The yuga of Sphujiddhvaja and the Era of the Kusanas

121 The yuga of Sphujiddhvaja and the era of the K~apas Harry Falk 1. The basis The succession of the early KW}~a kings can now be regarded as certain, particularly, since the Rabatak 1 inscription gave us information on the first four kings, i.e. Kujula Kadphises, Vima Taksuma 2 , Vima Kadphises and Kaniska. According to this text Kaniska initiated an era. His predecessors had used the Azes era of 58/57 BC and supposedly another era of around 155 BC, usually associated with Menandef. From Kani~a onwards his own era was used for the majority of inscriptions in his realm, with the particularity that after 100 years the counting started anew, e.g., the cipher for the hundreds were dropped. This dropping was nothing but a theoretical possibility until it was given a solid basis in 1949 by J. van Lohuizen-de-Leeuw4 , strengthened with different arguments by Hartel 1996: 102f. It is now almost generally5 accepted by art- historians as well as numismatists. Numbers referring to the first or second century of the KW}~as can be found on hundreds of pieces of plastic art, many from the capital :Mathura, written in Brahmi, and many from GandMra, written in Kharosthi script. Despite our increased knowledge about the dynastic succession, the question. about the beginning a of the so-called Kaniska-era has not met with a generally accepted answer. There was ,conference· on this topic in 1913 at the Royal Asiatic Society in London a strong fraction favoured dates in 6; the first century BC. A second conference met in 1960 at the School of Oriental and African Studies, London7, where the strongest fraction favoured an equation with the Saka era starting in AD 78. The third round took place in 1996 in Vienna8 , where the majority of proposals were in favour of dates younger than Saka, preferably in the early decades of·the 2nd century AD. The basic argument for the identification with the Saka era was from the start in the 19th century up to . our days that only a powerful ruler would be able to install such a lasting era. 2. The Yavanajataka All these debates would have been umi.ecessary if a single inscription had turned up with dates in both the KW}~a and Saka eras. Maybe one day such an inscription will be found. In the meantime we may be satisfied to have one single literary evidence linking the KW}~a with the Saka era, thus providing us with the means to definitely pin down the beginning of a KW}~a century. This means is known since 1964, when D. Pingree published a first survey of the contents of the Yavanajataka, written by one Sphujiddhvaja9 in the year 191, certainly of the Saka era, thus in AD  122 S.R.A.A., VII (2001) 269. Pingree cites its verse 79,15 from the commentary of Utpala on B{hajjataka 7,9 in the following form: gatena slidhyardhasatena yuktlipyekena kosi'u;agatlibdas(JJJlkhyli, klil(JJJl sakliniJI:n parisodhya tasmlid atftam anyadyugaval$aylitam. Pingree translates this as: "(If one takes) the number of years of the Ka;il1;las which have passed and adds 149, then, subtracting from this number the year in the Saka era, (one obtains a year in which) another yuga ended." In a note Pingree evaluated the contents of this stanza: "From this verse it is clear that the Era of the K~il1;las, (i.e., of Kani,ska?) is not the same as the Saka Era; but this, unfortunately, is the only certain information that it provides. As the tenus of reference are not sufficiently well defined,- any attempt to interpret this obscure verse further would be futile; it can be made to support dates from the first century B.C. to the second century A.D." In 1978 Pingree published the full text, based mainly on one old manuscript from Nepal, which he labels N. In a note (408:15) he say that "this verse cannot be used to date theK~il1;las if one accepts the reading of N. But if one modifies Utpala's reading of theIastpada slightly, one can interpret the verse in the following way: ka;il1;lagatabdasqkha is the number of years between the epoch of the K~il1;las era and any date; and kala!) sakanap. is the number of expired years between the epoch of the Saka era and that date. If the elapsed years of the yuga are regarded as a negative . quantity, the epoch of the K~aoa era fell ·149-66 = 83 years before A.D .. 