गुप्त अभिलेख संग्रह (जे. एफ. फ्लीट - गुप्त सम्राटों एवं परवर्ती नरेशों के ऐतिहासिक पाषाण व ताम्रलेख)

86 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
Kârttika (October-November). This, however, apart from the natural presumption, is
rendered absolutely certain by the concomitant mention of the Hijra year; from which,
since the month Âshâḍha answers ordinarily to June-July, it is plain,—as was pointed out
by Dr. Hultzsch in editing the inscription,¹ and by General Cunningham in commenting
on the date,²—that the English date which we have to look for lies in or about June
or July, A.D. 1264. This at once removes the possibility of any reference to the northern
Vikrama year; since the month Âshâḍha of the northern Vikrama-Saṁvat 1321 current,³
is represented by June-July of the preceding English year, A.D. 1263. Also, since the
period June-July, A.D. 1264, fell in Śaka-Saṁvat 1187 current, it removes any real neces-
sity of making calculations for Śaka-Saṁvat 1186 current and 1188 current; the results,
however, for these two years will be given, in order to help in setting the general question
entirely at rest.
So much, as regards the English equivalent of the given date lying in A.D. 1264, had
been clearly pointed out by Dr. Hultzsch and General Cunningham. But nothing fur-
ther, that I can find, was said about the details of the English equivalent, until just before
the time when I first had occasion to consider this date;⁴ when, in a letter dated the 3rd
December, 1885, General Cunningham announced to me that the equivalent English date
is Sunday, the 25th May,⁵ A.D. 1264.
This result, Sunday, the 25th May, A.D. 1264, is, as will be seen below, the cor-
rect one. But, in order to explain its full bearing, more is requisite, than simply to state
it; especially because it is necessary to shew clearly, in the face of what has elsewhere been
written about this date, that this result is not obtained from the mention of Vikrama-Saṁvat
1320, though it does answer the requirements of that mention; i.e. that it is not the
result for a year commencing on the first day of the bright fortnight of the month Kârttika
which fell in A.D. 1263; and, consequently, that, even apart from what I have already
established, this record proves that Valabhî-Saṁvat 945 commenced at any rate not with
that day.⁶ And here I would remark incidentally, that no argument, of identity between
¹ Ind. Ant. Vol. XI. p. 241. ² Indian Eras, pp. 50, 53, 63.
³ According to the present Tables, the given Vikrama year 1320 must be interpreted as an
expired year; and the indicated current year is, therefore, 1321. This, however, does not make the
given Valabhî year an expired year; just as it does not, and can not, convert the given Hijra year
into an expired year.
⁴ Ind. Ant. Vol. XVI. p. 147 ff.
⁵ The ordinary equivalent of Âshâḍha is June-July. But the dark fortnight of the northern
Âshâḍha corresponds to the dark fortnight of the southern Jyeshṭha, which month is ordinarily May-
June. And this, with the fact that Śaka-Saṁvat 1187, current, commenced rather early, on Saturday,
the 1st March, or Friday, the 29th February, A.D. 1264, is the reason why this dark fortnight of
Âshâḍha fell entirely in May.
⁶ As opposed to these results Gen. Cunningham (Indian Eras, p. 53) has quoted this record as
proving absolutely that Valabhî-Saṁvat 1=A.D. 319 (.20); which could only be by taking the epoch as

THE GUPTA INSCRIPTIONS; INTRODUCTION. 87

the two years, can be based on the mere fact that the record mentions both a Valabhî
year and a southern Vikrama year. It might just as well be asserted that the mention also
of the Hijra year 662, shews that the scheme of the years of that era, too, is identical with
the scheme of the southern Vikrama year; whereas,—even apart from the fact that the
Hijra year 662 commenced, as stated above, on Sunday, the 4th November, A.D. 1263,
while the southern Vikrama-Saṁvat 1320, as an expired year, i.e., more properly, southern
Vikrama-Saṁvat 1321 current, commenced on Friday, the 5th October of that same
year,¹— everyone knows that these two eras have absolutely nothing in common at all ;
the Hijra era being a purely Musalmân era. The Verâwal record is simply analogous
exactly to scores of documents that are still being turned out in India, in which the date
is expressed according to the English and also one or other of the Indian systems ; and in
which the principal record depends entirely upon the person by whom, and the circum-
stances under which, it is drawn up. Sometimes it will be the English date; sometimes
the Indian. We shall see immediately that, in this Verâwal inscription, the principal
record is the Valabhî date; and that the Vikrama date accidentally, as well as the Hijra date
naturally, was entirely subordinated to it. Possibly, we may hereafter obtain instances, in
which the reverse of this will be found to have been the case. But they will not avail to
disprove any of the pointed and unavoidable conclusions, regarding the epoch of the
Gupta-Valabhî era and the scheme of its years, which are absolutely forced on us by the
circumstances of the present Verâwal date.
