Henry Briggs: The Trigonometria Britannica: Britannica (1) (7 . B.) by Adrian Vlacq in Gouda The Following Year: The

HENRY BRIGGS: THE TRIGONOMETRIA BRITANNICA 457
Henry Briggs: The Trigonometria Britannica

IAN BRUCE
1. Introduction
In 1632, Henry Gellibrand, then the Professor of Astronomy at Gresham
College, London, arranged for the publishing of the Trigonometria
Britannica [1] (7*. B.) by Adrian Vlacq in Gouda the following year: the
work consisted of two Books, and sets of tables of natural sines in steps of
one hundredth of a degree to 15 places, as well as tables of tangents &
secants to 10 places, together with their logarithms. The explanatory Book
I was the last work of Henry Briggs (1559-1631), Savilian Professor of
Geometry at Oxford, and was devoted mainly to the construction of his table
of sines; while Book II, written by the youthful Gellibrand on the instigation
of the dying Briggs, his mentor, contained instructions and examples on the
use of logarithms in solving trigonometrical problems. This article is
concerned with Book I of the T. B., an unfortunately long overlooked and
forgotten work. For Briggs had produced an initial reference table of 20
sines of angles up to 62.5° in the first quadrant with a carefully chosen
spacing of 3£°, accurate to 22 decimal places, which was sufficient to fill out
the tables of sines in the first quadrant by interpolation on division by 5
repeatedly, and by making use of a simple trigonometrical identity. Even
now> without the aid of a computer, this exercise would be a daunting task
for a skilled mathematician. The mathematics involved is remarkable -
given when it was accomplished - because of the novel numerical methods
used - and should still be of interest today, even though such table making
is a thing of the past. Here we find, among other things: an array of
numbers later known as Pascal's triangle; a sophisticated numerical
procedure following Viete (1540-1603) - by means of which Briggs
successfully extracted the roots of quadratic, cubic, bi-quadratic, and quintic
equations to the 22 places mentioned; and an equally sophisticated
interpolation scheme apparently of Briggs' own making. Thus, the T. B.
encapsulated the contemporary mathematics of the early 17th century, and is
worthy of attention.
The initial approach by Briggs to solving the problem of generating the
sines is geometric, and then arithmetic, where proportionals are employed
extensively, and with the occasional small pieces of laborious algebra. Any
additional algebra shown in this article is for expediency, as we try to
maintain the character of the original exposition, for much of the original
arithmetical procedures would now be given an algebraic setting. All of the
Tables in Book 1 have been reconstructed using spreadsheets by this writer,
though only a small sample of these can be presented here. Though Briggs
followed the immediate lead of Viete, numerous mathematicians had
worked on the problem of generating tables of sines, and the origins date
from before Ptolemy's Table of Chords {circa 140 AD): As always, Hutton
[2] is a mine of information on such matters for the interested reader.
Downloaded from https://www.cambridge.org/core. University of Waterloo, on 31 Dec 2019 at 23:44:57, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0025557200176090
458 THE MATHEMATICAL GAZETTE
Briggs, however, wished to brush aside all previous labours: which were to
be supplanted by his own improved methods. Basically, arcs of angles that
subtend chords of known lengths are either doubled, tripled, etc., and the
new chord lengths found; or the original arc is subdivided into a number of
equal parts, and the chord length of the fraction of the arc or angle is found,
a more tedious business involving the root extractions. The odd dissections
are treated first by Briggs, as the even ones can all be generated readily from
bisections. Briggs had solved these section equations by the early 1600s by
a slight adaptation of Viete's scheme. Viete [3] had constructed an
algorithm for finding the root or latus of a 'pure power' such as the square,
cube, fourth power, etc. of natural numbers (essentially by inverting a
binomial expansion, which we consider later). It was not too difficult a step
to extend this procedure to what were termed by Viete 'adfected powers':
for he viewed cubic and higher order equations as being similar to the pure
powers (i.e. equations of the form x" = A), that had become affected by the
addition of extra terms raised to powers smaller than n. T. Richard Witmer
[4], an American lawyer with a long standing interest in translating Viete,
has rendered some of this work from the original verbose Latin and
occasional Greek into English; there are of course a number of translations
of Viete's works into French. In addition, Thomas Harriot - the late
Elizabethan mathematician/scientist/adventurer was a friend of Viete, and
they corresponded on mathematical matters [5]. From the mass of papers
left by Harriot on his death in 1621, there arose the Artis Analyticae Praxis,
... published posthumously in 1631, on which Harriot's enduring fame as a
mathematician has mainly rested. This work contains the first algebraic
account of Viete's numerical method. Next, an associate of Briggs, Will
Oughtred (1575-1660), set out an account of the method in a popular book
of the time - the Clavis Mathematicae [6], of which an edition in the
English of the day exists, as well as several Latin editions. Viete's method
was thus the workhorse for finding the roots of equations numerically until
the 1680s. However, as Gellibrand informs us in the Preface to the T. B.,
and also Hutton in his preamble, Briggs was using his own perfected
method, essentially a simplified form of Viete's. One is surprised perhaps to
learn that the same method was in use as early as 1208 by the Iranian
mathematician Al-Tusi [7].
In this short account we shall not follow the original order of
development in the T. B. Here we establish the method for dissecting a
chord first for odd and then even numbers of cuts, followed by the
procedure for extracting roots (Briggs puts his root extraction method
between the odd and even dissections); finally, we present the beginning of
Briggs' Table of Sines, from which intermediate sines can be found by
Briggs' interpolation scheme. The diagrams are copies of the originals,
keeping the same labelling where possible.
2. Beginnings
No excuse is offered for making some preliminary remarks about the
nomenclature associated with the beginnings of Trigonometry. The analogy
FIGURE 1(a) FIGURE 1(b)

