NPTEL C OURSE ON

M ATHEMATICS IN I NDIA :
F ROM V EDIC PERIOD TO M ODERN TIMES

L ECTURE 32
Jy
anayanam: Computation of Rsines

K. Ramasubramanian
IIT Bombay

Outline
I

Introduction: Motivation for finding sine values

Approach in earlier Siddh

antas

Construction of sine table: Aryabhat

. as approach
I Geometric & Analytic

I The terse verse in Aryabhat

.ya

Bhaskaras
formula (his invention?)
I
I

A brilliant rational approximation

Plausible derivation of the result

The sine table in S

uryasiddh
anta

Govindasvamins improvised tabular Rsines

Rule for obtaining the sine values in Tantrasangraha

(+ comm.)

Madhavas
series for sine

His simplified formula (vidv

anity
adi)

Introduction
Motivation for finding accurate sine values

The positions of the planets in

the background of stars forms
the basis for reckoning time.

Determination of their positions

crucially depends upon accurate
knowledge of sine function.

Sun (in conjunction

with Moon)

suklapaksa

S3
S2

M2

S1
A
Earth spinning
on its axis

Earth

krsnapaksa

M1

Introduction
Determination of time from shadow measurement

Figure: Zenith distance and the length of the shadow.

t = (R sin)


R cos z
R sin .
cos cos

If and are known ( = f (, )), then t is known.

Computing Rsines: Approach in earlier siddhantas

I Var
ahamihira has given the following Rsine values and relations in his
Pa
ncasiddh
antik
a (c. 505 CE):1
R sin(30 )

R sin(45 )

R sin(60 )

R sin(90 )

R
2
R

3
R
2
R

(1a)
(1b)
(1c)
(1d)

R sin()

R cos(90 )

(2)

R sin2 () + R cos2 ()
 

R sin
2

R2
 
1
1
[R sin2 () + R vers2 ()] 2
2

(3)

The above Rsine values (1) and relations (2)(4) can be derived using the
bhuj
a-kot.i-karn
aya and trair
asika (rule of three for similar triangles).
. a-ny
Equations (2)(4) can be used to compute all the 24 tabular Rsine values.
1

Pa
ncasiddh
antik
a of Var
ahamihira, Ed. by T. S. Kuppanna Sastry and
K. V. Sarma, Madras 1993, verses 4.15, pp. 7680.

(4)

Construction of the sine-table

A quadrant is divided into 24 equal parts, so that each arc bit

0
0
= 90
24 = 3 45 = 225 .

Aryabhat
. a presents
two different methods
for finding
R sin i, (Pi Ni ) i =
1, 2, . . . 24.

The Rsines of the

intermediate angles
are determined by
interpolation (I order
or II order).

Finding tabular sines: Geometrical approach

.sa;ma;vxa:a;pa;a:=+a;Da;pa;a;dM ;a;C+.nd;a;a;t,a ;
a:a;Bua:ja;a;a;tua;BRua:ja;a;Ea;va
.sa;ma;.ca;a;pa:$ya;a;Da;Ra;a;na tua ;
a;va;Sk+:}Ba;a;DeRa ya;Tea;;a;a;na

I We know, chord 60 = R = 3438

I From the triangle OBC,

D
F
C

R sin 30 = BC = R/2 = 1719

R

60
O

G
R/2

I Using bhuj
a-kot.i-karn
aya from 8th
. a-ny
Rsine (30 ) we can get 16th Rsine (60 ).
I In other words, R sin
R sin(90 ).
I Further, noting that
R sin

R cos

R sin & R vers

R vers

R sin ,
2

it can be seen that all the 24 Rsines can

be calculated.

Finding tabular sines: Geometrical approach (contd.)

I Most of the Indian astronomers have presented their sine tables by

dividing the quadrant (90 ) into 24 parts.

I By the principle outlined above, it can be easily shown that all the 24

Rsines can be obtained provided the 24th, 12th and 8th Rsines are
known.

8th
16th

4th

23rd

11th

13th

The circumference of the circle was

taken by Aryabhat
. a to be 21600
units.