The second point to be noted· is, that the month Âshâḍha which fell in A.D. 1264,—
i.e. both the Âshâḍha of northern Śaka-Saṁvat 1187 and northern Vikrama-Saṁvat 1322,
and the slightly different Âshâḍha of southern Śaka-Saṁvat 1187 and southern Vikrama-
Saṁvat 1321,—was an intercalary month.² The effect of the intercalation was as fol-
lows :—The initial day of Śaka-Saṁvat 1187, both northern and southern, and of the
A.D. 318-19, and by treating the years as commencing, from the beginning, on Kârttika śukla 1.—Also
(id. pp. 50, 63) he treats Vikrama-Saṁvat 1320 as the leading record of the date; and very clearly
implies throughout, though he does not actually state, the identity of the scheme of the Valabhî and
southern Vikrama years.
¹ Cowasjee Patell’s Chronology, p. 150.
² See Indian Eras, p. 179.—The fact is also proved by K. L. Chhatre’s Tables. In C. Patell’s
Chronology, p. 150, the intercalation is entered opposite Śaka-Saṁvat 1186 (expired) and southern
Vikrama-Saṁvat 1321 (expired). This is in accordance with a mistake that runs all through his
Table I. pp. 94 to 183. The intercalations are placed by him correctly for the Śaka years. But he has
omitted to point out that, in applying them to the Vikrama years, which, throughout his Table, are
the southern Vikrama years, they must, in consequence of the way in which the years of the two eras
overlap, be read off as far as the months Chaitra to Âśvina, both inclusive, are concerned, for the
Vikrama year preceding that opposite to which they are entered; at least, I can find no note in his
book to that effect.—A reference to Table III. at page 71 above, will shew at once that an intercalation
of any month from Chaitra to Âśvina inclusive, for Śaka-Saṁvat 1186 expired, northern or southern,
did occur for northern Vikrama-Saṁvat 1321 expired, but southern Vikrama-Saṁvat 1320 expired.

90 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
nesday, the 4th July, or Thursday, the 5th July, A.D. 1263, for the thirteenth tithi, but,
in either case, Friday, the 6th July, for the thirteenth solar day; with northern Śaka-
Saṁvat 1187 expired, Saturday, the 13th June, A.D. 1265, for both the thirteenth tithi and
the thirteenth solar day; and with southern Śaka-Saṁvat 1187 expired, Sunday, the 12th
July, A.D. 1265, for the thirteenth tithi, but Monday, the 13th July, for the thirteenth
solar day. The last is the only other case in which the result includes a Sunday at all.
But it is of no practical value; because it is obtained by the treatment of the Śaka year as
a southern year, whereas I have already shewn that this is not the treatment which is to
be applied; and, still more conclusively, because it would have to be allotted to the
southern Vikrama-Saṁvat 1321 expired, whereas the record specifically refers to the
preceding year, 1320 (expired).
The true English equivalent of the given date, therefore, really is Sunday, the 25th
May, A.D. 1264. This result, and this alone, answers all the requirements of the record.
It is the equivalent of a date which falls, as recorded, within the limits of southern
Vikrama-Saṁvat 1320 (expired); though it is not the equivalent of any day in the month
Âshâḍha as referred to that year; for, if it is treated as the equivalent of a date referred
to that year (and to southern Śaka-Saṁvat 1186 expired), then, of course, it represents
the thirteenth tithi and solar day of the dark fortnight of the month Jyêshtha, preceding
Âshâḍha. It answers to the specified day of the natural, not the intercalated, month;
as is expressly required, partly by the absence, in the record, of any specification of the
intercalated month, and partly by the general prohibition regarding intercalated months.
