ब्राह्मस्फुटसिद्धान्त (ब्रह्मगुप्त - शून्य, कुट्टक, बीजगणित एवं सम्पूर्ण २१ अध्याय सान्वय सटीक)

258 BRAHMAGUPTA AS AN ALGEBRAIST this type was a problem, write Datta and Singh, which could not be solved completely nor satisfactorily by Brahmagupta. In fact, Brahmagupta had to depend on trial. Success in this direc- tion was, however, remarkably attained by Bhāskara II. He evol- ved a simple and elegant method which assisted in deriving an auxiliary equation having the required interpolator ±1, ±2, or ± 4, simultaneously with its two integral roots, from another auxi- liary equation empirically formed with any simple integral value of the interpolator, positive or negative. This method has been technically known as Cakravāla or the cyclic method. This is so called because it proceeds as in a circle, the same set of opera- tions being applied again and again in a continuous round. For the details of this method, our reader is requested to consult the Algebra of Bhāskara II and the narrative on this method as given by Datta and Singh under the title "Cyclic Method" in their History of Hindu Mathematics: Algebra, 1962 Edition, pp. 161-72. Solution of Indeterminate Quadratic Equation It is remarkable to see that Brahmagupta was the first algebraist in the history of mathematics to find a general solu- tion of the indeterminate quadratic equation Nx²±c=y² in positive integers. We have the following verse in the Brāhmasphuṭasiddhānta in this connection: From two roots (of a Square-nature or varga-prakṛti) with any given additive or subtractive, by making (combination) with the roots for the additive unity other first and second roots (of the equation having) the given additive or subtractive (can be found).¹ Let us take the following two equations: a₁k=an+b; and b₁k=bn+Na From them we get : by eliminating n a₁b−ab₁=1

1. रूप प्रक्षेपपदे पृथगिष्टक्षेपशोध्यमूलाभ्याम् । कृत्वाऽऽन्त्याथपदे ये प्रक्षेपे तेने ॥ —BrSpSi. XVIII. 66

SOLUTION OF INDETERMINATE 259 Hence b₁ = (a₁b - 1) / a = a whole number. Now n² - N = ((a₁k - b)² - Na²) / a² = (a₁²k² - 2bka₁ + k) / a² = k(a₁²k - 2ba₁ + 1) / a² Therefore (k / a²)(a₁²k - 2ba₁ + 1) is a whole number. Since a, k have no common factor, it follows that (a₁²k - 2ba₁ + 1) / a² = (n² - N) / k = k₁ = an integer. Also k₁ = (n² - N) / k = (a₁²k - 2ba₁ + 1) / a² = (a₁²(b² - Na²) - 2ba₁ + 1) / a² = ((a₁b - 1) / a)² - Na₁². Thus having known a single solution in positive integers of the equation Nx² ± c = y², says, Brahmagupta, an infinite number of other integral solutions can be obtained by making use of the integral solutions of Nx² + 1 = y². If (p, q) be a solution of the former equation found empirically and if (α, β) be an integral solution of the latter, then by the principle of Composition x = pβ ± qα; y = qβ ± Npα will be a solution of the former. Repeating the operations, we can easily deduce as many solutions as we like. FORM Mn²x² ± c = y² : In this connection, Brahmagupta says : If the remainder is that divided by a square, the first root is that divided by its root¹. This seems to mean that if we have the equation Mn²x² ± c = y² (i) such that the multiplier (i.e. the coefficient of x²) is divisible

