40) . In
fig. 40 the margin is seen to bend round, not abruptly, but very gradually
upwards from an angle of 75° to one of only 20°.
A further but perhaps less frequent type of central intrusion is the
formation of radial dykes. Harker believed that a system of vertical
dykes exists in Skye, radiate to, and not extending to more than about
three miles from a centre within the Cuillins. This is presumably the
same centre which supplied the cone-sheets. The evidence is not, however,
very definite. A more certain example occurs in the Spanish Peaks
district of the United States. The dykes here radiate from a stock or
stocks of diorite; a map of the relationship has been reproduced by Daly
(1933, fig. 22).
The data about the different types of central intrusion and fracture may
now be summed up as follows. Cone-sheets and ring-dykes often surround
the same centre, but in some cases only the one or only the other may be
present. vVhere both are present the ring-dykes may succeed the cone-
sheets in time, but this rule is not invariable, and sometimes there is
alternation. Cone-sheets have an inward dip, which tends to increase as
the centre of intrusion is approached. The exact inclination of ring-dykes

Ii ;S. t, F H i . ,T ''*: . 41;: C!I!' # .... '! '


of the Format£on of
Cone-sheets, Ring-dy kes, an d Caldron-subsidences. 135
these steeply dipping surfaces cannot very often be seen, but sometimes they are vertical, while in other
arly horizontal (fig. 4). Thus cases there is a distinct outward tendency. It should be noted too that
granite No. 3, and No. I has ring-dykes are as a rule much the more massive. The majority of cone-
uggestion also that the outer sheets vary perhaps between IO ft. and 50 ft. in thickness (3 to 15m.),
me way. Thus granite No. 1 and ring-dykes from 50 yards to over a mile (say 50 to 2000 m.). The
erranean caldron-subsidence. fracture bordering the Glencoe caldron-subsidence was originally vertical.
rusion . After No. I had con- The outer margins of granite stocks may, it is suggested, in most cases be
. have formed within it, with explained as subterranean ring-fractures, and these, like some ring-dykes,
5 was again repeated, leading
have an outward tendency. .
5t in each case mark different Such then is a brief resume of the structures which it is the object of
:al boundary, in the manner the p.-esent paper to explain. For further particulars reference should
words, it must have been a be made to the various private papers and Survey memoirs which have
been quoted, and more especially to Richey's paper on "Tertiary Ring
often margined by rings of Structures in Britain" (1932), of '.vhich I have made the fullest use. I
rion from the interiors. The
wish also to express my debt for personal help received from Dr Richey.
ishing radius may therefore Scotland and Northern Ireland are in the meantime the type areas for
other known case in which this class of structure, and Dr Richey's part in their investigation has been
'an so well be seen .
a very large one.
,in is finely illustrated in the f A preliminary attempt at explanation was made by myself in the
bia, which has been figured Geological Survey memoir on the Tertiary Geology of Mull, etc. (Bailey,
e intrusion seems, however, 1924, pp. I I, 12). The theory there put forward was essentially the same
nethod different from cross- as that which is now developed, except that it was non-mathematical
diagrams (figs. 36, 40). In in its treatment. It is therefore convenient to quote this attempted
'lbruptly, but very gradually explanation almost in full.
_0°.
" A dynamical theory of the formation of cone-sheets and ring-dykes
of central intrusion is the may be developed on the following lines. One may disregard the probable
I that a system of vertical
existence of a broad lava-cone and other irregularities, and suppose that,
ending to more than about
speaking very generally, the surface of the ground, at the periods at which
s. This is presumably the
these features were developed, was horizontal. Underneath this hori-
he evidence is not, however, zontal surface, at a depth of several miles, was a magma reservoir. Its
urs in the Spanish Peaks shape may have been roughly that of a paraboloid of revolution.
re radiate from a stock or "One may suppose that at first the magma had a specific gravity equal
s been reproduced by Daly
to that of the surrounding rocks, and -that it was under a pressure just
high enough to have raised it to surface level if there had been an outlet.
intrusion and fracture may
Overlooking the fact that the rocks themselves were not quite homo-
ring-dykes often surround
geneous, one may further suppose that they were under the same hori-
e or only the other may be zontal and vertical pressures at every point as would obtain in a liquid
<es may succeed the cone- with the same surface and the same specific gravity. While such was the
e, and sometimes there is case no fractures could arise.
which tends to increase as
"Cone-sheets.-If, however, these conditions were modified by an
ct inclination of ring-dykes
increase of pressure in the magma-basin, the pressure-system in the crust


136 E. 111. Anderson, The Dynamics of the Formahon of
would have superimposed on it a system of tensions, acting across surfaces
which near the basin were roughly conical. The fine firm lines in fig. I "
(fig. 5 of present paper) " are intended to show the intersection of these
surfaces with the plane of the diagram . A superimposed system of pres-
sures would also act across surfaces which cut the former orthogonally
and are indicated in section by the fine broken lines. The superimposed
tensions, together with the increased pressure of the magma, might cause
a series of fractures to develop, along which the magma would intrude.
The opening fractures would follow the fine firm lines of fig. I, and thus

Magma
Reservoir

FIG. 5.-Stress-diagram to show supposed mode of formation of cone-sheets


and ring-dykes_ For explanation of lines, see text.
Reproduced from "The Tertiary and Post-Tertiary Geology of Mull, etc. ,"
Mem. Ceol. St.rv., fig. I, p. 12.

may have originated the cone-sheets of Mull. It will be seen from the
diagram that if the surface were denuded to a certain depth the cone-
sheets exposed might be expected to be steeper in the central parts of th e
area of intrusion . This is actually the ~ase.
"Rz"ng-dykes.-If the conditions were reversed and the pressure of the
magma fell. below that which was at fi~st assumed, the original 'hydro-
static' pressure in the crust would be modified in a different way. Super-
imposed pressures would act across the surfaces whose trace is shown by
the fine firm lines of fig . I, and superimposed tensions across thos e which
are indicated by the broken lines. It seems likely that, in this case, sur-
faces of fracture would originate inclined at an angle to the surfaces
across which there were maximum superimposed tensions, as in the case
of normal faults (Anderson, 1905). The angle may have been about 20°
--- -----
..

oj the FormaNon oj
Cone-sheets, R£ng-dykes, and CaZdron-subs£dences. · 137
_nsions, acting across surfaces!
The fine firm lines in fig. I " or 30° . . Such surfaces of fracture correspond, not to tension-cracks, but,
how the intersection of these in theory at least, more nearly to planes of maximum shearing-stress.
superimposed system of pres- They deviate from the directions across which this stress is an absolute
cut the former orthogonally maximum owing to certain considerations of friction. An attempt to
.en lines. The superimposed show the trace of such surfaces, taking the angle mentioned as about
e of the magma, might cause 25°, has been made in the diagram. It can easily be seen that the theory
"l the magma would intrude.
explains the tendency to an outward slope, which is perhaps a feature of
firm lines of fig. I, and thus ring-dykes. If the fractures formed curves that were closed in cross-
section, the rock inside them might tend to become detached and to
sink down into the magma. The gap between the subsiding mass and
the stationary walls would widen with the subsidence, and this perhaps
explains, in part, the greater width of ring-dykes when compared with
cone-sheets. It is uncertain whether, in the majority of cases, the ring-
fractures continued upwards to the surface of the ground. \Vhen they
did so they must have given rise to circular depressions. Such a sub-
sidence o£ the surface is known to have happened at Glencoe, and it
appears to have happened in Mull in the case of the supposed south-
eastern caldera."
There is bound, of course, to be a tendency to tangential tension
round any com'exly bounded enclosure containing fluid under pressure.
There must also be a tendency near it to a radial pressure. If the enclosure
I contains only vacancy and there is pressure in the surrounding solid, the
conditions are more or less reversed. There will then be a tangential
f formation of cone·sheets J' pressure along the walls with vanishing normal pressure; one has only
nes, see text.
'ertiary Geology of "rulI, etc.," t to consider the example of an arch. The mathematical formulation of
g. I, p. 12. the cone-sheet problem proved, however, somewhat difficult, and it \\'ould
be practically impossible to calculate the stresses corresponding either to
f
It wiII be seen from the a parabolodial or any other arbitrarily chosen shape of the underground
a certain depth the cone- caldron . Two solutions of the equations of equilibrium have neverthe-
r in the central parts of the less been found, satisfying the surface conditions and giving displacements
which might be caused by the presence of a magma basin. It is probably
led and the pressure of the of little consequence that the different possible surfaces of pressure which
umed, the original 'hydro- will produc~ these displacements are none of them paraboloidal. In one
n a different way. Super- of the solutions the surfaces open downwards, while the directions of stress
's whose trace is shown by seem to fit the facts of central intrusion fairly well.
ensions across those which
e<ely that, in this case, sur- 2. MATHEMATICAL SOLUTIOKS.
an angle to the surfaces
Use is made, in the first place, of two well-known systems of strain
ed tensions, as in the case
which may occur in a homogeneous, isotropic, and infinite solid. The
may have been about 20°
first of these is that which surrounds a "Centre of Dilatation." In this
138 E. M. Anderson, The Dynamics of the Form ation of