Also, as is seen from the particulars given above, it is obtainable only by treating the
expired Śaka year, which is the basis of the calculation, as a northern year. According-
ly, it not only bears out in all respects the results obtained from the date in the Êraṇ
pillar inscription of Budhagupta, but also goes beyond them. It definitely proves the
running difference of two hundred and forty-two years between current Gupta-Valabhî and
current Śaka years. It also proves that the true original scheme of the years of the
Gupta-Valabhî era,—viz. the regular Pûrṇimânta northern scheme, as established at
page 76 f. above,—was preserved, in Kâṭhiâwâḍ, up to at any rate A.D. 1264. It
gives Śaka-Saṁvat 1187 current (A.D. 1264-65), as the equivalent of the given current
Gupta-Valabhî year. And it definitely proves that the exact epoch of the era was Śaka-
Saṁvat 241 expired, or 242 current, corresponding to A.D. 319-20.
This result, therefore, will now be taken, in the place of the result obtained from the
date of the Êraṇ pillar inscription of Budhagupta of Gupta-Saṁvat 165, as furnishing the
definite standard, on the analogy of which we must calculate all the remaining dates in the
Gupta-Valabhî era that supply details for computation.
The Verâwal Inscription of Valabhî-Saṁvat 927.
The third and last date, with the mention of a week-day, that I have to comment on,
is contained in an inscription which has not as yet been published, but has been placed at

my disposal, for present purposes, by Dr. Bhagwanlal Indraji, the discoverer of it. The
inscription is on the pedestal of an old image, which is now built into the wall of the
modern temple of the goddess Harsaṭadêvî at Verâwal. The date, with some important
words in the context, (from Dr. Bhagwanlal Indraji’s cloth-rubbing) runs — (l. 1) śrîmad-
Valabhî-sa[ṁ*]vat 927 varshê Phâlguṇa śu di 2 Saumê || Ady=êha śrî-Dêvapattanê
. . . . . . . . . . . . . . . . . . . . (l. 4) śrî-Gôvarddhana-mûrtti[ḥ*] . . . . . . . . . . . . . . . . . . . . . .
(l. 5) kârâpitâ ;—in which, unfortunately, there is some doubt as to the proper rendering
of the first syllable of the word that gives the name of the week-day. The vowel au was
undoubtedly formed ; though, in the rubbing, the top-stroke is partially filled up, in con-
sequence either of want of depth in the engraving, or of want of care in making the
rubbing. And, the consonant presenting the appearance in the rubbing of being bh,
the natural inclination is to read Bhaumê, “on Tuesday.” Dr. Bhagwanlal Indraji, how-
ever, tells me that, in the original, the consonant is certainly s; and the appearance of bh,
therefore, is due to an imperfection in the rubbing. The reading of the original, accord-
ingly, is to be taken as Saumê. But this is not a real word ; and it requires to be corrected
into either Sômê, “on Monday ;” Bhaumê, “on Tuesday ;” or Saumyê, “on Wednesday.”
It is unfortunate that we should have to make any correction at all, in a point of such
importance ; especially when so very free a choice is open. But it has to be done. And
the calculated results favour the supposition that the reading intended was Sômê,
“on Monday.” Dr. Bhagwanlal Indraji was of opinion that the intended reading
was Bhaumê, “on Tuesday ;” which, of course, might be supported by assuming
carelessness on the part of the engraver, in letting his tool slip in such a way as
to give the bh a more or less complete appearance of s. But, from the appearance of the
rubbing, the reading Sômê is equally justifiable, on the assumption that the partial appear-
ance of bh in the rubbing, instead of s, is due only to a fault in the rubbing, and that the
mistake in forming au instead of ô was discovered before the stroke which turns ô into au was
completed ; this would account for this stroke being so shallow as to cause the blur which
almost entirely conceals it in the rubbing. Adopting the reading or correction of Sômê, the
translation will be — “the year 927 of the famous (city of) Valabhî ; in (this) year ; (the
month) Phâlguna ; the bright fortnight ; the (solar) day 2 ; on Monday ; to-day ; here
in the famous (city of) Dêvapattana . . . . . . . . . . (this) image of the holy Gôvardhana
. . . . . . . . . . has been caused to be made.”
This gives us, for calculation, Valabhî-Saṁvat 927, current ; the month Phâlguna
(February-March) ; the bright fortnight ; the second solar day of the fortnight, and pre-
sumably the second tithi ; and Sômavâra, or Monday. And, on the analogy of the
Verâwal inscription of Valabhî-Saṁvat 945, the given tithi should belong to Valabhî-
Saṁvat 927 + 242 = Śaka-Saṁvat 1169 current (A.D. 1246-47) ; and the calculation
should be made with the basis of Śaka-Saṁvat 1168 expired.
m 24

92 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.

