University of Ceylon Review

Vol. XVIII, Nos. 3 & 4. July-Oct. 1960

New LiBht on the Buddhist Era in Co/Ion

and Early Sinhalese ChronoloBY
san Appendix to my edition of the Tamgoda-vihara pillar-ills-

A c~iption, I have included al~ essay .on th~ Chronology.

kmgs from Mahascna to Mahinda V, m which I have examined all the
of Ceylon

data having a bearing on the subject available to me at that time, and en-
deavoured to fix the dates of these sovereigns as accurately as possible.! I am
now in a position to bring forward important new evidence which, while
supporting the position that I had taken up in that essay with regard to Sinhalese
chronology during the early centuries of the Christian era, particularly with
regard to the Buddhist era, makes it necessary to effect slight adjustments
in the dates of kings up to the sixth century, and imparts exactness to many
of these dates.

This new evidence is furnished by an inscription which, as a matter of

fact, had been discovered before I wrote the essay mentioned above, but
had not been fully and correctly read. The record is indited on a rock in
a private land about a quarter of a mile to the north of the spill of the Basa-
vakkularna at Anuradhapura. It covers an area of 4 ft. 7 in. by 2 ft. 9 in.
of the rock surface, and comprises eleven lines of writing, the height of indi-
vidual letters varying between 11 in. and 4 in. The discovery of the ins-
cription was made by Reverend Pandit T. Sri Dipananda Thera of Sri
Bharatindrasrama, Anuradhapura, who conveyed the information to me,
and I had an estampage of it prepared in 1952. Reference has been made to
the inscription at page 33 of the Administration Report of the Archaeo-
logical Survey of Ceylon for 1952. It is included as No.2 in the List of
Inscriptions copied during that year, forming Appendix II of that Report
(p.40). The first line of the inscription refers to the king who was the
donor as son of the great king Budadasa (Budclhadasa), after which the
1. Epiyraphiu Zeylan'icu (EZ), Vol. v, pp. su.n i.

129
 UNIVERSITY OI~ CEYLON REVIEW

letters forming the name of the king himself are not quite well preserved, but
the last two letters of the king's name, before the word maha-ra]a, occurring
at the beginning of line 2, can be distinctly read as uanii. It was therefore
assumed at that time that the record is one of Mahanama, who was a son of
Buddhadasa, and the indistinct letters were restored to conform to a title of
Mahanama known from that monarch's inscriptions.

The inscription is badly weathered in places, and some of the letters

in the last five lines have become altogether illegible. The execution of
the record has been done in a slovenly manner. The lines are not of equal
length, and the individual letters not of uniform size. The script is rather
r ,~
! · cursive and there is considerable variation in the forms of individual letters.
The record also contains, as is apparent even from a cursory examination,
words and phrases not found in other epigraphs of the period. All these
circumstances make the satisfactory decipherment of this inscription a very
laborious task, requiring many days of concentrated attention. At the time
of its discovery, I was not only busy with administrative duties of pressing
urgency, but was also engaged in the onerous task of seeing the Sigiri
GraJfiti through the press, and writing some sections of that voluminous
monograph. I had therefore to defer the study of this inscription to a more
leisurely time which I, however, did not get before I relinquished the
post of Archaeological Commissioner.

Recently, in the course of an epigraphical study in which I am at pre-

sent engaged, this record came to my notice once again, and I could afford
the time necessary for its satisfactory decipherment. At my request, Dr.
C. E. Godakumbure has been kind enough to furnish me with a fresh
estampage prepared by Mr. T. K. Jayasundara, in which some of the letters
are clearer than in the earlier one, while others, perhaps due to the deterio-
ration of the rock in the interval, are not so clear. As a result, it has now
been ascertained that the record is not of Mahanama, but of his elder brother
Upatissa, and that it is dated in the twenty-eighth year of that monarch. In
addition, the record also gives the year 941 from the Parinirvana of the
Buddha as the equivalent of the regnal year. The day of the grant is
given as Duratu new-moon and, though this detail is not quite so certain
as the others, this new-moon day is said to have fallen on a Tuesday.
The record, thus, is one of capital importance for the study of Ceylon
history. Before discussing these points, it is necessary to give the text
of the inscription, as deciphered, and its translation.

130
 THE BUDDHIST ERA IN CEYLON

TEXT

1 Sidam[I*]Budadasa-l11a[ha*]rajaha puta Upati [sa- Sirimekal--

2 nami-maharaja kirivita nekeri-dora t[o]ral).a-dora ata -
3 -la-dora kila koha -da-ga Upatisa-raja-maha-vihara kara-
4 -va Kana-ketehi varu-pota Nakara-gala Kclela-amitjilh-uta
5 di Dirati-gama Dasa-gama pohatakarata lea] Boya-geyata
6 Kabota-aganahi Mahanelaka-valina sari karihi-kubura akada-
7 tana gal).ayaDoraka-vaharata dinc [1*] Cata lagita-
8 -k -ata-visiya-javanaka-vasahi] _ cada-[ avamasi]
9 Duratu kalakara-pohata [-divasa] Kuja-[ varc] Ba [gavata- Bu] daha
piri-[nivita]-
10 kale nava-sata-cka- catari[sa]- vana-[vasahi] .
11

TRANSLATION

Success! The great king Upatisa, bearing the name of Sirimeka-, son
of the great king Budadasa.s having founded the Upatisa-raja-maha-vihara>
(at a place) half a krosa6 ahead of the city gate, the gate of the archway, the
gate of the watch-tower? and the monumental columns, which he himself
had caused to be constructed, granted to this Doraka-vihara", (thc villages
2. In this name, the first aksara u is clear enough; of the pa, the right-hand vertical
stroke is damaged, but what is preserved of it is enough to identify it. Of the aksara. ti,
the i sign has to be supplied conjecturally. Of the four letters read as Sirimeka, there
are enough traces to justify the reading. The name or title of Sirimeka (P. SiI'irnegha)
was borne by JeHhatissa II, the grandfather of Upatissa I, and there is evidence to
establish that the kings of Ceylon during this period at times bore the names of their
grandfathers (EZ, Vol. IV, pp. 122ff).
3. P. Upatissa Sirimeqha; Skt. UpatillYa Srimeqlia.
4. P. and Skt. Buddhad/isa,
5. P. Upatissa-riija-rnah/i-oihiira,
6. Koha-od+aqa :-Kohada= Skt. krosardh«, aqa = Skt. aqre. A krosa is half of
a gavyuti (S.gavu), four of which made a yojana. A Sinhalese gavu was about 2t miles;
half a krosa would thus be little over half a mile in length. See H.W. Codrington in Ceylon
Journal of Science, Section G (CJSG), Vol. II, p. 134, and J. F. Fleet, • Imaginative
Y ojanas,' in the Journal of the Royal Asiatic Society of Great Britain and he/and (J RA S)
for 1912, pp. 229 ff.
7. Nekeri-dora = Skt. nuqura-ilotiro, The form nekeri in the inscript.ion perhaps
goes back to Skt. naqari rather than naqara, Torana-dora = Skt. torana-diuira, Atala-dora :
atala = Skt. a?tala.
8. Kila is taken to be the same as khila in P. inda-kbil«, for which see P.T.S. Pali
Dictionary, 8.V. It was a pillar set up at a city gate.
9. Doraka-vihara = Skt, Dixiralca-viluira, This is obviously the same as Upatissa-
raja-maha-vihara, mentioned earlier, and was given the alternative designation as it was
in the proximity of the city gate.

131
•

UNIVERSITY OF CEYLON REVIEW

of) Diratigama and Dasagama for the benefit of the Hposatha-house1o and
sixty karisas of field from Mahanelaka-va]a in (the village of) Kabota-
agal~all for the benefit of the Bodhi-shrinet- (having had these) acquired
from the minister Nakaragala Kelelal3, giving him the varupota'» of
Kanaketa, and having (the grant) registered as perpetual in the administrative
offices'S, on Tuesday the fast day of the Duratu new-moon!c in the month
of in the twenty-eighth year of the raising of the umbrella,
(being) the year Nine-hundred and forty-one in the era of the Parinirvana
of the Blessed Buddha .

