fHA, xxxi (2000)

THE ASTROLABE AND SPHERICAL TRIGONOMETRY IN

MEDIEVAL INDIA

KIM PLOFKER, Brown University

1. SPHERICS AND THE INDIAN ASTRONOMICAL TRADITION

The Ptolemaic model of the universe relies on the techniques of Greek spherical
trigonometry to derive quantitative predictions from its complex mechanism
of uniform circular motions and nested spheres. Classical Indian astronomy,
on the other hand, while employing many of the same spheres and circles to
describe celestial motions, appears to have remained unaware of the corresponding
mathematical methods for the exact solution of spherical triangles. Their function is
for the most part amply fulfilled in Sanskrit astronomical texts by the use of plane
trigonometry in orthographic projections or analemmas, I in formulae apparently
derived by means of the relations between plane triangles inside the sphere,' and in
approximations assuming small spherical triangles to be plane.
The Islamic descendants of the Ptolemaic model inherited its spherical
trigonometry, which was further extended by the mathematicians of the medieval
Islamic world. Some of the rules thus developed were eventually incorporated
into Indian works dealing with western astronomy, particularly with that popular
astronomical device, the Islamic astrolabe.' But the earliest such work, which was
also perhaps the most influential, appears to misunderstand several of the rules in
question, blending the resulting incorrect formulae with various approximation
devices with an inclusive resourcefulness very characteristic of Sanskrit mathematics.
The consequent inconsistencies in the rules, their explanations, and the data indicate
that the elements of this new trigonometry upon the surface of the sphere were not
fully grasped by their Indian recipients.

2. THE YANTRARAfA

In A.D. 1370 a Jaina astronomer at the Delhi court of Firuz Shah Tughluq, Mahendra
Suri, wrote (in metrical Sanskrit stanzas, as was customary for medieval Indian
scientific works) the Yantraraja, a treatise on the astrolabe (Skt. "yantraraja", "king
of instruments"), for which his student Malayendu Suri later provided tables and a
commentary.' This appears to be the first Indian work to concern itself explicitly
with the methods of medieval Islamic astronomy. Mahendra claims to have written
the Yantraraja to provide an orderly and concise summary of the knowledge
concerning the theory and construction of astrolabes that was available to him in
a profusion of Islamic texts (although it does not seem likely, for reasons that will
become clear, that he himself possessed any great mastery of Arabic or Persian).

0021-8286/00/3101-0037/$2.50 © 2000 Science History Publications Ltd

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

38 Kim Plofker

In his opening verses, he states:"

apare samsare katikati babhuvur na caturah
param tair durbodham ganitam araci praudhamatibhih II
tatah svalpam saram visadam idam atyantasugamam
vitanve 'ham sastram suhrdayahrdanandakrtaye 11211
klrpas tatha bahuvidha yavanaih svavanya
yantragama nijanijapratibhavisesat II
tan varidhin iva vilodya maya sudhavat
tatsarabhutam akhilam pranigadyate 'tra 11311
In endless transmigration, some have not [yet] become clever; and [even] for
those with developed minds, mathematics is difficult to understand. Therefore
I rendered as a small epitome this brilliant, entirely complete treatise, to
gladden the hearts offriends.
There are very many treatises on instruments produced in their own language
by Westerners, each from the peculiarity of his own intelligence. So having
stirred them together like the oceans, I now proclaim the best part, as it were
the nectar, ofthat in its entirety. 6
Unfortunately, however, the nectar of knowledge that Mahendra was able
to chum from this ocean of learning remained somewhat contaminated by the
poison of ignorance; and the resulting procedures reveal a curious blend of
misinterpretation and re-interpretation according to approximative methods.
Furthermore, Mahendra's partial comprehension was evidently transmitted to his
student and commentator Malayendu Suri, and in the use of this widely popular
text during the subsequent centuries, so far as is known, these inconsistencies
were never remarked.

3. THE CALCULATION-BF EQUATORIAL COORDINATES

The particular application of spherics encountered in the Yantraraja involves

conversion from one system of celestial coordinates into another. To fix a star-
pointer in the astrolabe's rete, which contains a stereographic projection of the
celestial sphere onto the plane of the equator centred at the celestial pole(s), it
is necessary to calculate from the standard set of ecliptic coordinates (celestial
longitude and latitude, or Aand {3) two distances computed with respect to -the pole
of the equator: the 'true' declination r5 and (instead of the orthogonal right ascension
a) the polar longitude A, (see Figure I). Then the cosine of the star's declination r5
determines its radial distance from the centre of the rete (the segment OS in Figure
2), and the angular separation between this radius and the east-west line is a(A,),
the right ascension corresponding to the polar longitude.
Mahendra does not name any of the "very many" foreign sources for the
coordinate transformations in which the Yantrarajas rules originated, and I

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 39

equator
ecliptic

FIG. I.

have not been able to identify any Islamic text dealing specifically with astrolabe
construction that seems to be a likely candidate.' But the quantities 0 and A., are
employed in various contexts in zijes, and the formulae for computing them are
frequently stated and sometimes proved.

3.1. The Use % and L, in the Pre-Islamic Indian Tradition

In classical Indian astronomy, on the other hand, star coordinates are much less
rigorously determined;" and the methods routinely used for calculating them often
depend upon approximations. The true-declination calculations in general use in
Indian texts (for such purposes as the computation of conjunctions of stars and
planets) usually involve taking the sum or difference of some version of ecliptic
declination and latitude. Where the polar coordinates A., and f3, are known (for
example, in some star-lists), 0 is correctly computed? from them and from the
ecliptic declination 0 I of A., by

(1)

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

40 Kim Plofker

EI---~I----~Ff.+-"';"";:;';"'---+----1W

FIG. 2.