1. वर्गेच्छिन्ने गुणके प्रथमं तन्मूल भाजितं भवति । —BrSpSi. XVIII.70

260 BRAHMAGUPTA AS AN ALGEBRAIST by n², then we are justified in saying that if we put nx=u, the equation (i) becomes Mu²±c=y² (ii), and clearly the first root of (i) is equal to the first root of (ii) divided by n. The corresponding second root will be the same for both the equations. FORM a²x²±c=y² : We find Brahmagupta giving the following rule in this connection : This is a solution of a particular form of a varga-prakṛti or Square-nature. If the multiplier be a square, the interpolator divided by an optional number and then increased and decreased by it, is halved. The former (of these results) is the second root; and the other divided by the square-root of the multiplier is the first root.¹ Thus the solutions of the equation a²x²±c=y² are : x = 1/(2a) (±c/m - m) y = 1/2 (±c/m + m) where m is an arbitrary number. Bhāskara II and Nārāyaṇa have also given the same solutions as proposed by Brahmagupta. Rational Geometrical Figures In the days of the Taittirīya Saṁhitā and the Śatapatha Brāhmaṇa, Indian mathematicians got familiarity with the solution of such equations x²+y²=z² and the results were arrived geometrically on the basis of the law of rectangle as propounded by Baudhāyana in the Śulba Sūtras and which goes by his name. The reader is referred to the Chapter on Baudhāyana, the first Geometer in the author's "Founders of Sciences in Ancient India". Baudhāyana (c. 800 B.C.) gave a

1. वर्गे गुणके क्षेपः केनचिदुद्धृतयुतोनितो दलितः । प्रथमोऽन्यन्मूलमन्त्यो गुणकारपदोद्धृतः प्रथमः ॥ —BrSpSi. XVIII. 69

, is there a space or a matra? Let's compare त in कोटिः or तद्दलं: त has a left loop. In आयतचतुरस्रस्य: आ य त च तु र स्र स्य. No, it's आयतचतुरस्रस्य. Wait, let me re-examine: आयताचतुरस्रस्य? Wait, look at त: the right side of त goes down, then there is a च. It's आयतचतुरस्रस्य. Wait, what about चेष्टस्येष्टाधिका? Let's look at चे ष्टस्येष्टाधिका: चे has a space after it? Wait, look at the footnote line 1: 1. इष्टस्य भुजस्य कृतेर्भक्तो नेष्टेन तद्दलं कोटिः । Line 2: आयतचतुरस्रस्य चे ष्टस्येष्टाधिका कर्णः ॥ —BrSpSi. XII. 35 Wait, look at चे ष्टस्येष्टाधिका: There is चे, then a space, then ष्टस्येष्टाधिका. Wait, could it be चेष्टस्येष्टाधिका with just a broken space? Yes, in De

262 BRAHMAGUPTA AS AN ALGEBRAIST The sum of the squares of two unequal numbers is the side; their product multiplied by two is the altitude, and twice the difference of the squares of those two unequal numbers is the base of an isosceles triangle.¹ Thus if m,n be two integers such that m is not equal to n, the sides of all rational isosceles triangles with integral sides are given by m² + n², m² + n², 2(m² - n²) and the altitude of the triangle is 2mn. This method was also followed by Mahāvīra and other Indian mathematicians. In fact, their solutions are based on the juxtaposition of two rational right triangles, equal so that they have a common leg. It is remarkably a powerful device, for every rational triangle or quadrilateral may be formed by the juxtaposition of two or four rational right triangles. Isosceles Triangles with a Given Altitude Here we have a rule given by Brahmagupta for finding out all rational isosceles triangles possessing the same altitude : The (given) altitude is the producer (karaṇī). Its square divided by an optional number is increased and diminished by that optional number. The smaller is the base and half the greater is the side.² Thus if m be any rational number then for a given definite altitude a, the sides of the rational isosceles triangles are ½ (a²/m + m) each and the base is (a² - m²)/m . We shall illustrate it by an example taken from the commentary of Pṛthūdaka Svāmī The given altitude is 8; let us take any rational number m = 4 then the two equal sides of the isosceles are given by ½ ((8² + 4²)/4) = 10 each and the base is (8² - 4²)/4 = 12. Thus the three sides of the

1. कृति युतिर सदृशराश्योर्बाहुर्घातो द्विसंगुणो लम्बः । कृत्यन्तरमसदृशयोर्द्विगुणं द्विसमत्रिभुज भूमिः ॥ —BrSpSi. XII. 33
2. करणी लम्बस्तत्कृतिरिष्टहतेष्टोन संयुताऽल्पा भूः । अधिको द्विहृतो बाहुः संक्षेप्यो यद्वधो वर्गः ॥ —BrSpSi. XII. 37.