* * * * *
This inscription is the earliest document so far known in which a date
is given in the Buddhist era, reckoned from the Parinirvana. This date can
be made use of for chronological purposes as it is equated with the regnal
year of the king reigning at the time-the twenty-eighth of Upatissa I.
Furthermore, this date is sufficiently close to a synchronism with Chinese
history, namely the embassy of Mahanama!", the successor of Upatissa I,
which was received by the Chinese emperor in 428, so that the Sinhalese
chronology can be brought into relationship with the well-established
10. Pohatakara. See EZ, Vol. III, p. 168.
11. Kabota-aqana would be Kapouinqana in Pali. In modern Sinhalese it would be
Kobeyi-gane. A village of this name is mentioned in the long but fragmentary rock ins-
cription of Bhatika-Abhaya, close to the summit of Mihintale (Muller, AIC, No. 20).
12. Boya-geya = P. bodhi-geha. See University of Ceylon, History of Ceylon (UHC),
Vol. 1, p. 308.
13. Nalcaraqala is the title which occurs in inscriptions of the tenth and twelft.h
centuries as Nuvarayal. See Epigraphia Zeylanica, Vol. II, pp. 57 and 2540and Vol.
III, p. 325.
140. Varu-pota is obviously the same as the later varupeta. This word occurs in Ute
Anuradhapura slab-inscription of Lilavat.i (EZ, Vol. I, p. 180) and in the Piijaval'i, 340th
chapter, in the phrase 'I'alaoatu-tiia diya pava visi-dahasak kusnburu-kiri. karava Derui-
nakayehi sarhgha,ya{a dan-varupet karava (Mabopitiya Medhankara Thera's edition, p. 15).
From these contexts it is clear that the word denotes an irrigated tract of rice fields. Per-
haps it is derived from a compound of Skt. viiri and priipta,
15. Akada-tana garwya;- Akada (Skt. akha-rufmh), • not ceasing,' is taken to have
been used adverbially, modifying ga1Jaya, the absolutive of the verb gal'!-a.,Skt. Yrh1Jci/-i.
Tana (Skt. sthana) means an office or administrative ccntre in such words as de-kam-uin.
and kam-ton-ledarii occurring in tenth-century inscriptions (EZ, Vol. I, p. 96 and Vol. II,
p. 31). Similar phraseology is not uncommon in inscriptions of the third to fifth centuries.
An unpublished inscription of the reign of Bhat.iya II from a place called Nelugala in
Tamankadu has akarJct-tanahi-ga1Javaya.
16. Kalakara is the prototype of the later kaluuaro: (dark), which is the equivalent of
Skt. kr?rw in kr?rw-pak?a, 'the dark fortnight.' Pohata is P. uposatha, Duratu (Durutu)
is the same as Skt. Pausa. The origin of the Sinhalese name for this month has not yet
been ascertained.
17. Journal of the Ceylon Branch of the Royal Asiatic Society (JCBRAS), Vol, XXIV
(No. 68), p. 83.

132
University of Ceylon Review - Vol. XVIII Nos. 3 & 4
 THE BUDDHIST ERA IN CEYLON

chronology of China. And, what is more significant, the record itself

furnishes details of the date-if the reading Kujavarc in 1. 9 be accepted-
enabling its exact verification, so that the validity of the conclusions arrived
at from other lines of evidence with regard to the date can be tested.

Taking first the traditional date of the king mentioned in this inscrip-
tion and that of his successor for consideration in the light of this evidence,
Tumour, computing from data in the Mahdvamsa, gives 911 of the Bud-
dhist era as the date of U patissa' s accession, and 953 as that of Mahanama 18;
Sumangala and Batuvantudavc are in agreement with Turnour.t? Wije-
sinha's dates for the accession of these two monarchs are 370 and 412 A.C.
respectively,20which, according to the equation, date in the Christian era
+543, are 913 and 955 B.E., respectively. Wikremasinghe's date21 in the
Buddhist era of 544-3 B.C. for Upatissa is 909 and for Mahanama 951.
Geiger's date for Mahanama is 409 A.C. (952 B.E.)22 ; that scholar's treat-
ment of Upatissa will be referred to later. According to Mabopitiye
Medhankara Thera23, Upatissa came to the throne in 912 B.E. and Maha-
Dima in 954. My own dates in the Buddhist era for the accession of these
.two kings are 908 and 949 rcspectively.s+ The evidence of the present
inscription establishes that the real date of the accession of Upatissa I was
913, and of Mahanama 955 of the Buddhist era. It will be seen that Wije-
. sinha's dates for these events are in complete accord with the contemporary
evidence of the inscription. The others are from one to six years earlier
than the actual dates.

The discrepancies are due to adjustments for possible current years at

the dose of the reigns of some kings, and the variations in the received
texts of the Sinhalese chronicles in the sum total of the number of years that
is said to have elapsed from the Parinirvana to the end of Mahasena's reign.
According to B. Gunasekera's text of the Pujiivalf (34th chapter), 844 years
9 months and 25 days had elapsed between the Parinirvana and the close of
18. G. Turnour, The Muhuuamso, with the Translation subjoined. Cotta Mission
Preu, 1837, p. lxii.
19. Mahdvamaa. Sinhalese translation. Part, II. from the 37th ehaptcr, Colombo,
Government Press, 1917, p. xx.
20. Mahdoomaa. translation into English, (1overnm('nt Printer. Colombo, 1909 (lUt.
W).p.x.
21. EZ. Vol. III. p. 12.
22. Oii,laoomaa, English translation (Ov. T.), part ii, p. xi.
23.IPii,javaliya, 34th chapter, edited by Pundit. Mahopit.ivr- Modhnuknra Thera,
Hahabodhi Press, Colombo, 1932, (P\'. xxxiv, 1\1). p. iv.
24. EZ, Vol. IV, p. 109.

133
•
UNIVERSITY OF CEYLON REVIEW

Mahasena's reign25 and the later works, the Nikayasmigraha and the Raja/Jali,
are in accord with this26. Geiger has accepted this as ' one of those single
dates which rest on a sure traditional basis.'27 Wickremasinghe, too, has
made his chronology conform to it so far as his dates in the Buddhist era
are concerned.t" Ihave myself followed these two scholars in this matter-".
Some manuscripts of the Nyavali, indeed, give the reading ata-siva-susdlis
(844) in enumerating the number of years, while others have the reading
ata-siva-sasdlis (846). In Mabopitiye Medhankara Thera's critical
edition of the Piyavali (34th chapter), the reading adopted is the latter, as
it is supported by the majority of the manuscripts that he consulted.w And
this reading of sasdlis instead of susiilis is in accord with the present ins-
cription which equates the twenty-eighth year of Upatissa with 941 from
the Parinirvana, . The total lengths of the reigns of Sirimeghavanna, Jenha-
tissa I and Buddhadasa, who came between Mahasena and Upatissa I, add to
66 years which, together with 28 years of Upatissa up to the time of the
record, amount to 94 years. Subtracting this from 941, the year of the
Buddhist era corresponding to the 28th of Upatissa, we get 847, which
tallies with 846 years 9 months and 25 days of Medhankara Thera's text of
the Piljavali, rounding off the 9 months and 25 days as one year. It is also
noteworthy that, in order to arrive at this agreement, no allowance is neces-
sary for current years at the close of anyone reign, and that the twenty-
eighth year of Upatissa has also to be included. It is therefore possible
that, at the time the present record was indited, the 28th year was drawing
to a close, or that the year 941 of the Buddhist era had not run much of its
course.

If the Buddhist era of this record was computed from the same starting
point as that era has at present in Ceylon, Burma and Siam, and as it has
been among Sinhalese Buddhists at least as far back as 1200 A.C., and if the
year given is an expired one-! , the date is equivalent to 398 A.C. But it is
well-known that the present Buddhist era is in error for the time of Asoka
2ii. B. Gunasekara, A Contribution. to the His/my oj Oeqto», translated from. the P11jii-
')(tliya, Colombo, 1895, p. 25.
26. Nikaya-8angrahct, English translation by C. lVI. Fernando, Colombo, 1908, p. 14,
Riijiivctliya, English translation by B. Gunasekara, Reprint, 1954, p. 4.~. The Nikiiyn-
sangmha gives the date of Mahasena's accession as 818 H.E. This king reigned twentv-
seven years.
27. O». T., part ii, p. xvii.
28. EZ, Vol. III, p. It.
29. EZ, Vol. V, p. 88.
30. Pv. xxxiv, lVI, p. Hi.
31. E?5, Vol. II, p. 220_

134
 THE BUDDHIST ERA IN CEYLON

whose date can be ascertained within narrow limits from the references to
a number of Greek kings in his inscriptions.V Computing from this base
. of the true date of Asoka, various attempts have been made by distinguished
scholars to ascertain the correct date of the Buddha's Parinirvana. Of these
dates, 483 B.C., determined by J. F. Fleet after an exhaustive' study of the
evidence supplied by various sources, is the one that is accepted by most
Orientalists.f