of the bow and arrow was strong for early writers: in Figure 1(a) we have
the chord AB as the string, the sagitta EC as the arrow, and of course the
smaller arc AB as the bow. Only three ratios were used originally, the sine,
the tangent, and the secant. These were considered as lengths associated
with an arc, which in turn was synonymous with the angle for a circle of
unit radius: thus, BE is the sine of the angles BXE and BXD. In Figure 1(b),
the tangent touches the arc at A, and has length AB, while the secant BC is
cut by the circle at F, both relating to the same angle ACB.
Note: 'Sine' [8] is a curious word, said to come from sinus, usually
meaning an in-fold or cavity - also of course being a medical term, while
the other ratios are derived from tangere: to touch, and secare: to cut.
The Table of Chords (i.e. Sines) of Ptolemy's Almagest [9] moves the
chord AB along the diameter of a circle marked off into 120 equal parts, in
half degree steps, where the circumference of the circle is divided into 360
equal arcs or degrees. There are chords and angles of known magnitudes
from regular figures such as the square, pentagon, hexagon, and decagon
inscribed in the circle, with which to start the calculations. Ptolemy
skilfully uses a powerful theorem to find the chord associated with the
difference of two arcs: If a, b, c, d are the four sides of a cyclic quadrilateral
in order, while p and q are the diagonals, then a.c + b.d = p.q. Thus
(Figure 2) if a and p are the known chords of known arcs, while d is the
diameter, then c and q can be found from the right angles in the semi-circle,
and so the unknown chord b also found for the difference of the arcs. Thus
e.g., the chord of 12° can be found from the difference of the known chords
of 60° and 72°. Another theorem finds the chord associated with half the arc
of a known arc and chord. Ptolemy is then able to find the chord lengths for
6°, 3°, l£° and | ° . For the latter small angles and chords, Ptolemy notes the
approximate proportionality between them, and so can find the chord
associated with \°. Hence, using the main theorem again, but in an additive
sense, Ptolemy establishes intervals with the angle BXE increasing at half
degree intervals, and gives the length BE in sexagesimal notation, correct to
4 or 5 decimal places in modern notation. Briggs, however, pauses only
long enough in his very brief historical introduction to discuss Ptolemy's
theorem, and to give the names of some of his predecessors in the table
making business, before beginning his own treatment.
FIGURE 2: Ptolemy's Theorem
3. Briggs' method for finding odd sections of chords
FIGURE 3: Trisecting an arc and chord

3.1 Trisection: Here is the beginning (Ch. Ill) of Briggs' elegant yet
surprisingly simple scheme for producing sub-multiples and multiples of
angles and chords, applied here to trisection, see Figure 3:
'If the straight line AC drawn through the Centre should cut the
Base of the section EB [the chord associated with the given
minor arc, of known length] in the point G, and the Arc in the
point C, these being equidistant from the vertex B [i.e.
BC = BG]; it [i.e. the line AC] will cut the Arc in the duplicate
ratio; and the Arc EDC will be double the Arc CB.'
[Italics in this kind of context mean we are quoting from a translation of the
original; the logical development is hence that of Briggs.]
FIGURE 4: Nested proportional isosceles triangles

This theorem can hardly be a geometric prescription for the impossible
task of trisecting an angle, as neither BG or BC is known from geometry, so
one has to regard the construction as hypothetical. On the other hand, one
can readily triplicate an angle, and relate the length of the resultant chord BE
to the original BC.
For the angle at the centre subtended by the arc CDE is double the angle
at the circumference. Also, if the radius is 1, and BC = BG =p<\, then
GC = p1, AG = 1 - p2, GF = p - p3, and from symmetry, the triple
chord BE = 2BG + GF = 3p - p3. [We have illustrated the idea of
nested similar isosceles triangles in Figure 4, for the case p < 1]. Hence
we immediately have the means of finding the chord BE for the triple angle,
given p. On the other hand, to find p, given the length of the chord BE, a
cubic equation must be solved. Note: Viete, in his Ad Angulares Sectiones,
as set out posthumously by Anderson (1615), had produced all of the same
section equations but by different means. Briggs elaborates upon this as an
alternative method in Chapter 11 of the T. B., but does not name his source
[3]. As an illustration of his method, to produce a chord of known length,
Briggs considers a regular decagon inscribed in a circle (Figure 5). The
chord length for a side BE is 2 sin 18°; now, cos 36° is known from the
pentagon to be r / 2 , where r is the golden mean, £(l + V5): hence,
2 sin 18° = V2 - T = 0.61803398875.... In the notation of the day,
Briggs can then write down the cubic equation:
3 © = 061803398875 + 1(f) o r 3 ® - 1(3) = 061803398875
where a radius of 100000000000 has been taken, thus avoiding the need for
the decimal point. It will be convenient for us to write:
/ ( * ) = x3 - 3x + A, (1)
where A = 0.61803398875. However, we defer discussing Briggs' solution
of this equation to trisect the chord until section 6, proceeding instead to
discuss other sections of the chord.
FIGURE 5: The regular decagon
3.2 Fivefold section
FIGURE 6: Quinquisection of arc and chord.