the' displacement is radially outward from a given point and inversely


proportional to the square of the distance. This is expressed by
x
u=A-; (I)
r3
where tt, v, and ware the displacements parallel to x, y, and z. These
equations represent the strain which would be produced in such a solid
if a tiny sphere at the origin were to expand equally all round .
The other strain was investigated by Lord Kelvin, and was named
by Boussinesq the "premier type simple d'integrales" (1885). It is given
by a set of relations such as
XZ
u=B , (2)
r3'
where ,\ and fJ- are Lame's constants. These displacements would be
the type produced if an imaginary finger could be inserted into a homo-
geneous solid, from some fourth dimension, at the origin, and were to
exert on it a force in the direction of z. Some simple term for the stress
system appears to be rather necessary, and it is suggested that it may be
appropriately named the "point-push." (I) and (2) both satisfy the
equations of equilibrium
a a 0)
(,\ + fJ-) (ax' oy' OZ 6. + fJ-V 2(u,
. w)
V, = 0,

ou
where ~=-+-+-.
aVOw
ax oy oz
These solutions refer, however, to an infinite solid. To apply them
to the case of central intrusions one must find the effect (I) of a centre of
dilatation, and (2) of an upward point-push in the "semi-infinite" solid
",.hich is otherwise unlimited, but bounded upwards by a horizontal plane.
This plane may be denoted by z=o, z being measured downwards, while
x and yare horizontal. The solutions must have the property that near
some point 0, 0, a the other terms which they contain must become
negligible in comparison with those which denote the point dilatation or
the point-push. They must satisfy the equations of equilibrium, and in
addition the components of traction must vanish across the plane z=o.
I shall use the symbols P 1/Z> etc., to denote the tractions, with the usual
convention that P lm is the force per unit area, in the m direction, acting
on the I face of the element dx, dy, dz from the side on which I is greatest.
P .,,'" P !1U' and p .. will then be positive when· they denote tensions and
negative when they denote pressures.
To obtain the solution for the case of the point dilatation one must
---
r the Formatz"on of Cone-s heets, Ring-dykes, and Caldron-subsz'dences. I39

given point and inversely revert for the meantime to the supposition that one is dealing with an
his is expressed by . infinite solid. One must then suppose that the dilatation at 0, 0, a is
balanced by a similar dilatation at 0, 0, - a. The resulting displacements
(r) are
z-a z+a)
lleI to x, y, 'and z. These w 1 =A ( -3-+--3 '
1'1 1'2
Ie produced in such a solid
where 1'1 and 1'2 are the distances from the points 0, 0, a and 0, 0, - a
::rually all round.
respectively. It can easily be shown that where z = 0 the tractions P %:z:
-d Kelvin, and was named
and P %11 vanish, while ,
grales" (rSSS). It is given OWl 3a2)
Pu =2}.1-Tz+M=4}.1-A ;a --;s ,
(I
I
f
where l' denotes either 1'1 or 1'2'
One may endeavour next to obtain a displacement u 2, V2, W2, \vhich
e displacements would be shall satisfy the field equations, and give rise to the same values of P Z%
d be inserted into a homo- I across the plane z=o, where it likewise makes P::Z:=P'1I=0. If such a
at the origin, and were to I displacement can be found, containing no singularities where z is positive,
~ simple term for the stress I and vanishing at infinity, then by subtracting its components from ul> VI'
s suggested that it may be
and (2) both satisfy the I WI we shall arrive at a resultant strain which is a solution of the problem.
This result may be attained by using a theorem of Hertz, which may
be stated as follows (rSSr). If to the plane z=o which bounds as before
a semi-infinite solid we apply any general pressure p =f(x, y) without
, w)=o, associated shearing stress, then if P be the electric potential of a density

L of electricity at every point on the plane, the displacements which are


47T

caused by the pressure may be derived from this potential. The quantity
ite solid. To apply them II is first defined as follows;-
the effect (r) of a centre of
1 the "semi-infinite" solid
zp
Il=--+-- r
}.I- A+}.I-
[J"" Pdz-J]
z '
ards by a horizontal plane. where J is a constant so chosen as to make II finite. We then have
easured downwards, while
oIl oIl _ oIl 2(A + 2}.1-)p
ave the property that near Uz = ox; Vz = oy ; W 2- oz + ,>_ + }.I- .
ley contain must become
lOte the point dilatation or One has next to consider what is the potential of a distribution
,ms of equilibrium, and in - }.I-A(~ _ 3~2) of electricity, over an infinite plane, where r is the distance
7T l' l'
ish across the plane z=o.
, tractions, with the usual from a point whose perpendicular distance from the plane is a. This
in the m direction, acting can be found by the method of electric images. 'Where z is positive the
ide on which I is greatest. potential will in fact be that of an electric dipole, of direction 02 and
they denote tensions and strength ,2}.1-A, situated at 0, 0, - a. Its value is thus
2}.1-A(z+a).
·1 3
point dilatation one must 1'2
,J . ' '.

140 E. M . Anderson, The Dynatnz"cs of tlte Formatz"on of

We have therefore
~.

from which we obtain


U _--L~}
=2A{3 XZ (Z +a)
2 A+ fL 1"23 '
1"2
5

v =2A{3yz(z+a) _~L}
2 1"2
5
A+ fL 1"23 '
W = 2A{3 (Z + a)2 + --L .!.. + A + 2fL .!!....}.
Z
5 3 3
, 2 1"2 A + fL 1"2 A + fL 1"2

Hence the displacements, which are a solution of the problem, may be


obtained as follows;-
__
U - U1 U2 -
-A{~3 + A+3fL
1"1
, ~3 _ 6xz(z+a)}
1\
5'
+ fL 1"2 1"2

v
=
VI
_
V2 -
-A{z.... + A+3fL
,
3
L 3
_ 6yz(z+a)} 5'
1"1 1\ + fL 1"2 1"2

W=W _ ow =A{Z -a + A-fL ~ _ A +3fL .!!.- _ 6z(z+a)2}


1 2 1'13 A+ fL 1"23 A+ fL 1"23 1"25 .