Devanampiya Tissa of Ceylon being a contemporary of Asoka, the date

of that monarch has to be adjusted to be in conformity with that of Asoka,
and not in terms of the Buddhist era now current in Ceylon. And this
. adjustment was continued by Fleet right down to the time of Mahanama,
. the successorof the king mentioned in our inscription.s- Wickremasinghe
"{ullyendorsed Fleet's views and claimed to have discovered evidence in the
ili,scriptionsof South India to prove that a Buddhist era with 483 B.C. as
~ starting point prevailed in Ceylon up to the eleventh ccntury.s> This
vie~ of Wickremasinghe was accepted by Geiger in his translation of the
Mahava/ilsa published in 1912)6 E. Hultzsch, however, pointed out that
the premise on which Wickremasinghe based his theory of a Buddhist era
of 483 B.C. prevailing up to the eleventh century, namely, a supposed
synchronism with South Indian history, was faulty, as he had relied on an
erroneous translation by Wijesinha of a passage in the Cii lavamsa (Chap. 53,
v.44)37. Wickremasinghe, thereupon, modified his view, and was not
certain whether a Buddhist era with 483 B.C. as its epoch was prevalent in
Ceylon in the eighth to eleventh centuries, but clung on to the position that
such an era prevailed up to the beginning of the eighth century, in spite of
the evidence against it from the Chinese synchronisms pointed out by
SylvainLevi.38 I have demonstrated in detail the untenability ofWickrema-
singhe's position, and need not recapitulate the arguments hcrc.s?
32. P. H. L. Eggermont" The Chronoloqs] of the Reign of Asoka Moriuo, Leiden, )!lr;r.,
p.180.
33. J. F. Fleet, 'The Day on which the Buddha Died,' in the Journal oj the Royal
.Asiatic Society of Great Britain arul Ireland (J RAB) for) !l09, pr. ) ·34.
34. J. F. Fleet, 'The Origin of the Buddhavarsa, tho Ceylonese Reckoning from the
Death of Buddha' in J RAS for 1909, pp. 323-3;')6.
35. EZ, Vol. I, pp. 1155 ff.
36. Mahiivamsa, translated into English by Wilhelm Geiger, Reprint, Colombo, 19;')0,
pp. xxviii ff.
37, E. Hultzsch, "Contributions to Sinhalese Chronology' in J RAS, 1913, pr.
517·531.
38. Sylvain Levi in Journal Asiatique, I !lOO, pr. 297 ff, 401 ff, English translation
by John M. Seneveratne of relevant parbs of Sylvain Levi's papPI' in ,JC'BRA8, Vol, XXIV,
(No. 68), pp. 82-102.
39. EZ, Yol. V, pr. ss ff.

135
 UNIVERSITY OF CEYLON REVIEW

Geiger was alive to the significance of the Chinese synchronism refer-

ring to the reign of Mahanarnaw ; in his translation of the Culavamsa, he
therefore accepted the position that the traditional Buddhist era from the
time of that monarch onwards has to be taken as starting from 544 B.C.
But, Geiger insisted that up to the end of the reign of Mahasena, the era was
to be computed as starting from 483 B.C. In order to provide for the
transition from one era to the other, he left the chronology between the
end of Mahasena's reign and the accession of Mahanarna in an indefinitive
manner, and proposed that the interval between these two events, which,
according to the Cii!avmnsa and the Pi"yiivalt, comprised 108 years, be re-
duced to 49 years.s! He even proposed a measure so drastic as to entirely
discard Upatissa I with his 42 years. "He may be purely fictitious or per-
haps a prince who reigned along with his father and either never came to the
throne himself or if so, only for a short period,' says Geiger. But, in addition
to the present inscription, Upatissa has left another record dated in his 24th
rcgnal year, where he is called Je~atisa Upatisa.t? With regard to Geiger's
reduction of the period between the death of Mahasena and the accession
of Mahanama to 49 years, the present record is evidence that Upatissa
reigned at least for 28 years, and the 20th year of his father Buddhadasa is
referred to in another record.O Three inscriptions of Jenhatissa II, the
father of Buddhadasa, are known ; one of these is dated in the second year. 44
There are many records of Sirimeghavanna ; the twentieth regnal year is
mentioned in one of them.s> It is thus clear that the number of years
allowed by Geiger for the interval between the death of Mahasena and the
accession of Mahanarna is quite inadequate. There is nothing incredible
in the lengths of reigns, given in the chronicle, of Sirimeghavanna, Jenha-
tissa II, Buddhadisa, Upatissa I and Mahanama. These five monarchs, be-
longing to three generations, reigned, one after the other, for 130 years.
Mahanama, most probably, was past the fiftieth year of his age when he was
raised to the throne.

As has been stated already, Sinhalese envoys of Mahanarna were at the

Chinese court in 428 A.C. If the date of the present inscription is converted
into the Christian era on the basis of 483 as the starting point of the Buddhist
40. Cv. T., part. II, pp. v ff.
41. Cv. T., part, II, p. xi.
42. Cv. T., part II, p. vii, Ceylon Journal oj Science, Section. (j (('J sa), Yol. II,
p. 1O;~.
43. Ceylon Antiquary and L'iterary Register (('ALR). Vol, III, p. 20i.
44. CJ SG, Vol. II, p. 102. "Brahm) Insoript.ions in the Yala East Wild Life Reserve .
h~- r-. \V. Xicholas in Soil' Paul Pier'i .• Felicitation. I'olwne, Colombo, 19Mi, p. 67.
4,). In fin unpublished insr-r ipt ion at Kiil'arilhagnln in the Hambant ota Distriet..

136
 THE BUDDHIST ERA IN CEYLON

- era, it would be equivalent to 458 A.C., which would mean that Upatissa's
e reign of 42 years ended and Mahanama ascended the throne in 472 A.C.,
:! 44 years after the envoys of the latter had reached China. It is therefore
clear that the theory of 483 B.C. as the epoch for the Buddhist era does not
hold good for this period. On the other hand, if the Buddhist era had the
same starting point as it has today, the date of our inscription, assuming
that the year is an expired one, would be 398 A.C., and Mahanama's access-
sion would have been in 412 A.C. He reigned for 22 years; his envoys
could therefore have been in China in 428 A.C.

But, from this Chinese evidence, one cannot say with certainty that the
epoch of the Buddhist era on the date of this inscription was exactly the
same as it was in the beginning of the thirteenth century. Even if the epoch
of the Buddhist era was six years earlier or sixteen years later than 544 B.C.,
Mahanama could have come to the throne on a date enabling his envoys
to be in China in 528 A.C. Such circumpection is all the more justified
when we consider that it is only after the lapse of nearly 800 years from the
date of this record that we find in Ceylon another inscription in which
reference is made to the Buddhist era,46 though the prevalence of the era
in Burma is known from inscriptions of Kyanzittha+? whose reign began
in 1084: that, too, nearly 700 years after the date of our epigraph. It is quite
conceivable that the Buddhist era went out of vogue some time after the
time of Up atissa, and that it was re-introduced in the eleventh century either
in Burma or in Ceylon. And the computation of details on which the
initial point of the era was then decided might have given a result differing
from that in the time of our inscription. It is on these considerations that
the mention of the week-day coupled with the tithi, enabling the exact
verification of the date given in our record, is of capital importance.

It is therefore very regrettable that the inscription has not been well
preserved atthepoint where the week-day is mentioned. But there is no rea-
sonabledoubt about the identity of the two syllables reading Kuja, the name
of the week-day; only the sign for the medieval vowel u has to be read
conjecturally. And these two letters cannot form a word either taken
together or separately, with those which precede or follow them. Between
ku and the last word legible before it, there is room for three ak~aras which,
as they follow the word pohata (P.uposatha), must have formed, on the
46. Inscription from Madagama in the Vau~lavili Hatpattu, O.JSG, Vol. II, pp. IR6
and 212. This inscription, dated in the 17th year of Parakramabahu I, gives the year
1696 of the Buddhist era as the date of that monarch's accession.
47. EpigrapMa Birmanica (Ep. Rh·.), Vol. I. pp. iH and nr,

137
 UNIVERSITY OF CEYLON REVIEW

analogy of other inscriptions of the period, the word davasa, 'day.' Fol-
lowing Ku]« is a letter which is somewhat blurred, but can be read as Vii;
the next letter is clearly re, The reading Kujavare is thus quite certain.
After re, the akeara ha is quite clear. Three letters following this are illegible.
The next letter, bu, is partly damaged, but the four alisaras after them admit
of no doubt, and read-Jaha piri. Two or three letters at the end of this
line, the ninth, have been completely worn away. The first two letters of
the tenth line read kale; these are followed by the enumeration of the date
in words. Taking into consideration the stage of phonetical development
of the Sinhalese language represented by the rest of the inscription, that
portion of the ninth line following Kuiauarc, with the fIrSt two letters of the
tenth line, may be read Ba[gavataBII]daha piri[l1hJita] kale, 'in the era of the
Parinirvana of the Blessed Buddha'.
,1

"i It might also be questioned whether the name of a planetary week-day

could have been mentioned in a Ceylon inscription of the close of the fourth
century, when the earliest known epigraph from India itself, containing the
name of a week-day, is the Eran inscription of Budhagupta dated in the
(Gupta) year 165 (4H4 A.C.), over ninety years after the date of our rccord.v
The adoption by the Hindus of the Jewish-Christian calendrical week
beginning with Sunday was a result of the contact of India with the West,
and it cannot be definitely stated when tills took place. There is literary
evidence for the use of week-days in India"? in about the first or second
century A.C. As Ceylon and South India had direct relations with the
Roman empire from the first to fourth centuries, the names of the week-
days could possibly have been adopted by the Sinhalese and the Tamils
directly from the traders of the Roman empire, and not introduced from
North India. Consequently, there is nothing improbable in the mention
of a week-day in an ancient Sinhalese document of the last decade of the
fourth century A.C. To meet with a week-day again in a Sinhalese
document, we have to wait until the reign of Vijayabaliu £.50

Having thus concluded that there is justification for taking that, ac-
cording to this record, the new-moon of Durutu in the Buddhist year 941
48. J. F. Fleet, Guptn Inscriptions (Corpus lnscriptionuni Irulicarum. Vol. [II), p. 89
i'\ee also 'The Use of the Planetary Weck in India' by J. F. Fleet in J HAS for 1!)12,
pr. IO:H)·I046. I am not nwaro whether an inscription of an earlier dute mentioning n
week-day has ber-n disooverr-d in India allot' t ho pnblication of Fleet's art icle ill Ill!:!.
4!l. K, P. JayaswHI, 'The Week-days and Vickruma' in Indian Ant iquaru, Vol.
Xr.YIJ, 1!l!2. p. IIZ.
!i0. An unpublished inscription from Periynkulnma or Xiitnn;ir·kr)vil. Ammoi Report
of thr Arrlutroloqirnl 81I1'~'PII of ('Pllion (A.':;('A R) 1'0" !!I'i3. p. 2:!.