This beautiful piece of elementary geometry is an extension of the set of
similar isosceles triangles that Briggs has considered for trisection. Again, it
is based on a number of similar isosceles triangles: OE = BV = p;
AO = 1 - p2; AO/VO = lip gives VO = p - p\ as before. Then
BO = BV + OV = 2p - p3 = BP\ PO = 2p2 - p4, and AP = AO-PO
= 1 - 3/J2 + p4. Again, API 1 = PQIp gives PQ = p - 3p3 + p5, and
finally:
BG 2BP + PQ = 5p - 5p + / . (2)
We note initially a customary relation as a check: on setting p = sin#/2
in (2) we obtain
BG = 5(2 s i n - ) - 5(2 s i n - ) + (2 s i n - )
a n r\ en
= 10 sin 40 sin3 - + 32 sin5 - = 2 sin —,
2 2 2 2
a result that may be verified by an application of the Theorem of de Moivre

of later days; (2) is the equation solved by Briggs in Ch. 6, where the
variable is (2 sin \6), leading to very accurate values of the sine of the 5 th
part of the original angle of 36°, from the regular pentagon.
Higher order odd sections follow: Briggs deals with the sevenfold
section in detail, though we only summarise the result here due to lack of
space:
The total length of the subtending chord in terms of the
sevenfold section p is given by lp - 14p3 + Ip5 - p1. (3)
For 7(2 sin£0) - 14(2 sin J6»)3 + 7 (2sin \G)5 - (2 sin \d)\ and
7sin£#-56sin 3 f0 + 112sin5£# -64sin 7 £#; again in agreement with the
result from de Moivre's Theorem.
If p = 2 sin (70/2) < 0, then this has the effect of changing the signs
in the equation for the chord; otherwise the equation is used as it stands.
Even Briggs does not pursue the roots of this equation, as the method has
become unwieldy, although it can be used to give 7 times multiples of a
given chord. By now other means of an inductive nature had become
apparent to Briggs using his Abacus Panchrestus - which we examine next.
As indicated, section equations were already well-established through the
work of Viete and others, though the method of development presented here
belongs to Briggs.
4. Briggs' Abacus Panchrestus
-(2) (D "(3) CD
1 1 I 1 I 1 1 I I
10 9 8 7 6 5 4 3 2
55 45 36 28 21 15 10 6 3
220 165 120 84 56 35 20 10 4
715 495 330 210 126 70 35 15 5
2002 1287 792 462 252 126 56 21 6
5005 3003 1716 924 462 210 84 28 7
11440 6435 3432 1716 792 330 120 36 8
24310 12870 6435 3003 1287 495 165 45 9
48620 24310 11440 5005 2002 715 220 55 10
92378 43758 19448 8008 3003 1001 286 66 11
TABLE 1: Part of Briggs' Abacus Panchrestus
The expansion of (a + A)" by Briggs, for some number a and a small

correction A involved in numerically solving the above equations, leads
inevitably to a working knowledge of such expansions for positive integers
n. This will be illustrated later in Table 4. In addition, the coefficients that
arise have interesting properties of their own: for example, sums of these
give rise to the coefficients in the polynomial equations (1) - (3) and other
equations, as will be explained later. We are not suggesting that Briggs
discovered any of this, merely saying that these things were 'common
knowledge' at the time. However, it is of some interest to display a portion
of Briggs' table, which he called his Abacus Panchrestus, or 'wonderfully
useful' table. In modern notation, the entry for the r th row and k th column
(k, r > 1) is given by the binomial coefficient k + r~lCr-\, and is a form of
Pascal's Triangle. The elementary properties of these coefficients had been
found by Briggs, amongst others he says he has found, but which are not
enlarged upon. For mere is apparent a gathering sense of urgency in his
work, as ill health presumably was forcing him to curtail digressions and
unnecessary explanations.
® -CD -<D © (D -© -CD © CD