It can be independently verified that these expressions satisfy the equations


of equilibrium and the boundary conditions.
One may next investigate the case of the upwardly directed point·push,
which will be taken as before to be situated at the point 0, 0, a . In an
infinite solid its displacements ,\'ould be given by
Bx(z -a)
u=----·
1"13 '

W= _B{(Z-:)2 + A+3fL
A+fL
~}.
1"1 1"1

This stress may be supposed to be balanced by an equal but downward


point push, situated at 0, 0, - a. vVe thus obtain

VI =etc.;

These displacements as before cause p.", and P' II to vanish along the
plane Z =0, while the pressure p across this plane (p = - P zz) has the value
4fL2 a a3 }
B{ A+ fL ;3 + 12fL;S .
Applying the same method as before it is necessary to find what is the
.
- ........... """ .............~
---'....."""'. """-'.:.;..:..--"'-"""'-'

oj the Formation oj Cone-sheets, Rz'ng-dykes, and Caldron-subsz'dences.

potential of a surface density of electricity on z =0, equal to the above


quantity divided by 41T. U sing the theory of electric images it may be
shown that where z is positive this potential is equal to that of a charge
2 BfL\(>' + 2fL) together with a dipole of strength 2BfLa in the direction OZ,
l l\+fL
J' acting together at the point 0, 0, - a. Its value is therefore
fL(A + 2fL)":' fLa(z + a)}
2B{ \ + 3 •
l\+fL 72 72
. A+2fL~}
A+fL ti . Using the formulce (3) and (4) the values of U2, V2, and W2 appropriate
to the case of the point-push may now be calculated . We obtain in the
tion of the problem, may be
first place
II = 2B{~ !!. - >. + 2fL !.. _ fL(A + 2fL) (r. + z + a) _ az(z + a)} 100'.,
~}
3
A+fL 72 A+fL 72 (>'+fL)2 - 72 '
r 5 ,
2 from which it follows that
:;+ a)}
~
2
5 , u
2
=2B{ _~
>'+fL
ax + >.+ 2fL xz _fL(>'+ 2fL)
>'+fL 72
3
(>'+fL)2 7 23 7 2(7 2
X
+z+a)
+3 axz(Z +a)}
7 25 ,

v2=etc.,

ressions satisfy the equations W2= 2 B{(A(>'+fL)


+ 2fL)2
2
-
I
1'2
+
(>. + 2fL)Z2 + (>. + 3fL)az + (>. + 2fL)a 2 3az(Z + a)2}
(>'+fL)7 23
+ "'----'-~'-
5
72 '

We therefore have
pwardly directed point-push,
at the point 0, 0, a. In an U = 11 _ U = B{_ax
_ -_x_z + >'_+_3_fL a_x_'_- _x_·'z + 2fL(>. + 2fL) x _ 6axz(z + a)}
'n by
1 2 7 3
1
>'+fL 7 3
2
(A+fL)2 2 2+z+a) - - ' 72'-;-5-":""
1' (7

V=V 1 -v2=etc.,
By(z - a)
7 13 W=W -W =B{_>'+3fL ..:._>.2+4>.fL+5fL2 ..:._(a-z)2
3
1 2 >'+fL 71 (>'+fL}2 72 1'1
1
J' (A + 3fL)Z2 + 4fLaz + (>. + 3fL)a 2 _ 6az(z + a)2}
(A + fL)r 23 1'25 '
by an equal but downward
btain As before, it may be independently verified that these expressions satisfy
the equations of equilibrium and the boundary conditions.
=etc. ; It has next to be shown how these two sets oJ results may be applied
to the cone-sheet and ring-dyke problem. In the first place it may be
-=- + A+ 3fL .:...} noted that the displacements, and therefore the stresses they correspond
71 >'+fL 72 •
to, have radial symmetry about a vertical axis. It fo11o'ws that the stresses
d P Z1/ to vanish along the may be represented as regards direction by the const~uction of isostatic
ane (p = - P .Z) has the value

}. , surfaces, which are everywhere perpendicular to the principal directions


of pressure and tension. (The most general types of stress do not of course
possess this property.) The term "isostatic surface" is in the present

, cessary to find what is the i connection a rather unfortunate one. It has nothing to do with isostasy
as understood in geology.
t
l
:

o•
_--..J... ..... . _ _ ~.--...:

142 E. M. Anderson , The Dynamz'cs of the Formatz'on of

One set of isostatic surfaces will be in each case vertical and radial
to the axis . . The other two systems may be determined by fixing
their intersections with the plane of x and z. vVe have, in fact,

tan 28=(:; + ~~) l(:: -~;). where 8 is tan- 1 ~:. This equation leads
to two sets of curves which are orthogonal. The pressures or tensions
acros~ the corresponding surfaces are given by the formula

p = Au + CA + fL)(OU + OW) ± fLJ(ozt + OW)2 + (OU _OW)2


nn X ox oz oz ox ox OZ
which applies only to the plane XOZ. To obtain the pressure or tension

across a surface along whose intersection with this plane ~; is positive,


that value of the root must be selected whose sign is contrary to that of

(~: + ~~). The other value refers to the surface which is at right angles
to the first, and also to XOZ. Across the plane XOZ itself the tension is
given by

These formulce, ' and those for the displacements themselves, may be
simplified by assuming A= fL. This corresponds to a value of Poisson's
ratio a which is accurately i , and the assumption is on the whole justified
both by experimental evidence and the facts of seismology. In drawing
up figs. 6 to 9, and the table of principal pressures which is given below,
this simplification has been used.
Figs. 6, 7 show the directions of the displacements for the point-
dilatation system and the point-push system respectively. The amounts
of displac~ment are shown approximately by the lengths of the arrows.
Figs. 8, 9 show the corresponding networks formed by the isostatic
surfaces. The plane of the diagrams is in all cases XOZ.

3. ApPLICATION TO CONE-SHEETS.

The application of these results to the problem of central intrusions


is no\v fairly simple, and I shall deal first with the case of cone-sheets.
The broken lines in figs. 8, 9 show isostatic surfaces across which there is
a pressure, if, that is, the constants A and B are taken to be positive.
One may select any such surface and suppose that the part of the solid
within or below it is replaced by a molten magma. If this magma exerts
the same pressure on the enclosing or superincumbent rock as the substance
replaced, the stresses in the remaining solid will be unaltered.
' ,1>

oj the Forma#on oj Cone-sheets, Ring-dykes, and Caldron-subsidences . 143

each case vertical and radial One may, for instance, select that surface in each of the two cases
lay be determined by fixing which is distinguished by a shaded margin. It will be noticed that in
and z. We have, in fact, fig. 8, which shows the type of solution arising from a point dilatation,
one is dealing with a closed nearly spherical cayity. It must be supposed
;m- 1 ~:. This equation leads
11. The pressures or tensions
by the formula
:! + aw)2 + (aU_aW)2
:r ax ax az
obtain the pressure or tension

n.t h thoIS pane


1 dz
dx is positive,
'se sign is contrary to that of
FIG. 6.-Point dilatation beneath horizontal surface. Dotted lines show
u face which is at right angles direction, arrows proportionate amounts of displacement.

lane XOZ itself the tension is that the magma is introduced into this cavity by a pipe of inappreciable
width. It must be under such pressure as to cause it to push outwards
, aw)
oz . in all directions. These conditions seem unlikely to be realized in fact,
and it is fig. 9, giving the solution corresponding to an upward point push,
cements themselves, may be which is put forward as the more probable approximation to reality.
onds to a value of Poisson's
,
tion is on the whole justified
of seismology. In drawing
It
\
ot
\
, '\
\
i
t
\
i t t r t r
\ T T r l' r /r

t \

\ t t t i, if tr f,
ssures which is given below, t \
i

t i t i i f J
\
t
.lisplacements for the point- ! T
respectively. The amounts tI t ;
t
1 \ i t t 1-,
t l' 1
i

r t tr t t 1i +t
I

t
y the lengths of the arro\vs.
ks formed by the isostatic
I cases XOZ .
t
i
t
I
I

+
t r
t t
1 \
~ t
t i
~ t t
ti ,t
~
it

-SHEETS . FIG. 7.-Upward point·push beneath horizontal surface. Dotted lines show
direction, small arrows proportionate amounts of displacement.
roblem of central intrusions
:ith the case of cone-sheets. The thin firm lines in figs. 8, 9 indicate isostatic surfaces across which
' rfaces across which there is there is a tension, and along which one may therefore expect the formation
I3 are taken to be positive. of cone-sheets. In accordance with fact, these increase in steepness
e that the part of the solid towards the centre. 'Within a certain radius the sheets should also
rna. If this magma exerts theoretically steepen upwards, and approach the surface in a dyke-like
Imbent rock as the substance manner (see, however, p. 149). It is unknown whether any of them
ill be unaltered. actually reached it, as the superficial layers have, of course, been removed

I
f
f

I
144 E. M. Anderson, The Dynamics oj the For'm ation oj

by erosion. Beyond the radius they should flatten upwards, and become
sill-like. This is perhaps the explanation of certain horizontal intrusions,
%
o ----~--~~~--~--.-~T__,rr----

0';;-

/ ·s_
%.
FIG. S.-Isostatic surfaces produced by point dilatation beneath free horizontal
surface. Further explanation, see text.