138
 THE BUDDHIST ERA IN CEYLON

,n on a Tuesday, we have to ascertain the year in the Christian era which

, . es these details. Fortunately, the scope of our research is restricted
.,'ithin narrow limits by the circumstance that this year, the twenty-eighth
. Upatissa I, can be roughly ascertained by the evidence of the dates of the
, inhalese missions which arrived in China in 428 and 527. The Sinhalese
'" bassy to China in 428 was despatched in the reign of Mahanama who
ucended the throne fourteen years after the date of the present inscription,
;and reigned for twenty-two years. The envoys could have left Ceylon
"pen in the last year of Mahanama, and arrived in China actually after the
,~th of that king. Assuming that the voyage from Ceylon to China took
year, the envoys would have left Ceylon in any year between 405 and
A.C., within which dates must have fallen the first year of Mahanarna's
. The 28th year of Upatissawould thus have fallen between 391 and
3 A.C. The Sinhalese mission which reached China in 527 was des-
ched by a ruler named Kia-che Kia-Io-ha-li-ya, who has been identified
SylvainLevi with Kassapa, the son of Up atissa 11.51 I have given reasons
'. r taking him as Silikila52 who succeeded Upatissa II. The reign of
';. -kala, according to the chronicle, lasted for 13 years ; it must have com-
Jlenced between 513 and 527 if the envoys sent by him arrived in China
,'in 527. The interval which separated the accession of Silakala from the
)wenty-eighth year of Upatissa J, according to the details in the Cii!alJatiISa,
;1\'as 130 years 3 months and 25 days, i.e. 130 years in a round number.
; According to the evidence of this synchronism with Chinese history, the
;~~e 28th year of Upatissa J was between the years 513 -130 and 527 -130,
ii.e. 383 and 397 A.C. Taking the years common to the calculations based
~on the dates of the two missions, we get 391 and 397 as the limits of the
.~period in which the 28th year of Upatissa I fell. And, if the above argu-
..ments with regard to the mention of the week-day in the present inscrip-
',tion and the calculations combining the details of the CiiJavan"lsa with the
Chinese synchronisms are valid, there should have been a Durutu (Pausa)
: new-moon which fell on a Tuesday within these six years.

Before we apply this test, some remarks on another matter are also
,necessary. The Sinhalese month Duratu (modern Durutu), the new-moon
of which is referred to in the details of the date, is the equivalent of the lunar
month called Pausa in Sanskrit. Lunar months in Ceylon today, as they
have been for some centuries past, follow the amiinta system, i.e. they end
with the new-moon, as is the usage in South India. But the other system,
51. .JCBRAS, Vol. XXIV (No. fiil). r- m f.
52. EZ. Vol. V, pp. !lR,!)!).

139
 UNIVERSITY OF CEYLON REVIEW

piir~irlliitlta, according to which the lunar month ended with the full-moon,
was not unknown in ancient Ceylon. The Alutvava pillar-inscription
furnishes evidence of the use of a piir~imiitlta month in the tenth century, 53 1
and the Panakaduva copper-plates indicate its usage in the eleventh. 54 The
definition of the seasons in the SikhavalaMa-vinisa of the tenth century»
is on the basis of piirnimdnta lunar months. 'In the ptlnJimanta system each
month beginning with a full-moon is named after the next amdnta month,
but takes in the dark fortnight preceding each new moon.'56 Thus, if the
piimimdnta system was followed by the writer of this epigraph, the new-
moon (amiivasyii) of Durutu (Pau~a) would have been the new-moon of the
month of Margasir~a (Sinhalese Uiiduvap) according to the amdnta system.
, The new-moon is called sometimes by the name of the month of which
it marks the end, and sometimes by the name of the following month.'57
Our record, in addition to calling the new-moon by the month of Duratu,
has the mention of a month preceding it. The name of this month is not
legible, but it is possible that the writer of the record followed the practice
of calling the new-moon by the name of the month which followed it.
In effect, this would result in the same day as if the piiTlJimanta system of
naming lunar months were followed. We thus have to test the date having
in view both the amiinta and the piir~imiillta systems.

Referring now to the Table of Solar Years and New-moons from 1 B.C.
to A.D. 500 in Swamikannu Pillai's Indian Ephcmcris,58 we find that between
391 and 397 A.C., between which the 28th year of Upatissa I must fall
according to Chinese synchronisms, there was no amiinta month of Pa- sa
(Durutu) of which the new-moon fell on a Tuesday. But if the month
be taken as piirnimiinta, the new-moon of the month of Pausa in 396 A.C.
was on the 16th of December, which was a Tuesday. If the practice of
calling the new-moon of an amiinta month by the name of the following
month be adopted, this day was a Pausa alllavasyii, and Swamikannu Pillai,
in fact, following this practice, has called it by that name. Consequently,
the equivalent in the Christian calendar of the date given in our record has
to be taken as Tuesday, 16 December, 396 A.C.
53. EZ, Vol. II, p. 231.
54. EZ, Vol. V, pp. 9·10.
55. Sikhavalanda.vini8a, edited hy Sir D. B. Jayatilaka, p. 15. The rainy season,
for instance, began with the first of the dark fortnight of Asii!hi and ended with the full-
moon of Kattika.
56. Diwan Bahadur L. D. Swamikannu Pillai, Irulian. Ephemeris, (lnd. Eph.l, Vol. I,
part I, p. 52.
57. Ind. Eph., Vol. I, part I, p. 31.
5R. Ind. Eph., Vol. I, part I, pr. 214-21; •.

140
 THE BUDDHIST ERA IN CEYLON

We now proceed to consider the bearing which the date in the present
1 inscription has on the Buddhist era as it has been in use in Ceylon from the
\
twelfth century up to modern times. The Polonnaru inscription of
Sahasamalla gives Wednesday, the 12th day of the waxing moon in the
month of Binara (Sanskrit Bhadrapada) after the expiration of 1743 years,
3 months and 27 days of the Buddhist era as the date of the accession of that
monarch. 59 The details work correctly to Wednesday, 23 August, 1200
A.C.60 This document establishes that, in the Polonnaru period, the years
in the Buddhist era, as those of the Saka era, were expired ones (gata), and
that the year began on the full-moon day of the month of Vaisakha, The
starting point of the era, according to these details, was in 544 B.C., so that
the formula for the conversion of a date in the Buddhist era to one in the
Christian year is-543 if the day was before the end of December and-542
if it was after January 1, and before the day of the Vaisakha full-moon. H.
w. Codrington has shown that the dates in the Buddhist era available Irorn
Kandy times conform to this formula.v' The details in the only date in
the Buddhist era found in India, Karttika waning moon 1, Wednesday, in
the Buddhist year 1813, given in an inscription of a king named Asokavalla
found at Buddhagaya arc correct, as pointed out by J. F. Fleet, for Wednes-
day, 1 October, 1270 A.C., if the Buddhist era is taken as having its initial
point in 544 B.C.62

If the Buddhist era had the same starting point in the fourth century
as in the twelfth century and later, the year 941 given in our record, taken
as expired, would have run from 17 April, 398 to 6 April, 399. If
the year be taken as current, its limits would have been between 28 April,
397 and 16 April, 398. In neither of these two years was there a Durutu
new-moon, whether according to the aindnta or the pur~/il'lliinta reckoning,
which fell on a Tuesday. 63 If the year in our epigraph was an expired one,
there is thus a difference of two years between the starting point of the
Buddhist era in the fourth century, and that in the twelfth. The wording
of our inscription makes it more likely that the year was current (vartamiina);
in that case, the discrepancy in the starting point of the era was only
one year.
59. EZ, Vol. II, pp. 223 and 22H. Buddha-oarsa ek-tlaluis sat-siqu. tesdlis-luururiuiu.
tun-mas sat-visi daoasak yiya tena Binerii pura doios.cak: lada Bad/i-daoas,
60. See .Fleet, in JRAS for 1909, p. 331 and Ind. Eph., Vol. IV, p. 3.
61. H. W. Codrington, 'The Buddha-varsha in the Kandyan Period,' in GALR,
Vol. II, pp. 51-53.
62. JRAS for 1909, p. 347, The inscription in question is published in the Indian
Antiquary, Vol. X, p. 342.
63. Ind. Eph., Vol. I, part I, p. 215.