ISOSCELES TRIANGLES 263 rational isosceles triangle with altitude 8 are (10,10, 12). Rational Scalene Triangles: Brahmagupta lays down the following rule in the case of rational scalene triangle : The square of an optional number is divided twice by two arbitrary numbers; the moieties of the sums of the quotients and (respective) optional numbers are the sides of a scalene triangle; the sum of the moieties of the differences is the base.¹ In other words, if m, p, q are any rational numbers, then the sides of a rational scalene triangle are : ½ ( m²/p + p ), ½ ( m²/q + q ), ½ ( m²/p - p ) + ½ ( m²/q - q ) Here the altitude (m), area and segments of the base of this triangle are all rational. Thus putting m=12, p=6, and q=8 in Brahmagupta's gene- ral equation, Pṛthūdaka Svāmī derives a scalene triangle with sides (13,15) and (14) altitude (12), area (84 and the segments of the base (5) which are all integral numbers. ½ ( m²/p + p ) = ½ ( 12²/6 + 6 ) = 15; ½ ( m²/q + q ) = ½ ( 12²/8 + 8 ) = 13 B' A C' A' B' C' B H C H B C Fig. 19 Fig. 20

1. इष्टद्वयेन भक्तो द्विधेष्ट वर्ग फलेष्टयोगार्धे । विषमत्रिभुजस्य भुजाविष्टोनफलार्धयोगो भूः ॥ —BrSpSi. XII, 34.

264 BRAHMAGUPTA AS AN ALGEBRAIST Thus the two sides of the rational scalene triangle are 15 and 13. The base is . ½ ( 12²/6 - 6 ) + ½ ( 12²/8 - 8 ) = 9 + 5 = 14 The altitude is m = 12; area is equal to (base × altitude) / 2 = (14 × 12) / 2 = and the segments are √(13² - 12²) = 5 and √(15² - 12²) = 9. Thus they are all integers. Rational Isosceles Trapeziums Brahmagupta has given us a method of obtaining such isosceles trapeziums whose sides, diagonals, altitude, segments and area are all rational numbers. His rule is as follows : The diagonals of the rec- tangle (generated) are the flank sides of an isosceles trapezium; the square of its side is divided by an optional number and then lessened by that optional number and divided by two; (the result) increased by the upright is the base and lessened by it is the face.¹ Fig. 21 Here in the figure, we have the isosceles trapezium ABCD of which C D is the base and A B is known as the fase. Accord- ing to Brahmagupta's rule, we have ( p being the optional number). CD = ½ ( (4m²n² - p) / p ) + ( m² - n² ) (base) AB = ½ [ 4m²n²/p - p ] - [ m² - n² ] (face) DH = ( m² - n² ) (upright)

1. आयतकर्णौ बाहू भुजकृतिरिष्टेन भाजितेष्टोना । द्विहृता कोट्यधिका भूर्मुखमूना दिसमचतुरस्रे ॥ —BrSpSi. XII. 36

RATIONAL TRAPEZIUMS 265 AD = BC = m²+n² (the sides of the trapezium) HC = base-upright = ½ [ 4m²n²/p - p ] (segment) AC = BD = [ 4m²n²/p + p ] (diagonal) AH = 2 mn (altitude) ABCD = mn [ 4m²n²/p - p ] (area) By chosing the values of m n and p suitably, the values of all the dimensions of the isosceles trapezium can be made integral. Pṛthūdaka Svāmī starts with the rectangle (5, 12, 13) and suitably takes p as 6; then he calculates out the dimensions of the trapezium : flank sides (AD and BC) = 13, base =14, and base = 4, altitude (AH) = 12, segments of base (DH and HC) = 5, and 9, diagonals (AC and BD) = 15, area ABCD = 108. All these values are integers. In this example, the rectangle chosen is (5, 12, 13) which is AA' DH, where AD = m² + n² = 13 and DH = m² - n² = 5 whence by adding the two we have 2m² = 18 This gives the value of m = 3, and hence n = 2. Pṛthū- daka Svāmī has taken the value of p = 6 by choice. Putting these values of m, n and p, the values for the dimensions of the isosceles trapezium follow from the expressions given by Brahma- gupta. CD = ½ ( (4.3².2² / 6) - 6 ) + ( 3²-2² ) = 9+5 = 14 (base) Face = 9-5=4 Sides AD = BC = 3²+2² = 13 and so on for the other dimensions. Rational Trapeziums With Three Equal Sides This problem is very much the same as one of the rational isosceles tpapezium with the only difference that in this case one of the parallel sides is also equal to the slant sides. We