Mr. Sh. B. Dikshit’s calculations, however, made, as before, for a year before and a
year after the resulting year, as well as for that year itself, give the following results, in
each case for both the second solar day and the second tithi, both by Prof. K. L. Chhatre’s
Tables and by the Sûrya-Siddhânta; viz. with Śaka-Saṁvat 1167 expired, **Monday,**¹
the 19th February, A D. 1246; with Śaka-Saṁvat 1168 expired, **Saturday,**² the 9th
February, A.D. 1247; and with Śaka-Saṁvat 1169 expired, **Wednesday,**³ the 29th Janu-
ary, A.D. 1248.
The result for Śaka-Saṁvat 1169, which is the year in which the week-day should
presumably prove correct, does not answer at all. If we could understand that the reading
intended was Saumyê, "on Wednesday," then the result for Śaka-Saṁvat 1170 might be
accepted; subject only to the considerations that Saumyavâra, though perfectly allowable,
is not often used as a synonym for Budhavâra, 'Wednesday;' and that the result is later
by a year than what it ought to be, and can be arrived at only through accepting an altera-
tion in the reckoning of the Gupta-Valabhî era, precisely the opposite of the alteration
which, as shewn at page 72 f. above, was certainly made, in Gujarât, before the date of the
Kaira grant of Dharasêna IV. of the year · 30. If, on the other hand, we take Sômê, "on
Monday," as the intended reading, then the result for Śaka-Saṁvat 1168 may be accepted,
subject only to the consideration that it is earlier by a year than what it ought to be.
This result would be quite intelligible, if we could refer the given date to a year com-
mencing with Mârgaśîrsha; for then, belonging to Valabhî Saṁvat 927, it would belong
quite regularly to Śaka-Saṁvat 1168, and its English equivalent would fall quite regularly
in A.D. 1246. This, however, as I have stated at page 78 above, is distinctly prevented
by a perfectly conclusive obstacle. The only resource that remains, is to hold that, for
some reason or other, the date given in this record, like the date in the Kaira grant of
Dharasêna IV. of the year 330, was taken from a Gujarât almanac, and belongs to a year
the initial day of which was the Kârttika śukla 1 preceding the true commencement of
Gupta-Valabhî-Saṁvat 927. This, again, would be perfectly intelligible, if we could only
assume that the image, which is probably portable enough, was fashioned, together with
the engraving of the inscription, at some place in Gujarât; and was then transported by
a pilgrim to Verâwal. But the objection to this is, that the inscription seems to record
distinctly that the image was made at Dêvapattana; and Dêvapattana is well known as
another name of Sômnâthpâṭan, i.e. the modern Verâwal itself; and it is difficult to
understand how the corrupt Gujarât reckoning of the Gupta-Valabhî era can have been
introduced at Verâwal in A.D. 1246, when, as we have already seen from the other Verâwal
inscription, of Valabhî-Saṁvat 945, the true original reckoning was used there up to at

¹ The tithi ended, by Prof. K. L. Chhatre’s Tables, 29 ghaṭîs, 59 palas, and by the Sûrya-Siddhânta, 30 ghaṭîs, 2 palas, after mean sunrise at Bombay.
² Here the times are respectively 8 ghaṭîs, 33 palas, and 9 ghaṭîs, 25 palas.
³ Here the times are respectively 31 ghaṭîs, 57 palas, and 34 ghaṭîs, 43 palas.

least eighteen years later. The explanation, however, is perhaps to be found in the sup-
position that the inscription was prepared under the personal direction of a pilgrim from
Gujarât, who had brought a Gujarât almanac with him.
On the whole, be the explanation what it may, there seems no doubt that the proper
result is Monday, the 19th February, A.D. 1246. This, however, does not support
the running difference of exactly two hundred and forty-two years between current
Gupta-Valabhî and current Śaka years; nor, as the given date belongs to a bright fort-
night, does it prove anything as to the northern or southern nature of the Gupta-Valabhî
year. If it is taken as the result for a date belonging to a year the scheme of which was
identical with the scheme of the Śaka year, northern or southern,—i.e. as the result for
a date in a year which commenced with the Chaitra śukla 1 belonging to Śaka-Saṁvat
1168 current,—it requires a running difference of two hundred and forty-one years only.