141
 r?

UNiVERSITY OF CEYLON REVIEW

It may be argued, particularly in view of the fact that no document

dated in the Buddhist era is available, either from Ceylon or any other
Buddhist country, for nearly 700 years after the date of our inscription, that
the Buddhist era went out of use some time after the date of our record,
and was re-introduced in the eleventh century, calculating the starting point
from. details found in the Dipavamsa and the Mahiivaiiisa, and the
lengths of reigns of the kings who came after Mahasena, But, if a rc-
introduction of the era based on fresh calculations was effected, one would
expect the discrepancy between the starting points to be much more than a
year or two. On the other hand, C. O. Blagden has pointed out that certain
of the dates in the Buddhist era occurring in Burmese inscriptions before
1300 A.C. indicate a starting point for the era some two years earlier than
544 B.C.64 One of these dates quoted by Blagden refers to Thursday, the
6th waning of Tazaungon (Karttika) in 654 of the Burmese Sakkaraj era
and in 1837 of the Buddhist era. The Burmese Sakkaraj era begins in
638 A.C.; therefore, the year 654 of that era is equivalent to 1292 A.C.,
between which number and that of the Buddhist year given in the inscrip-
tion, the difference is 545 and not 544 or 543. Swamikannu Pillai who
has verified this date has pointed out that, in this instance, the Sakkaraj
year has to be taken as expired and not current, as usual with that era, in
order to make the 6th waning of Tazaungon fall on a Thursday. His
equivalent in the Christian era for this date is Thursday, 22 October, 1293
A.C.65 If the year of the Buddhist era was an expired one, the initial
point of that era would fall in 545 B.C. As this document uses expired
years for an era which normally quotes current years, the probability is
that the Buddhist year referred to was also an expired one.

Two dates in the Buddhist era with a difference of two between them
have been given in Burmese inscriptions for the same event. In an inscrip-
tion set up by his son some time after the death of the famous Burmese king
Kyanzittha, that monarch is said to have begun his reign when 1628 years
had elapsed after the Parinirvana of the Buddha,66 while in an inscription
of Kyanzittha himself, that event is said to have taken place in 1630 B.E.67
This difference has been explained by taking 1630 B.E. as the date of the
king's coronation, and the earlier date as that of his acccssion.v" But pre-
cisely the same difference will be found between the date of the accession
64. 'The Revised Buddhist Era in Burrna ' in the JRAS for 1910, pp. 474·6.
65. Ind. Eph., Vol. I, part 2, p. 132, No. 54.
66. Ep. s«; Vol. I, p. 51.
67. Ep. Bi1'., Vol. I, p. 141.
68. Ep. Bir., Vol. I, p. 4.

142
 THE BUDDHIST ERA IN CEYLON

of Sahasamalla as stated in that monarch's inscription, and if it were stated

in the Buddhist era continued down to that time, according to the reckoning
:in this inscription. It is therefore very probable that, among the Burmese
Buddhists in the eleventh century, there were two different methods of the
,reckoning of the Buddhist era, between which a difference of two years was
possible for some dates.

Of these two years, one can be explained as the difference between

,expired (gata) and current (vartallliil1a) reckonings. A consideration of the
interval between the death of Mahascna and the twenty-eighth year of U pa-
.tissa Ihas led us to infer the possibility of a new year in the Buddhist era
,li,aving started shortly before the date given in the present inscription, i.e.
~ new moon of Duratu, in other words that the beginning of the Bud-
/~lllst year was not then on the full-moon day of Vaisakha, And, in
;t~~, Fleet has established that, in the Dipavamsa, there is evidence to provc
i~. the .Buddhist year at the time o~~he writin$ of,th~t chronicle di~ not
"~ WIth the full-moon day of Vaisakha.69 It is said 111 the Dipavamsat"
,~t Devanarnpiya Tissa was anointed as king 236 years after the death of
Jibe Buddha, that is in the 237th year, seventeen years and not quite six
.;JPonthsafter the anointment of Asoka. The reference here is clearly to his
~~t coronation, for the statement is followed by an account of the precious
}~gs which miraculously appeared after his first coronation. The first
1:;oronationwas in the month of Maggasira (Skt. Margasir~a, Sin. Uiiduvap).
<The Dipavamsa also states that Mahinda arrived in Ceylon on the full-moon
:'~y of jettha (Skt. jyesrha, Sin. Poson), 236 years after the Parinirvana of
~.;'flae Buddha," that is in the year 237 of the Buddhist era. If an year of the
~lluddhist era ended on the day before the the full-moon of Vaisakha, and
",anotherbegan on the full-moon day, the year of Mahinda' s arrival in Ceylon
,:shouldhave been 238 B.E. current, for there was a Vaisakha full-moon inter-
"~ between that event and the first coronation of Devanampiya Tissa.

'. ,The Dipavamsa informs us that Devanampiya Tissa was anointed for
:.the first time on a day in the second month of the Hemanta Season (Magga-
::Iira) when the constellation Asalhi (Skt. A~a~ha) was in the ascendant (1st or
;;2ild of the waxing moon) in the seventeenth year of Asoka, not quite six
'months after the commencement of that year. Full seven months after the
,6rst anointment of Devanampiya Tissa, the same authority states, and in the

'f, 69. JBAS for 1909, pp. 10 ff.

70. Chapter XVII, v. 78.

143

;s-
.{:
k~·
iL_
--
UNiVERSITY OF CEYLON REViEW

eighteenth year of Asoka, Mahinda arrived in Ceylon on the full-moon day

of Jegha.71 From these statements it can be inferred that the anniversary of
the anointment of Asoka was shortly after the first day of the bright fortnight
of Jegha. Thc Dipauamsa also informs us that Asoka was anointed 218
years after the death of the Buddha, that is in the 219th year. Then, on the
basis of the full-moon of Vaisakha as the commencement of the Buddhist
year, 'the seventeenth anniversary of the anointment of Asoka carne
about three weeks after the end of the year 235; and the first anointment
of Dcvanampiya Tissa, when the 236th had elapsed, was not six months, but
eighteen months after that.'72

These seeming errors in the chronology of the D ipavati, sa, Fleet has
pointed out, disappear if the commencement of the Buddhist year were
taken to have been on Karttika, sukla 8th, which, according to the Sarvasti-
vadins, was the day of the Buddha's Parinirvana, not the full-moon day of
Vaisakha. Fleet has further argued that the details given in the Maha-
parinibbana-sutta are more in favour of Karttika, sukla 8, than of Vaisakha
full-moon as the day of the Buddha's death,73 The Mahaparinibbana-sutta
certainly does not support the belief that the Buddha's Parinirvana took
place on a Vaisakha full-moon day, but the account of the First Council
at Rajagaha given in the ClIllavagga apparently takes for granted that this
great event fell on that day.74 Buddhaghosa in the Salilalltapasadika and the
Mahiivamsa categorically state that the full-moon of Vaisakha was the day
on which the Buddha died.75 It is possible that there were two schools of
opinion among the Buddhists of ancient Ceylon on this matter. The
Sa11lalltapasadika and the Mahiivantsa give of course the view of orthodoxy
held by the Mahavihara. The Abhayagiri and the Jetavana sects possibly
held a different view, and the chronology of the Dtpavatissa could have been
influenced by that.

But there is evidence to indicate that the commencement of the Bud-

dhist era in ancient Ceylon did not tally with what was believed to be the
exact date of the Buddha's death. In his sub-commentary on the Vinaya,
entitled Siiratthadipani, the erudite Siiriputta-mahasami, commenting all
the statement of the Smllalltapasadika that Mahinda-thera arrived in Ceylon
71. Chapter XU, vv, 42·43.
72. JRAS for 190!l, p. It.
73. JRAS for 1909, pp. is ff.
74. T. W. Rhys Davids and H. Oldenberg, Viltttyu Puako, t.ranslat.ion, part, III,
(S,B,E. Vol. XX), pp. 376 ff.
7;). Samanta-pdsdtlikii, P.T.S. Edition, \'01. 1, p. 4, Muluicamsu, Chapter Ill, VV.
I and 2.