However, the polar coordinates themselves are generally produced by approximate

formulae. Occasionally, they are even used interchangeably with ecliptic
coordinates.'? Somewhat more correctly, A* can be derived from the so-called
"ayanadrkkarma" or visibility correction due to the deviation of the ecliptic: namely,
the determination ofthe ecliptic arc M" the longitudinal difference between the
ecliptic and polar longitudes. I I A typical formula for this quantity" is

M, '" [eYers (A + 90)· (Sin E)· f3]1R 2. (2)

This approximation can be explained with reference to Figure I. If a small circle
is drawn through B parallel to the equator, it will intersect arc SF in some point H,
and the arc BH is taken approximately equal to BA = M •. Then if the right triangle
SHB is considered to be plane, BH/f3 = Sin LHSB/R. It is clear that HSB is zero
when A = 90, and it is assumed that it is equal to the ecliptic obliquity E when A =
o (true only if f3 = 0). The Versine function is then selected to give interpolated
values between these extrema: Vers (A + 90)/R '" (Sin LHSB)/(Sin E), from which
Equation (2) follows. From this, f3. can be derived by assuming that f3* = SA '" SH,
and solving for SH by the Pythagorean theorem in the plane triangle SHB; and
finally, Equation (I) can be applied to find a correspondingly approximate value of

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 41

o. In other situations, it is even less rigorously assumed!' that

° "'OI(A)±/3.

3.2. The Yantraraja's Approacli to True Declination (0)

Mahendra's discussion of these coordinate transforms in his chapter on constructing
an astrolabe begins with the determination of the stellar declination 0, for which two
rules are given. The first of these (Yantraraja 1,41-45) is as follows:
mesadav uttaro golas tuladau daksinah smrtah II
trirasiyuktabahujya banakotijyaya hata 114 III
bhajyamtyaya tatas cape navateh sodhite punah II
gunena bhakta banajya praptam bhagadikam tatah 114211
syad amtaram parakramtis tena yukta tathonita II
gole bhabanayor aikye bhede sadhyas tato gunah 114311

hatah purvadhanurmaurvya bhaktas camtyajyaya tatah II

labdhac capam yutam hinam kartavyam ca sarena tat 114411

gole bhapattrinoraikye bhede syad apamah sphutah II

yamyasaumyavibhagena madhyarekhasamasritah 114511

The northern hemisphere is understood [to be] in [the six signs] beginning
with Aries, the southern in [the six signs] beginning with Libra. The Sine of
the arc [of longitude] increased by 90 degrees is multiplied by the Cosine of
the latitude, and divided by the last [Sine, i.e. R). When the arc from that is
subtracted from 90, the Sine ofthe latitude is divided by the Sine [of that arc).
The result from that is in degrees, etc.; it should be the "difference" [i.e., an
additive correction). The maximum declination [ecliptic obliquity, f] is increased
or diminished by that when [the longitude of] the star and the latitude are in
the same or different hemispheres [respectively). The Sine produced from that
is multiplied by the Sine of the previous arc and divided by the Radius. Then
the arc from the quotient is to be increased or diminished by the latitude
when [the longitude of] the star and the latitude are in the same or different
hemispheres. [That] should be the correct declination, either south or north
with respect to the equator.
This four-step procedure defines two new quantities here termed A' and /3', which
may be expressed in terms of the arcs shown in Figure I as follows:
Cos A' = [(Cos A) . (Cos /3)]/R, (3)
/3' = (Sin /3)/(Sin A'). (4)
Then it uses these quantities to obtain the declination by means of an additive

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

42 Kim Plofker

correction to the latitude:

Sin L\b = [(Sin (e ± {3'» . (Sin A')]/R, (5)
b={3±M. (6)
Malayendu's commentary provides a worked example of the procedure which
confirms this interpretation."
Mahendra's second rule for b follows immediately in verses 46-48. With slight
emendations to the sandhi of the printed text, it reads:
athavaniya bhakramtim naksatradhruvakat tatah ll
tatha sarasya samyogo viyogo 'traikyabhedatah 114611
tasmaj jiva hatotkrstakramtikotijyaya tatha II
uktabhakramtikotijya bhakta praptad dhanur bhavet 114711
etavat kramtibhagah syuh saumya yamya yatha vidhi II
kramtisayakayor yoge viyoge sesatas ca dik 114811

Or else, when one has found the declination of the star from its longitude,
then the sum or difference of the latitude and that [is made). depending
on the sameness or difference [of their directions]. Then the Sine of that
is multiplied by the Cosine of the maximum declination and divided by the
Cosine of the declination of the given star. The arc should be [computed}
from the quotient. Those are the degrees of declination [b]. north or south
according to the direction ofthe result ofthe sum or difference ofthe declination
and the latitude.
Malayendu's remarks and worked example accompanying these verses make it
clear that these instructions are equivalent to the expression
Sin b = [(Sin (f3 ± bl(A») . (Cos e)]/(Cos b/A». (7)

3.3. The Relation ofthe b-rules to Islamic Counterparts

The Yantraraja, although it claims to be an exposition of "Yavana" or Western
astrolabe theory, actually intermingles to some extent Islamic and traditional
Indian approaches, not always intentionally and not entirely successfully. The
first example of this is provided by the rule in Equations (3)-(6), which bears
an obvious similarity to the following procedure given by Ibn Yunis in his
Hakimi Zlj of 1003: 15
Cos A' = [(Cos A) . (Cos {3)]/R, (8)
Sin {3' = [(Sin {3) . R]/(Sin A'), (9)
Sin b = [(Sin (s ±{3'». (SinA')]/R. (10)
Mahendra's presentation accurately reproduces the first step from Equation (8), but
distorts the following ones. Evidently, the initial source of the corruption of the

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 43

Islamic rule was the loss of a factor of R from Equation (9). (Such computations
in zljes employing the standard value R = 60 often omit to specify explicitly
multiplication or division by R, merely lowering or raising the sexagesimal place
of the answer instead; and some such technique in Mahendra's ultimate source
may have assisted the misunderstanding.) The smallness of the result Sin p/Sin A'
then led Mahendra, or perhaps some earlier recipient of the corrupted formula, to
interpret it as an arc rather than a Sine. Similarly, the consequent smallness of the
result of Equation (5) suggested that it should be taken not as the desired final result
~ but as an additive correction to p. (The confusion may have been increased by
the deceptively similar use of the sum of p and the second declination ~2 in the rule
described below; also by the combining of latitude and ecliptic declination ~ I in the
traditional Indian approximation discussed above.) Having more or less saved the
phenomena in this fashion, the tinkerer left it at that, and it was this version that
was eventually preserved and studied by Malayendu.
The formula in Equation (7) is also identifiable with, but not identical to, a
"Yavana" counterpart. The version presented by Ibn Yunis" and later by TU~1
in the Zij-i Ilkhani" states:
Sin ~ = [(Sin <p ± ~/A») . (Cos e)]/(Cos ~/A», (II)