From that using the approximation for

given by him, we get
R = 24th Rsine 3438.

Once this it known, it is noteworthy

that in the proposed scheme of

18th

3rd
10th

22nd

1st

6th

20th

2nd

I
12th

14th

21st

9th

15th

5th

7th

19th

17th

constructing the table, all that is

required is extraction of square root,

for which Aryabhat

. a had clearly
evolved an efficient algorithm.

Finding tabular sines: Analytic approach

:pra;Ta;ma;a;a;a;pa:$ya;a;Da;Ra;t,a .yEa:+.nMa Ka;a;Nq+.
/ /  tMa ;
a;d
;ta;a;ya;a;DRa;m,a
ta;tpra;Ta;ma:$ya;a;Da;a;ZEa;~tEa;~tEa:+.na;a;a;na Zea;Sa;a; a;Na

Let Bk = R sin (k 2250 ), (k = i, 2, . . . , 24) , be the twenty-four Rsines, and

let k = Bk Bk 1 , (k = i, 2, . . . , 24)... be the Rsine-differences.
Then, the above rule may be expressed as2
2

k +1

B1
B1
(B1 + B2 + . . . + Bk )
B1
B1

B1

(5)
(k = 1, 2, . . . , 23).

(6)

This second relation is also sometimes expressed in the equivalent form

k +1 = k

(1 + 2 + . . . + k )
B1

(k = 1, 2, . . . , 23).

(7)

From the above the discrete version of the harmonic equation follows
k +1 k =

Bk
B1

(k = 1, 2, . . . , 23).

Aryabhat.a is using the approximation 2 1 10 .

(8)


Aryabhat
. as table for computing Rsines

I Using either/both the approaches, Aryabhat

. a having obtained Rsine

values has presented a table in Gtik

a-p
ada of Aryabhat
.ya (verse 12).

I This verse3 lists the 24 first order Rsine-differences (in arc-minutes):

ma;a;Ka
 Ba;a;Ka
 :P+.a;Ka
 ;Da;a;Ka
 :Na;a;Ka
 Va;a;Ka


.z+.a;Ka
 h;~.Ja .~k+:
a;k ;
a;k+:Sga .~.Ga;
a;k ;
a;k+:Gva
;Gl+.
a;k ;
a;k+:g{a h;k+.a ;Da;
a;k ;
a;k+:.ca

.~ga ZJa z
*: a :P C k+:l+.a;DRa:$ya;aH

225, 224, 222, 219, 215, 210, 205, 199, 191, 183, 174, 164,
154, 143, 131, 119, 106, 93, 79, 65, 51, 37, 22, and
7these are the Rsine-differences [at intervals of 2250 of
arc] in terms of the minutes of arc.

I In Aryabhat
. as notation:
3

ma

25;

&

;a;Ka


200;

This verse is one of the most terse verse in the entire Sanskrit literature
that I have ever come across. Only after several trials would it be ever
possible to read the verse properly, let also deciphering its content.

Bh
askaras approximation to sine function
In his Mah
abh
askarya, 4 while presenting planetary parametersthe

dimension of epicycles, their application, etc. Bhaskara

(c. 628)
presents a brilliant rational approximation to sine function.
ma;K.ya;a;
a;d:=+
a;h;tMa k+:mRa va;[ya;tea ta;t,a .sa;ma;a;sa;taH
. ca;k
+:ea;Da;a;Za;k+:sa;mUa;h;a;t,a ;
a;va;Za;ea;Dya;a yea Bua:ja;Ma;Za;k+:aH
ta;.cCe+.Sa;gua; a;Na;ta;a ;
a;d
;+aH

Za;ea;Dya;aH

Ka;a;Brea;Sua;Ka;a;a;b.
/ /
Da;taH

. ca;tua;Ta;a;Zea;na Zea;Sa;~ya ;
a;d
;M A;ntya;P+.lM h;ta;m,a
ba;a;hu;k+:ea;fa;eaH

x +:t=+:
:P+.lM k
Ma k
+:ma;ea;tk
+:ma;gua;Na;~ya va;a

l+.Bya;tea . ca;nd+ta;a;[Na;Ma:(;a;eaH

ta;a:=+a;Na;Ma va;a;
a;pa ta:va;taH

This in modern notation translates to

sin x =

x(180 x)
1
[40500 x(180 x)]
4

Mah
abh
askarya, VII.17-19.