144. or in A.D. 61. If one accepted the late chronology for Kaui,ska, this K~il1;la era of A.D. 61 might have;been founded by Kujula Kadphises. But this line of interpretation is to the highest degree hypotheticaL" The topic was taken up by B.N. Mukherjee in 1982 where in the JAOS, in an article titled':The Yavanajataka of Sphujidhvaja, The Sakakala and the Kaui,ska Era", Mukherjee followsPingree's reasoning in its most essential aspect and expects "the epoch of this Ka;il1;la era ( ... ) "X" number of years earlier than (A.D. 145-82) = A.D. 63. Since the value of "X" is not known, we cannot fix more precisely the epoch of this Ka;il1;la era" (1982:362b). Nonethe1ess, he sees a possibility, - beyond my comprehension -, to maintain "the theory identifying the KaJ)i,ska.era with theSaka era of A.D. 78" (l982:363b). In the sequel, despite these publications none of the many authors dealing with KusaJ;la chronology even hinted at the existence of this text: Although text and ttanslation leave much to be desired Pingree's correct observation that Kusaua and Saka cannot be identical - if one is subtracted from the other - should have found its way into the permanently ongoing discussion. The Yavanajataka is a remarkable work, being the oldest extant text on astronomy in India, dealing very detailed in 78 chapters about the influence of the planets on the life of a' human being, generally at the time of its birth, and subsequently at any given point in time. Humans are regarded as completdy dependent on their personal horoscope and they can find their way through life best  Falk : The yuga of Sphujiddhvaja and the era of the K~~as 123 only if they act or don't act according to the advice of an astrologer. The last and 79th chapter is different. It presents a yuga of 165 years, in order to allow the astrologer to compute the position of the planets in retrospect or for future dates. This yuga of 165 years is one of the last of its kind; its raison d'etre is the belief that the major heavenly bodies meet at the same place again after a certain given time. Well-known is the Vedic yuga of 5 years, current even in the early centuries of our era. Another yuga is based on the Jupiter cycle, assuming similar constellations every 12 or 60 years. For Sphujiddhvaja a yuga completes after 165 years, exactly when sun and moon in conjunction enter the sign of Aries at sun-rise (mesasm:nkranti) thus marking the beginning of spring and the beginning of a new year. The text of all the chapters, and particularly of this last technical chapter, shows that its author formulates very carefully. He knows what he is talking about and avoids any ambiguity. Therefore it is remarkable that he should have given us such a useless formula to compute the beginning and present stand of his own yuga in terms of Saka and K1.1SW;la dates. The suspicion arises that the text could be corrupt, either improperly preserved by tradition or incorrectly presented in the edition. The latter consideration gains instant momentum when reading a review of Pingree's text and translation of some verses from the 79th chapter from the pen of K.S. Shukla, published in 1989. Shukla had to show that the numbers defining the Juga as given by Pingree are all wrong. I present some of the differences in the following table: Verse topic Pingree Shukla 79,6 tithis in a yuga 60.265 61.230 ~ ! , 79,7 civil days in a yuga 61.230 (tithis) 60.272 79,34 civil days in a solar year 365+14,47/60 365+4-7/165 79,11 civil days in a solar month 30;26,9,52,4 30;26,25,27,16 79,12 civil days in a synodic month 30;3,55,34 29;31,50,14,24 79,13 civil days in a sidereal month 27;17,10,34 27;19,18,39 The reason for the mistakes are found in the way Sanskrit prose expresses numbers aridPingree's unfamiliarity with some of the expressions. Shukla had to note several times that the correct reading of Ms N was found in the footnotes, whereas the text as printed was unnecessarily , emended. With this knowledge we can start to have a fresh look at the two stanzas in the Yavanajiitaka which allow us to co-ordinate a) our era, b) the Saka era, c) the K1.1SW;laera and d) the yuga of Sphujiddhvaja. The first relevant stanza is 79,14ab where b) Saka and d) yuga are synchronized.Pingreereads and translates as follows: gate sa4eke 'rdhasate saman&:n kCilakriyatattvam idOlll sakCinam,  1 i i· 124 S.R.A.A., VII (2001) "When 66 years of the Sakas have elapsed, that is the truth (i.e., foundation) of the calculation of time." In all the stanzas preceding this line the text deals with the details of the yuga. So, when asking what exactly is gata, "gone", it might either be years of the yuga or the years of the Saka era. Pingree expects the latter possibility without any attempt to discuss the first. There is another oddity. The ms. N - and all its offshoots - reads !ia4agre, which Pingree changed to !ia4eke for a technical reason: If we add 56 (i.e. sa4agre 'rdhasate) to the beginning of the Saka era in spring AD 78, we end in the year 134 AD, and there the moon is far away when the sun enters Aries. Pingree (1978:408) was looking for a likely year and found as the closest following date 22 March 144 AD. To adjust the possible date to the stanza he changed the text-typical -agre into an -eke, although Sphujiddhvaja never expresses decimals this way. If we blindly follow the text as it runs naturally in its contexts, and if we preserve the reading of the old manuscript, a completely different picture emerges from a verbal translation: "When 56 years (of the yuga) have gone, this is the state of (the sky leading to) the epoch of the Sakas." That means, the beginning of the yuga is located 56 years earlier than the creation of the Saka