1 I 1 1 I 1 l 1 I
11 10 9 8 7 6 5 4 3
65 54 44 35 27 20 14 9 5
275 210 156 112 77 50 30 16 7
935 660 450 294 182 105 55 25 9
2717 1782 1122 672 378 196 91 36 11
7007 4290 2508 1368 714 336 140 49 13
16445 9438 5148 2640 1254 540 204 64 15
35750 19305 9867 4719 2079 825 285 81 17
72930 37180 17875 8008 3289 1210 385 100 19
140998 68068 30888 13013 5005 1716 506 121 21
20384 7371 2366 650 144 23
1
TABLE 2: Finding the coefficients of section equations; part of Briggs second table
If we look for the coefficients of equations (2) or (3), or any other
section equation in Table 1, then we are initially disappointed; however, the
numbers in Table 2 (the top part of which we show) are formed from the
vertical sums of two adjacent numbers in the column of the same power in
Table 1 above: such as 55 + 66 (near the bottom of column 2) and 121 in a
similar place Table 2.
To use Table 2 to find the coefficients for the l.h.s. of odd section
equations, such as (3), -lp1 + lp5 - 14p3 + lp: Start with the coefficient
of the 7 th power in the first row, - 1 , move down one row, and across two
columns to reach +7 for the coefficient of p5, and similarly for the p3 and p
HENRY BRIGGS: THE TRIGONOMETRIA BRITANNIC A 465
coefficients, -14 and 7 respectively. These have been underlined and
highlighted in Table 2 for convenience. The coefficients for the l.h.s. of the
11th power section can be found in the same way:
-pn + Up9 - 44p7 + 77p5 - 55p2 + lip.
Consider Adriaan van Roomen's equation of the 45 th degree, by means
of which the chord for 2° (= 2 sin 1°) could be found by extending the table
to 45 rows and 45 columns: as given by Briggs, Ch. 8:
Thus, if the Arc of the Periphery is cut into 45 equal parts, the equation will
be
45 (D —3795 ® + 95634 ® —1138500 Q) + 7811375 @ —34512075 <Q) +... = y/Z
See Goldstine [10] for a discussion of this famous equation proposed as a
challenge by Roomen, and solved by Viete in a devastating manner. Note
that we are back to Ptolemy's small chords again!
FIGURE 7: Bisection of arc and chord.
5. Briggs' method for finding even sections of chords

5.1 Bisection
For even sections, Briggs shows that the square of the chord is given by
the section equation. Let BC, CD, DE be equal chords, and the chord BE
continued to F so that EF = ED: the lines BD and DF are drawn, see
Figure 7. Now, following the trisection case, BC = BG = p, and ABCG
being similar to AABC, GC = p2 - HD; hence, AH = AG =l-p2; from
AG I AC = HG/DC, it follows that HG = (l - p2)p = p - p3. Hence
BE = 3p - p 3 and BF = 4p - p3. Also, the Triangles BCD and BDF are
readily shown to be similar; hence BCIBD = BDIBF: giving
BD2 = BC.BF = Ap2 - p4. Thus, the section equation gives the square of
the chord to be bisected.
FIGURE 8: The fourfold section

5.2 Fourfold section
According to Briggs, (Figure 8): If the Periphery be cut in four equal
parts, as OV, VY, YX, XS; the Square of the line OS is thus computed.
The square of the line OY (as we showed above), is equal to four of the
Square of the line OV, less the Biquadratic [i.e. 4 th power] of the same OV.
For the same reason, the Square of the line OS, is equal to four of the
square of OY, less the Biquadratic of the same line OY. The Square of the
line OY is 4 ^ * ^ , which going into itself makes the Biquadratic of the
lineOY;^®-^© +1®.
In modern terms: OY2 = 4p2-p4; OS2 = 40Y2 - OY4 =
2 6 2
16p - 4 / - 1 6 / + 8/> - p* = 16P - 2 0 / + 8 / - / . Six and eight fold
sections are subsequently presented; these coefficients, and all even sections
for the square of the chord, can also be found from Table 2.
6. Briggs' method for finding roots of polynomials