While, if it is taken as the result for a date belonging to a year the scheme of which
was identical with the scheme of the southern Vikrama years,—i.e. as the result, obtained
through Śaka-Saṁvat 1167 expired, for a date in a year which, with the southern
Vikrama-Saṁvat 1303 current, commenced with the following Kârttika śukla 1, still
belonging to Śaka-Saṁvat 1168 current, and falling five months before the true com-
mencement of Gupta-Valabhî-Saṁvat 927,—then it requires a running difference lying
between two hundred and forty-one and two hundred and forty-two. As in the case of the
Kaira grant of Dharasêna IV. of the year 330, which has been referred to at page 72 above
and is now to be explained in detail, the latter is the way in which I apply the result.
And it gives, therefore, the southern Vikrama-Saṁvat 1303 current (A.D. 1245-46), as the
nominal equivalent of the given current Valabhî year. But I have to point out distinctly
that the date is not a satisfactory one, since an important correction of some kind or
another has to be made, in order to interpret it intelligibly at all; and that in no way
does it give a conclusive result, like that of the other Verâwal inscription, of Valabhî-
Saṁvat 945.
The Kaira Grant of the year 330.
The preceding is the last instance, as yet obtained, of the mention of a week-day in a
record dated in the Gupta-Valabhî era. There are, however, three inscriptions, containing
other interesting details for computation, which will now be noticed, before dealing with the
question of the Twelve-year Cycle of Jupiter.
The first of them is a Kaira grant of Dharasêna IV. of Valabhî; in which
the date (from Dr. Bühler's published text; Ind. Ant. Vol. XV. p. 339, line 57) runs —
saṁ 300 30 dvi-Mârggaśira śu 2,— "the year 300 (and) 30; the second (month) Mârgaśira;
the bright fortnight; (the lunar day) 2.
This gives us, for calculation, Gupta-Saṁvat 330, current; an intercalation of the
month Mârgaśira or Mârgaśîrsha (November-December), as shewn by the reference

94 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
in the text to two months of that name; the bright fortnight of the second Mârgaśirsha;
and the second tithi or lunar day. And, on the analogy of the Verâwal inscription
of Valabhî-Saṁvat 945, the given intercalated month should belong to Gupta-Saṁvat
330 + 242 = Śaka-Saṁvat 572 current (A.D. 649-50); and the calculation should be
made with the basis of Śaka-Saṁvat 571 expired.
General Cunningham, however,¹ shews no intercalation in that year; but, in the pre-
ceding year, Śaka-Saṁvat 571, an intercalation of the month Kârttika, which would fall in
A.D. 648; and this appears to be quite correct, in accordance with the regulation of inter-
calations by the actual place of the sun. Looking further into the matter, Dr. Schram, as
reported by Dr. Bühler,² found that in A.D. 648 there certainly was an intercalated month,
which, according to the present method would be Kârttika, but according to the rule
for mean intercalations, would be Mârgaśîrsha. So, also, Mr. Sh. B. Dikshit finds, by
actual calculation from the Sûrya-Siddhânta, that, by mean intercalation, in A.D. 648
there was an intercalated month between the natural Mârgaśirsha and the natural
Pausha, which would be named Mârgaśîrsha according to the verse Mêsh-Adi-sthê
savitari &c., that is quoted as belonging to the Brahma-Siddhânta;³ though it would
be named Pausha according to the present practice. In either case, the two intercalated
fortnights are, of course, the same lunar period; the only difference is in respect of
the name by which that period should be called. And, in finding that period coupled
with the name of Mârgaśîrsha, there is the fact, which must be admitted, that, in contra-
vention of the usual rule, the grant recorded in this inscription was made in the intercala-
ted month; a deviation for which I cannot find any reason in the record itself. Having
regard to this and other points, I asked Mr. Sh. B. Dikshit to also make the necessary
calculations on the possibility of the abbreviation dvi for dvitîya, ‘the second,’ qualifying,
not only the word Mârgaśira, but the whole expression Mârgaśira śu 2; i.e. on the possi-
bility of the intercalation or repetition referring to the tithi or lunar day, not to the month;
in which case the date might have been referred, quite regularly, to Śaka-Saṁvat 572. He
finds, however, that the second tithi of the bright fortnight of Mârgaśîrsha of Śaka-Saṁvat
572, falling in A.D. 649, was not a repeated tithi, either by Prof. K. L. Chhatre’s Table, or
by the Sûrya-Siddhânta; and that, on the contrary, by the Sûrya-Siddhânta, there is the
possibility of this tithi having been expunged, in the far eastern parts of India. It is certain,
therefore, that the intercalation refers to the month; not to the tithi. And it is equally
certain that in Śaka-Saṁvat 571 there was an intercalated month, falling in A.D. 648,
which might be named Mârgaśira or Mârgaśîrsha, and which evidently was actually so
named in the almanac that was consulted by the drafter of this record. This being the
case, there cannot have been an intercalation of the same month, or in fact of any month
at all, in the following year, Śaka-Saṁvat 572. Therefore, the Mârgaśîrsha of this record
¹ Indian Eras, p. 158.