era, not 56 or even 66 year later, as in the interpretation of Pingree. This way we end in AD 22 for the beginning of the yuga; 56 years:lllter the Saka era was installed. When we look at the sky we see that on 21 March 22 AD the year starts with sun and moon and the first star of Aries all on one level [fig. 1]. Sphujiddhvaja seems to be right when he speaks in 79,14ab of the "state of the epoch", i.e. of the Sakas, and not, as he would have done otherwise, of "the beginning of the yuga". The following stanza 79,15 links b) the Saka era, c) the Kusfu.1a era and again d) the yuga.It comes as follows in Pingree's edition and translation: gatena sadhyardhasatena yukta vyekena kO!iGJ:tagatabdasaJ}1khya kO/am sakOndI:n parisodhya tasmad atftavmsa yugavmsajatah. "Take the number of years that have elapsed of the Kos fu.1as , add 149, and subtract from this (sum) the time of the S&kas (i.e., the year in the Saka era); (the remainder) is the number of years in the yuga which have elapsed." Pingree has seen that the formula according to his own interpretation leads us nowhere, because of the too many uncertainties involved. If we take a closer look at the inner structure of his own formula we see that it produces a completely wrong sort of result. As a result of the computation the text promises the number of elapsed years of a current yuga, and this number should increase by 1 every year. That means, we need an increasing number. According to Pingree's interpretation we start from an unknown number of years, elapsed, in the Kusfu.1a system. This number increases by 1 every year. To this number we have to add a fixed amount, i.e. 149. From the result we are --  Falk : The yuga of Sphujiddhvaja and the era of the Kll1i~aS 125 I 5" I en us Cet set Ari I 15" '15" 90 120 i Ir - - - - - - - - - - - - " ' ' - - - - r - - - ' - - - - - - ' - - - - - - - - - - - L - STARS SYMBOLS N"wD,,/hi • <1 • 3.5 *o Multiple star .~: Dark nebula 6 Radio source Saturday 21. March 22 AD • • • 1.5 2 2.5 4 >4.5 1 !fo. o o Variable star Comet Galaxy Bright nebula Efl Globular cluster )( X-ray source ~) Open cluster ~ Planetary nebula is) Quasar o Other object • 3 i--------- ··~--------·---- .. -----·-·-..---~-----------'-----------I Local Time: 06:25:37 21-Mrz-22 UTC: 00:55:36 21-Mrz-22 Sidereal Time: 17:48:54 Location: 28° 22' 12" N 77" 13' 11" ERA: 22h21 m52s Dec: -4°43' Field: 80.0° Julian Day: 1729172.5386 Fig. 1  126 S.R.A.A., VII (2001) supposed to subtract an unknown number of years, elapsed, in the Saka system. Now, the Kru;fu;la years and the Saka years increase at the same speed, by 1 every year. When we add 1 or 10 or 100 on the Kru;fu;la side we will have to subtract from the sum the same 1 or 10 or 100. The result then will be a never changing fixed amount, through all centuries to come, and we will never know where a given date is located in the yuga. That means, either Sphujiddhvaja wrote nonsense, or the text is not correctly presented.Jfwe go back to the old manuscript we see that Pingree conjectured quite extensively. The text can be reconstructed from the footnotes; additionally, I compared the original as present in a microfilm copy at the St,aatsbibli.oth~k zu.Berlin (reel no .. A 31/16). It reads, with the. 9.ifferencestQ Pingree~s emendations highlighted: gatena sadhyardhasatena yuktya vyekena kO!j{iJ;lagatabdasco:nkhya kala/J sakiindJ:n parisodhya tasmad atltam anyad yugavmsayatah. The second of Pingree's changes is the crucial one: thems. N as well as Utpala in his commentary clearly read kiila/J sakiindJ:n at the beginning of the second line. Usually, an editoris not entitled to dismiss a reading without a full consideration and discussion of the syntax of the stanza. If we accept the nominative kiila/J , we end up with a sentence containing two nominatives: Such sentences are nothing uncommon; they usually express that one thing changes intosornething, else IO , not infrequently the means behind the change occurs in the instrumental. This is precisely the construction of the beginning of our stanza: "The elapsed years of the KU$ fu;l as (nom., kosafJagatabdasalJlkhya)" change into "the time of the Sakas (nom., kala/:t sakanalJl)\ "in' combination with 149, (instr. sadhyardhasatena yuktya vyekena). That means: the distance between the Saka era, starting spring 78 AD, and the K1l')ao.a era is exactly 149 years,which leads us to the begin of a K1l')ao.a reckoning in the year 227 AD. With regard to the Kru;aua chronology some further considerations are necessary, but first the rest of the stanza shall be read and interpreted. Pingree saw that parisodhya requires an object in the accusative. Therefore he changed kiila/Jto kiilco:n. By maintaining the reading of the old ms we have lost kiilco:n as an accusative; We have to look for the words to come and find a sole candidate, atftam. To understand this term we have to ~Td;mf9I;;~M"~~<D~~{lt(t!f:rT*lT~~~4l1 r ~;{~l'J~ti~~fi~~~'~ Fig. 2  Falk : The yuga of Sphujiddhvaja and the era of the K~fi11as 127 