We now retrace our steps back to Chapter 4, where the use of the
'algebraic' method (Briggs' terminology) for locating the root of an equation
is demonstrated. Let it be said at the outset that Briggs presents neither
statement or proof of his method, nor indicates its origins: the first has to be
discerned from his arithmetical calculations, while we have to look at
Viete's work to find the origins, which we do briefly, and which also are
reflected in the work of Harriot and Oughtred.
In Viete's original development [3], alluded to in the introduction, the
roots of powers of natural numbers are sought, up to the 6 th. Later he
extends his method to any number, and calculates \/2 to 15 places. We
consider his treatment of the cube root of 157 464 (each digit of the root
being called a 'side' or latus) to illustrate the method.
In modern terms, Viete sets (a + bf as an initial approximation for
N = \51464, which is known to be a natural cube, and by inspection finds
the largest 'side' involving a single digit to be a = 50, leading to
3a2b + 3ab2 + b3 = N - a3; the second single digit side or approximation will
be b = (N - a3 - 3ab2 - b3)l{3a2) = (N- a3)/{3a2) - b2la - b3l{3a2), of which
the integral part is taken to be (N - a3)/(3a2) = 32 464/(3 x 2500) ~ 4, to
one significant figure. Thus, 54 is the next approximation, and turns out to
be the actual cube root in this case, on working out the whole cube. As
Viete points out, the above scheme does not always work, as a negative
correction may be needed for some numbers: e.g. for N = 6859 (= 193),
where a = 10 generates b = 19, which is not a single 'side'; but instead
a = 20 leading to b = - 1 should be chosen. (We mention in passing a
reluctance by mathematicians of the age to use negative numbers in general,
and negative roots in particular.) Viete next turns to his 'affected cubes', or
cubic equations, which he analyses in a related way. We will illustrate this
from Harriot's Praxis [5, p. 132], where algebra is used with the method for
the first time.
Harriot considers a cubic originally written in the form:
aaa + daa + ffa = ggg; for the coefficient of a was considered to be a
square, and the constant term ggg a cube, corresponding to Viete's idea that
each term should have the same dimensions: we accommodate for this
unnecessary complication by writing the cubic in a in the form:
a3 + da2 + f2a = g3. Then set a = b + c, to give (in modern notation):
(b1 + db + f)b + (c2 + 3bc + 3b2 + f + dc + 2db)c = g\
Harriot writes this equation in the two-part form Ab + Be = g3.
A specific cubic with d = 68, f = 4352 and g3 = 186 394 079 is
considered. In a table not presented here, following Viete, Harriot partitions
g3 by placing a point over each 3rd digit, from the left: 186394079. The first
approximation is b = 500, as b3 = 125 000 000, while Ab = {b2 + db +f)b =
1 441 760; subsequently, Be = g3 - Ab = 42 218 079 has to be resolved.
Harriot's divisor is B = c2 + 3bc + 3b2 +f2 + dc + 2db: and c is chosen as the
single digit multiple of 10 to make Be the nearest to 42 218 079. This turns
out to be c = 40, and now b becomes 540, while the units value of c is next
found to be 7, giving 547 as the root. This is a tedious process, as B has to
be evaluated for each trial of c. Oughtred presents a similar picture [6].
Now we give an outline of the method used by Briggs, who streamlined
the above method, applied here to a cubic: x3 + Bx2 + Cx = D. If, by trial
and error, an approximate value A of the root is found to 1 significant figure,
then a better approximation (to two or three figures) will be given by
L = A + E, where the magnitude of E is found on the division of a residual
term by a term that depends on known numbers. If (A + E)3 + B(A + E)2 +
C(A + E) s D, then on expansion and collecting terms,
E=(D-A3-BA2- CA)/(3A2 + 2A5 + C), where the small terms in E2 and
E3 have been ignored. With E established, A + E becomes the next
approximation, and a new correction Ef found, the process being continued
to refine the root. In general, each application of the method increases the
number of accurate places by more than one, and it is often found that the
denominator is a very slowly varying function, so that an ordinary division
can be used at last. A very useful method for someone doing calculations by
hand, as errors show up at once. Briggs used the method skilfully for
equations up to the 5 th power: he presumably stumbled on the refinement
when he realised the higher order terms contributed little, yet substantially
increased the time involved in extracting the next place.
We show how Briggs used the method to extract the root closest to zero
foriiy.x3 - 3x + 0.61803398875 = 0.
0 61 RfV^
1st estimate: A = 0,andL = A + E = 0 + — — = 0.2 (to 1 sig. fig.);
3
„ J • . „„ JF . A - 3A + 0.618033... nnnn
2nd estimate: A = 0.2, and L = A+ ; = 0.209.
3 - 3A2
We will spare the reader the rest of the numerical details in this calculation,
though Tables 3 and 4 give an illustration of Briggs' actual workings.
3)061803398875(02
8 Cube to be added
0626
6 Triple of Quotient to
be taken away
26033 Remainder
26033988(020
TABLE 3: The algorithm begins with xx = 0.2
26033988 Remainder to be divided. 40000 3 Root found 20.
3 . . . . Divisor. 200 3 Square root found 400
120000 Taken from Divisor 120000
288 . . Divisor corrected. 600
By having found the Correct Divisor, I go on and ask how many times 288 [goes]
into 2603.
And with finding the number, 9 should be placed in the Quotient.
1200
26033988(0209 °0 9
1122322 With the gnomon of the Cube to be added 1080000 729
27163317 48600
27 . . . Product from root number 9fry3 taken from the divisor ^29
163317750 Remainder 1129329
TABLE 4: It continues with x2 = 0.209
We can merely give an indication here, due to space constrictions, of

one of the great early triumphs of numerical analysis, namely, the solution
of the quintic equation for four of the five roots as outlined in Chapter 6:
/ ( * ) = Xs - 5JC3 + 5x - A, where A = 1.17557050458 = 2 sin 36° is the
known chord.
We illustrate the convergence for the smallest positive root:
jcx = 0.2 x2 = 0.25 x3 = 0.2506 x4 = 0.25066
x5 = 0.250666 x6 = 0.2506664 x1 = 0.25066646;
The other roots represent possible angles associated with the given chord, of
which Briggs evaluates three.
7. Towards the final table of sines