(·0 / ."). I'~ a.

,:,
...............
.
, ,
,

,
, ,\,
.. ... ........ ........ . , '!
'\
,
, . '\.: " ..... : . . ... ,

,:
:-
,'"
"

,
........ --; .. : ..... .
\ . \ :
\ \.
'\

"4 ":'- , ... ': ...... ,

ycf;. .......:........ ..
FIG. 9.-Isostatic surfaces produced by vertical point-push beneath free
horizontal surface. Further explanation, see text.

of typical cone-sheet type, which I have mapped, on the outer fringe of


the area invaded from Centre I in Mull. The change in behaviour of the
outer members ,vas anticipated in the diagram which was published in
the Mull memoir (fig. 5).
..

••
~
oj the Forma#on of
Cone-sheets, Ring-dykes, and Caldr on -subsidences. 145
flatten upwards, and become
Figs. 8, 9 also show proportionally at certain points the amounts of
f certain horizontal intrusions,
magma pressure which are implied in the two solutions. In fig. 9 these
~---r-i---
% diminish rapidly in depth, and farther down it is evident that they would
decrease asymptotically to zero. This is importa nt as showing that the
shape of the caldron, beyond a certain point, is probably inessential.
The same thing might be deduced from the likeness of the two sets of
results. All that is necessary for a cone-sheet tension seems, in fact, to
be an upward pressure on some type of dome.
In speaking of the pressures and tensions which result from the for-
mulre, it has so far been assumed that there is a w eightless medium. In
reality, however, the medium is not weightless, an d the matter is further
complicated by the possibility that there may have b een pre-existing lateral
pressures or tensions. The simplest possible supposition is that which
atation beneath free horizontal was made in the Mull memoir, namely, that the density of the intruded
n, see text.
rock is uniform, and that before the period of intrusion there was every-
where a "hydrostatic" pressure, equivalent to the s uperincumbent weight.
This extra pressure has no effect on the directions of the isostatic surfaces,
ou ow
,:, .. ... .. ... ... . . I . . . OU ow
as It leads to no alteratIOn In the values of OZ + ox or ox - oz' It has,
: ,, : however, to be added to the pressures which are indicated in the magma
.. ..... .. . -.- -,... -... - basins, and combined with the stresses correspond ing to the chosen valu e
"\ . ~
\ "\ . of A or B. It will be instructive to enquire \\-hether the approximate
"\ :
.: .)... .... . .. .\.: .... ... .- -. _ .. . amount of B can be calculated for a particular case, assuming these simple
: \ ,:
': \ ~ conditions, and the validity of the solution which corresponds to the
\ \
. ~ \... . .
:',
\ . .. ..: ... "'~.~- .. ... point-push .
The value arrived at \\-ill depend altogether on the nature of another
\;
:\
\
\
\
, :
\.
assumption. It is sometimes supposed that dyke formation, for instance,
can only occur under actual tension. This need in any case only be small,
as the tensile strength of a rock is always much less than its compressive
strength. It might be taken as from 30 to 50 bars, which are the limits
given for granite in the I nternahonal Cn'hcal Tables (vol. ii, New York,
1927). This is, however, not the only possible hypothesis, as when
! molten fluid is intruded into a crack under pressure it must exert a wedging
action. It is possible that tension only exists a cross the edges of such
al point· push beneath free I cracks, and that in this way dykes may propagate themselves perpendi-
nation, see text.
I,
;
cularly to what are not actually directions of tension, but only of minimum
pressure. The necessary condition might be that the magma itself was
apped, on the outer fringe of
he change in behaviour of the at a pressure a certain amount greater than this minimum. This has to
'ram which was published in be taken account of in estimating the constant, and it will be found that
as a matter of fact the cone-sheet theory will itself afford a criterion which
tt will help to decide between the two hypotheses.
P.R.S.E .-VOL. LVI, 1935-36, PART u.
, 10
146 E. M. Anderson, The Dynamics of the Fonna#on of

To evaluate B one may use in the first place a datum given , by Richey
(1932 b, p. 129), that the cone-sheets round Centre 2 in Ardnamurchan
would converge, if they continued downwards \yith their surface inclina-
tion, at a depth of about 3 miles. The fine firm lines in fig. 9 radiate
roughly from a point a little below 0, 0, a, in the lower part of the diagram.
The latter point may therefore at present be less than 3 miles in depth,
but allowing for denudation a may well be taken as approximately 3 miles
or S km. (S x 105 cm.). The subterranean reservoirs were situated well
below the surface basaltic lavas, and probably in some type of crystalline
metatnorphic rock. The density will be assumed to be 2'7, which is
~lightly above the specific gravity of granite.
It will now be of advantage to compare fig. 9 with the table below,
which gives the principal pressures and tensions due to an upward point-
push at a given depth, at different points in the surrounding rock. These
are equivalent to the stresses caused by the upward and outward pressure
of the assumed magma: They refer, however, to a weightless medium,
and do not take into account any" hydrostatic" pressure. The figures
shown are. to be multiplied by B~,
a p. being taken to be equal to.t a IS

in the present case 5 x 105 cm., and B the constant to be evaluated.


From an inspection of fig. 9 it will be seen that, if cone-sheets are
to be formed, as they have been in the Ardnamurchan instance, at angles
of less than 45°, the process of intrusion must have spread outward to at
least a certain distance from the central axis of the diagram. The present
ground-level may be at a depth of from 1 to 2 km. belo\\O the original
surface, that is from '2a to '4a. On the level 'Sa intrusion must have
taken place at least as far as '7a from the axis. From the table on p. 147
it will be seen that at the point '7a, 'Sa the factor by which B~
a
has to be
multiplied to obtain the cone-sheet tension is 4'37.
We shall now make the assumption, which mayor may not be justified,
that this tension was large enough at least to overcome the" hydrostatic"
pressure, which existed independently of the system of intrusion. At
2'S km, in depth this pressure would be about 6·6 x 108 dyne/ cm. 2 , or
Bp.
660 bars. We thus get 4'37 - 2 =6,6 x 108 , or Bp.=Y78 x 10 19 dynes.
a
The upward force of the point-push which gives the displacements which
have been assumed is expressed by I 27T B,u. or I '42 x 1021 dynes. This is
equal to the weight of 1450 cu. km. of water, or 480 cu. km. of basalt,
and is only slightly greater than the total superimposed upward force,
exerted by the magma on the walls of the caldron,

" IJ
oj the Form.at£01i oj
Cone-sheets, Ring-dykes, and Caldron-subsidences. 147
ace a datum given, by. Richey TABLE OF PRINCIPAL TENSIOXS AND PRESSURES,
Centre 2 in Ardnamurchan
tls with their surface incIina-
o 'Ia '3 a 'sa '9a -+ x
12'50 12'07 9'13 5'3 2 0'3 8
e firm lines in fig, 9 radiate o 12'5 0 12'27 10'62 8'28 4'37
he 100yer part of the diagram, 0'00 0'00 0'00 0'00 0'00