(0 x 180)

Bh
askaras approximation to sine function
ma;K.ya;a;
a;d:=+
a;h;tMa k+:mRa va;[ya;tea ta;t,a .sa;ma;a;sa;taH
. ca;k
+:ea;Da;a;Za;k+:sa;mUa;h;a;t,a ;
a;va;Za;ea;Dya;a yea Bua:ja;Ma;Za;k+:aH
ta;.cCe+.Sa;gua; a;Na;ta;a ;
a;d
;+aH

Za;ea;Dya;aH

. ca;tua;Ta;a;Zea;na Zea;Sa;~ya ;
a;d
;M

ma;K.ya;a;
a;d:=+
a;h;tMa k+:mRa
va;[ya;tea
.sa;ma;a;sa;taH
. ca;k
+:ea;Da;a;Za;k+:sa;mUa;h;a;t,a
Bua:ja;Ma;Za;k+:aH
yea ;
a;va; a;Za;Dya;aH
ta;.cCe+.Sa;gua; a;Na;ta;a

(Bua:ja;Ma;Za;k+:aH )

;
a;d
;+aH
Za;ea;Dya;aH

(O;;k+.a)

Ka;a;Brea;Sua;Ka;a;a;b.
/ /
Da;taH
. ca;tua;Ta;a;Zea;na
;
a;d
;M

(;
a;va;Ba:jea;t,a)

Ka;a;Brea;Sua;Ka;a;a;b.
/ /
Da;taH

...

operations without ma;K.ya;a;

a;d
is being stated
completely

from 180
arg. of sine in degrees (x)
that which is to be subtracted
x(180 x)
kept in two places
have to be subracted (in one place)
from 40500

with

the value placed in the other place

1 th
4

of the result (divide)

How good is Bhaskaras approximation?

Bh
askaras

Modern

value

value

0.000

0.000

15

0.260

0.259

30

0.500

0.500

45

0.706

707

60

0.865

0.866

75

0.966

0.966

90

1.000

1.000

The formula gives exact value for

certain arguments.

However, it is noted that for the

entire range [0 90 ] it is almost
99% accurate

This clearly speaks of:

I

The brilliance of the discoverer

in arriving at an excellent
approximation.
a novel attempt to obtain a
rational approximation to sine
function.
the beauty and sophistication of
ancient Indian mathematics

Properties mirrored by Bhaskaras approximation

I We know that sine function
1. is symmetric about 90 point

Bhaskaras
approximation

2. is concave over the range

0 180
I The formula given by Bh
askara

sin x =

4x(180 x)
40500 x(180 x)

clearly satisfies these properties

I Isnt a mathematicians delight to arrive at an expression for sine

function that at once captures the properties as well as serves as a very

good approximation ( 99%) for the entire range (0 180 )?
I Here I may quote the statement made by Hardy5

The Greeks were the first mathematicians who are still real
to us today. Oriental mathematics may be an interesting
curiosity, but Greek mathematics is the real thing. . . .

G. H. Hardy, A Mathematicians Apology Cambridge, 2nd ed. (1967) p. 80.

Proof of Bhaskaras formula

In the figure, the area of the triangle ABC can

be expressed in two ways:
A =

or

1
1
AB.BC = AC.BD
2
2

1
BD

AC
AB.BC

(9)

Since the length of the chord < that of the arc, (9) may be expressed as an inequality

or

1
BD

>

1
BD

=
=

or

R sin

AC
_

AB . BC
x.AC

_ +y
_
AB . BC
2xR
+y
(180 )
(180 )
2xR + (180 )y

(10)

Proof of Bhaskaras formula

Substituting = 30 and = 90 in (10) we have,
9000
R
8100
2xR + 8100y =
R
2xR + 4500y =

(11)
(12)

Soving the above equations for x and y we have,

y =

1
4R

and

2xR =

40500
4R

(13)

Using the above values in (10), we have

R sin =

4(180 )R
40500 (180 )

(14)

which is the same expression given by Bh

askara.6

The above proof has been given by K. S. Shukla in his edition of the text
Mah
abh
askarya with translation and annotation.