7.1. A table of known sines
There is one last ingredient in our exposition of the Trigonometria: the
interpolation scheme developed by Briggs, though we forgo a discussion of
the evaluation of the logarithms of the sines, tangents, and secants. Briggs
had hopes, inspired by Viete, of finding the sines of thousandths of degrees,
and of introducing a decimal system of degrees and parts of degrees: none of
this was ever to materialise, though his tables use 1/100 th degree steps, and
he gave a table of sines for a circle divided into 100 equal parts in Chapter
14. We note the beginning remarks of Chapter 13, where he explains how to
get the sines at 3.125° intervals up to 60°, correct to 22 places:
Really we should return to the Sines [he has just completed a tour
de force demonstration of his interpolation scheme]: if for which, it were
wished to make a whole Table of Sines, for thousandths of a Degree; For a
given Chord of 60: 0': Degrees the Chord for 20:0': should be sought by
Trisection, which by five-fold multiplication shall give the 100: 0': Chord.
Then by Bisection the squares of the Chords are sought 50: 0':\25: 0': \12:
30': 16: 15':\. These Squares will give the 50: 0': \25: 0': \12: 30': 16: 15':
Chords themselves. And these Chords shall give by Triplication the Chords
18: 45': \56:15' \37: 30':\75: 0'\ 112: 30'\.
And by 5-fold multiplication the Chords 31: 15'\62: 30':\125: 0':\93:
45'. And by 7-fold multiplication the Chords 43: 45':\87: 30':\. For by
multiplication [of 6: 15':] by 11. 13. 17. 19. to give the Chords 68: 45': IS/:
15':\106: 15':\118: 45':\ Of these Chords the halves, namely the Sines of
3\.6\.9\.12^ Degrees, etc. have been placed here below.
Briggs uses the identity sin(60° + x) - sin(60° - x) = sin* to

complete the sines in the quadrant from this table. Thus we have the
beginnings of Briggs' answer to Ptolemy's Table of Chords.
7.2 The interpolation scheme

Briggs indicates in the Preface to the Logarithmica that he had laboured
long to arrive at the interpolation scheme used in the Trigonometria. This
method was used subsequently as a means of interpolation between logs to
complete his tables, but not by Briggs himself. Briggs merely showed how
it should be done and hoped that someone else would take up the challenge.
It appears that Walter Warner, the editor of Harriot's Praxis, was attending
to this matter, see Shirley [5]. However, independently Vlacq took up the
challenge, but not in the way that Briggs had intended, as he rounded off the
numbers and used the easier first interpolation scheme in the Arithmetica
[1], thus upsetting Briggs greatly - as one can gather from a paragraph
directed at Vlacq (who, by a twist of fate, was to publish the book!).
Essentially, the intervals between known sines at (^) ° spacing are divided
Deg. 1/lOOOth Sines of Degrees and Deg. Min. Sec.

parts thousandth parts
3 125 545145016380063218757 3 1 30
6 250 1088668748519645759679 6 15 0
9 375 1628954733945887394808 9 22 30
12 500 2164396139381028797595 12 30 0
15 625 2693400539532260178082 15 37 30
18 750 3214394653031615807011 18 45 0
21 875 3725829021441029702515 21 52 30
25 000 4226182617406994361870 25 0 0
28 125 4713967368259976485564 28 7 30
31 250 5187732581605214495200 31 15 0
34 375 5646069259743004405407 34 22 30
37 500 6087614290087206394161 37 30 0
40 625 6511054499119490930562 40 37 30
43 750 6915130557822693777612 43 45 0
46 875 7298640726978356573501 46 52 30
50 000 7660444431189780352024 50 0 0
53 125 7999465651001296901396 53 7 30
56 250 8314696123025452370788 56 15 0
59 375 8605198338560790089653 59 22 30
62 500 8870108331782217010546 62 30 0
TABLE 5: Briggs' Table of Reference Sines
into five parts, and the intervening sines are found approximately and
subsequently very accurately from a correcting table for the successive
orders of differences. Let us follow this process briefly: in Table 6, the
'raw' differences up to the 5th order are found.
[Note that Briggs never uses a decimal point; also, he was able to deduce
that the even order differences were proportional to the complement of the
original sine, while the odd orders were proportional to the original sine
with a sign difference, as established in Chapter 12 of the T. B.]
These differences in Table 6 are then divided by 5 successively: see
Table 7. Briggs never divulged the origin of his table for correcting the
differences, though of course the method was rediscovered later by others.
See Goldstine [10] for information on this aspect of Briggs' work, who
shows how the correcting coefficients in Briggs' Table (originally to be
5th Diff 4th Diff 3rd Diff 2nd Diff 1st Diff Sine Deg
000 000000 0
483 162129 54514502
483 162129 54514502 34
481 161646 54352373
964 323775 108866875 6|
474 160682 54028598
1438 484457 162895473 9|
TABLE 6
Mean Differences
5th Diff 4th Diff 3rd Diff 2nd Diff 1st Diff Sines Deg
0 0000 0 00000000 0
15 1297 0 10902900 4
77 6485 2 54514502 34
15 1293 2 10870474 6
1 54 12951 0 108866875 6*
15 1285 5 10805719 6
2 30 19378 3 162895473 9 |
15 1274 0 10708828 2
TABLE 7
found on p. 29 of the Arithmetica) can be obtained. A great simplification is