, less than 3 miles in depth, ,S 7'3 0 4'78


'en as approximately 3 miles 'Ia 7,86 6'27
- 0'94 - 1'00
eservoirs were situated well
ly in some type of crystalline 5'94 6'06 2'74
,sumed to be 2'7, which is 5'94 5'16 2'54
7'08 - 5,89 - 1'76

fig, 9 with the table below f 'sa


6' 14
6'14
6'20 3'24
1,8 I
ns due to an upward point~
5'9 2
-21'93 - 20'66 -2'3 0
e surrounding rock. These
12'3 8 I I '30 7,66 5,64 4'33
ward and outward pressure 12'38 10'74 5'14 2,64 1,68
cr, to a weightless medium , -72'78 - 62'02 - 25'43 -9,60 -4'26
ttic" pressure, The figures 9,60 5,80 4'41
3'9 2 1'5 0 1'00
aken to be equal to~, a is - 23'97 -6'14 - 2'34
, tant to be evaluated, 22'40 9'40 5'34
een that, if cone-sheets are l ' la - 2'14 0'18 0'5 0
~ - 4'34 - 1' 14 -0'47
murchan instance, at angles
z
have spread outward to at
f the diagram, The present Figures when multiplied by ~~t denote principal tensions and pressures caused under
o 2 km, below the original free horizontal surface (z =o) by p oint. push of amount 121TB/t at point 0, 0, a, upward
el 'Sa intr usion must have if B is positive, A positive product denotes tension, a negative pressure, First figure
, From the table on p , 147 in each square refers to traction across direction of cone·sheets (see text and fig, 9),
second across plane of table (y= o), third across surface perpendicular to two former,
·tor by which BtL has to be
a2 B itself has the dimensions of an area; if j.L is assumed to be 2'7 x lOll
dyne/cm,2 B is about 1'4 x ro S cm,2. The best way to gain some idea of
ay or may not be justified, the magnitude of the forces involved is, however, as follows, The pressure
vercome the" hydrostatic"
BfL
system of intrusion. At at the summit of the caldron shown in fig, 9 WQuid be 2 I '93- or about
a2
ut 6'6 x lOs dyne/cm;2, or 3300 bars, equal to that of a column of 34 km, of water. One has therefore
or BtL ==3'78 x 10 19 dynes. to consider the mechanism by which such a pressure may have been
produced,
cs the displacements which
The most obvious cause of upward pressure is a defect in density of
I '42 x 10 21 dynes , . This is
the magma, in comparison to the surrounding rock , Liquid basalt at
or 480 cu, km. of basalt
the temperature of fusion was found by Day, Sosman, and Hostetter to
perimposed upward force',
rOn, have a density of about 2'59 (1914), and if the magma, through a process
of differentiation, was more acid than basalt, it might be considerably

" ·A. e F

. ",

148 E. M. Anderson, The Dynamics of the Format£on of

lighter. One supposes the liquid to have been at the "hydrostatic"


pressure corresponding to its depth at some point much farther down in
its reservoir than the levels with which we are dealing. If then it was
less dense than its walls, the pressure would decrease less quickly upwards.
than it would in the main body of the rock. There would thus come to
be an excess of pressure in the magma, increasing towards the surface.
The effect is analogous to that which is noticed in the distribution of
coal gas, which, as is well known, has a higher effective pressure in the
upper-levels of a town. In the case of cone-sheets, however, there is a
limit to the excess of pressure, imposed by the possible variations in
density, and the size of the caldro~s. Even if the density of the liquid
were as low as 2'4, to produce 3300 bars the caldron would need to have
been 112 km. in depth.
Such a depth is extremely improbable, and the upward excess of
pressure was certainly only a small fraction of the amount which has
been deduced. The same must apply to the cone-sheet tension. The
discrepancy would hardly be removed, even if a difference in shape of the
caldron caused the relative intensity of the latter to be increased. The
assumption on which the figures have been calculated is therefore probably
untrue. It is unlikely that any actual tension can have preceded the
formation of the cone-sheets.
This result is perhaps of some importance in the general mechanics of
intrusion, but it need not have been unexpected. \Vhile dykes may
conceivably in some cases be produced under tension, the same cannot
apply to horizontal sills. The latter must, in general, split apart the
strata, in spite of a vertical pressure corresponding to the weight of the
overlying rock.
It appears, therefore, that there was no actual tension, but only a
diminution of pressure, across the directions which have been followed
by the cone-sheets. These took the path of least resistance; that is, they
opened fissures along planes across >yhich the transverse pre~sure was
least. In the circumstances the value of B becomes difficult to estimate.
The upward pressure in the caldron must still have been considerable,
though much less than that which was calculated at first.
A further consequence of the particular solution corresponding to the
point-push remains to be discussed. If the cone-sheet tensions on any
horizontal level are examined in the table they are seen always to be
large, and sometimes to reach their maximum, where x =0; that is, directly
above the centre of the caldron. This does not, however, correspond with
reality. The sheets of any particular system occupy a belt one or two miles
in width. The central area is often obscured by later ring-dykes, which
s of the Formation of Cone-sheets, Ring-dykes, and Caldron-subsidences. 149

ve been at the "hydrostatic" make observation difficult, but the sheets are never present there so far
ne point much farther down in as can be judged. This suggests that the actual outline of the caldron is
Ie are dealing. If then it was a little different from that which corresponds to the point-push. If the
d decrease less quickly upwards real shape were something intermediate between that of fig. 9 and an
k. There would thus come to inverted flowe~-pot, the concentration of the cone-sheets could perhaps
creasing towards the surface be explained. Under increased magma pressure the sharp edge might
noticed in the distribution of cause a local intensification of the tension. This would die out rapidly
ligher effective pressure in the upwards, but might nevertheless effect the localization of the intrusions
,m e-sheets, however, there is a in the manner which has been observed.
by the possible variations in The second figure which is given for each point in the table refers to
en if the density of the liquid the tangential tension, which, if it had been greater than the cone-sheet
e caldron would need to have tension, would have led to the formation of radial dykes. In this respect
the solution which has been arrived at corresponds to what, in most cases,
e, and the upward excess of is the reality, as the cone-sheet tension is almost everywhere the greater.
on of the amount which has It has to be noted, however, that the rule is reversed at the surface and at
the level' la, which in the example chosen would have been at a depth of
the cone-sheet tension. The
if a difference in shape of the
I half a kilometre. The rocks to beneath this level have undoubtedly been
. latter to be increased. The I removed by erosion, so that it is impossible to say whether this is a real
al~ulated is therefore probably t effect or due only to an imperfect solution of the problem . A slight change
nSlOn can have preceded the in the shape of the caldron might in any case have caused a more general
reversal, and in some such manner the Spanish Peaks exa mple of radial
'e in the general mechanics of dyke formation may be explained.
~xpected: While dykes may The different inclination of the cone-sheets of different systems was
er tenslOn, the same cannot referred to in the introductory section, but has still to be discussed. Thus
" in general, split apart the the sheets of Centre 2 in Mull are, partly at least, steeper than those which
~ onding to the weight of the surround' Centre I. Round Centre 2 in Ardnamurchan Richey gives the
average angle of the earlier sheets as from 35 ° to 40°, and that of the later
actual tension, but only a sheets as from 65° to 70°. This may perhaps be explained as follows.
s which have been follo\ved In treating the two solutions it has so far been assumed that a "hydro-
least resistance; that is, they static" pressure only is superimposed. If in addition there is a relative
the transverse pre~sure was regional tension, say in all horizontal directions, the results will be some-
becomes difficult to estimate. what modified. The directions of principal pressure in the plane XOZ
still have been considerable are given by the formula, which has been already quoted
lated at first. '
olution corresponding to the tan 2(} = ( OU OW)/(OU
OZ + ox ox - OW)
OZ •
. cone-sheet tensions on any
The tension would leave unaltered the values of such differentials as
they are seen always to be
OU ow . . OU d OV . . h -ow T he
d d'lmlnlS
where x =0; that is, directly OZ and ox' but It would Increase - an - an
ot, however, correspond with
ox oy OZ
numerator of the fraction would thus be unaltered; the denominator
ccupya belt one or two miles
would either be increased o'r given a smaller negative value. It follows
j by later ring-dykes, which