where ~2(A) is the so-called "second declination" or arc of the secondary to the
ecliptic passing through A between the ecliptic and the equator. The only difference
in Mahendra's version is the substitution of ~I for ~2 as the "declination of the
star" (an interpretation confirmed not only by Malayendu but by the fact that the
verses nowhere give a rule for computing ~). Did Mahendra Suri (or whoever was
originally responsible for its formulation) simply mix up the two declinations, or
did he deliberately replace the second by the first to simplify the computation?
The modified formula has the advantage of not requiring the calculation of
~2' which in any case is usually close in value to the ecliptic declination ~ ,.
Hence the originator of the rule in its present form made either a convenient
approximation or a very natural mistake, and it is probably not possible to say
with certainty which it was.
Given these discrepancies between the original formulations of the above b-rules
and their representations in the Yantraraja, it seems unlikely that Mahendra
was himself entirely competent to read and understand such sources as the
abovementioned zijes. Rather, he probably depended at least partially on the
assistance of translations, Muslim colleagues, and/or interpreters," one or more
of whom inadvertently misled him about some details of the procedures. In
short, these spherical trigonometry formulae were adopted, via any number of
intermediary interpretations, by astronomers whose own tradition supplied little
or no background in the underlying theory, and some corruption of the rules
was a natural result.
Possibly adding to the confusion was the fact that in many cases, the results
produced by the corrupted versions are not glaringly wrong. The rule in Equation

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

44 Kim Plofker

(7) is of course exact at the extrema A = 0 and A = 90 (since there 0 1 = 0), and
0
usually within 1 of the true value of 0 elsewhere. The first approximate formula,
from Equations (3)-(6), is exact only for f3 = 0, when 0= M = 0 1, Asf3 increases, the
accuracy of the rule deteriorates rapidly for A « 90. These discrepancies, though
uncorrected, did not go entirely unnoticed: Malayendu's commentary compares the
results of the Yantraraja's two formulae in the case of the star Rohint (ex Tauri)
and acknowledges the small difference between them, but adds that "because of its
smallness, it is not wrong". (This is by no means always the case; Malayendu has
been fortunate (or perhaps diligent) in selecting an example where the agreement
between the rules is good.) This remark seems to indicate that one if not both of
the formulae is expected to be imprecise. The value of 0 in the worked example
illustrating the method of Equation (6) also appears in Malayendu's star table;
therefore, at least some of these numbers were worked out by one of Mahendra's
methods rather than borrowed from a foreign table (as is the case with the ecliptic
coordinates).'? A recomputation of all the coordinates is required to determine
exactly how the calculations were made; but in any case it is clear that whatever
imprecision was noted by Mahendra or his pupil, this was not considered bad enough
to inspire a stricter examination of the rules' theoretical soundness.

3.4. The Yantrarajas Procedures for Polar Longitude (A,), and Their Relation
to Islamic Astronomy
The next step in the coordinate conversions of the Yantraraja is the determination
of the polar longitude A., traditionally accomplished in Indian astronomy by
the "ayanadrkkarma" procedure described above. Again, Mahendra gives two
different rules for this computation: the first is an accurate rendition of a solution
from spherics, while the second is a standard Indian formula that, however,
shows evidence of Islamic influence. The first iYantraraja 1,49-52) is explained
as follows:
drkkarmany ayanam marva mrgadyam atha karkajam II
satribhajyam parakramtimaurvya hatva vibhajitah 114911
uktodukramtikotyakhyamaurvya praptam kaladi yat II
tena hatvesukotijyam vibhajyamtyajyaya tatah 115011
labdhat kodamdakotijyam utpadya bhasarad gunah II
saqltadyoktaphalenoktadhanubkotijyayoddh[tab 115111
praptac capam nu drkkarmaphalam purvoktayuktitah II
When one has determined the ayana beginning with Capricorn or Cancer [i.e.,
whether 270 < A::; 90 or 90 < A::; 270J, [and] when one has multiplied the Sine
[of L] plus three signs by the Sine of the maximum declination, [the product]
is divided by the Cosine of the declination of the aforesaid star. Whatever the

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 45

quotient is in minutes etc., when one has multiplied the Cosine ofthe latitude by
that and divided by R, and then produced the Cosine ofthe arc from the result,
Sin fJ is multiplied by the previously stated result and divided by the Cosine
of the aforesaid arc. The arc from the quotient is the drkkarma correction.
According to the aforesaid rule, it is negative when the [ayana ofthe] star and
the latitude are in the same direction, positive when different.
This rather complicated but correct procedure directs the reader to make use of
two new quantities here called x and y, in a calculation equivalent to the following
sequence of equations, where the symbols-refer to arcs in Figure 3.
Sin x = [(Cos A.) . (Sin f)]/(Cos 15), (12)
Sin y =[(Sin x) . (CosfJ)]IR, (13)
Sin M. = [(Sin fJ) . (Sin x)]/(Cos y), (14)
A.. =,1, ±M•. (15)
The quantity A.. is well known in the zIj literature as the "culminating degree", or
the point on the ecliptic that crosses the meridian simultaneously with the given star,
but there it seems to be more usually computed by means of its right ascension;" I
have not yet encountered this particular formula in an Islamic text. Its validity can

equator
FIG. 3.