obtained if the differences in the final column are considered to be equal,
which is the case if the results are being calculated to a finite degree of
accuracy, by setting equal the different row levels in the final results above.
In the case considered with constant 5 th order differences, to the accuracy
required, it follows that the 3 rd order differences are easily corrected for a
given row from Briggs' correcting table, and so the 1 st order also; the 4 th
order is also correct for a given row, and the 2nd order easily corrected.
Hence, to ease the arithmetical work, there is a great deal of sense in using
differences that obey these rules.
The result of using these corrected differences (C. D.) is shown below in
Table 8 using a spreadsheet, where of course we only verify Briggs' results,
and where we need go no higher than 5 th order.
(Deg.) Sine 1st U./C.D. 2nd (/./CD. 3rd U./C.D. 4th 5th
O.OOOE+00 O.OOOOOOOOE+00 1.09067936E-02 O.OOOOOE+00 1.2975E-06 O.OOE+00
1.09080913E-02 1.2977E-06 1.54E-10
6.250E-01 1.09080915E-02 1.09054958E-02 1.29764E-06 1.2793E-06 1.54E-10
1.09067934E-02 1.29795E-06 1.2977E-06 1.54E-10
1.250E+00 2.18148850E-02 1.09029003E-02 2.59514E-06 1.2970E-06 3.08E-10
1.09041977E-02 2.59575E-06 1.2974E-06 1.54E-10
1.875E+00 3.27190828E-02 1.08990075E-02 3.89232E-06 1.2966E-06 4.62E-10
1.09003045E-02 3.89324E-06 1.2970E-06 1.54E-10
2.500E+00 4.36193874E-02 L08938179E-02 5.18904E-06 1.2959E-06 6.16E-10
1.08951142E-02 5.19027E-06 1.2964E-06 1.54E-W
3.125E+00 5.45145016E-02 1.08873320E-02 6.48514E-06 1.2952E-06 7.71E-10
1.08886276E-02 6.48668E-06 1.2956E-06 1.54E-10
TABLE 8
8. Conclusions
Sadly, Briggs' remaining chapters in the Trigonometria show a tailing
off, to end rather abruptly in Chapter 17, due to the author's increasing ill
health. The task performed by Gellibrand in finishing the book and
preparing it for publication was quite monumental, a fact that he bewailed in
the preface, for Briggs had left no plan of how he intended to proceed with
the applications, as he had done in the Arithmetica. A number of
'Gentlemen at Arms' and other dignitaries are mentioned here by
Gellibrand; perhaps they contributed financially to enable the book to be
published.
The book itself does not appear to have been a commercial success.
Originally printed in Gouda in 1633 under the auspices of Adrian Vlacq,
whose own tables in traditional minutes and seconds, in steps of 10",
appeared at about the same time. According to Hutton, it is probable that
decimal division of degrees would have replaced the traditional sexagesimal
approach if this latter set of tables had not appeared. Thus Newton, some 30
years later appears to have made no use of the T. B., or even been aware of
its existence, though he studied Viete in some depth; see Whiteside [11] for
details. Indeed, Whiteside appears to be the only person to have noted that
the algorithm used by Briggs to solve the section equations was none other
than that rediscovered by Newton and now called either Newton's method,
or the Newton-Raphson method [12].
Thus disappeared from view a man whose immense labours over his
lifetime was largely responsible for the final shape of log and trig tables,
who enabled succeeding generations to perform otherwise impossible
calculations and who gave explanations of his methods. To which Briggs
himself might well have applied that old adage: sic transit gloria mundi.
Acknowledgement
I am greatly indebted to the reviewer, Jackie Stedall, Queens's College,
Oxford, for the considerable effort, patience, and time given, in helping to
make this article something that hopefully reflects a little of the state of
mathematics in the early 17th century, and the contribution made to it by
Henry Briggs. This has involved pointing me towards references of which I
was unaware, strongly advising me to use contemporary mathematics only
in explanations, and supplying me with the encouragement to 'get it right' as
far as possible.
References
1. Henry Briggs, Arithmetica Logarithmica, etc., London (1624). A
microfilm of a copy of Briggs' book in the Bodlean Library is available,
produced by University Microfilms International, Ann Arbour,
Michigan 48106.
Henry Briggs and Henry Gellibrand, Trigonometria Britannica, in two
books. The present writer obtained a photocopy of the one held by the
Rare Books Department at Cambridge University Library, for which he
is duly grateful. See also:
Ian Bruce, The agony and the ecstasy. Math. Gat. 86 (July 2002)
pp. 216-227.
2. A good summary of Briggs' achievements is given by Charles Hutton in
the extensive preamble to the earlier editions of his Mathematical
Tables, London (1811 edition) pp. 75-84.
3. Francois Viete, Opera Mathematica, Schooten, Olms (1970).
De Numerosa Potestatum Purarum Resolutione, pp. 163-172.
De Numerosa Potestatum Adfectarum Resolutione, pp. 173-228.
Ad Angulares Sectiones, pp. 287-404.
A good introduction to Viete, and his relation to the ancients, can be
found in:
Jacob Klein. Greek mathematical thought and the origin of algebra.
Dover, N.Y. (1968).
4. T. Richard Witmer, The Analytic Art, Kent State University (1983)
pp. 311-370.
5. Thomas Harriot, Artis analyticae praxis, ad aequationes algebraicas
nova methodo resolvendos, London (1631). Available on microfilm
from University Microfilms, Inc. Ann Arbour, Michigan.
For a biography of Harriot, see:
John W. Shirley, Thomas Harriot; a biography. Clarendon Press,
Oxford (1983).
The friendship between Viete and Harriot is mentioned on p. 3 of this
book, the original reference being:
Jon V. Pepper, 'A letter from Nathanial Torporley to Thomas Harriot'.
Brit. J. Hist. Sc, III (1967), pp. 285-290.
6. William Oughtred, The Key of the Mathematics New Filed, London