² Ind. Ant. Vol. XV. p. 338.
³ See page 88 above, note 5.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 95
undoubtedly fell, not in A.D. 649, as should be the case according to the true Gupta
Valabhî reckoning, but in A.D. 648; and belonged, for astronomical purposes, to Śaka-
Saṁvat 571, or, according to the popular usage of Gujarât, to the southern Vikrama-Saṁvat
706. And, since the localities mentioned in the charter connect it absolutely with a pro-
vince of Gujarât,¹ the year 330 of this record must have commenced, like the southern
Vikrama-Saṁvat 706, with the month Kârttika that preceded the true commencement of
Gupta-Saṁvat 330, with Chaitra śukla 1 of Śaka-Saṁvat 572, according to the original
scheme of the years of the era.
The result, therefore, gives the southern Vikrama-Saṁvat 706 current (A.D. 648-49),
as the nominal equivalent of the given current Gupta year. And this date, with that of the
Verâwal inscription of Valabhî-Saṁvat 927, must be allotted to a class of cases, in which the
running difference of exactly two hundred and forty-two years between current Gupta-Valabhî
and current Śaka years was not maintained, in consequence of a local adaptation of the
Gupta-Valabhî year to the scheme of the southern Vikrama year, made in such a way that,
as pointed out at page 73 above, each subsequent Gupta-Valabhî year, where this
adaptation was applied, commenced five lunations before the true commencement of the
year according to the original scheme. The anomaly, however, is natural enough in the
present instance, because of the locality to which the record refers itself.
The Nêpâl Inscription of the year 386.
The next inscription that I have to notice, is the Nêpâl inscription of Mânadêva,
of the Sûryavaṁśi or Lichchhavi family of Mânagṛiha, on the lower part of a broken pillar
placed to the left of the door of the temple of the god Chângu-Nârâyaṇa, about five miles
to the north-east of Khâṭmâṇḍu; in which the date (from Dr. Bhagwanlal Indraji's publish-
ed text and lithograph; Ind. Ant. Vol. IX. p. 163, line 1 f.) runs — saṁvat 300 80 6
Jyêshṭha-m[â]sa-śukla-pakshê pratipadi 1 [Rô]hiṇî-nakshattra-yukt[ê*] chandramasi
m[u]hûrttê praśastê=Bhijiti,—" the year 300 (and) 80 (and) 6; in the bright fortnight of
the month Jyêshṭha, on the first tithi or lunar day, (or in figures) 1; the moon being in
conjunction with the Rôhiṇî nakshatra; in the excellent muhûrta (named) Abhijit."
The Nêpâl inscriptions were first brought to notice by Dr. Bhagwanlal Indraji, in the
Ind. Ant. Vol. IX. p. 163 ff.; and his view of the historical results of them was given in
the same Journal, Vol. XIII. p. 411 ff. My own view, originally published in the same
Journal, Vol. XIV. p. 342 ff., will be found explained in detail in Appendix IV. below.
And all that it is necessary to state here, is, that the earliest inscriptions disclose the use
of two eras; viz. the so-called Gupta era, and the Harsha era. At the time, however,
when Dr. Bhagwanlal Indraji wrote, the fact that the Gupta era was used in some of these
¹The charter was issued from the victorious camp at Bharukachchha, i.e. the modern Broach
(Bharuch); and it records the grant of some lands in the Khêṭaka âhâra, or the territorial division of
which the chief town was Khêṭaka, i.e. the modern Kaira (Khêḍâ).