reconsider: Sphujiddhvaja wants to find the actual year, elapsed, of the yuga. He wrote in 191 Saka, i.e. in AD. 269. In verse 79,14ab he told us that the yuga started 56 years before Saka, i.e. in 22 AD. The yuga lasts for 165 years. The yuga that started in AD 22 came to an end in AD 187, i.e. almost a century before the composition of our text; or, with other words, Sphujiddhvaja wrote in the second yuga which started in AD 187. If we follow his instructions, we have to add up: 56 years from the beginning of the first yuga to the beginning of the Saka era; then we add the current Saka year, e.g. 191, the time when the Yavanajfitaka was composed. Thisleads to 247 years in the. yuga-system. To get the actual date of the second yuga we have to subtract from 247 the amount of years of theoneyuga already elapsed, i.e. 165. This leads to an elapsed year 82 of the current or second yuga, exactly as the ms. says: anyadyugavar.sayatQh, the result are "the elapsed (yatQh) [years] of the years ofthe second yuga" . This way, by preserving the readings of the manuscript, we get a fully understandable text meanmg: "The elapsed years of the Kusao.as in combination with 149 (change into) the time of the Sakas. Subtracting from this (Saka time [plus 56]) the elapsed (yuga, i.e. 165 years) (produces) the elapsed years of the second yuga." Sphujiddhvaja seems to have been familiar with the Saka era for which the method runs: la) Take the current Saka year, elapsed, add 56 and the result gives the yearssiIlce the beginning of the firstyuga (Y179,14ab). 1b) Subtract 165 (yugam atftam) and the result is the current state of the secondyuga, current in . his time (Yl 79,15cd). For those living in Kusao.a dominated territories, the same procedure is prescribed only preceded by a synchronisation: 2a) Take the current Kusao.a year, elapsed, and add 149; the result is the current Saka year; elapsed (Yl 79,15abc). 2b) continue with Saka la). Checking the stanzas has shown that they contain meaningful statements, which are interrelated. The statements about the yugas and their beginnings are in accordance with what we know about the position of the stars at the time concerned. For the question of the absolute chronology of the Kusao.as just one sentence proved decisive: a date Kusao.a elapsed plus 149 produces Saka elapsed. The text is quite explicit: although the Kusao.a years given on pieces of plastic art may be current . - I 128 S.R.A.A., VII (2001) years, the Yavanajfitaka explicitly asks for elapsed years, kosilF;agatiibdascu:nkhyd. The Saka years are usually elapsed years, although current years may also be found in inscriptions ll. The difference between elapsed years and current years arises from the different use of the numbers. Astronomers, on the one side, count in heavenly revolutions. As long as a revolution is not complete it may be expressed mathematically by a prefixed zero; 0.75 e.g. would mean a revolution of three forth, or of 270 in the sky. Kings, on the other hand, would speak of their first 0 regnal year while it is current, and speak of the second year once the first year has ended. I interpret the word yata at the end of YJ 79,15 as equivalent to gata l2 , so thatall three numbers refer dearly tb elapsed years. For astronomers nothing else is to be expected. For us' iUs, difficult: to imagine the result of comparisons in elapsed and current years. A graphical presentation of the possibilities and their consequences is given below, presupposing that gala systems start with 0, while current systems start with 1: AD related Saka numbers gata curr. a: gata b:citrr. c:curr. d:gata 78 o 1 79 1 2 80 2 3 81 3 4 82 4 5 83 5 6 84 6 7 85 7 8 86 126 226 8 9 o 87 127 227 9 10 o 1 1 88 128 228 10 11 1 2 1 89 11 12 2 3 2 The comparison shows that there are, in principle, four possibilities: a: if Kusao.a gata 0, i.e., the absolute starting point of a Kusao.a century, is linked via the fixed amount of 149 to Saka gata 149 the result would be the actual start of Sphujiddhvaja's Kusao.a. reckoning in the spring of AD 227. b: If the date Kusao.a curr. lis changed via 149 to Saka current, the resulting year Saka current 150 is likewise AD 227. c: If Kusao.a current 1 is changed via 149 to Saka gata, then the resulting Saka gata 150 would indicate a begin in AD 228 d: If Kusao.a gata 0 is changed via 149 to Saka current 149, then the resulting Saka current would mean a begin in AD 226. Since there is no hint in the text that the systems should be used crosswise, and· since  FaIk : The yuga of Sphujiddhvaja and the era of the KU?~as 129 ~~~~==// li /f ~/! . j) -__ . \ k5' ___------- _i~~1 1<1 I ~- I~ I I I/ / ~'\' \,/ \\ \ r ) .\ 11. I- j i / \ " Peg \ \ 1 ' ! , " I \ I I r-------~r:-- I 60/ -----_ \ \ " " I I \3QC'~----\------'1' -~OO\ I ' , II ~-______, ____ ~' - _ - \I \ ! ! ,- ~--- ~ Id! -lj ,\ \ 90, '\ t~---'~ ! / \ . V I "I :y1--. ! .---------_ . _.------ -.____\ ~. -~\=----------90 15_..._ ..~ ___ .___ 'I- /iJ ". \ : ------ 115' --..- - ,i ---.