The sine table given in S

uryasiddh
anta

(II. 1722)

The values presented are in Bh

utasany
a system and the same as that of Aryabhat
.ya
;a;naH
ta:va;a;a:(


x +:ta;a .+.pa;BUa;a;ma;Da:=;Ra;vaH
A;*:+.
/ /
Da;k
a;a;b.

*.a;ZUa;nyea;Za;a ba;a;Na:+.pa;gua;Nea;nd;vaH
Ka;a;*:+.
a;;Ea :pa:

*.Ea;k+:aH
ZUa;nya;l+.ea;.ca;na;pa:

;a;C+.d;+.pa;mua;na;a;nd;vaH

;a;naH
;
a;va;ya;a;nd+a;a;ta;Dxa;ta;ya;ea gua;Na:=+nDra;a;}ba:=+a;a:(


x +:ta;d;~:a;k+:aH

+;a;k

mua;a;na;Sa;q:a;ma;nea:a;a; a;Na . ca;nd+a;a;

;a;na;ga;a;a:(
;a;naH
*+:=+a;

a:(


*.a;a;;
a;va;Sa;ya;a;[a;a; a;Na k
u +.
:pa:

.=+nDra;pa:
*.a;a;;k+:ya;ma;aH

va;~va;d ;a;*:+.
ya;ma;a;~ta;Ta;a

x +:ta;a;;ZUa;nya:$va;l+.na;aH
k

+.$va;l+.na;aH
ya;ma;a;
a;d+va;
a;

.+.pa;a;a;
+;a;sa;a;ga:=+gua;Na;aH

+:yaH
na;a;ga;a;
a;d+Za; a;Za;va;

+:yaH
. ca;nd+nea:a;a;a;
+;a;va;

:Sa;f
, :pa:
*.a;l+.ea;.ca;na;gua;Na;aH

.=+nDra;ZUa;nya;a;NRa;va;a;
+;a;yaH

x +:ta;va;
+:yaH
va;~va;a;
+;a;k

225, 449, 671, 890, 1105, 1315, 1520, 1710 . . . 3178, 3256, 3321, 3372, 3409, 3431,
3438.

How to find intermediate sine values?

Prescription of first order interpolation (based on trair
asika-ny
aya)

;a;l+.a;a;Bya;~ta:va;nea:a;a;a;aH

ga;ta;a .$ya;aH

ga;ta;ga;}ya;a;nta:=:*+;a;a
&+.ta;a;~ta:va;ya;mEaH

Zea;Sa;taH
;a;[a;pea;t,a

:pua;naH
2

d;eaH k+:ea;a;f:$yea na;yea;de;vMa .$ya;a;Bya;(

a;a;pMa ;
a;va;pa;yRa;ya;a;t,a

Suppose we want to find the

jy
a of where
i 225 < < (i + 1) 225.

Let = i 225, whose jy

a is
known (Ji ).

Now, jy
a is given by
jy
a = Ji +

(Ji+1 Ji )
.
225

Improvised values of tabular Rsines

I Govindasvamin (9th cent.) in his commentary on Mah
abh
askarya

presents a set of corrections to the values of the tabular Rsine

presented by Aryabhat
. a.
I He observes:
ma;K.ya;a;d;ya;ea ;
a;h nyUa;na;a;a;Da;k+:a;va;ya;va;aH .tea;na .$ya;a;.cCe+.d;
a;va;Da;a;na;a;t,a
A;va;ya;va;a;va;ga;a;taH ta;Ta;a . ca A;va;ga;ta;a O;;tea ta;tpa:=+a;d;a;aH