266 BRAHMAGUPTA AS AN ALGEBRAIST have the following solution of this problem from Brahma- gupta : The square of the diagonal (of a generated rectangle) gives three equal sides; the fourth (is obtained) by subtracting the square of the upright from thrice the square of the side (of that rectangle). If greater, it is the base; if less, it is the face.¹ As before, the rectangle generated from m, n is given by (m²–n², 2mn, m²+n²), that is these are the three sides of the right triangle, which correspond to the two sides and the diagonal of the rectangle generated by them. Let us suppose, we have a trapezium ABCD whose sides AB, BC and AD are equal, then AB = BC = AD = (m²+n²)² CD = 3(2mn)² – (m²–n²)² = 14 m²n² – m⁴ – n⁴ or CD = 3(m²–n²)² – (2mn)² = 3m⁴+3n⁴ – 10 m²n². Pṛthudaka Svāmī has taken an illustration, where m=2, n=1 and he deduces two rational trapeziums with three equal sides (25, 25, 25, 39) and (25, 25, 25, 11). The segment (CH), altitude (AH), diagonals (AC, BD) and area of this trapezium are also rational, and given by : CH (segment) = 6 m²n²—m⁴ – n⁴ AH (altitude) = 4 mn (m²—n²) AC = BD (diagonals) = 4 mn (m²+n²) ABCD (area) = 32m³n³ (m²—n²). Rational Inscribed Quadrilaterals We find in the Brāhmasphuṭasiddhānta a remarkable proposition formulated by Brahmagupta : To find all quadrilaterals which will be inscribable within circles and whose sides, diagonals, perpendi- culars, segments (of sides and diagonals by perpendi- culars from vertices as also of diagonals by their intersection), areas, and also the diameters of the

1. कर्णकृतित्रिसम भुजास्त्रयश्चतुर्थो विशोष्य कोटि कृतिम् । बाहुकृतेस्त्रिगुणाया यद्यधिको भूर्मुखं हीनः ॥ —BrSpSi. XII. 37

RATIONAL INSCRIBED QUADRILATERALS 267 circumscribed circles will be expressible in integers. Such quadrilaterals we shall call as Brahmagupta Quadrilaterals. The solution of this formidable problem has been given by Brahmagupta as follows : The upright and bases of two right-angled triangles being reciprocally multiplied by the diagonals of the other will give the sides of a quadrilateral of unequal sides : ( of these ) the greatest is the base, the least is the face, and the other two sides are the two flanks.¹ Taking Brahmagupta's integral solution, the sides of the two right triangles of reference are given by : (1) m²-n², 2 mn, m²+n²; (ii) p²-q², 2 pq, p²+q²; where m, n, p, q are integers. Then the sides of the Brah- magupta Quadrilateral are. (m²-n²)(p²+q²), (p²-q²)(m²+n²), 2mn(p²+q²), 2pq(m²+n²) (Arrangement A) Pṛthūdaka Svāmī has- illustrated the rational inscribed quadrilateral by taking an example of the right angle triangles. (i) (3,4,5) (m²-n²=3, m²+n²=5, whence m=2, n=1) (ii) (5,12,13)(p²-q²=5, p²+q²=13, whence p=3, q=2) Fig. 22 Substituting these values in the above equations, we get the sides of the quardilateral as ( 39, 25, 52 and 60).²

1. जात्यद्वय कोटिभुजाः परकर्णगुणा भुजाश्चतुर्विषमे । अधिको भूर्मुखंहीनो बाहुद्वितयं भुजावन्यौ ॥ —BrSpSi. XII. 38 2 The diagonals of this quadrilateral are given by Bhāskara II as 56 (=3.12+4.5) and 63 (=4.12+3.5). (Cont. on page 268)