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

46 Kim Plofker

be confirmed in the context of medieval spherics by consideration of the spherical

triangles in Figure 3 centred on the star S (drawn so that the arcs SW, SX, and AY
all equal 90°). The two acute angles of the right triangle SBA containing the arc
.u.. are given by LASB =x and LSAB =y. In the spherical quadrilateral SXVP',
the Menelaos Theorem" implies that the product of the sines of three arc segments
with no endpoint in common equals the product of the sines of three other arcs
without a common endpoint, e.g., that
(Sin x)/(Sin VW) = [(Sin e)/(Sin VP')] [(R/(Cos b)].
The "rule of four quantities" (stating that for two right spherical triangles having
equality of one acute angle, the ratio of the triangles' hypotenuses equals the ratio
of their sides opposite that angle) implies that in the two right triangles VP'Wand
BP'(Cancer 0), whose acute angles at the vertex P' are equal,
(Sin VP')/(Sin VW) = R/(Cos A);
so by substitution
Sin x = [(Cos A) (Sin e)]/(Cos b).
By the same rule applied to other pairs of triangles,
(Sin y)/(Sin x) = (Sin P'S)/(Sin SW) = (Cos [3)/R,
(Sin [3)/(Cos y) = (Sin SA)/R, and
(Sin .u..)/(Sin x) = (Sin SA)/R;
these are easily combined to produce Equation (14). A somewhat similar approach
to computing a difference arc of the ecliptic from another set of coordinates is
given in the zIj of Habash al-Hasib."
The second rule for this quantity appears in Mahendra's two subsequent
verses:
purvavat satrirasijya taccharena hata hrta II
krtabhranarndavasubhir labdham bhagadikam tatah 115311
hatam niraksalagnena taddhisnyasthitarasinah II
hrtam khakhagnibhir bhagadikam drkkarma purvavat 115411
As before, the Sine [of A] plus three signs is multiplied by its latitude and
divided by 8904. The quotient is in degrees etc. Then it is multiplied by the
rising time of the sign in which that [star] stands, and divided by 300. The
drkkarma is in degrees etc., as before.
In other words,
.u.. = [((Cod) . [3)/8904] . [a/3OO], (16)

where a is the right ascension in vinadis (1/3600 of a day; thus one sign or 30°
corresponds to 300 vinadis) of the sign in which A stands. This rule is one of many
variants of the traditional "ayanadrkkarma" procedure illustrated by Equation (2)

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 47

above." Mahendra (or his source) has substituted the constant 118904 for the factor
Sin e/R 2 , and multiplied the whole by a/300. This second factor appears to be
intended to adjust the arc M, slightly by stretching it near the solstices and shrinking
it near the equinoxes, in accordance with the size of the right ascension of the
corresponding sign. Other authors use instead 1800/a, where is a in asus (sixths of a
vinadi);" it is not clear why Mahendra has inverted the factor.
Thus far, Equation (16) represents a standard Indian solution to this problem; but
the value 118904 reveals Islamic influence. It is not unusual" to replace the factor
Sin e/R 2 by a numerical constant depending on a specific value of the standard
radius R, but any of the typical Indian values" of R will produce a factor somewhere
between 118500 and 118000. It seems likely, therefore, that this modified formula
was adapted for use with the Islamic value R = 3600, which Mahendra uses in the
Yantraraja. Curiously, the value of the factor 1/8904 is still somewhat inaccurate
for Mahendra's value of e = 23;35°; it should be 1440;15/36002 "" 1/8998. It is
interesting to note that if e is taken as 23;51, Sin e /R2 = 1455;36,36/36002 "" 118903.
Using Sin e = 1455;36 or even 1455! would give the 118904of Mahendra's formula.
Did Mahendra adopt the rule of some earlier synthesizer who used the Ptolemaic
value" of e instead of the more common 23;35, without noticing the discrepancy
with the parameters of the Yantraraja't Another roughly contemporary Sanskrit
astrolabe text, the first chapter of the Yantraratnavali of Padmanabha (A.D. 1423),
also uses parameters implying a comparable value for e - in this case, about
23;50 - but like Mahendra, Padrnanabha names no Arabic or Persian source,
and seems not to notice that his stated value of e is at variance with this hidden
parameter." Again, the Yantraraja' s commentary does not concern itself with
tracing or proving the given formulae, confining itself to the usual gloss and worked
example. The results of Malayendu's application of each of Mahendra's two rules
to the coordinates of Rohini differ slightly, but as before, Malayendu concludes
that the discrepancy is small enough to neglect.

4. The Yantraraja's Rules in an Eighteenth-century Text

Later Sanskrit treatises on the astrolabe, as far as is now known, add little to the
exploration of these issues. The majority of such works discuss the mechanical details
of making and using the astrolabe rather than the trigonometric formulae underlying
its construction." These formulae seem to have been of such limited interest that
even some copies of the Yantraraja itself omit the chapter containing them." One
notable exception to this trend is the Yantrarajakalpa (or Yantrarajaghatanay'
written by Mathuranatha Sukla in 1782. Mathuranatha, like Mahendra, provides
rules for determining the star positions," and in fact, they are almost precisely those
given by Mahendra in Equations (3)-(6), and (12)-(15).
esam spastarasyadm asmac cakrad avadharya granthestavarsety uktaritya
yantranirmanakalikan vidhaya tatah krantim aniya drkkarrnasuddhan vidadhyat I

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

48 Kim Plofker

tadartham ca prathamatas tatkalikad istarksat krantim anayet Itadanayanaprakaro

yatha II
naksatradhruvakat trirasisahitad dorjyesukotijyaya I
nighna trijyakaya hrta dhanur atah sodhyam khanandamsakat II
bhajyas tasya gunena bhesujaguno labdhena yuktonita I
syatkrantih paramaikyabhinnahariti tryadharksatadbanayoh I1I11

tato gUQaQ purvaharajyaya hato hrt trijyaya labdhagunasya yaddhanuh II

hinanvitam bhe!?ulavairbhabaQayorbhinnaikadiktve 'pama Ihitodunah 11211

iti krantyanayanavidhih II atha drkkarmasodhanaprakarah II

tryadhyarksadhruvasifijinfha paramakrantijyaya ghni hrta I
svaspa!?tapamakoticapajaguQairaptena koterisoh II
jivasamgunita hrta paramaya labdhaddhanuh syat tatah I
kotijya hyanaya puraptahatabhesujya hrtaptaddhanuh 11111
idam bhabanayanadigbhidaikye dhanarnamuktam bhaganadhruvesu II
drkkarmasuddhani bhavanti bhani yantraya yogini tadakhilani 11211
Having determined the [longitudes] ofthose [stars] beginning with the correct
sign with respect to that [ecliptic] circle, [and] having established themfor the
time of the construction of the instrument by means of the rule beginning "The
year given by the text... ", 33 [and] having calculated the declination[ s] from
that, one should correct them by the drkkarma. For that purpose, one should
first calculate the ecliptic declination for the given star at that time. Then the
procedure is as follows: The Sine from the longitude of the star increased by
three signs is multiplied by the Cosine ofthe latitude and divided by the Radius.
The arc from this is to be subtracted from ninety degrees. The Sine of the
star's latitude isto be divided by the Sine of that; the maximum declination is
increased or decreased by the quotient when the directions of[the longitude of]
the star plus three signs and its latitude are the same or different, [respectively).
The Sine from that is multiplied by the Sine from the previous divisor and
divided by the Radius. The arc of the resulting Sine is diminished or increased
by the degrees oflatitude when the directions of[the longitude of] the star and
its latitude are different or the same. [That] is the declination ofthe given star.
Thus the rule for the calculation ofdeclination.
Now, the methodfor the drkkarma-correction. The Sine of the star's longitude
increased by three [signs] is multiplied by the Sine ofthe maximum declination
and divided by the Cosine of its own corrected declination [0]. The Cosine
of the latitude is multiplied by the quotient and divided by the greatest [Sine,
i.e. R). The arc from the quotient should be [made]. The Cosine from that