144
 THE BUDDHIST ERA IN CEYLON

in the year 236 from the Parinirvana of the Buddha, says as follows: ' "Pari-
,nibbiQa" means the year of the Parinirvana. The year of the Parinirvana
which is the limit should be left out, and the meaning should be taken as that
it was in the year 236 after that. '76 The reckoning of the Buddhist era
according to the reigns of the Ceylon kings, the first of whom Vijaya,
,tradition asserts, arrived in Ceylon on the very day of the Buddha's death,
is thus explained by Sariputta i=- 'Reckoning that the first year of Vijaya
here was the year of the Buddha's Parinirvana, that year should be left out
.and the two hundred and thirty-six years after the Parinirvana should be
made as follows: 37 years of Vijaya, 1 year of interregnum after that, 30
yearsof Panduvasudeva, 20 years of Abhaya, 17 years before the consecra-
tion of Pandukabhaya, 70 years after the consecration, sixty years of Muta-
siva and the first year of Devanampiyatissa.T' Sariputta also states that
24 years from the 32 years' reign of Ajatasattu have to be reckoned for the
'Buddhist era, leaving out eight years, as the Parinirvana took place in the
eighth year of that Magadhan king.78

From these statements it is clear that there was a tradition among

Ceylon scholars to the effect that the commencement of the first year of the
','Buddhist era coincided with the commencement of the year following that
in which the Parinirvana took place. When the Parinirvana itself took
'place,there was of course no reckoning from that event, and what is meant
must have been the commencement either of the regnal year of the king
reigning at the time, or of the civil year. In ancient India the civil year
commenced either with the month of Caitra or with Karttika.I? There is
also evidence that, in mediaeval Ceylon, the religious year of Buddhist
monasteriesbegan with Karttika, probably to conform to the civil year.80
If this practice had come down from the time of our inscription, as seems
likely, the commencement of a New Year in the Buddhist era would have
been on the first tithi of the waxing moon of Karttika if the amdnta system

""Mm 76. Pariniboo1J<lw ti parinibbiitllt-vassato

VCl88a·satiinam wpari chauimsatime
; tam auadlii-bh.iuam. niuiicittxi taio 'add/uti;'
vasse ti attho qahetabbo. Sdruttluuliparii,
edited by Siri Nal.lissa.ra-thera, Colombo, B.E. 24,j8, part I, p. 131.
77. Sammiisambuddha88a parinibbtina-oassum. idlui Vijallussu puthumarh VU81>(l/i! Ii
IIomi tam apanetvii parinibbtina-oassato uddham. Vijayct8/;a saitatimsu-uussccm. tato anijilann.
eblm VCl88amParu!uviisudeva8sa timsu-txcssdrui AbhaYCt8s(t uisaii-uasstin: P([(~(lukab//(/y(j,88((
aMNekato pubbe sattarasa uassiini. abh.isittassa. sauati va8siini 1'.1utasinxissa. 8a!?hi-va88ani
DeodnampiYa88a pcuhanuui» oassam. t'i eoash. parinibbiinato dvinrwrh uassusauincun. upari
""fMa'VCI88iini veditabbom. SiirattlwdIpani, op. cit., p. 134.
, 78. Commenting on Ajii.tasattu catuuisati-sxissarn. in the Salnantapclsadika (P.T,:-:.
Edition, p. 72), the SiiratthadIpan-i says , Parin'ibba~w.-va8su-8arhkhat(/'/h auhuma-caescnii
~ vuttam. Siiratthadipani, op. cit., p. 134.
79. Ind. Eph., Vol. I, part I, pp. 53 ff. The ValabhI year, for cxarnplc, was Kart.t.ikad i.
80. EZ., Vol. III, p. 268, f. n. 4.

145
 UNIVERSITY OF CEYLON REVIEW

was followed, or the first tith; of the preceding dark fortnight if the piirni-
manta system was followed, i.e. seven or twenty-two days before Karttika,
sukla 8 which, according to Fleet, was the exact day of the Buddha's death
and the commencement of the Buddhist year. The chronological diffi-
culties created for the Dipavamsa by taking Vaisakha full-moon as the com-
mencement of the Buddhist year will be removed just as well by taking the
first day of the Buddhist year to have been Karttika, sukla 1 of the auuinta,
or bahala 1 of the pijr~1itnal1ta month. But, in our subsequent discussion,
we take sukla 8, as suggested by Fleet; if one of the other two tithis sug-
gested above be preferred, the necessary adjustment can be made easily.

Now, according to Sahasamalla's inscription, the year 1743, expired,

of the Buddhist era ran from the Vaisakha full-moon in 1200 to Vaisakha,
sukla 14 in 1201 A.C. This same period was the Buddhist year 1744, cur-
rent. Calculating backwards on this basis, 16 December, 396 A. C.
would have fallen within B.E. 939 expired, or 940 current, which com-
menced on 9 April, 396, and continued up to 27 April, 397 A. C,ll!
But, if the Buddhist year began at that time on sukla 8 of Karttika,
a new year, i.e. 941 current, would have commenced on 26 October of
that year. The date of our inscription, being in December, would thus
fall within the Buddhist year 941, and this is what is stated in our inscription.
Again, calculating forwards from the date of our epigraph to the year of
Sahasamalla's inscription, if August 396 fell within the Buddhist year 940
current, the Buddhist year current in August 1200 would have been 1744,
which, as an expired year, would be 1743, just as it is given in the inscription
of Sahasamalla, Thus there is no difference between the two reckonings
for the greater part of the Christian year if the current year is considered.
The difference of one is for that part of the Christian year beginning with
the day corresponding to Karttika, sukla 8, up to its cnd, i. e. roughly
form October to December.

This difference has been brought about by the shifting of the com-
mencement of the year from Karttika, sukla 8, to Vaisakha full-moon,
thereby also shifting the previous starting point of the era from a day in
September or October, 545 B.C. to a day in March or April, 544 B.c'
Thus it will be seen that, apart from the usual difference created by citing
expired or current years-a complication common to many Indian eras-
and the shifting of the starting point six months forwards, the Buddhist era
81. These dates have been calculated from the new-moons given in Ind. Eph., Vol. I,
part I, pp. 214-215.

146
 THE BUDDHIST ERA IN CEYLON

in our inscription is identical with that used in the Sahasamalla inscription.

When this shifting of the starting point from Karttika to Vaisakha was
effected, we have no means of ascertaining, but the unusual procedure
adopted in the Sahasamalla inscription of giving the number of months and
days in B.E. 1743, expired, up to Binara (Bhadrapada), sukla 12, indicates
that the writer of the document wished to leave no room in the mind of the
reader as regards the day on which the year commenced. If there was only
one day of commencement for the year universally accepted, such emphasis
on the Vaisakha full-moon would have served no purpose. The mere
mention of Binara, pura 12, would have been quite sufficient. The inference
therefore is that, even by the end of the twelfth century, there were schools
which postulated a day other than the Vaisakha full-moon as the commence-
ment of the Buddhist year. That this was so in Burma is proved likely
from the evidence we have already discussed.

It is not impossible that, even in the time of our epigraph, there were
those who computed the Buddhist year from the full-moon day of Vaisakha,
side by side with others who preferred a day in Karttika. If the adoption
of Vaisakha full-moon as the commencement of the year was a later deve-
lopment, the change, when decided upon, could have been effected in one
of two different methods. The Vaisakha full-moon of the year current
could have been fixed as the commencement of a new year, which was
given the number consecutive to that of the current year. Or the current
year may have been allowed to run its full course and the Vaisakha full-
moon of the following year could have been reckoned as the commence-
ment of the next year. Apparently, the second method had been adopted,
shifting the starting point of the era some six months forward.

How far prior to the time of our inscription a Buddhist era was in use
we have no means of determining. Our inscription cannot be far removed
in date from the time that the Dipauaihsa took the shape in which it has
come down to US.82 And, in that chronicle, many events connected with
the history of Buddhism and the political history of India and Ceylon arc
dated in years after the Parinirvana of the Buddha, though this dating has
not been continued after the time of Dcvanarnpiya Tissa. It is possible that
the adoption of the era of the Buddha's Parinirvana was due to that same
interest in the history of the Buddhist religion in this Island which resulted
in the compilation of the Dipavamsa.
82. The Di/pauamsa gives an account of Muhasena : it was therefore finally redacted
after the reign of that king. It has been quoted by Buddhaghosa who wrote 1Il t-he reign
of Mahii.nama, the successor of U patissa 1.

147
 UNIVERSITY OF CEYLON REVIEW

It is quite possible that the fixing of the starting point of the era was
based on data which were not accurate. But, once the era was fixed, the
dates in it would be quite reliable for the period after that, if it continued in
use without a break. That the Buddhist era continued without a break
from the time of Sahasamalla, or even of Parakramabahu I, is admitted by
all competent scholars. And we hope that we have succeeded in establish-
ing that the era in use in the Polonnaru period was identical with that made
use of in the present record. Chinese synchronisms have established that
the dates in this era givc accurate results for the fifth century. For the period
preceding the date of the present inscription, the accuracy of a date in the
era would depend on the reliability of the material utiliscd when it was
first computed, namely the regnal years of Ceylon kings as given in the
Chronicle. I have elsewhere given reasons for the trustworthiness of these
details up to the time ofDughagamal).i.83 The details of the reigns of kings
between Dutthagarnani and Devanarnpiya Tissa are manifestly unreliable ;
hence the date that has to be assigned to Devanarnpiya Tissa based on dates
in the Buddhist era does not tally with that assignable to him on the basis
of Asoka's date deduced from the mention ofGrcekkings in his edicts. Thus,
the prevalence of the Buddhist era in Ceylon at the close of the fourth cen-
tury A.C. by no means vouches for the accuracy of the date of the Buddha's
Parinirvana that might be arrived at by the determination, from the data given
in our record, that the year 941 from that event corresponds to 396 A.C.