----------~ \ i I j \! /. \ i · i \\ fSC / -Venus \! r---::=::>- \ ./ \, \ I \\; ~ , , A / /~ \ \ \1 ' j! I! \t(/ 11 Trl Y-' \. i { I 1 r:\:~' -~7i\t~--=· r1-===J'ti · \ A~'./ ercury. I \j I" I \ I. I I ! I ' J \ \ i 1// //\ ! \ I!' i~ 1.'_______ / !/ _____ ... (\ ~--i ,t.15' /~..~ __ -------~----- \ ___ I ~15'._ .?f" · t ars ///~/ /30' /r--,_~ / : )..~o. --~.....--- I / - - \ \ /) /! 1~, I j: I I -- -- ---~-- STARS -- - I - - - - - - SYMBOLS . - - - - . - - - - -..., ....- - - - -.. --.. , New Delhi .... -~--r -. --~-1I • <1 • 3.5 '[ •.' Multiple star r ' Dark nebula Le, Radio source I 28. March 167 Tuesday ! • 1.5 4 I () Variable star EEl Globular cluster )( X-ray source I • 2 >4.5; ~ Comet ~) Open cluster 0 Other object i • 2,5 0 Galaxy I -<'- Planetary nebula I e I ~ __~~i~~t neb~_a_~ 3________ ~_L "', __~u_~_sar _____ .--LII __ . _ _ _ _ _ ._.____ ._ _..Ji Local Time: 06:15:37 28-Mrz-187 UTC: 00:45:37 28-Mrz-187 Sidereal Time: 18:10:28 • Location: 28° 22' 12" N 7r 13' 11" ERA: 23h08m38s Dec: +6° 23' Field: 80,OQ Julian Day: 1789445.5317 I ________________ .. ___ _____ J .~ -  130 S.R.A.A., VII (2001) Sphujiddhvaja is diligent to mention the gata state of all dates, we have little choice but to accept a beginning of his Kusaoa reckoning in AD 227. Now, apart from G6bp3 hardly anyone would accept AD 227 as the date for the accession of Kani,ska I to the throne. To understand this date we have again to consider the time when the Yavanajataka was composed: in terms of his own yuga system Sphujiddhvaja was living in the second revolution, which correctly started 28 March, AD 187 [fig. 3], Julian day 1789445, exactly 60272 days after the first yuga on Julian day 1729172 in AD 22. In terms of Kusaoa dates, on the other hand, he may have been living in the second Kusfu;la century, provided the theory of van Lohuizen-de-Leeuw holds good that the Kusaoassimply started again with 1 after their first century had come to its close. In other words: either the Kanilika era started in AD 227 or Sphujiddhvaja's arithmetics supply the proof that there was a second Kusaoa century with dropped hundreds! We cannot have both, as Bivar (1970: 13ff.) preferred, a Kanilika era in AD 128 and all Kusaoas in 98 years. If we accept the dropped hundreds then me:sascu:nkranti of 127 AD is the real starting-point of the Kanilika era. We have dates of Vasudeva I in the 90s, and dates of his successot Kani,ska II from a year 5. Given the advanced state of account keeping, lending on loan and other systems based on currency it seems very unlikely that a first "century" ended in a year e.g. 99 or 103: only a number complying to the decimal system would be usefup4. We know from the Rabatak inscriptiori that Kanilika has something to do with a year 1; he has "inaugurated the year one as the gods pleased" according to the translation of Sims-Williams, presented with a note (Sims':Williams/Cribb 1995/96:82) stressing the preliminary nature of this translation. It seems important to note that the beginning of the Indian year on Caitra 1, (March 23) AD 127 is not a year beginning with a conjunction of sun and moon. This might be another argument to regard this date as a political date, an actual relict of the inauguration of Kanilika 1. 3. The Chinese chronicles A date of 127 AD for Kanilika 1 is in full accordance to what we know from Chinese chronicles about the dynasty of the Kusaoas. In around 92 AD, or slightly later, only two "Great Ytie-Chi" kings were known, both with double names not sharing a single member. This might well refer to Kujula Kadphises and Vima Taksuma. From another source we know that on January 26, AD 230 an ambassador from the Great Ytie-Chi arrived in Peking, whose name is given as Po-t'iao,l> phonetically too close to Vasudeva to be disputed. This embassy was not singular: "After the tottering Han dynasty had finally been superseded by Wei in A.D. 220 a regular succession of embassies is recorded from western countries" (Pulleyblank 1968:256). Of Vasudeva we have regnal years from inscriptions ranging from 64 to 98, in current years equal to 191 to 225 AD. The next known date is from Kanilika II with the date 5, equal to 232 AD. That means, first, that the embassy to China might well have taken place in the last years of  Falk : The yuga of Sphujiddhvaja and the era of the KU'?1q1as 131 Vasudeva, and second, that Vasudeva reigned right into the second century of the Kaniska era which did not start with Kaniska II himself. 