270 BRAHMAGUPTA AS AN ALGEBRAIST Put in other words, this means that one has to solve the following equations : (i) 5x—25 = y² (ii) 10x—100 = y² (iii) 83x—7635 = y² Pṛthūdaka Svāmī, the commentator on the Brāhmasphuṭa- siddhānta proceeds to solve these equations as follows : (1.1) Suppose y = 10; then x = 125. Or put y = 5; then x = 10. (2.1) Suppose y = 10; then x = 20. (3.1) Assume y = 1; then x = 92. He then remarks that by virtue of the multiplicity of suppositions there will be an infinitude of solutions in every case, But no method has been given either by Brahmagupta or his commentator to obtain the general solution. Double Equations of the First Degree Perhaps we have the earliest reference of the simultaneous indeterminate quadratic equations of the type x ± a = u² x ± b = v² in the Bhakaśālī Manuscript (Folio 59, recto). Brahmagupta gives the solution of such simultaneous inde- terminate quadratic equations of a general case as follows : The difference of the two numbers by the addition or subtration of which another number becomes a sq- uare, is divided by an optional number and then incre- ased or decreased by it. The square of half the result diminished or increased by the greater or smaller (of the given number) is the number (required).¹ Expressed in the language of algebra, shall have : = ½ { ½ ( (a - b)/m ± m ) }² ∓ a

1. याभ्यां कृतिरधिको नस्तदन्तरं हृत युतो न मिष्टेन । तद्दल कृतिरधिकोऽधिकयो रविको न यो राशिः ॥ —BrSpSi. XVIII. 74

DOUBLE FIRST DEGREE EQUATIONS 271 or x = { ½ ( (a - b) / m ∓ m ) }² ∓ b where m is an arbitrary number. Datta and Singh has given the rationale of this method as follows : u² = x ± a; v² = x ± b, From them, we have u² - v² = ± a ∓ b Therefore u - v = m and u + v = (± a ∓ b) / m , where m is arbitrary. Hence u = ½ ( (± a ∓ b) / m + m ) = ± ½ ( (a - b) / m ± m ) Sincc it is obviously immaterial whether u is taken as posi- tive or negative, we have u = ½ ( (a - b) / m ± m ) Similarly v = ½ ( (a - b) / m ∓ m ) Therefore x = { ½ ( (a - b) / m ± m ) }² ∓ a, or x = { ½ ( (a - b) / m ∓ m ) }² ∓ b, where m is an arbitrary number. Now we shall take up another particular case, for which Brahmagupta has given a rule : The sum of the two numbers tne addition and sub- traction of which make another number (severally) a square, is divided by an optional number and then diminished by that optional number. The square of half the remainder increased by the subtractive number is the number (required)¹. In the algebraic notations, we shall express it as follows :

1. यैरूनो यैश्च युतो रूपैर्वर्गस्तद्द्वयमिष्ट हृतम् । इष्टोनं तद्दल कृतिरूनाऽभ्यधिका भवति राशिः ॥ —BrSpSi. XVIII. 71

274 BRAHMAGUPTA AS AN ALGEBRAIST y = 1/a ((ad + bc)/m + b) if b > c and m > (ad + bc)/m. If these conditions be reversed then x and y will have their values interchanged. Datta and Singh have given the following rationale of these solutions : axy = bx + cy + d, or a²xy − abx − acy = ad, or (ax − c) (ay − b) = ad + bc. Suppose ax − c = m, a rational number; then ay − b = (ad + bc)/m. Therefore x = 1/a (m + c) y = 1/a ((ad + bc)/m + b) Or, we may put ay − b = m; in that case, we shall have ax − c = (ad + bc)/m; whence x = 1/a ((ad + bc)/m + c), y = 1/a (m + b). Brahmagupta’s own rule. Whilst the rule given above is ascribed to an unknown author, Brahmagupta’s own rule for the solution of a quadratic indeterminate equation involving a factum is as follows : With the exception of an optional unknown, assume arbitrary values for the rest of the unknowns, the product of which forms the factum. The sum of the product of these (assumed values) and the (respective) coefficients of the unknowns will be absolute quantities. The continued products of the assumed values and of the coefficient of the factum will be the coefficient of the optionally (left out) unknown. Thus the solution