I
- .. . . ." , .
ISO E. ·M. Anderson, The Dynamz'cs of the Form ation of

that the surfaces of pressure perpendicular to the cone-sheets would be


flattened, ",hile the cone-sheets themselves would be steepened. This
would affect somewhat the shapes of the underground caldrons vvhich
correspond to the two simple solutions. If one adhered to the old shapes
the stresses would become more difficult to calculate, but at a distance
from the caldrons they would certainly be altered in the way which has
been indicated.
It is thus suggested that the latest cone-sheet formation, both in Mull
and Ardnamurchan, took place under a condition of regional tension,
which was less marked, or perhaps absent, during an earlier period. The
chronology of the two cases may be the same. The fact that an epoch
of ring-dyke formation intervened in both districts between the earlier and
later sheets may also possibly have some significance.
The earlier sheets in Mull form a horse-shoe which, if completed,
would be roughly circular. The area invaded by the later sheets is, on
the other hand, somewhat elongated. It measures about 8 miles by 6,
and the longest direction is from south-east to north-west. This suggests
a relative regional tension which was greatest in amount from south-west
to north-east. This explanation has already been put forward by Richey,
and such a tension is known to have occurred both before and after the
entire period of surface vulcanicity and central intrusion, giving rise at
both epochs to north-western dykes. These dykes form the Mull and
Skye "swarms" among others, and extend for long distances to north-
west and south-east.

4. ApPLICATION TO RING-DYKES, STOCKS, AND SUBSIDE NCES.


Every text-book on the strength of materials recognizes the existence
of two distinct types of fracture. There is, in the first place, shearing
fracture, oblique to the direction of maximum tension, or least pressure,
and forming angles with this direction of from 45° to 70°. There is,
secondly, what may be termed tensile fra.cture, at right angles to a tension.
A mainly tensile stress does not always, however, give rise to this variety,
and so far extremely little is known about the circumstances which cause
sometimes the one type and sometimes the other to be produced.
It may be regarded as certain that both types occur in the earth's
crust, and have to be taken account of in geology. Faults are certainly
shear fractures, while ordinary dykes are always assumed to be tensile,
on account of their direction, which, except in the case of subsequent
tilting, is seldom very far from the vertical. There can be no doubt that
this assumption is right, though subject to the proviso which has been
4 the Format£on of
- . Cone-sheets, Rz"ng-dykes, and Caldron-subsz"dences . I 5I
to the cone-sheets would be
mentioned in the last section. There are, in fact, cases ,,,·here inclined
\wuld be steepened. This
normal faults and vertical dykes have almost certainly arisen under the
tnderground caldrons which
same horizontal tension (Anderson, 1905).
m e adhered to the old shapes
The subject is important from the present standpoint, as in explaining
calculate, but at a distance
central intrusions it is necessary to invoke both types of rupture, the
Ifered in the way which has
tensile type to produce cone-sheets, and the shearing type to account for
the production of ring-dykes, by first giving rise to nearly vertical ring-
heet formation, both in Mull
fractures. If there had been tangential movement along a cone-sheet
ndition of regional tension,
fracture, it seems necessarily to follow that this must have been an upward
ring an earlier period. The
and not a downward movement of its hanging wall-that is, of the side
e. The fact that an epoch
which is next the centre. Otherwise it is obvious that there would have
ricts between the earlier and
been "jamming," especially if one considers the great number of these
ificance.
fractures. This fact rules out one set, and limits consideration to the
'-shoe which, if completed,
steeper of the two sets of possible shearing planes which might have
'd by the later sheets is, on
resulted from the pressures and tensions considered in the last section.
' asures about 8 miles by 6,
But if these steeper shearing planes had been selected by the cone-sheets,
north-west. This suggests
the latter could not have been so little inclined as they sometimes are.
in amount from south-west
This is true at least with regard to the point-push solution. It is therefore
een put forward by Richey,
fairly certain that the cone-sheets follo-wed the actual planes across which
both before and after the
there was minimum transverse pressure.
al intrusion, giving rise at
In the case of ring-dykes the matter is different. To explain these at
dykes form the Mull and
all one must suppose, as has been already mentioned, that there was a
or long distances to north-
defect instead of an excess of pressure in the magma caldron. This
can be mathematically expressed by supposing that A or B is negative.
The centre of dilatation is replaced by a centre of contraction, or the
5, AND SUB5IDEKCE5. upward point-push by a down-w ard one. The fine firm lines of figs. 8,
9 then denote surfaces of pressure, while the broken lines correspond to
.Is recognizes the existence
surfaces across which there is a relative tension. Their directions will not
in the first place, shearing
be altered, and with change of sign the table of principal pressures can be
tension, or least pressure,
applied to the case of the dO\\'nward point-push. vVe may select the same
m 45° to 70°. There is,
shaded boundaries for the magma caldrons, and the defects of pressure
t right angles to a tension.
will vary along these like the excesses of pressure.
er, give rise to this variety,
It will be seen at once that the broken lines are in general too little
circumstances which cause
inclined to represent ring-dykes. These must therefore, if they are to be
er to be produced.
explained at all, have been produced in the first place by shear fracture.
ypes OCCur in the earth's
In rocks as a whole such fractures are inclined at an average of about
ogy. Faults are certainly
65° to the directions of tension or minimum pressure. This is illustrated
lys assumed to be tensile
by normal faults, produced when the minimum pressure was horizontal,
n the case of subsequen~
and usually dipping at about this angle. There are thus in the present
here can be no doubt that
instance two possible sets of shearing planes at each point, one of which
: proviso which has been
is every\vhere inclined about 25° more than the broken lines, and in an

I
-:=) ,(
, '",

152' E. M. Anderson, The Dynamics of the For-mation of

outward direction. This is obviously the set which we should select.