*.a;vea;d;aH
/ /
Da;nea:Ma mua;a;na;pa:
+;a:=+nDra;a; a;Na ;
a;va;ya;
.sa;a;a;a;
u +Na;a;gMa nea:=+a;a;b.
d
;a;[ya;;yaH

x +:Za;a;nuaH
:Sa;NNa;ya;na;
a;d
+;a;BUa;tMa .=+
a;va;;k
:=+a;ma;a :vea;d;a;a;

.=+nDra;a;Bra;pa;[Ma gua;Na;a;pa;a;va;k+:a;;Ea . ca;[ua;

a;vRa;ya;tsa;a;Ka;.ca;nd+sUa;ya;RaH

.+:d+a;a;
+;a;.ca;nd+a ma;nua;sa;a;sa;ea;ma;a d;~:a;a;Br
a;nea:Ma na;ya;nMa ;
a;d
;sUa;yRa;m,a

A;a;[b.
/ / / / / / /
Da;pa;[Ma va;sua;nea:a:=+nDrMa . ca;nd+a;a;
+;a;
a;va;d;a;a va;sua;Ka;a;;.ca;nd+m,a

.=+nDrea;Sua :vea;dM na;va:+.pa;a;ma;D}.Ma Ka;a;Bra;a;

+;a;ya;ssa;a;gua;Nea;D}.a;sMa;K.ya;m,a

I The values listed in the verses are: 900 37000 , 700 30000 , 200 42000 , 400 57000 . . .

Improvised values of tabular Rsines

in min.
225
450
675
900
1125
1350
1575
1800
2025
2250
2475
2700
2925
3150
3375
3600
3825
4050
4275
4500
4725
4950
5175
5400

Aryabhat
.ya
225
449
671
890
1105
1315
1520
1719
1910
2093
2267
2431
2585
2728
2859
2978
3084
3177
3256
3321
3372
3409
3431
3438

R sin according to

Govindasvami
Madhava(also
Modern)
224 50 23
224 50 22
448 42 53
448 42 58
670 40 11
670 40 16
889 45 08
889 45 15
1105 01 30
1105 01 39
1315 33 56
1315 34 07
1520 28 22
1520 28 35
1718 52 10
1718 52 24
1909 54 19
1909 54 35
2092 45 46
2092 46 03
2266 38 44
2266 39 50
2430 50 54
2430 51 15
2584 37 43
2584 38 06
2727 20 29
2727 20 52
2858 22 31
2858 22 55
2977 10 09
2977 10 34
3083 12 51
3083 13 17
3175 03 23
3176 03 50
3255 17 54
3255 18 22
3320 36 02
3320 36 30
3371 41 01
3371 41 29
3408 19 42
3408 20 11
3430 22 42
3430 23 11
3437 44 19
3437 44 48

Computation of tablular Rsines (Tantrasangraha)

;
a;va;a;l+.a;a;d;Za;k+:ea;na;a .$ya;a .=+a;Zya;;Ma;Za;Da;nuaH k+:l+.aH

;a;a;Ba;~ta;taH
A;a;d;a:$ya;a;Da;Ra;t,a ta;ta;ea Ba;
e .sa;a;DRa;de;va;a;a:(


tya;
e ;
a;d
;ta;a;ya;Ka;Nq+$ya;a ;
a;d
;ta;a;ya;a .$ya;a . ca ta;d;au;a;taH
ta;ta;~tea;nEa;va h;a:=e+Na l+.b.DMa Za;ea;DyMa ;
a;d
;ta;a;ya;taH

Ka;Nq+.a;t,a txa;ta;a;ya;Ka;Nq+$ya;a ;
a;d
;ta;a;ya;~ta;d;au;ta;ea gua;NaH
txa;ta;a;yaH

.~ya;a;t,a ta;ta;(
Ea ;vMa . ca;tua;Ta;Ra;d;a;aH

k
+:ma;a;d
, gua;Na;aH

I The first pin

a is often taken to be 2250 based on the approximation,
. d.ajy
R sin R = 2250 .