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 49

[is computed], and the Sine of latitude, multiplied by the previous quotient.
is divided by that. The arc from the quotient [is made]. This is stated [to be
applied] positively or negatively to the stellar longitudes. as the latitude and
ayana of the star are in different directions or the same direction. [Thus]
the stars are corrected by drkkarma; [when] affixed to the instrument [in the
reteJ, they are complete.
The only substantial difference between these rules and those of the Yantraraja
is that Mathuranatha (or his intermediary source) has managed to distort Equation
(5) even further by making the addition or subtraction of E and f3' dependent on the
direction not of Abut of A + 90°. Moreover, although he omits the rule in Equation
(7), he retains the otherwise unnecessary requirement of computing 0). In the
worked examples immediately following the above explanation, Mathuranatha
also follows Malayendu in choosing the coordinates of RohiIJ.I to demonstrate
these procedures. Thus, although the Yantrarajakalpa must have had other Islamic
or Islamic-derived sources as well," it is clear that at least some of it is merely
borrowed from the Yantraraja, and borrowed without any very close scrutiny of the
material. This is very much in keeping with the pragmatic approach of Mahendra
Suri and his pupil four centuries earlier when confronted with the new problems
and methods of astrolabe construction: the necessary rules were appropriated even
if not well understood, and any accompanying inconsistencies lost in the blurring
of the distinction between approximation and error. The theoretical justification
of the rules was a matter of much less concern to the practical astronomers who
first transformed them into Sanskrit, and apparently it did not gain significantly in
importance during the subsequent four hundred years.

APPENDIX: MALA YENDU' S COMMENTARY ON THE QUOTED VERSES

Yantraraja 1,41-45
When the epoch position of the star at [the time of] the beginning of [this] book [as listed in I, 22-37]
stands in the six signs beginning with Aries, [that] is the northern hemisphere; when standing in the
six signs beginning with Libra, [that] is the southern hemisphere, From that, when one has set down
the longitude of the star on a tablet, and added three signs [900 ] in the signs place, the Sine should be
calculated as before. That is the "star-Sine". And it is called the "Sine of [longitude plus] three" and the
"Sine of longitude", It is multiplied by the Cosine of the latitude and divided by the Radius, 3600, When
one has calculated the arc from the result, and subtracted [it] from 90, then the quotient from the Sine of
the star's latitude divided by the Sine produced from that remainder is called the "difference", beginning
with degrees. The degrees of the maximum declination, 23;35,0, are increased or diminished by that, the
sum [being] when the aforesaid epoch longitude is in the same direction as the latitude standing beneath
it, and the difference in the case of difference in their directions, [respectively], Then after one has
computed the Sine of that [and] multiplied it by the Sine computed from the previously calculated arc, it
should be divided by the Radius, 3600. Then whatever is the quotient is the Sine of the declination. Then
after one has computed the arc from the quotient, that arc is added or subtracted, as before, when there is
similarity or dissimilarity of the hemispheres of stars, [i.e. their location in longitude and in latitude] in
the part north or south of the equator; [thus] the degrees of declination are corrected.
Now an example: In the year Samvat 1427 the longitude of Rohini (in the northern hemisphere on
account of its standing in [the six signs] beginning with Aries), corrected for precession, [was] 2,1;33,52,

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

50 Kim Plofker

[its] south latitude 5; 10. And when the longitude, 2, I;33,52, is increased by three signs, there results
5,1;33,52. From this, according to [the rule] beginning "Up to three [signs], [an arc] is not to be altered
[i.e., is already reduced to the first quadrant] ..." [Yantraraja I, 10], the Sine of the arc is computed.
Since the arc [reduced to the first quadrant] is 0,28;26,8, the result found beneath the degrees of arc (28)
in the Sine table is 1690;6. The [interpolation-]difference beneath that is 55;13. When the [fractional
part] in minutes and seconds (26,8) standing beneath the degrees of arc is multiplied by that difference
and divided by 60, and the obtained result 24;<4>3 is added to the previous result 1690[;6], the resulting
Sine of the arc is 1714;9. (And it is always [done] in this way.)
Now the Cosine of the latitude: As the south latitude of Rohini is 5; 10, this is subtracted from
90; there results 84;50. The Sine is computed from this as before: 3585;25 [error for 3585; 18]. The
previously-computed Sine of the arc, 1714;9, is multiplied by that, and there results 6145941 ;58. When
it is divided by the last Sine [R], there results 1707;12. Here it is assumed that this itself is a Sine.
Now the calculation of [its] arc: Since the result standing beneath 28° in the Sine-table is 1690;6, it
is subtracted from the previously-computed Sine beginning with minutes (1707;12). The remainder is
17;6. The 28° of arc thus obtained are accompanied by zero [as a fractional part]. After the remainder
(17;6) is multiplied by 60, it is divided by the difference standing beneath 28° (55;13). The resulting
quotient in minutes etc. is 18;34 [error for 18;35]. This is the resulting arc from the Sine: 28; 18,34. (The
arc is always to be calculated in this way.) When this arc is subtracted from 90, the remainder beginning
in degrees is 61;41,26. From this, as before, the Sine is 3169;20. When the previously-computed Sine
(334;10 [misprint for 324;10]) of the star's latitude (5;1[0]) is divided by this, the difference in degrees
etc. is 0;6,8. Then, because of the difference in direction of [the longitude of) the star Rohint (added to
<or subtracted from> three signs) and its latitude, there is dissimilarity of its hemispheres. Therefore
the degrees of the maximum declination 23;35,0 are decreased by that difference 0;6,8 - there results
23;28,52. From this, as before, the Sine is 1434;22. This is multiplied by the Cosine of the arc [i.e., of
the first arc found, the complement of A'] (3169;20), and there results 4545986;15. It is divided by the
final Sine (3600). The quotient is the Sine of the declination, 1262;45 [should be 1262;46], and its arc is
calculated as before: 20;33, I [33 is a misprint for 32; the seconds value should be 5 or 6 depending on
which value of the above Sin b is used]. The degrees of declination are to be known as [equal to] this
much. Now when the latitude 5;10 is subtracted from the arc (because of the difference in direction of [the
longitude of) the star Rohini increased by three signs and of its latitude), the remaining degrees of north
declination of the star Rohini are correct: 15;22, I. By so many degrees, the star Rohini stands to the north
of the equator. The calculation of the declination of all stars is done by this method.
[When the above procedure is repeated using correct values in place ofMalayendu's slightly erroneous
ones, and following his example in rounding all arc-calculations to two sexagesimal places and all
Sine-calculations to one, it yields a final value of 15;22,7° for the declination of Rohini.]