The thick firm lines of figs. 8, 9 have been drawn at the appropriate
angles, and in their lower parts they may well form a representation of
ring-dykes.
These lines have an inward curvature, flattening considerably upwards,
and such a flattening has not been observed, at least with regard to ring-
dykes. The inclination of these is, however, somewhat difficult to measure.
An individual dyke can seldom be followed as much as 'one kilometre in
height, and the parts actually seen were originally a fair distance below
the surface of the ground. What may be an example of such a deflection
is shown by the boundary of the Castle Peak granite stock (Daly, 1912,
fig. 40). This is at least suggested by the curve which Daly has figured.
The first stage in the production of a ring-dyke was always, one may
suppose, the formation of an arcuate fault, with dO\\'nthrow towards the
centre. The further process need not, however, in every case have been
identical. If the arc was incomplete the downthrow must have been a
relatively small one, and the segments may have been separated at first
only by a fault breccia. Into this fault breccia the intruding magma may
have forced its \\'ay, without causing much further subsidence. Such a
sequence of events is suggested by the behaviour of the apophysis given
off from the Glen More ring-dyke. It is unlikely, hO\\'ever, that any of
the wider dykes can be explained in this manner. In these cases it seems
necessary to suppose that, even if the dyke itself does not form a closed
curve, the fault must have done so. The included block must also have
been free to sink. It may have extended to the surface. It may alterna-
tively have tapered to a point, or have been separated from the rocks
above by the formation of a cross-fracture. If the original fault had an
outward inclination a gap would be formed, on one or all sides, during the
subsidence, and it is by the filling of such gaps, it is suggested, that most
ring-dykes have been produced.
In cases like that of the dykes surrounding the Beinn Chaisgidle
centre, in Mull, a separate fault may have been responsible for each ring-
dyke. These intrusions here sometimes adjoin, but often they are separ-
ated by "screens" of the intruded rock. With regard to some of the
Ardnamurchan ring-dykes, however, such a conclusion is improbable.
Screens are here almost absent. It seems, in fact, as if the filling of a
fissure had not prevented it from remaining as a plane of weakness. After
each of the dykes had formed, there was subsidence again along its inner
margin, ' leading to the production of a further ring-dyke. In this way
the radius of the intrusions would gradually become less, and the inner
dykes would be the later, in accordance with the field evidence. A
I?f the Formation of Cone-sheets, Ring-dykes, and Caldron-subsidence:: 153
set which we should select. sincrle fault may thus have been responsible for a whole series of ring-dykes.
On~ may express this differently, and say that these were successively
,,
' n drawn at the appropriate
vell form a representation of formed by the downw-ard movement of a subsiding block.
-
It is uncertain, however, whether the broader of the Ardnamurchan
ening considerably upwards, ring-dykes can be explained in this manner. If they can there must
at least with regard to ring- have been very large downward movements of the subsiding block, or an
omewhat difficult to measure. abrupt tapering. These intrusions may, on the other hand, have been
as much as one kilometre in originally parts of flat sheets which were intruded along cross-fractures.
inally a fair distance below Their inner margins may be later ring fractures, which gave rise in turn
example of such a deflection to further intrusion . This is at least a third manner in which ring-like
- granite stock (Daly, I9 I 2, masses may have been produced.
rve which Daly has figured. The Beinn a' Ghraig granophyre in Mull may indicate a fourth method.
; -dyke was always, one may This slightly arcuate intrusion has mainly vertical sides, and seems to
-ith down throw towards the have been arched over by an irregular roof. It has been suggested that
'er, in every case have been in this case there was subsidence between two ring fractures. Both faults
wnthrow must have been a would according to theory be formed with an inward downthrow, but
lave been separated at first the direction of motion of the inner one may afterwards have been reversed.
the intruding magma may Ring-dykes do not extend so far from their centres in a horizontal
rther subsidence _ Such a direction as do cone-sheets. In the case of Centre 2 in Ardnamurchan,
'our of the apophysis given which has already been quoted, cone-sheets are found up to 7 km.
ikely, however, that any of from the centre, while ring-dykes cease within 4 km. In the table
r. In these cases it seems of principal pressures given on p. I47, the first figures may now be
self does not form a closed taken to refer to the pressures which are trans\-erse to the original cone-
uded block must also have sheet directions, and the last figures to the superimposed tensions. Both
e surface. It may alterna- these have to be combined with the" hydrostatic" pressure, and, as in the
separated from the rocks case of cone-sheet formation, it is certain that there can be no absolute
f the original fault had an tensions. Fractures are most readily formed where a certain quantity
one or all sides, during the with a linear relation to the greatest and least pressure attains a maximum .
., it is suggested, that most If the t\VO pressures be denoted by P and R, this quantity is not, ho\\-ever,
simply the difference P - R. It has the form P - nR, where n is a constant
ing the Beinn Chaisaidle b which may be as large as 5-8 (Anderson, 1934, p. 6). '
responsible for each ring- This quantity will be influenced by the" hydrostatic" pressure which
, but often they are separ- depends on the depth, but for any particular level the position of its
th regard to some of the maximum may be calcuiated from the superimposed pressures and tensions
conclusion is improbable_ which are recorded in the table. The maximum will be seen to occur
fact, as if the filling of a at each level where x=o, or directly above the centre of the caldron.
plane of weakness. After This, again, does not correspond to reality, as in the Ardnamurchan
ence again along its inner example, for instance, a fracture must have been formed at one depth
r ring-dyke_ In this way about 4 km. away. To overcome this difficulty one may suggest, as
ecome less, and the inner before, that the outline of the caldron had something of the nature
h the field evidence. A of a projecting angle. The maximum superimposed pressures might ·

I,i
I

v'
154 E. M. Anderson, The Dynamics of the Formation of

then have been localized above this, as one supposes the tensions to have
been during the formation of cone-sheets.
For the shape of caldron shown by the shaded margin in fig. 9, and the
corresponding magmatic pressures, the maximum tendency to fracture
will occur at the point 0, ·sa. Assuming as before that a = 5 km., the
Bp.
total lateral pressure is here 660 bars + 6· I4 The superimposed
-2 .
a
vertical tension cannot, however, exceed the" hydrostatic" pressure, as
the magma cannot exert an actual suction upon the rock; in other 'words,
Bp.
its pressure cannot fall below zero. One has therefore 2 I ·93 -21>660 bars,
a
and with this as a limitation one finds that the horizontal pressure cannot
exceed 845 bars. Even if there is no transverse pressure, however,
fracture will not take place unless the main pressure exceeds the ordinary
crushing strength of the rock, \vhich for granite is about 800 bars. Two
conclusions may apparently be drawn from this calculation. The magma
in the basin must have been at a pressure much below the" hydrostatic"
pressure for any particular level, and at the top this may have fallen
nearly to zero. There is also an additional reason for supposing that
the shape of the caldron is somewhat different from what has been repre-
sented. The point-push solution sho\\"S how some of the conditions of
central intrusion can be represented mathematically, but it can only be a
first approximation to reality.
The steeply dipping outer margins of granite stocks may in general
have been initiated as ring-fractures, but the present theory does not
give an explanation of these, which can be regarded as complete. It
might be objected that to explain a stock in terms of the point-push
solution one had to assume a previously existing basin which was itself
more or less of a stock. The origin of the basins will be discussed in the
last section, but the point is not fully met. A difficulty of another type is
involved in the explanation of surface subsidences. The ring-fracture
directions flatten upwards, and if such ~ fracture were initiated close to
the surface its angle of dip would be only about 25°. The fault which
bounds the Glencoe subsidence was, however, originally vertical. It
may have originated at some depth and gradually extended upwards.
In doing so it may have retained its original direction, but it must be
admitted that the dynamics of the process has not been worked out.