(15)

I In contrast to the above, it is stated to be 2250 1000 = 2240 5000 .

I This is based on the approximation sin

R sin

21600
2

3
. Thus we have,
3!



3

= 224.83890 2240 5000 .

6

(16)

Computation of tablular Rsines (Tantrasangraha)

;
a;va;a;l+.a;a;d;Za;k+:ea;na;a .$ya;a .=+a;Zya;;Ma;Za;Da;nuaH k+:l+.aH

;a;a;Ba;~ta;taH
A;a;d;a:$ya;a;Da;Ra;t,a ta;ta;ea Ba;
e .sa;a;DRa;de;va;a;a:(


tya;
e ;
a;d
;ta;a;ya;Ka;Nq+$ya;a ;
a;d
;ta;a;ya;a .$ya;a . ca ta;d;au;a;taH
ta;ta;~tea;nEa;va h;a:=e+Na l+.b.DMa Za;ea;DyMa ;
a;d
;ta;a;ya;taH

Ka;Nq+.a;t,a txa;ta;a;ya;Ka;Nq+$ya;a ;
a;d
;ta;a;ya;~ta;d;au;ta;ea gua;NaH
txa;ta;a;yaH

.~ya;a;t,a ta;ta;(
Ea ;vMa . ca;tua;Ta;Ra;d;a;aH

k
+:ma;a;d
, gua;Na;aH

I The above verses essentially present the following equations for generating the
successive jy
a values.
Ji+1 = Ji + i+1
i+1 = i

Ji
233 12

(0 i 23)
(1 i 23),

I Since 1 = J1 , is known, all the jy

as can be generated using the above
equations recursively.

(17)
(18)

Evolution in the recursion relation

I From the time of Aryabhat

. a (499 CE ), the astronomers have been
using the recursion relation, to obtain the values of the 24
Rsine differences.
I Either these values themselves, or the jy
a values have been listed

explicitly in the form verses.

I The exact recursion relation for the Rsine differences is:

i+1 = i Ji 2(1 cos )

( = 225)

(19)

I While, 2(1 cos ) = 0.0042822, the different. approximations that

have been employed are:

1
225
1
2(1 cos )
233.5
1
2(1 cos )
233 32
60
2(1 cos )

0.0044444

(Aryabhat
. a)

0.0042827

(Nlakan.t.ha)

0.00428204

nkara
(Sa

Variyar).


Rsine table due to Madhava
The values presented are in Kat.apay
adi system and are correct to thirds
(rea;M na;a;ma va;a:=++a;na;Ma ;
a;h;ma;a;
a;d+veRa;d;Ba;a;va;naH
ta;pa;na;ea Ba;a;nua;sUa;+.:]a;ea ma;Dya;mMa ;
a;va;a:;d
d;ea;h;na;m,a 1

;a;Da;ga;a:$ya;ea na;a;Za;nMa k+:M C+.a;Ba;ea;ga;a;Za;ya;a;a;}ba;k+:a

//

mxa;ga;a;h;a:=+ea na:=e+Za;ea Y;yMa va;a:=+ea .=+Na:ja;ya;ea;tsua;k H 2

mUa;lM ;
a;va;Zua:;d
M na;a;l+.~ya ga;a;nea;Sua ;
a;va:=+l+.a na:=+aH
A;Zua;a:;d
;gua;a;a . ca;ea:=+(ra;aH

Za;*:
u +.k+:Na;eRa na;gea:(;a:=H 3

;Da;a:=+ea yua;va;a k+:Ta;a;l+.ea;l H

:pUa:$ya;ea na;a:=+a:ja;nEa;BRa;gaH

k+:nya;a;ga;a:=e na;a;ga;va; +:a :de;va;ea ;

a;va:(;a;~Ta;l+.a Bxa;guaH 6

ta;tpa:=+a;
a;d;k+:l+.a;nta;a;~tua ma;h;a:$ya;a ma;a;Da;va;ea;
a;d;ta;aH
.~va;~va;pUa;vRa;
a;va;Zua:;d
e tua ; a;Za;;a;~ta;tKa;Nq+.ma;Ea;
a;vRa;k+:aH 7

I+.a;ta

I The first and the last values are: 2240 5000 22000 and 34370 44 48000 .
I These values have been arrived at by considering terms up to 11 in the series

expansion of sin ascribed to Madhava.