4. An earlier solution For the beginning of the Kaniska era almost every possible year has· been put forward in the course of time, from 68 BC to AD 240. Therefore, it is no wonder that AD 1271128 or something very close to it has been named before. Most notable is AD 1281129, first propounded by van Wijk in 1927, popularized in Konow's edition of the then known KharO)t;hi epigraphs ,in 1929,accepted by Marshall (1951,1:71) for archaeological reasons, and still favoured by Bivar (1963:502, 1970: 14). The arguments behind this date are noteworthy: Konow had the idea that dates, containing year numbers, months and tithis, and even na~atras could be used to find out the beginning of their common era. The first data he found in the Taxila silver scroll of the year 136 ayasa asa4asa masasa divase 15, i.e. l36 of the Azes [ayasa, i.e., Vikrama] era, on the 15th of month Asadha. Now, Konow explained ayasa as adyasya (Konow/van Wijk 1924:76) and interpreted this as refering to an intercalary month. Intercalary As3<;lhasare not very common, and therefore van Wijk ended up with a year AD 117, to the dislike of Konow who still preferred a date posterior to AD 125 (Konow/van Wijk 1924:74,78, 91). Three years later van Wijk started anew, on the assumption that "the Hindus have derived their system for counting thena~atras from the Chinese, whose astronomical system is purely equatorial and not ecliptical, which renders the equality of the 'lunar mansions' an impossibility" (1927: 168). On this basis and again on the unfounded assumption that the months concerned ended with the full-moon and the year in October he arrived at the year AD 1281129 for the two dates of theZeda and the Ohind inscriptions. With admissable changes in the length of thena~atras, with a .shift toinonths ending with the new-moon (as does the Yavanajataka), or with an irregularly introduced intercalary month many more calculations would be possible, leading to again differing results. It has been observed often before: Indian dates can be explained once their epoch is known, butthey are not useful to reconstruct an unknown epoch. Van Wijk's result of AD 128/129 was good enough for the . archaeologist and historian, but it cannot be regarded as the result of an indisputable method. Now that we have contemporary information, some specialist might feel challenged to compare thefew na/qatra dates with the epoch reached through the Yavanajataka, 5. The Saka era There is one last problem to solve: if the Saka era is not identical with the Kaniska era, how then did it originate and why did it last right into our time?16 It has often been stated that the Western Saka era is best associated with the Western Saka dynasty of Ujjain, particularly with Cast;ana as the first firmly settled ruler of the dynasty himself? When looking for the beginning of the Indian year in 78 AD we find April 1, [fig. 4], and this is truly a singular day: not only do sun and moon  132 S.R.A.A., VII (2001) ;-;-------~'~----------r__-- I • ~ , I /---~~ I \ \\ 1~~-1----__j----- __ ~ 1 /1 /1 90 -~ _-----4~ J I! ! I , ' ~ f' /, 1 \ / \ ' ; I ' , I Peg I, ,, I 6r-----_~\. - " I,l~ fJ \\ / __---~ , I i..1 ! I I \\\ ',904.' "tl' """ ", I / Venus ~"-- 1 -~ ! >-------. And . . I' 30 0 , ~I' __ ___\ ". \ I I~ ~ I l - ~,~",\\1"- 0 , I ' 00 ~ , I \ I r---:.:~ l--------- •Mercury \ ~\ ~I : i I1 Tri ii 11 I /~j------- ~~I !f f-----~\----- Ii i / \ : I , .. I ~ I Cet I \ \ I r\ i . _____--~ \ 'I \\ \\ -----~ ! ! I I 1 I 115 0 ;,: ----/-".~ / 1-' _______________ ._ _ _ , _ _ _ _ _ _ _ _ _ " -1-201',·1, I. 0 Kept"""--~~_,,_, t=_~ ~_ /30 . 90 ! i • I. STARS <1 • 3,5 ' ! SYMB~~----"--- Multiple star r' Dark nebula 6 Radio source __ I I New Delh Wednesday 1,Apri178 I J I I • 1 1,5 . 4 I 9 Variable star e Globular cluster )( X-ray source , • 2 >4,5 ~- Comet :) Open cluster 0 Other object • 2,5 , 0 Galaxy(;- Planetary nebula 1 . • 3 [J Bright nebula is) Quasar j ,i - - - - - - - - - - '~-----'~---~-~.----'.---~--_' __'_'______ '--'--'------. Local Time: 06:06:37 1-Apr-78 UTC; 00:36:36 1-Apr-78 Sidereal Time: 18:14:54 : Location: 28° 22' 12" N 77° 13' 11" ERA: 23h13m04s Dec: +6° 23' Field: 80,0° Julian Day: 1749637,5254 I ,__, _______,_.______ .. ____ ,_.. _ _ _ ,~ _ _ _ _ _ ""'_,,__ , ____,,_ .,.. _ _ _ _ _ _ _ _ J Fig,4  Fa1k : The yuga of Sphujiddhvaja and the era of the Ku~~as 133 meet jointly in Aries, but with them is Jupiter, a very rare event indeed. From AI-Blnlnl we know that the Saka era was mainly used by astronomers 18. From later South Indian inscriptions we know that Jupiter years are often combined with a Saka date l9 • Therefore it seems highly probably that the Saka era is an astronomer's era, maybe once the starting point of yet another yuga, or of one of the Jupiter yuga systems 20 • The rare event, useful for astronomers through many centuries to come, does much better explain why this era was kept alive in diverse circles than any reference to kings of foreign decent. 6. Results 6.1 History of astronomy The begin of the yuga of the Yavanajataka was corrected from AD 126, emended by Pingree to AD 136, to AD 22. This new starting point is important since it gives us a date when people from the West (yavana) spread their astronomical knowledge in South Asia. Since their yuga lasts for 165 years the starting point in AD 22 could not have been defined through experiences of a single human life-time. There must have been generations before, watching the stars and noting remarkable constellations. Historically, this date coincides with the early years of Gondophares' rule in India. Maybe it was he who made Western astronomers teach their science in Gandhara and beyond. The fatalistic view of human life as displayed in the Yavanajatakaig absolutely new in its rigidity. When we see which role the Pancailgas play in present-day India We realize that these new, thoughts fell on fertile ground. 