BRAHMAGUPTA'S OWN RULE 275 is effected without forming an equation of the factum Why then was it done so ?¹ Datta and Singh think that the reference in the latter por- tion of this rule is to the method of the unknown author : “Kiṁ kṛtaṁ tadataḥ” ? The principle underlying Brahma- gupta's method is to reduce, like the Greek Diophantus (c.275 A.D.), the given indeterminate equation to a simple determi- nate one by assuming arbitrary values for all the unknowns ex- cept one. So undoubtedly it is inferior to the earlier method. We now take an illustrative example from Brahmagupta : On subtracting from the product of signs and degrees of the Sun, three and four times (respectively) those quantities, ninety is obtained. Determining the Sun within a year (one can pass as a proficient) mathema- tician. If we presume x to denote the signs and y the degrees of the Sun, then the equation would be : xy - 3x - 4y = 90 Pṛthūdaka Svāmī solves it in two ways : (i) Let us assume the arbitrary number to be 17. then x = 1/1 ( (90.1 + 3.4) / 17 + 4 ) = 10 y = 1/1 (17 + 3) = 20 (ii) Let us assume arbitrarily y = 20. On substituting this value of y in the above equation, we get 20x - 3x = 170 whence x = 10.

1. भावितके यद्घातो विनष्टवर्णेन तत्प्रमाणानि । कृतेष्टानि तदाहत वर्णैक्यं भवति रूपाणि ॥ वर्ण प्रमाणभावित घातो भर्वातेष्ट वर्ण संख्यैवन् । सिध्यति विनाऽपि भावित-समकरणात् किं कृतं तदतः ॥ —BrSpSi. XVIII. 62-63
2. भानो राश्यंशवधात् त्रिचतुर्गुणितान् विशोध्य राश्यंशान् । नवतिं दृष्ट्वा सूर्यं कुर्वन्नावत्सराद् गणकः ॥ —BrSpSi. XVIII. 61. — : o : — Reference H.T. Colebrooke : Algebra with Arithmetic and Mensuration from the Sanscrit of Brahmagupta and Bhas- cara, London, 1817. B. Datta and A.N. Singh : History of Hindu Mathematics Pt. I and II, 1962.

The provided image is a blank page containing no text or readable content.

CHAPTER X Arabic and Indian Divisions of the Zodiac I It has long been a debated question whether the Indian and Arabian divisions of the zodiac had a common origin. Sir William Jones thought that they had not; but Colebrooke holds a contrary view. The coincidence, in the two systems of division is so exact that he thinks, it could not be due to chance. Coleb- rooke has discussed this point in details in one of his Papers entit- led "On the Indian and Arabian divisions of the zodiac", Asiatic Researches Vol. ix.. p. 323-376, reproduced in the Miscellaneous Essays, Vol. II., p. 321-373, 1872.

1. Aśvinī, now the first nakṣatra, but anciently the last but one, probably obtained its present situation at the head of the asterisms, when the beginning of the zodiac was referred to the first degree of Meṣa (the Ram). As measuring a portion of the zodiac, it occupies the first 13°20' of Meṣa; and its beginning follows immediately after the principal star in the last nakṣatra Revatī, reckoned by some exactly, by others nearly, opposite to the very conspicuous one, which forms the fourteenth asterism. As a constellation, Aśvinī comprises three stars (Aries α, β, γ) figured as a horse-head; and the principal, which is also the northern one, is stated by all ancient authorities, in 10°N and 8°E. from the beginning of the Meṣa. According to Arabs, the first manzil or lunar mansion is entitled Sheraṭan (by Persians, Sheratain), and comprises two stars of the third magnitude on the head of Aries. in lat. 6°36' and 7°51' N and long. 26°13' and 27°7'. With the addition of a