Yantraraja 1,46-48
"Or else": by another method, when one has ascertained the hemisphere as before and has computed
the arc from the star's epoch longitude, the [ecliptic] declination is computed from that. The sum of that
declination with the latitude, when they are in the same direction, or their difference when in different
directions, is made. And when one has computed the Sine as before from that sum or difference in
degrees, and multiplied that by the Cosine of the maximum declination, and found the Cosine of the
previously computed declination, this [product] is divided by that [Cosine]. Then whatever the quotient
is, its arc is obtained as before. That arc is just the degrees of declination of the star in question.jwhich]
are north or south. How is their direction [determined]? It is said: When the declination and the latitude
have the same direction, the direction of the declination of the star is to be known from the sum. When
the latitude and declination have different directions, when the lesser is taken from the greater, the
direction of the declination of the given star is from the remainder.
Now an example: The epoch longitude of the star Rohini, corrected by precession (in the northern
hemisphere, because of its being in [the hemisphere beginning with] Aries), is 2, I;33,52. The latitude is
southern, 5; 10. Now from its being less [than 90°], this longitude is just the arc; the degrees of the arc
are 61;33,52. Their declination is computed as [follows]: The result obtained from what is beneath the
degrees of arc (61) in the tables of <Sine and> declination-results is 1228;57. The difference beneath

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 51

that is 12;14. When the [amount] in minutes and seconds standing beneath the degrees of arc, 33;52, is
multiplied by that difference and divided by 60, the quotient in minutes etc. is 6;54. When [that] is added
to the previous result, 1228;57, there results the declination in minutes, 1235;51. (The declination is
always to be found in this way.) When these are divided by 60, the quotient is the northern declination,
beginning with degrees: 20;35,51. The southern latitude, in degrees etc., of the star Rohini is 5; 10. Now
when the difference of these two is taken (because of their difference in direction), the latitude being
subtracted from the larger declination, the "declination-remainder" is northern: 15;25,51. From this the
Sine is found as before: 957;50. This is multiplied by the Sine (3299; 12) of the complement (66;25) of
the maximum declination (23;35), by making a "gomutrika" [multiplication procedure]: 3160083;44
[is the result]. When it is divided by the Cosine (3359;54 [error for 3369;45]) of the declination found
for that (20;35,51), the quotient in two [sexagesimal] places is 937;52. The ere from that, as before,
is 15;5,51. This arc is just the degrees of declination; and they are to be understood as northern,
because the previously-computed declination-remainder was northern. In the procedure for declination,
there is some difference in minutes and seconds between the method previously stated and this one;
because of its smallness, it is not wrong.
[Malayendu's last two calculations are rather puzzling; if the erroneous Cosine 3359;54 is used to
calculate Sin 0, the final answer ought to be 15;8,42. If the entire procedure is performed accurately,
the resulting value of 0 is 15;5,59.]

Yantraraja 1,49-52
In the calculation of the drkkarma, when one has found the ayana - north or south, [depending on the
star's position] in [the semicircle of the ecliptic] beginning with Capricorn or Cancer [respectively]-
of the star's epoch longitude; and, as before, has calculated the Sine of the arc after it is increased by
90 0 ; and multiplied it by the Sine (1440;15) of the maximum declination (23;35); and divided that by the
Cosine of the previously-computed [true] declination; and multiplied the Cosine of the latitude by the
result obtained; [then the resulting amount] should be divided by the Radius, 3600. Then, when one has
computed the arc and its Cosine (for the purpose of [future] division) from the number [thus] obtained,
and computed the Sine of [the star's] latitude and multiplied it by the aforesaid result, it is divided
by the Cosine of the previously-computed arc. Then the arc is to be calculated from the quotient,
and that is the drkkarma result. As before, it is negative in the case of similarity between the star's
ayana and its latitude (because of [their being] in the same direction), and positive in the case of
their dissimilarity (because of difference in direction). When that result is applied negatively or
positively, as stated, to the star's epoch longitude at the commencement of the book in the given
year, that star is drkkarma-corrected.
Now. an example: Since the epoch longitude of Rohini, corrected by precession, in the northern
ayana (from its beginning with Capricorn) is 2, I ;33,52 and its southern latitude 5; 10, the Sine of the arc
(computed as before) from Rohini's epoch longitude added to three signs is 1714;9. This, multiplied
by the Sine (1440; 15) of the maximum declination (23;35), becomes 2468804;32. By dividing by
the Sine (3471;1 [0]) of the complement (74;37,59) of the previously-computed true declination of
Rohini, [one obtains] the quotient 711; 15. The Sine (3471;1) of the complement (84;50) of the latitude
(5;10) of Rohini is multiplied by that; there results 2550067;36. When it is divided by the final Sine
3600, the quotient is 708;22. The arc from that, as before, beginning with degrees, is 11;21,5. The
complement is 78;38,[5]. From this, as before, the Sine is 3529;27. This Cosine of the arc is set down
separately for use in a subsequent division.
Now the latitude is 5;10. 'From this, as before, the Sine is 324;10. This is multiplied by the previous
result 711 ;15, and there results 230563;32. When it is divided by the Cosine of the previously-calculated
arc, 3529;27, the quotient is 65;28. The arc from this, as before, is 1;2,21 [error for I ;2,31]. This is just
the drkkarma-result, When one has made it positive in view of the difference of the two (on account of
the southern latitude and the northern ayana of the star Rohini), it is added to Rohini's epoch longitude.
There results Rohinis drkkarma-corrected epoch longitude, 2,2;36,23. The drkkarma-calculation
of all stars is [done] in this way.
[The arithmetic errors in this example are numerous and inconsistent, but if the procedure is done
correctly, the value of M, is 1;2,22].