6.2 Theepoch of the K~fu).as The material on the K~fu).a era presented above will not produce a theory. All there is is the statement given in a text, providing a means to change a K~fu).a date into a Sakadate: "take the K~fu).a year and add 149". This leads to the 2nd K~fu).a century, startillg in spring AD 227. Forthe first century we get the year AD 127. Because of its accurateness with regard to heavenly constellations, because of its early being quoted and because of the antiquity of the only reliable manuscript it is impossible to suspecrthe text to be a late fabrication. All numerical data come in prose numbers, preserved inside the. brackets of the meter. If my translation of the original verse is accepted, together with its interpretation, then we would have found the answer to a question which has baffled the students • of Indian history individually and collectively in three conferences for the last one hundred years. Acknowledgment Due to the efforts of Elizabeth Errington the gist of this paper could be presented to the audience of a celebration in honor of Francine Tissot at the BritishMuseum, on September 12, 2000. The ensuing discussion, involving A.D.H. Bivar, J.C: Harle, D.W. MacDowall, N. Sims-Williams, D. Wujastyk and others led me to considerations of neglected aspects of the problem. Hearty thanks also to Joe  134 S.R.A.A., VII (2001) Cribb for granting access to tbe K~fu.!a coins collection in the British Museum, and for pointing out the latest literature on tbe subject. Notes 1) For the original publication see Sims-Willams & Cribb 1995/96; a French version in Sims-Williams 1996; a critical evalutation in Fussman 1998; cf. also Mukherjee 1995 [sic; appeared 1999]; for an encompasing bibliograph¥ of recent research cf. Fussman 1987. 2) After inspecting the originals it seems safe to assert that the name appears as vema takoma in Kharo~~hi on three of the king's coins (Sims-Williams/Cribb 1995/96:140b,c, 141a), read as takto by Cribb. On some others, unpublished so far, the name ends in takho, aspirated, without the ma. The Ma~ pedestal clearly has [--]ma tak<;1;lma (Fussman 1998:607). It is obvious that takho and tak<;u are adaptations to the same phonetic original; the unaspirated variety, however, seems nearer to the Bactrian form uemo taktu, which Sims-Williams holds ascertained (1995/6:96, n.3; but cf. Fussman 1998:615). There is an unpublished Buddha statue at the Army Museum, Rawalpindi, not on display; its inscription is found as the "Nowshera pedestal inscription" in Konow 1929:134. Konow read dhivhakarasa takhtidre/}-a karide. Some characters are written rather carelessly. Nonetheless, the end of the text can be restored as takhto danamukho. More difficult is the beginning, containing the title. Konow's dhivhakara is impossible, since the second character is not vha, and the third can bqa, or spa or ka. Maybe the title ends in isparasa. At least, in the proper name the ligatur~ khto is absolutely clear. Could the base be something like tak-tsu (plus a nasal) ? In a recent article Mukherjee (1997-98) rejects any form of takto and presents readings of his own which are hard to digest. Finally, he identifies Vima Ta~uma and Vima Kadphises. 3) According to Bivar 1963:501, n. 1 this association goes back to A.K. Narain, The Indo-Greeks, Oxford 1957:144 and R.M. Smith, On the ancient chronology ofIndia, JADS 78.1958:178. 4) J. van Lohuizen-de-Leeuw 1949:235ff.; suitably summarized in de Lohuizen-de-Leeuw 1986:passim; cf. Sillomon' 1998:183, n.87. The idea of dropped hundreds is as old as 1874, when E. Thomas interpreted the K~~a dates a'sihose of the Seleucidan era of 312 BC or of the Parthian era of 248 BC (Fleet in JRAS 1913:981); cf. Konow in Konow/van Wijk 1924,71 n. 2. 5) Notable exceptions are Bivar 1970: 13f. and Mitra 1996: 72f., who expect all Ku~~a kings from Kani~ka !to Kani~ka 11 to have lived in one single century. 6) The proceedings are published in the JRAS 1913 as "The date of Kan~ka", with contributions by F.W. Thomas (627-650, 1011-1042), EJ. Rapson (911-913), J.F. Fleet (913-939, 965-1011), V. Smith (939-942), CD. Bamett (942- 945), L.A. Waddell (945-952), L. Dames (953-960), W. Hoey (960-965). A systematic analysis of the theories advanced is found in Banerji 1908: 27f. For personal memories cf. R.B. Whitehead 1968: 1-3. 7) For the proceedings see Basham 1968. 8) For the proceedings cf. Alrarn/Klimburg-Salter 1999. 9) The mss. write Sphujidhvaja following the common rule that of two similar stops before a third consonant only one is written. Skt spujit seems to be a variant of asphujit, the planet Venus. 10) E.g. Mbh 5. 