(1655) p. 121 onwards. Also on microfilm as above.
7. Roshdi Rashed, Resolution des Equation Numeriques etAlgebre: Saraf -
al-Din al-Tusi, Viete. Archive for history of exact sciences (1974) 12,
pp. 244-290.
Also by Rashed: The Development of Arabic Mathematics: Between
Arithmetic and Algebra, Kluwer, Boston (1996).
8. Alfred Hooper, Makers of Mathematics, Faber & Faber (1961) p. 127.
This author has an interesting explanation that hinges on the
misinterpretation into Latin by a medieval translator of a Hindu word
for 'half-chord' present in an Arabic manuscript.
9. Ptolemy, Britannica Great Books, Vol. 16 (1975) pp. 14-21.
10. Herman H. Goldstine, A History of Numerical Analysis, Springer-
Verlag N.Y. (1978) p. 33.
11. D. T. Whiteside (ed.), The mathematical papers of Isaac Newton,
Cambridge (1967), Vol. I, Ch.2; Vol. 2. pp. 221-222.
12. T. J. Ypma, Historical Development of the Newton-Raphson Method,
Siam Review, Vol. 37,4 (1995) pp. 531-551.
IAN BRUCE
Dept of Physics & Mathematical Physics, University of Adelaide,
S. Australia. P.C.5005
e-mail: ibruce@physics.adelaide.edu.au
To rcri'ivi' ;t ™nf(-rem'i: bntcfwre anil hooking form (without tibHg:!tioi))< pleiiw wild your name ttml addr<-«t 10 III* BCME Di'lrgnte leuni t !a
The Mathtimiiiira) A s e d a t i o n , 25*J Uutdnn K»sid L r i w t e r 1.E2 ?&£ «<• fa« (II16 211 Iti.15 or «n.ail (IWi'K«lci« bi:«i«6.nr(!

Henry Briggs: The Trigonometria Britannica: Britannica (1) (7 . B.) by Adrian Vlacq in Gouda The Following Year: The

Uploaded by

Copyright:

Available Formats

Henry Briggs: The Trigonometria Britannica: Britannica (1) (7 . B.) by Adrian Vlacq in Gouda The Following Year: The

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Henry Briggs: The Trigonometria Britannica: Britannica (1) (7 . B.) by Adrian Vlacq in Gouda The Following Year: The

Uploaded by

Copyright:

Available Formats

HENRY BRIGGS: THE TRIGONOMETRIA BRITANNICA 457

Henry Briggs: The Trigonometria Britannica

FIGURE 1(a) FIGURE 1(b)

FIGURE 2: Ptolemy's Theorem

3. Briggs' method for finding odd sections of chords

FIGURE 3: Trisecting an arc and chord

FIGURE 4: Nested proportional isosceles triangles

FIGURE 5: The regular decagon

3.2 Fivefold section

FIGURE 6: Quinquisection of arc and chord.

a result that may be verified by an application of the Theorem of de Moivre

4. Briggs' Abacus Panchrestus

The expansion of (a + A)" by Briggs, for some number a and a small

® -CD -<D © (D -© -CD © CD

FIGURE 7: Bisection of arc and chord.

5. Briggs' method for finding even sections of chords

FIGURE 8: The fourfold section

6. Briggs' method for finding roots of polynomials

We can merely give an indication here, due to space constrictions, of

7. Towards the final table of sines

Briggs uses the identity sin(60° + x) - sin(60° - x) = sin* to

7.2 The interpolation scheme

Deg. 1/lOOOth Sines of Degrees and Deg. Min. Sec.

found on p. 29 of the Arithmetica) can be obtained. A great simplification is

6. William Oughtred, The Key of the Mathematics New Filed, London

You might also like