Comparing the jy
a values of different texts

Aryabhat
uryasiddh
anta, Tantrasangraha,

Laghuvivr.tti, Madhava
& modern
.ya, S

Dhanu or C
apa
Symbol
Length
used
(min)
S1
225
S2
450
S3
675
S4
900
S5
1125
S6
1350
S7
1575
S8
1800
S9
2025
S10
2250
S11
2475
S12
2700
S13
2925
S14
3150
S15
3375
S16
3600
S17
3825
S18
4050
S19
4275
S20
4500
S21
4725
S22
4950
S23
5175
S24
5400

Notation
used
J1
J2
J3
J4
J5
J6
J7
J8
J9
J10
J11
J12
J13
J14
J15
J16
J17
J18
J19
J20
J21
J22
J23
J24

As in
AR/SS
225
449
671
890
1105
1315
1520
1719
1910
2093
2267
2431
2585
2728
2859
2978
3084
3177
3256
3321
3372
3409
3431
3438

Value of the jy
a (in minutes, seconds and thirds)
From
From
Given by
Modern
TS
LV
M
adhava
224 50
224 50 22
224 50 22
224 50 21
448 42
448 42 58
448 42 58
448 42 57
670 39
670 40 16
670 40 16
670 40 16
889 44
889 45 17
889 45 15
889 45 15
1105 00
1105 01 41
1105 01 39
1105 01 38
1315 32
1315 34 11
1315 34 07
1315 34 07
1520 26
1520 28 41
1520 28 35
1520 28 35
1718 49
1718 52 32
1718 52 24
1718 52 24
1909 51
1909 54 46
1909 54 35
1909 54 35
2092 42
2092 46 19
2092 46 03
2092 46 03
2266 35
2266 40 10
2266 39 50
2266 39 50
2430 45
2430 51 40
2430 51 15
2430 51 14
2584 32
2585 38 37
2584 38 06
2584 38 05
2727 14
2727 21 31
2727 20 52
2727 20 52
2858 15
2858 23 42
2858 22 55
2858 22 55
2977 02
2977 11 30
2977 10 34
2977 10 33
3083 03
3083 14 23
3083 13 17
3083 13 16
3175 53
3176 05 07
3176 03 50
3176 03 49
3255 06
3255 19 50
3255 18 22
3255 18 21
3320 24
3320 38 11
3320 36 30
3320 36 30
3371 27
3371 43 24
3371 41 29
3371 41 29
3408 05
3408 22 20
3408 20 11
3408 20 10
3430 07
3430 25 35
3430 23 11
3430 23 10
3437 27
3437 47 29
3437 44 48
3437 44 48

Infinite series for the sine function

I

The verses giving the series for the sine function is7

;a;na;h;tya . ca;a;pa;va;geRa;Na . ca;a;pMa ta:a;tP+.l+.a;a;na . ca

h:=e+t,a .sa;mUa;l+.yua;gva;gERaH ;
a:a:$ya;a;va;gRa;h;tEaH k+:ma;a;t,a
. ca;a;pMa :P+.l+.a;a;na . ca;a;Da;eaY;Da;ea nya;~ya;ea;pa;yRua;pa;a:= tya:jea;t,a
.ja;a;va;a;yEa, .sa;*:" +.h;eaY;~yEa;va ;
a;va;d;a;a;na;tya;a;
a;d;na;a kx+:taH
I

N0 = R

D0 = 1
2

Ni+1 = Ni (R)2

N1 = R (R)

D1 = R 2 (2 + 22 )

.ja;a;va;a =

.ja;a;va;a;yEa = For obtaining the jva (Rsine)

N0
D0

Di = Di1 R 2 (2i + (2i)2 )

N3
N2
1
[N
D1 ( D2 { D3 . . . })]

Yuktidpik
a (16th cent) and attributed to Madhava
(14th cent. AD).