5. ORIGIN OF THE MAGMA RESERVOIRS .


The theory of central intrusion which has been discussed explains ,
some of the facts, and something may be learnt from what it fails to
of the FO'rmatz'on of Cone-sheets, Rz"ng-dykes, and Caldron-subsidences. 155
supposes the tensions to have explain. In particular it seems that no uninterrupted dome could have
been acted on by a magma so as to explain the distribution of either
aded margin in fig. 9, and the class of intrusions. The reservoirs must have had some,,,hat the shape
tximum tendency to fracture of an "inverted flower-pot."
as before that a = 5 km., the The existence of some type of basins has so far been taken for granted,
BtL but no theory will be complete which does not account for their origin.
'14~. The superimposed
At present, however, this is mainly a matter of conjecture. One may
e "hydrostatic" pressure, as imagine the original reservoir to have had the form of a flat sheet, of no
on the rock; in other words , great thickness, but at a considerable depth. If the method of intrusion
BtL of ordinary dykes is vertical, such a sheet must have extended in early
therefore 2 I '93 - 2 .} 660 bars
a ' f Tertiarv times from the Hebrides at least as far as Yorkshire.
. e horizontal pressure cannot The magma must then have worked its way up into the crust in
ansverse pressure, however, certain areas in more or less stock-shaped forms. This could not
ressure exceeds the ordinary probably have been by selective melting. This hypothesis would require
nite is about 800 bars. T,vo a large degree of horizontal inhomogeneity in the crust, and inhomo-
l is calculation. The mao-rna <> geneities seem not to be in the horizontal direction so much as the vertical.
uch below the" hydrostatic" It would appear rather that at certain spots there must have been selective
)e top this may have fallen stoping, or subsidence of blocks into the liquid layer on a larger scale.
I reason for supposing that It is difficult in most cases to see why, in what must have been large
t from what has been repre- areas, only a few particular spots were selected. But the liquid layer
.. some of the conditions of could not have been formed without irregularities, and some of these
tically, but it can only be a would probably be local, though maybe only gentle projections into its
roof. Suppose no\\' that after these were formed there \vas a period of
anite stocks may in general cooling, with partial solidification of the magma. The projecting areas
he present theory does not might become more or less separated as lenticular pockets, and the roof
regarded as complete. It of the layer might every here and there be in contact with its floor. It
in terms of the point-push is easy to see what would then have happened if there had been further
ting basin which was itself consolidation. The diminution of volume which this entailed would cause
ins will be discussed in the a very great reduction in the pressure of the magma. This might
. difficulty of another type is originally have corresponded roughly to the weight of the overlying rock.
Idences. The ring-fracture If it became small enough the roofs of the pockets would be bound to sink,
' ture were initiated close to and possibly at least they might sink by fracture.
bout 25°. The fault which In this' way the upward penetration of the liquid layer into the crust
.-er, originally vertical. It may have begun. One may suppose that there were many periods of
adually extended up\vards. consolidation, each followed by remelting, and each of them may have
1 direction, but it must be carried the process a little further. A similar partial solidification may
. not been worked out. have caused the loss of pressure in the magma basins which was referred
to in the last section. If the basins still communicated, the influence may
ES ERVOIRS.
have extended from one to another, and so, as has been suggested, the ring-
!s been discussed explains dykes of the Beinn Chaisgidle centre in Mull may be contemporaneous
'a rnt from what it fails to with the earlier of those which surround Centre 2 in Ardnamurchan.

.
156 E. JVJ. Anderson, The Dynamics oj the Formatioll oj i
These suggestions are, of course, purely speculative, and the theory of 1•
central intrusion is still far from complete. An attempt has only been
I
made to sketch the outlines of this very complex problem.
I am indebted to the Director of the Geological Survey and the Con-
!
j

I
tr~ller of the Stationery Office for permission to reproduce figs. I, 3, and 5; j
I
and in the case of fig. 2, to the above and to the Council of the Geological
Society of Glasgow, as well as to Dr Richey. With regard to fig. 4, the I
j
Council of the Geological Society of London and Dr Richey have kindly
given similar consent. I
6. SUtiIMARY.

It has been shown how the displacements due to a point dilatation


are modified, when the point is in the neighbourhood of a free plane
I
I
surface. The same has been done in one case for the strain corresponding
to the" premier type simple d'integrales," here renamed the point-push.
I
The conclusions arrived at might possibly have other applications, besides \
those in geology.
The geological application is as follows. An upward point-push at a
certain depth might have caused the relative tensions which have given
rise to a system of cone-sheets . It is unnecessary, hO\\"ever, for the thrust
to have been concentrated at a single point. The same solution \yill
apply to the upward and outward pressure exerted by a magma in a
certain shape of basin. This, as a first approximation, probably repre-
sents the facts, though the actual contour of the basins must have been a
little different from those which are deduced from the formula:.
A downward point-push, at a corresponding depth, might h?-ve
caused shear fractures to develop in the directions now followed by any
system of ring-dykes, and this might also have been effected by a falling
off of pressure in the supposed basin. The greater \~idth of ring-dykes
:
may be explained, if the fractures were closed in cross-section. The
blocks which they surrounded might then have become completely
detached, and they widened downwards. If they sank into the magma,
they would leave large vacant spaces along their sides, to be filled by the
intrusions .

.
!

- - -----
,
e>eF Ii

llze Formation of
Cone-sheets, Ring-dykes, and Caldron -subsidences. 157
eculative, and the theory of
.-\n attempt has only been
7. REFERENCES TO LITERATURE.
lex problem.
ANDERSON, E. M., 1905. "The Dynamics of Faulting," Trans. Geo!. Soc. Edin .,
gical Survey and the Con- vol. viii, pt. 3, p. 38 7.
'0 reproduce ~gs . I, 3, and 5; ANDERSON, E. M., 1934. "Earth Contraction and Mountain Building (II),"
Ie Council of the Geological Gerlands Bdtrage zur Geophysik, vol. xliii, p . 1.
With regard to fig. 4, the BAILEY, E. B., and others, 1924. "Tertiary and Post-Tertiary Geology of l'.-Iull,
lnd Dr Richey have kindly Loch Aline, and Oban," Mem. Geol. Surv.
BAIN, A. D. N., 1934. "The Younger Intrusive Rocks of the Kudaru Hills,

I
Nigeria," Quart. jount. Geol. Soc., vol. xc, p. 201.
BOUSSINESQ, J., 1885. AppNcat£on des potentiels a l'itude de l'iquiNbre et du
mouvement des soNdes ilastiques, Paris.
, due to a point dilatation
CLOUGH, C. T., MAUFE, H. B., and BAILEY, E. B., 1909. "The Cauldron-
Ibourhood of a free plane Subsidence of Glen Coe, and the Associated Igneous Phenomena," Quart.
or the strain corresponding ,
'i
jOllnt. Geol. Soc., vol. lxv, p. 6II.
I
'e renamed the point-push. DALY, R. A., 1912. "Geology of the North American Cordillera at the Forty-
other applications, besides ninth Parallel," l11em. iVo . 38, Dept. Mines, Canada . (Figs. 39, 40 are
reproduced as figs. 46 and 5 I in volume next quoted.)
n upward point-push at a DALY, R. A ., 1933. Igneous Rocks and the D epths of the Earth, New York.
DAY, A. L., SOSMAN, R. B., and HOSTETTER, J. C., 1914. "The Determination
ens ions which have given
of Mineral and Rock Densities at High Temperatures," Amer. jount. Sci.,
ry, however, for the thrust
vol. x.xxvii, p. I.
The same solution will HARKER, A., 1904. "The Tertiary Igneous Rocks of Skye," Mem. Geol. Surv.
xerted by a magma i~ a HERTZ, H., 1881. "Ueber die Beruhrun'g fester elastischer Karper," journ.
ximation, probably repre- reine u. angewandte fila th ., vol. xcii, p. 156. Translated in 1liscellaneous
e basins must have been a Papers, London, 1896.
m the formulce. KINGSLEY, L., 1931. "Cauldron-subsidence of the Ossipee Mountains," Amer.
ding depth, might h~ve j ount. Sci., vol. x..'cii, p. 139·
RICHEY, J. E., 1928. "The Structural Relations of the Mourne Granites
'ions now followed by any
(Northern Ireland)," Quart. journ. Geol: Soc., vol. lxxxiii, p. 653·
been effected by a falling RICHEY, J. E., and others, 1930. "The Geology of Ardnamurchan, North-west
reater width of ring-dykes Mull and Coll," JII£em. Geo!. Surv.
ed in cross-section. The RICHEY, J. E., 1932 a. "The Tertiary Ring Complex of Slieve Gullion (Ireland),"
have become completely Quart. journ. Geol. Soc., voL 1x.;·O{viii, p. 77 6 .
1ey sank into the magma, RICHEY, J. E., 1932 b. "Tertiary Ring Structures in ~ritain," Trans. Geol. Soc.
ir sides, to be filled by the Glasgow, vol. xix, pt. I, p. 4 2 •
SHAND, S. J., 1928. "The Geology of Pilansberg (Pilaan's Berg) in the Western
Transvaal: A Study of Alkaline Rocks and Ring-Intrusions," Trans . Geol.
Soc. S. Africa, vol. x.x.,'"i, p. 97·

i
I
J
(£ssued separately August 7, 193 6 .)

You might also like