We now proceed to arrange the dates of kings of Ceylon from Duttha-

gamal;i to Dathopatissa II in accordance with the evidence supplied by this
inscription, i.e. that 16 December, 396 A.C., fell during the currency of the
2Hth year of Upatissa I. The inscription does not state how far the twenty-
eighth year of the king had run its course on that day. The two extremes
of possibility are that this day, the new-moon of Durutu in that year, was
the first day or the last day of the twenty-eighth year. The twenty-eighth
could therefore havc commenced on any day from the day after the Durutu
new-moon of the preceding ycar, i.c. 29 November, 395, up to the day
recorded in this inscription, 16 December, 396; the end of the ycar could
have been any day from 16 December, 396, to the day beforc the Durutu
new-moon of the following year, i.e. 4 December, 397. The accession
of the king would consequently have been any day between the day after
the Durutu new-moon of 368 and the day before the Durutu new-moon
of 370, i.c. 27 November, 368 and 3 December, 370 A.C.BI We shall
take the mean between these extremes of possibility and fix the date of
83. EZ, Vol. V, p. 03.

148
 THE BUDDHIST ERA IN CEYLON

the accession of Upatissa I as 369 A.C., and assuming that this monarch's
42nd year was not completed at the time of his murder, take the first year
of his successor Mahanama as 410 A.C.

For fixing the reign periods of the kings who came before Upatissa I,
we shall, following Geiger and Wickremasinghe, give due consideration
to the number of years which the P/-Uiivali and other Sinhalese chronicles
state to have elapsed from the Buddha's Parinirvana to the end of Malia-
sena's reign, but adopting the number 846, instead 0£844, as it accords with
the evidence of this inscription. Though Geiger and Wickremasinghe both
adopt this number as trustworthy, the periods of reigns which they assign to
thekings of Ceylon from Vijaya to Mahasena do not add up to that number.
This is mainly due to the fact that out of the readings of the text of the
Chronicle found in manuscripts, in the statements referring to the lengths of
the reigns of certain kings, Geiger, in his edition, has adopted readings
giving a lower number. An instance is the duration of the reign of Kanittha
Tissa, which according to Geiger is 18 years. But the commentary to the
Mahiivatilsa has expressly noted that the reign lasted for twenty-eight years.
This authority states that there is also a reading of the text giving the num ber
of years as eighteen, but that this is an error and that the correct
reading is the word meaning twenty-eight found in the ancient Anhakatha.R4
Geiger has paid no heed to the commentator, and has placed 1110rereliance
on modern manuscripts than on his statement. All the manuscripts of the
PUjavali consulted by Mabopitiya Medhankara Thera also agree in giving
a word meaning eighteen as the length of Kanittha Tissa's reign. But the
lengths of the reigns of kings from Vijaya to Mahasena as given in the
Pujaval; add to a number short by thirteen of the total given in that work
for the period.

Even when we give twenty-eight years, instead of eighteen, to Kanittha

Tissa, there is yet a deficiency of three in the total. This is due to the (1Ct
that the Piijiivali gives only six years to Ilanaga, i.c. the period of his rule
after his return from exile, and does not take into account the three years of
rule of the Lambakannas and the six months he was on the throne before
he was forced into exile, as stated in the Mahiiva/ilsa.85 These three years
have to be separately shown as an interregnum, as done by Wijesinha, or
added to the six years of Ilanaga, as done by Geiger and Wickremasinghe.
84. Vamsatthappakasini, P.T.S. Edition, p. 6,,9.
Aithaoiea-sanui rajjam. ti auhuuisati-vaestini Lankiidi pe rajjam ak/iraui : katthaci
atthiirasa soma ti likhanti, tarn purmida-Leklurm: eram hi .V!Ii"kalhr7ym;1 rutt arh.
8,i. Mnhiiocnnsa, chapter XXXV, YY. 1;j·4;i.

149

,.
, r
UNIVERSITY OF CEYLON REVIEW

With these two necessary corrections, the details in the P'Uiivali up to the
end of Mahasena's reign, when added up, give the total as stated in that
authority.

The Pujiivali gives for Anu!a and her paramours 5 years and 4 months,
while in the Pali chronicles the total period of their power was 4 years and
three months. The number in the PiiJavalf has to be adopted to get the total
up to the end of Mahasena given in that text. Though Geiger and Wickre-
masinghe, and following them Mendis, have adopted numbers which do not
add up to the total required at the close of Mahasena's reign, they adjust
matters to tally with that number by making allowances in the case of some
reigns for what are called fractions of years after completed years, i.e. in
effect they have made the lengths of some reigns exceed by one the numbers
given for them in the chronicles.

Just like the total at the end of the reign of Mahasena, the number of
years that had elapsed from the Parinirvana to the accession of Duttha-
gamal)i-382-is an important link in the traditional chronology based on
the Buddhist era. All previous systems of chronology have adopted this
number and we have also followed the same procedure, making adjustments
at the ends of three reigns for possible fractions of years after completed
years. In order to get this number, it is necessary to allot twenty-two years
to Sena and Guttika, as given in some manuscripts of the Mahdvamsa, and
not twelve, the number of years assigned to them in the Dfpavafilsa.86
Assuming that these details in the chronicles are trustworthy, there is always
the possibility, as we go further from the time of this inscription, of an error
of one to five due to the fractions of years in current years, or after expired
years, at the close of reigns. According to the Mahiilla/ilsa,87 the Abhaya-
giri-vihara was founded by Vanagamal)i Abhaya in the first year of his
restoration, when 217 years, 10 months and 10 days had passed since the
founding of the Mahavihara. The Gal-vihara inscription88 states that this was
454 years after the Parinirvana, i.c. in 89 B.C. if the epoch of the Buddhist
era be taken as 544 B.C. The present revised chronology has taken this
into consideration and has given 89 B.C. as the year of Vattagamani's
restoration on the assumption that the year of the Buddhist era was current.
But, as the dates for this period can only be approximate, it has not been
considered necessary to make adjustments to suit the slight discrepancy in
the epoch of the Buddhist era indicated by the present inscription.
86. Dipaoams«, chapter XVIII, v. 4;.
Ht, Chapter XXXIII, v. 80.
RR. "JZ. "01. II, p. :?i3·

150
 THE BUDDHIST ERA IN CEYLON

For the period before Mahascna, up to Dcvanariipiya Tissa, there are

no synchronisms with established systems of chronology by means of which
the accuracy of the dates assigned to Ceylon kings can be tested. Contact
with Rome in the first century A.C., however, appears to be of some
significance for early Sinhalese chronology. The commentary of the
Mahiivmilsa tells us that Bhatika Abhaya (or Tissa) sent envoys to Rorna-
nukha-rattha and obtained coral thercfrom.s? There is no doubt that the
Roman empire is meant by this name. As is well-known, Pliny has told
us the story of the freedman of Annius Plocamus, who farmed the Red Sea
taxes, being carried out of his course to the coast of Ceylon, hospitably
treated by the king of the Island and returning to Rome with Sinhalese
envoys in the time of Emperor Claudius.w Inscriptions, or rather graffiti,
have recently been found in thc Eastern Desert of Egypt, containing the
name of Lysas, a slave of Publius Annius Plocamus. One of these graffiti,
in Greek, is dated in the thirty-fifth year of an Emperor who could have
been none but Augustus. Sir Mortimer Wheeler remarks on this: 'Identity
of this Annius Plocarnus with Pliny's is not proved, and that of the two
freedmen is not suggested, but the coincidence of the name in so appropriate
a geographical setting amounts to near-proof in respect of Plocamus, and
it would be wise to consider the date of his errant freedman in Ceylon as
likely to have been appreciably earlier than the reign of Claudius.'?' If the
view so cautiously put forward by Sir Mortimer Wheeler be accepted, it
would mean that trade relations were established between Ceylon and
Rome in the first or second decade of the Christian era. And Bhatika
Abhaya who, according to the Sinhalese historical tradition, sent his envoys
to Rome and obtained coral therefrom, reigned, according to the chrono-
logy based on our record, from 18 B.C. to 10 A.C. He was thus a con-
temporary of the great Roman Emperor Augustus, in whose thirty-fifth
year (6 A.C.) is dated the graffiti of the freedman of Annius Plocamus.
The year of Mahanama's accession, according to the evidence of this
inscription, has now to be 410, and not 406 A.C., as I have previously taken.
This necessitates a corresponding alteration in the dates, given by 111e,of
the kings who succeeded Mahanama. As I have pointed out in the essay
referred to at the beginning of this paper, the dates given by Wijesinha,
Wiekremasinghe and Geiger to kings from Kitti Sirimegha have become
erroneous as these scholars have assigned to that ruler only nineteen days,
whereas, in point of fact, his reign extended to nineteen years. In the essay
89. T"ari!saUlwppakiisilli. P.T.S. Edition. p. ti30.
90. Pliny, VI, 84·!H.
91. Sir )Iortimer Wheeler. Rome Beyond Lm pcrial Frontier. •. p. 128.