Infinite series for the sine function

I

Expressing the series using modern notation as described as

described in the above verse
Jv
a = R

R (R)2 (R)2
R (R)2
+ 2
...
2
2
R (2 + 2 )
R (2 + 22 ) R 2 (4 + 42 )

Simplifying the above we have

Jv
a = R

(R)3
(R)5
(R)7
+ 4
6
+ ...
2
R 6 R 6 20 R 6 20 42

Further simplifying


3
5
7
Jv
a=R
+

+ . . . = R sin
3!
5!
7!

Thus the given expression well known sine series.

How to use the series to evaluate Rsines?

I Though Madhava

came up with the infinite series, it would certainly be

impossible to use the series expansion

R sin = R

(R)5
(R)7
(R)9
(R)11
(R)3
+

+ ...,
2
4
6
8
3!R
5!R
7!R
9!R
11!R 10

for obtaining the values of Rsines.

I However, since the series converges pretty fastbecause of the

factorial in the denominatorit would suffice to use a few terms in it.

I Noting this, Madhava

gave explicit values of the magnitudes of the

terms starting in the reverse from sixth and up to the second in (the
RHS of) the above equation when the arc R = 54000 = 90 .
I These values are mentioned in Yuktidpik
a using the kat.apay
adi

notation in the following verse.

;
a;va;d
;Ma;~tua;a;ba;lH

k+:va;a;Za;a;na;.ca;yaH

.sa;va;Ra;TRa;Za;a;l+.a;~Ta:=H
///

;a;na;
a;vRa:;d
;a;*:+.
*:+.ga;
a;d;tea;Svea;Sua k
+:ma;a;t,a :pa:*.
na:=e+nd;+:
a;
a ;sua

How to use the series to evaluate Rsines?

Madhavas

Term
no. in

Sanskrit equivalent

value according

RHS

in kat.apay
adi

to Yuktidpik
a

VI

;
a;va;d
;a;n,a

tua;a;ba;lH

IV

k+:va;a;Za;a;na;.ca;yaH

III

.sa;va;Ra;TRa;Za;a;l+.a;~Ta:=H
///

II

;a;na;
a;vRa:;d
;a;*:+.
,
na:=e+nd;+:k

0 00

0 0 44

000

00

16 05 41

0 0 44.54000
00 3300 5.6000

000

2730 5700 47000

0

00

value
0 00

00 3300 6000
0

Modern

000

2220 39 40

160 0500 40.87000

2730 5700 47.11000
22200 3900 40.10000

We find that the values given by Madhava

are indeed very accurate. Thus for
finding arbitrary value of sine we use the equation
R sin

R 3 (22200 3900 40000 ) + 5 (2730 5700 47000 ) 7 (160 500 41000 )

+ 9 (00 3300 6000 ) 11 (00 000 44000 ),

where =

R
.
5400

Concluding Remarks
I

It was quite interesting to know the evolution of different

techniques in India to evaluate the sine function which is
ubiquitous as accurately as possible.

Broadly speaking, the approach taken by Indian astronomers

and mathematicians can be classfied as:
I

I
I

to improve the accuracy of the sine table (which forms basis

for interpolation as well as sin(A + B) )
to obtain a good rational approximation
arrive at infinite series.

The absolute logical rigor with which the results have been
arrived at is indeed remarkable.

Why were they worried about very accurate values of sines ?

Accuracy of Trijy
a R Accuracy in the computation of sines
Accuracy in planetary positions Accuracy in the determination
of tithis, and so on, Avoid incompleteness.8

na;a;~tea k+:a;l+.a;va;ya;va;k+:l+.na;a

. . . (ra;Ea;ta;sma;a;tRa;v.ya;va;&+.a;ta:=+
a;pa

;a;C+.d;a;te
a ta:a ;Da;ma;RaH

Thanks!

T HANK YOU