278 ARABIC AND INDIAN DIVISIONS OF THE ZODIAC third, also in the head of the Ram, the asterism is denominated Āshrait. The bright star of the second or third magnitude which is out of the figure of the Ram, according to Ulugh Beg, but on the nose according to Hipparchus, cited by this author from Ptolemy, is determined Nātih : It is placed in lat. 9°30'N and long. 1 0° 43', and is apparently the same with the principal star of the Indian asterism; for Muhammad of Tizin, in his table of declination and right ascension, expressly terms it the first star of the Sheratain. 2. Bharaṇī, the second asterism, comprises three stars (35, 39, 41 Aries) figured by the yoni or pudendum muliebre and the principal and southern star of this nakṣatra is placed in 12°N. On the Arabian system, the second manzil, entitled Butain is placed by Ulugh Beg in lat. 1°12' and 3°12', and this cannot possibly be reconciled with the Indian constellation. But Muhammad of Tizin assigns to the bright star of Butain a decli- nation of 23°N exceeding by nearly 2° the declination allotted by him to Nātih or his first star in Sheratain. This agrees with the difference between the principal stars of Aśvinī and Bharaṇī; and it may be inferred, that some among the Mohammadan astrono- mers have concurred with the Hindus in referring the second constellation to stars that form Musca. 3. Kṛttikā, now the third, formerly the first, nakṣatra consists of six stars figured as knife or razor, and the principal and southern star is placed in 4½ or 5°N and in 65 sixths of degrees (or 10°50′) from its own commencement (cf. the Sūryasid- dhānta), or 37° 28′ to 38° from the beginning of the Meṣa (the Siddhānta-Śiromaṇi or the Graha Lāghava) respectively. This longitude of the circle of declination corresponds nearly with that of the bright star in the Pleiades, which is 40° of longitude distant from the principal star of Revatī. The stars indicated by Ulugh Beg for Thurayyā, also corres- pond exactly with the Pleiades. 4. Rohiṇī, is the fourth nakṣatra, the Arabic name for the fourth mansion is Debarān (or with the article Aldebarān). It corresponds to the bright star called the Bull's eye, and which is unquestionably the same with the principal and eastern star of Rohiṇī, placed in 4½° or 5°S and 49½°E by the Hindu writers on Astronomy. This nakṣatra is

figured as a wheel cart, and comprises five stars, out of the seven which the Greeks named the Hyades. The Arabs, however, like the Hindus, reckon five stars only in the asterism. Sir William Jones supposes them to be in the head and neck of the Bull; they probably are α, ρ, γ, δ, ε Tauri, agreeably to Mons. Bally's conjecture. 5. Mṛgaśirā, the fifth nakṣatra, represented by an antelope head, contains three stars; the same which constitute the fifth lunar mansion Hakāh; for the distance of 10°S assigned to the northern star of this nakṣatra, will agree with no other but one of the three in the head of Orion. The difference of longitude (24° to 25½°) from Kṛttikā corresponds with sufficient exactness; and so does the longitude of its circle of declination (62° to 63°) from the end of Rewatī; since the true longitude of λ Orionis, from the principal star in Revatī (ζ piscium) is 63½°. 6. Ārdrā, the sixth nakṣatra, consists of a single bright star, described as a gem, and placed in 9°S (by some in 11°) and at the distance of 4⅓ to 4° in longitude from the last asterism. This indicates the star in the shoulder of Orion (a orionis). The sixth lunar mansion is named by the Arabs as Hanāh; and comprises two stars in the feet of the second twin, according to Ulugh Beg, though others make it to be a shoulder. Mohammad of Tizin allots five stars to this constellation; and the Kāmūs, among various meanings of Hanāh, says, that it is a name for five stars in the left arm of Orion; remarking also, that the lunar mansion is named Tahāyī, comprising three stars called Tahyāt. Obviously here the Indian and Arabian asterisms are irreconcileable. 7. Punarvasū (used in a dual number) is the seventh nakṣtra, and is represented by a house, or even a bow, and it includes four stars, among which the principal and eastern one is 30° or 32° from the fifth asterism; but has been placed by all authorities in 6°N. This agrees with (β Geminorum) one of the two stars in the heads of the Twins, which together constitute the seventh lunar mansion ziraā, according to Mohammad of Tus and Mohammad of Tizin and other Arabian authorities. The seventh lunar mansion of Arabs is named ziraā ul ased according to Jauhari and other cited by Hyde in his Commen-