62, 25 vrddho bhavati vai yuva; for the predicative nominative cf., e.g., Sen 1926-27:348. 11) Salomon 1998: 184. 12) It has this meaning in the date yate'fu catur!fukJ:te'fu sate'fu ( .. ,) vatsare'fu in the Gandghar stone inscription, Fleet 1888:75. 13) Last in Gob11999: 162 "third century A.D.", earlier he spoke of 225 A.D. (1999: 164). 14) An odd ending and a deliberate new era with Kani~ka 11 is expected by Bivar (1963:498, n. 4). 15) For the source text see Ziircher 1968: 371. 16) Fussman 1998: 640, while defending Kani~ka as the originator of an era starting in AD 78, maintains that "cette date seule, a laquelle rien ne s'oppose, permet de resoudre l'enigme de l'ere saka". 17) Last by A.M. Shastri 1996: 77, who himself sees no way to identify the Kani~ka and Saka eras.  Falk : The yuga of Sphujiddhvaja and the era of the KU'iiip.as 135 18) Sachau 1914, II: 6; these astronomers start the saka year with caitra, Sachau 1914, II: 8. 19) Salomon 1998: 184,198. 20) Cf. Salomon 1998: 197f. References Alram, Michael & Deborah E. Klimburg-Salter (eds.) 1999, Coins, Art, and Chronology, Essays on the pre-Islamic History of the Indo-Iranian Borderlands. Wien. Banerji, R.D. 1908, The scythiim period of Indian history)ndian Antiquary 37: 25-75. A.L. Basham (ed.) 1968, Papers on the date of Kan~ka submitted to the conference on the date of Kani-!ka, London, 20-22 April, 1960. London. Bivar, A.D.H. 1963, The Kani~ka dating from Surkh Kotal, Bulletin of the School of Oriental and African Studies 26: 498-505. 1970, Haritl and the chronology of the K~fu.1as, Bulletin of the School of Oriental and African Studies 33: 10-21. Fleet, John Faithfull 1888, Inscriptions of the Early Gupta Kings and their Successors, (CIi/3), Calcutta. Fussman, Gerard 1987, Chronique des etudes Kouchanes. Journal Asiatique 275:333-400. 1998, L'inscription de Rabatak et l'origine de l'ere saka. Journal Asiatique 286:571-651. Gobl, Robert 1999, The Rabatak inscription and the date of Kanishka, Michael Alram & Deborah E. Klimburg-Salter (eds.), Coins, Art, and Chronology, Essays on the pre-Islamic History of the Indo-Iranian Borderlands, Wien, 151-175. Hiirtel, Herbert 1996 The term devaputra in the inscriptions of Kapardin Buddha images from Mathura --An incidentalobservation.Fr. Wilhelm (ed.), Festschrift Dieter Schlingloff, zur Vollendung des 65. Lebensjahres dargebracht von Schiilern, Freunden und Kollegen, Reinbek:99-108. Konow, Sten 1926, Kharoshthf Inscricptions - With the exception of those of Asoka, Calcutta. Konow, Sten & W.E. van Wijk 1924, The eras of the Indian Kharoghi inscriptions, Acta Orientalia, 3, 52-91. Marshall, John 1951, Taxila, an illustrated account of archaeological excavations carried out at Taxila unter the orders of the Government of India between the years 1913 and 1934 in three volumes, Cambridge. Mukherjee, B.N. 1982, The Yavanajataka of Sphujidhvaja, the Sakakala and the Kani~ka Era, Journal of the American Oriental Society 102: 361-364. 1995, The Great Kushfu.1a Testament, Indian Museum Bulletin 30: 1-105. 1997-98, Vima Taktu, an alleged Kushana king, Numismatic Digest 21-22: 5-10. Pingree, David 1961-2, The Yavanajataka of Sphujidhvaja, Journal of Oriental Research, Madras 31: 16-31. Pingree, David (ed.) 1978, The Yavanajataka of Sphujidhvaja, edited, translated, and commented on (HOS, 48).2 vols., CambridgelLondon. Pulleyblank, E.G. 1968, Chinese evidence for the date of Kan~ka, Basham, A.L. (ed.), Papers on the date of Kaniska, London: 247-258. Sachau, Edward C. (trans!.) 1914, Alberuni's India, An account of the religion, philosophy, literature, geography, chronology, astronomy, customs, laws and astrology of India about A.D. 1030, ed. with notes and indices, Two vols. in one, popular edition, London, Kegan Paul, Trench, Triibner & co. Sen, Sukumar --  136 S.R.A.A., VII (2001) 1926-27 The use of the cases in Vedic prose. ABORI 8:347-378. Shastri, Ajay Mitra 1996, Saka Era, Indian Journal of History of Science 31, 89-87. Sims-Williams, Nicholas 1996, NOLlveaux documents sur l'histoire et la langue de la Bactriane, Contes Rendus de l'Acadbnie des Inscriptions et Belles- Let/res: 633-654. Sims-Williams, Nicholas & Joe Cribb 1995/6, A New Bactrian Inscription of Kanishka the Great, Silk Road Art and Archaeology 4:75-142. Shukla, K.S. 1989, The yuga of the Yavanajataka, David Pingree's text and translation reviewed, Indian Journal of History of Science 24:211-223. van Lohuizen-de-Leeuw, J.E 1949, The "Scythian" Period, An approach to the history, art, epigraphy and palaeography of North Indiafram the 1st century B.c. to the 3rd cenl1lryA.D., Leiden. 1986, The second century of the Kani?ka era, South Asian Studies 2: 1-9. van Wijk, W.E. 1927, On dates in the Kaniska era, Acta Orientalia 5: 168-170. Whitehead, R.B. 1968, The first conference on the date of Kani?ka. A.L. Basham (ed.), Papers on the date of Kani~ka. London 1968: 1-3. Ziircher, E. 1968, The Ylie-Chih and Kani?ka in the Chinese sources, A.L.Basham (ed.), Papers on the date of Kani~ka. London: 346-390.