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

52 Kim Plofker

Yantraraja 1,53-54
When one has ascertained the ayana as before, the "Sine [of longitude] increased by three signs"
derived from the epoch longitude of the star is multiplied by its latitude and divided by 8904. The
quotient from that is in degrees etc. When that is multiplied by the "ascension at zero latitude" (the
vinadis of the risings at Lanka of Aries etc. are 278, 299, 323, and so forth), [i.e.,] by the amount of
the rising at Lanka of [the sign] beginning with Aries associated with the given star, and divided by
300, the quotient in degrees etc. is just the drkkarma-result. That is applied negatively or positively,
as before, to the given epoch longitude.
Now, an example: Since the longitude of Rohini corrected by precession, 2, I ;33,52, is in the northern
ayana beginning with Capricorn, and [its] latitude is southern (5;10), the Sine of the arc [plus] three
signs, as before, is 1714;9. This is multiplied by the latitude, 5;10, and there results 8856;26. This is
divided by 8904, and the quotient beginning with degrees is 0;59,40. That is multiplied by the amount of
the right ascension of Gemini (because of the star Rohini's standing in the sign Gemini), 323, and there
results 321; 12,20. That is divided by 300, and the quotient beginning with degrees is 1;4,14. That is just
the drkkarma-result, As before, it is added to [the longitude of] the star Rohinl, because of the difference
[in direction] between its latitude and its ayana. The drkkarma-corrected [longitude of] the star Rohini is
2,2;38,6. Here there is some difference between this and the previously-computed drkkarma-corrected
[longitude], I minute and 43 seconds; because of its smallness, it is not wrong.
[The recomputation of Malayendu's example yields a value for M, of 1;4,16.]

Acknowledgements
I am grateful to Professors S. R. Sarma of Aligarh Muslim University and David
Pingree of Brown University for introducing me to the study of the "yantraraja" and of
Mahendra's text, and in particular to Professor Pingree for his helpful comments on an
earlier draft of this paper. srtgurudevacaranarpanam astu.

REFERENCES
I. The Indian development of this Hipparchan technique is discussed in O. Neugebauer and D. Pingree,
The Paiicasiddhantika (Copenhagen, 1970-71), II, 41--44.
2. Several such formulae, particularly as represented in the work of the Madhava school in late-medieval
Kerala, are discussed by R. C. Gupta in "Solution of the astronomical triangle as found in
the Tantrasangraha (A.D. 1500)", Indian journal of history of science, ix (1974), 86-99, and
in "Madhava's rulefor finding angle between the ecliptic and the horizon and Aryabhata's
knowledge of it", in History of oriental astronomy, ed. by G. Swarup et al. (Cambridge, 1987),
199-202. As Gupta points out, many of these rules are indistinguishable from ones derived via
spherics per se, and the typical lack of any proof in the authors' presentation of them renders it
impossible to determine exactly how they were worked out by their inventors. But the absence
of explicit reference to the solution of spherical triangles in Indian mathematical works makes
it most likely that such rules were estabished from various ingenious manipulations of triangles
within the sphere rather than on its surface.
3. Some aspects of the Indian reception of the astrolabe are discussed in S. R. Sarma, "Astronomical
instruments in Mughal miniatures", Studien zur Indologie und Iranistik, xvi (1992), 235-76.
4. See D. Pingree, "History of mathematical astronomy in India", in Dictionary of scientific biography,
xv, 533-633, p. 626; Sarma, op. cit. (ref. 3), 238-9; and Y. Ohashi, "Early history of the
astrolabe in India", Indian journal of history ofscience. xxxi (1997),199-295.
5. Yantraraja I, 2-3. The edition used is The Yantraraja, ed. by K. V. Raikva (Bombay, 1936). The
work was previously edited by S. Dvivedi (Benares, 1883).
6. The reference is to the Puranic legend of the gods' churning of the oceans to obtain the divine nectar
of immortality; see, e.g., Visnupurana IX, 81-98.