250 ARABIC AND INDIAN DIVISIONS OF THE ZODIAC tary on Ulugh Beg, and that the Kāmūs makes this term to be the name of eight stars in the form of a bow. 8. Puṣya, the eighth asterism, is described as an arrow, and consists of three stars, the chief of which being also the middle most, has no latitude, and is 12° to 13° distant from the seventh asterism, being placed by Hindu astronomers in 106° of longitude. This is evidently δ Cancri; and does not differ widely from the eighth lunar mansion Nethrah, which according to Ulugh Beg and others consists of two stars, including the nebula of Cancer. The Indian constellation comprises two other stars besides δ Cancri, which are perhaps γ and β of the same constellation. 9. Aśleṣā, the ninth asterism, contains five stars figured as a potter’s wheel, and of which the principal or eastern one is placed in 7°S, and according to different tables, 107°, 108° or 109° E. This appears to be intended for the bright star in the south- ern claw of Cancer (α Cancri), and cannot be reconciled with the lunar mansion Tarf or Tarfah, which comprises two stars near the lion’s (siṁha) eye, the northernmost being placed by Mohammad of Tīzīn in 24° of N. declination. 10. Maghā, the tenth asterism, contains like the last, five stars, but which are figured as a house. The principal of the Southern one has no latitude; and according to all authorities, has 129° longitude. This is evidently Regulus (α Leonis) : which is exactly 129½° distant from the last star in Revatī. The tenth lunar mansion of Arabians is Jebhah, which comprises three (some say, four) stars, nearly in the longitude of the lion’s heart. In this instance, therefore, the Indian and Arabian divisions of the zodiac coincide. This nakastra consists of α, γ, η and ν Leonis. 11. Pūrva-Phālgunī is the eleventh nakṣatra and is repre- sented by a couch or bedstead; it consists of two stars deter- mined by the place of the chief star (the northernmost, accord- ing to the Sūrya-Siddhānta) in 12°N and 144°E., or according to Brahmagupta, the Śiromaṇi and the Grahalāghava 147° or 148° E. They are probably δ and θ Leonis. The Arabian name for this lunar mansion is zubrah or Khertan.

ARABIC AND INDIAN DIVISIONS OF THE ZODIAC 281 It may be mentioned here that Brahmagupta and Bhāskara selected the southern for the principal star; while the Sūrya- Siddhānta took the northern. Hence the latitude stated by several Hindu authorities is the mean between both stars; and the difference of longitude, compared to the preceding and subse- quent asterisms, may be exactly reconciled upon this suppo- sition. 12. Uttara-Phalgunī, which is the twelfth nakṣatra, con- sists of two stars, and is figured as a bed or cot. These stars are ascertained by the place of one of them (the northernmost) 13°N. and 155° E. This indicates β Leonis; the same which singly constitutes the Arabian Lunar mansion Serfah, though Moham- mad of Tīzīn seems to hint that it consists of more than one star. 13. Hasta, the thirteenth nakṣatra, has the name and figure of a hand; and is suitably made to contain five stars. The principal one towards the west, next to the north-western star, is placed according to all authorities in 11° and 170° E. This can only belong to the constellation Corvus ; and accordingly five stars in that constellation (α, β, γ, δ and ε Corvi). The thirteenth lunar mansion of Arabs is Awwā, which is also described to contain five stars, situated under Virgo and so disposed as to resemble the letter Alif. They are placed by Ulugh Beg in the wing. Here obviously, there is nothing common between the Hindu and Arabian specification of the asterism. The agreement is only in the number of stars and in the longi- tude. 14. Citrā, the fourteenth nakṣatra, is figured as pearl. It is placed by the Sūrya-Siddhānta in 2° S and 180° E. and by Brahmagupta, the Śiromaṇi and Grahalāghava in 1¾ or 2°S and 183°E. This agrees with the Virgin's spike (α Virginis). The same star constitutes the fourteenth lunar mansion of the Arabs named from it Simāc ul aāzi. 15. Svāti, the fifteenth nakṣatra, is represented by a coral bead. The Sūrya-Siddhānta, Brahmagupta, the Śiromaṇi and Grahalāghava, all concur in placing it at 37°N. They differ one degree in longitude of its circle of declination, three of them