151
 UNIVERSITY OF CEYLON REVIEW

in question, I have suggested a method of distributing this error of eighteen

years without affecting the dates of the kings of the ninth and tenth
centuries, which can be decided within narrow limits by the evidence
of Cola inscriptions. These establish that the thirty-sixth year of Mahinda
V, when he was captured by the armies of Rajcndra-co]a, should be 1017.92
In agreement with this, it can be deduced from the data in the Kaparararna
Sanskrit inscription that the reign of Mahinda V began in 982 A.C.93 From
this date, the chronology can be reliably worked upwards, according to the
lengths of reigns given in the Chronicles and the Chinese and South Indian
synchronism, to Aggabodhi IV. The difference of four years, referred to
above, has therefore to be adjusted before the reign of Aggabodhi IV. And,
in IIIy opinion, this can be effected in that period of political confusion
between Aggabodhi II and Aggabodhi IV.

The chronology of the Cij !a/!(//ilsa for this interval is not so hopelessly
confusing as it appears at ftrst sight. Aggabodhi II reigned for ten years,
and was succeeded by Sariighatissa whose right to the throne was challenged
and who was ousted by Moggallana III. The Cii lavamsa docs not say how
long Sariighatissa remained in power, but its narrative indicates that he was
vanquished by his rival very soon after he came to the throne. According
to the PI-Uiil'a/i, he reigned for two months only. Moggallana III reigned
for six years, and his successor, Silameghavanna, for nine years. Aggabodhi
llf, the next king, was forced to flight by Dathopatissa I in the former's
twelfth year. Aggabodhi III returned with an army from South India, but
was defeated by Dathopatissa I, and fled to Rohana where he died in his
sixteenth year. The actual period of the rule of Aggabodhi III at Anuradha-
pura was therefore twelve years. The last four years of Aggabodhi III and
the ftrst four years of Dathopotissa I were therefore concurrent. The
yuvarii]« of Aggabodhi III, Kassapa (the second of that name), forced Da-
thopatissa l, in his turn, to flight, but the Chronicle docs not say exactly how
long after his coming to power Dathopatissa was forced to abandon his
throne, Dathopatissa returned from India, but was defeated and slain by
Kassapa II. At this point, the Chronicle says that twelve ycars had passed
since he became king.94 Dathopatissa did not actually rule for twelve
years fr0111 the time he captured power. From the Chronicle itself, it can-
not be known for certain when Dathopatissa was forced to flight, and how
long he remained in India before he returned to be finally vanquished by
!l:!, .t RA."" 1'01' I !ll:!. p. ;;:!3,
!I:~, I,'?;, Yo!. v, "I', lii:1 ff.
')4, Ciilurnms«, r-huptr-r 44, vv, I:?(i, I,U and 14,;,

152
 THE BUDDHIST ERA IN CEYLON

, his rival. But this seems to be the obvious point at which should be de-
. 4ucted the four years which overlap when we calculate backwards from
_the certain date of the thirty-sixth year of Mahinda V, and forwards from the
: equally certain date of the twenty-eighth year of Upatissa I. Dathopatissa I,
~uently, has been taken in this revised chronology, as having fled in
',the eighth year of his reign .
• ' ? In accordance with the prcccding discussion, the dates of the Ccylon
.;~ from Dutthagamani to Darhopatissa II arc given below:

B.C.
Dunhagamat;li (S. Dutugamunn) 161-137
Saddhatissa (S. SadHtis) 137-119
Thiilatthana (S. Tulna) 119
Laiijatissa (S. Lamani Tis) 119-109
Khallata Naga (S. Kalun-na] 109-103
Vanagama~ll Abhaya (S. Valagam Aba) 103
Pulahattha '\
Bahiya I
Panayarnara r Pafica-Dravida 103- 89
Pilayamara J
Dathika
Va~ragamat;li (restored) 89- 77
Mahacli!i Mahatissa (S. Mahasiju Mahatis) 76- 62
Coranaga 62- 50
Tissa (S. Ku~a Tissa) 50- 47
Siva (S. Balat Sivu)
Vatuka
Da~ubhatika Tissa
Niliya (S. Purohita Bamuna, Vasukhi]
Anu!a (Queen) 47- 42
Kutakanna Tissa (S. Makalan-tis or Kiilakanni Tissa) 41- 19
Bhatika Abhaya (also called Bhatika Tissa or
Bhiitiya Tissa) B.C. 19-9 A.C.

A.C.

Mahadathika Mahanaga (S. Mahadaliya Mana) 9- 21

Anlat;l~a-gamal)i Abhaya (S. Ada Gamlll)u) 22- 31
Kanirajinu Tissa (S. Kinihirada]a) 31- 34
Ciilabhaya (S. KU9a Abha) 34- 35

153
 UNIVERSITY OF CEYLON REVIEW

Queen Sivali (Rcvati) 35

Ilanaga (S. Elunna) 35- 44
Candamukha Siva (S. Saiidamulrunu] 44- 52
Yasalalaka Tissa (S. Yasasilu) 52- 59
Sabha (Subha) 59- 65
Vasabha (S. Vahap) 65-109
Vankanasika Tissa (Vaknahf Tis or Vaknasi Nambapn) 109-112
Gajabahuka-gamani (Gajabahu I, S. Gajaba) 112-134
Mahallaka Naga (S. Mahaluna or Mahalumanfi] 134-140
Bhatika Tissa (S. Batiya) 140-164
Kanittha Tissa (S. Ciilatissa) 164-192
Khl~janaga (S. Kuhunna) 192-194
K lliicanaga (S. Kudana) 194-195
Sirinaga I (S. Sirina or Kllc;la Sirinii) 195-214
Voharika Tissa (S. Vera Tissa) 214-236
Abhayanaga (S. Abha Scn or Abha Tissa) 236-244
Sirinaga II (S. Sirina) 244-246
Vijaya-kumara (S. Vijayiildu) 246-247
Sarnghatissa I 247-251
Sirisariighabodhi (S. Dahami Sirisal\gabo) 251-253
Gothabhaya or Meghavanna Abhaya (S. Go]u Aba) 253-266
Jcnhatissa I (S. Kalakan Dctatis or Makalan Dctatis] 266-276
Mahascna (S. Mahascn) 276-303
Sirirncghavanna (S. Kit Sirimcvan) 303-331
Jcnhatissa II (S. Detatis) 331-340
Buddhadiisa (S. Bujas) 340--368
Upatissa I 368-410
Mahiinfima 410-432
Chattagiihaka Jantu (S. Liim'ini Tis) 432
Mitrascna (Mitscn) 432-433
Pandu ') 433-438
Pa~i;lda I 438-441
Khudda Pfirinda I S . S a d - D ravt
- id a 441-456
T·mtara
- ~
I· . . 456
Dathiya I 456-459
Pithiya J 459
Dhatusena (S. Dasen-Kdliya) 459-477
Kassapa I (S. Sigiri Kasubu) 477-495
Moggallana I (S. Mugalan) 495-512
K umara-Dhatusena (S. Kurnjirndiisa, K umaradas) 512-520

154
 THE BUDDHIST ERA IN CEYLON

Kittisena (S. Kirttiscna) 520-521 t,

Siva (S. Mandi Siva) 521
Upatissa II (S. Lamani Upatissa) 521-522
Silakala, Ambasamanera (S. Lamani Al11bahcralfa Salamevan) 522-535 .,
t,
Dathapabhuti (S. Dapulu-Sen) 535 !;

Moggallana II (S. Dala-Mugalan) 535-555

Kitti Sirimegha (S. Ku<;laKitsirimcvan) 555-573
Mahanaga (S. Senevi-Mahana) 573-575
Aggabodhi I'(S. Akbo) 575-608 H
I'
,.
Aggabodhi II (S. Kuda-Akbo) 608-618
Sarnghatissa II 618
Moggallana III (Dalla Moggallana, S. Lamiini Bona
Mugalan) 618-623
Silameghavanna (S. Asiggahaka) 623-632
Aggabodhi III, Sirisarnghabodhi (S. Sirisaiigabc) 632
Jcnhatissa III (S. Lamani Katusara Detatis) 632
Aggabodhi III (restored) 633-643
Dathopatissa I (S. Lamani Dalupatis) 643-650
Kassapa II (S. Pasulu Kasubu) 650-659
Dappula I (S. Dapulu) 659
Hatthaddtha (Dathopatissa II, S. Lamani Dalupatis) 659-667

S. PARANAVITANA

155