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

 The Astrolabe in India 53

7. The concise treatises of Pseudo-Masha'allah (in a Latin version; see R. T. Gunther, "Messahalla on
the astrolabe", in Early science in Oxford, ed. by R. T. Gunther, v (Oxford, 1929), 137-231,
and P. Kunitzsch, "On the authenticity of the treatise on the composition and use of the
astrolabe ascribed to Messahalla", Archives internationales d'histoire des sciences, xxxi
(1981),42-62), ibn 'Isa (c. Schoy, '''Ali ibn 'Isa, das Astrolab und sein Gebrauch", Isis, ix
(1927), 239-54), al-Biruni (A. Dallal, "al-Biruni's Book of pearls concerning the projection
of spheres", Zeitschrift fur Geschichte der arabisch-islamischen Wissenschaften, iv (1987/88),
81-138), and Tusi (whose Btst bab, in the Sanskrit version of Nayanasukha, is published in
Yantrarajavicaravimsadhyayt, ed. by V. Bhattacarya (Benares, 1979)) confine themselves
to practical procedures for making, marking, and manipulating the astrolabe, rather than
calculating coordinates. al-Btrunt's longer work, Istt'ab al-wujuh al-mumkinah fl san 'al
al-asturlab, may prove more comprehensive.
8. See D. Pingree and P. Morrissey, "On the identification of the yogataras of the Indian naksatras",
Journalforthe history of astronomy, xx (1989), 99-119.
9. Brahmagupta, for instance, prescribes this rule in Brahmasphut asiddha nta 10, 15-16
(Brahmasphutasiddhanta, ed. by S. Dvivedi (Benares, 1901-2)).
10. Pingree and Morrissey, op. cit. (ref. 8), Ill.
II. It is generally employed in the calculation of heliacal risings and settings of the planets;
Mahendra uses the term to include the A,-calculations for stars, whether by Indian or
foreign methods.
12. This is the standard form of the rule in the Aryapaksa or school of Aryabhata - see, for example,
Aryabhattya 4, 36 (Aryabhattya. ed. by K. S. Shukla and K. V. Sarma, i (New Delhi, 1976)). A
version frequently seen in the rival school of Brahmagupta, the Brahrnapaksa, substitutes the
Sine function for the Versine in this formula, as in Brahmasphutasiddhanta 6, 3.
13. C/. Brahmasphutasiddhanta 7, 5 and Lalla'sSi!iyadhlvr:ddhidatantra 9,1-2 (Si!iyadhlvr:ddhidatantra
of Lalla, ed. by B. Chatterjee (New Delhi, 1981)).
14. See the Appendix for translations of this and the following glosses and examples by Malayendu,
as they are rendered in the published edition; square brackets indicate editorial insertions or
comments, and angle brackets deletions.
15. See D. A. King, "The astronomical works of ibn Yunus", PhD. dissertation, Yale University,
1972, 295-6. Although, as was pointed out earlier, the Yantraraja has not been directly
linked to this or any other particular Islamic source, it is interesting that both Ibn Yunis
and Malayendu Suri choose (l Tauri (Aldebaran, SkI. Rohini) as an example to demonstrate
the o-calculations.
16. King,op. cit. (ref. 15),293-5.
17. See Javad Hamadani-Zadeh, "Nasir ad-Din on determination of the declination function", in History
of oriental astronomy, ed. by G. Swamp et al. (Cambridge, 1987), 185-9.
18. Little is known at present about the specific works used to teach Islamic astronomy in India in this
early period; see, however, D. Pingree, "Islamic astronomy in Sanskrit", Journal for the history
of Arabic science, i (1978), 315-30, and S. R. Sarma, "Yuntraprakara of Sawai Jai Singh",
Studies in history of medicine and science, x-xi (1986,1987), (supplement) 1-139.
19. Pingree and Morrissey, op. cit. (ref. 8), liS.
20. See, e.g., King, op. cit. (ref. 15), 302-5, and M.-Th. Debarnot, Kitab maqalid 'ilm al-hay'u:
La trigonometric spherique chez les Arabes de l'est a la fin du X" siecle (Damascus,
1985),216-18.
21. A more detailed explanation of the spherical trigonometry used here is given in, e.g., Edward
S. Kennedy, 'The history of trigonometry", in E. S. Kennedy et al., Studies in the Islamic
exact sciences (Beirut, 1983), 1-29.
22. Debarnot, op. cit. (ref. 20), 52-53.
23. In this form, it more closely resembles Brahmagupta's version (see ref. 12); Mahendra's use of Sin
O. + 90) = Cos Awould therefore seem to place him among the followers of the Brahmapaksa,

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015

54 Kim Plofker

who tend to use the Sine as the interpolating function in this procedure in preference to the
Versine favoured in the Aryapaksa,
24. See, e.g., Srtpati's Siddhantasekhara, 9, 6 tSiddhantasekhara, ed. by B. Misra (Calcutta, 1932,
1947»; Bhaskara's Siddhantasiromani, Ganita 7, 4 (Siddhantasiromani, ed. by B. D.
Sastri (Benares, 1967; repro as Kashi Sanskrit Series 72, Benares, 1989». Munjala uses
an equivalent expression in Laghumanasa 52 (Laghumanasa, ed. and transl. by N. K.
Majumdar (Calcutta, 1951».
25. See, e.g.,SiryadhlV1:ddhidatantra 8, 3, and Vatesvara's vatesvarasiddhanta. 6,10 (Vatesvarasiddhl1nta,
ed. and transl. by K. S. Shukla (New Delhi, 1986».
26. Most commonly R = 3438 (cf. A ryabhatt ya Gitika 12); other well-known values are 3270
tBrahmasphutosiddhanta 2, 5) and 3415 (Siddhantasekhara 3, 6). If the standard Indian
obliquity, e = 24°, is used in place of the Yantrarajo's smaller value, then the increased amount
of the factor Sin fIR' diverges still further from Mahendra's number.
27. This is the obliquity used in the zjj of al-Khwarizrni (c. A.D. 840; see E. S. Kennedy, A survey of
Islamic astronomical tables, in Transactions of the American Philosophical Society, ns., xlvi/2
(1956),123-77 (p. 148», but I know of no other zfj that uses it as a table parameter, although
Ptolemy's value was certainly well-known to Islamic astronomers.
28. Y. Ohashi, "A note on some Sanskrit manuscripts on astronomical instruments", in History of
oriental astronomy, ed. by G. Swarup et al. (Cambridge, 1987), 191-5; see also Ohashi,
op. cit. (ref. 4), 225.
29. For example, the discussions of the astrolabe in the works associated with the court of the
eighteenth-century ruler Sawai Jai Singh - the Yantrarajaracano (ed. by K. Jyotirvid (Jaipur,
1953», the Yantraprakara (Sarma, op. cit. (ref. 18», the Yantrarajavicaravimsadhvayt
(Bhattacarya, op. cit. (ref. 7», and the Yantraprabha (published with the Yantrarajaracana
in Jyotirvid, op. cit.) of Srmatha - all neglect these theoretical considerations, as does the
earlier Yantrasiromani of Visrarna (published with the Yantraraja in Raikva, op. cit. (ref.
5)). It would be interesting to know whether the work of Padrnanabha mentioned above
gives a fuller treatment of them.
30. Such abbreviated versions are preserved in the manuscripts Benares Hindu University B.3318
and B.521.
31. See Y. Ohashi, "Sanskrit texts on astronomical instruments during the Delhi Sultanate and Mughal
periods", Studies in history of medicine and science, x-xi (1986-87),165-81, p. 175.
32. All references to the Yantrarajakalpa are based on the manuscript Benares 35245; the rules in
question appear on ff. 21v-22v.
33. This refers to the for-mula for correcting celestial longitudes for precession given earlier by
Mathuranatha on f. l3v.
34. Witness Mathuranatha's use of e = 23;30,17 (f. 23r) instead of Mahendra's value.

Downloaded from jha.sagepub.com at NORTHERN KENTUCKY UNIV on June 17, 2015