96 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
records, was not apparent. And it only became clear on Mr. Bendall's discovery of the
Gôlmâḍhiṭôl inscription of the Mahârâja Śivadêva I. of Mânagṛiha, which he published
originally in the Ind. Ant. Vol. XIV. p. 97 f., and has given again, with a slight correction
in the reading of the date, in his Journey in Nêpâl and Northern India, p. 72, and Plate
viii. This inscription is dated in the year 316, without any specification of the era. But
the clue to the interpretation of the date is given by its mention of the Mahâsâmanta
Aṁśuvarman, as the contemporary¹ of Śivadêva I. Aṁśuvarman's approximate date, viz.
about A.D. 637, was very well known from Hiuen Tsiang's mention of him.² And, as the
Nêpâl series included three inscriptions of Aṁśuvarman himself, dated in the years 34, 39,
and 44 or 45 of an unspecified era, and another, of Jishṇugupta, dated in the year 48, and
mentioning Aṁśuvarman, Dr. Bhagwanlal Indraji very properly referred these dates to the
era running from the accession of Harshavardhana of Kanauj, and commencing³ in A.D.
606. This much being quite cert in, it follows that the date the year 316 for Śivadêva
I., the contemporary of Aṁśuvarman, must of necessity be referred to a starting-point just
about three hundred years before the Harsha era. And the era which exactly meets the
requirements of the case is the Gupta era; for, 316 + A.D. 319-20 = A.D. 635-36; which
is in due accordance with the recorded dates that we have for Aṁśuvarman in the Harsha
era, representing from A.D. 639 to 649 or 650.
As regards the present inscription of Mânadêv its palæography, as well as a general
consideration of the historical results, shews that the year 386, quoted in it, belongs to the
same series with the year 316 that is quoted in the Gôlmâḍhiṭôl inscription of Śiv. lêva I.
And, accordingly, this record gives us, for calculation, Gupta-Saṁvat 386, current; the
month Jyêshṭha (May-June); the bright fortnight; the first tithi or lunar day; the Rôhiṇî
nakshatra or lunar mansion; and the Abhijit muhûrta or thirtieth part of the day and
night. And, on the analogy of the Verâwal inscription of Valabhî-Saṁvat 945, the given
tithi should belong to Gupta-Saṁvat 386 + 242 = Śaka-Saṁvat 628 current (A.D. 705-
706); and the calculation should be made with the basis of Śaka-Saṁvat 627 expired.
Making the calculations by the Sûrya-Siddhânta, and applying the results to the
longitude of Khâṭmâṇḍu, Mr. Sh. B. Dikshit finds that, with the basis of Śaka-Saṁvat
627 expired, the given tithi, belonging to Śaka-Saṁvat 628 current, ended on Tuesday,
¹ This is also recorded in the other inscription of Śivadêva I., No. 5 of Dr. Bhagwanlal Indraji's
Nêpâl series, Ind. Ant. Vol. IX. p. 168 ff. But, unfortunately for the general chronological results
arrived at by him from those inscriptions,—which involved the application of the Vikrama era for the
interpretation of such of the Nêpâl dates as belong really to the Gupta era,—the date of Śivadêva I.
is there broken away and lost.
² See Beal's Buddh. Rec. West. World, Vol. II. p. 81; also Ind. Ant. Vol. XIII. p. 422, and Vol.
XIV. p. 345.
³ On this point see a note in Appendix IV. below.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 97
the 28th April,¹ A.D. 705, at 57 ghaṭîs, 12 palas, after sunrise; that there was the Kṛittikâ
nakshatra up to 11 ghaṭîs, 3 palas, after sunrise, and then the Rôhiṇî nakshatra, which
continued up to 11 ghaṭîs, 18 palas, after sunrise on the next day, Wednesday; and that,
consequently, the Abhijit muhûrta, being the eighth in order among the muhûrtas, and
beginning after the expiration of fourteen ghaṭîs after sunrise, occurred, as required by the
record, while the Rôhiṇî nakshatra was current. He also finds that the same conditions
of the nakshatra and the muhûrta did not occur, if the given tithi is treated as belonging
to Śaka-Saṁvat 627 or 629 current.
The result, therefore, answers, as is required, to the running difference of two hundred
and forty-two years between current Gupta-Valabhî and current Śaka years; and to the
treatment of the Gupta year as a northern Śaka year. And it gives Śaka-Saṁvat 628
current (A.D. 705-706), as the equivalent of the given current Gupta year. But, like the
result for the date in the Êraṇ pillar inscription of Budhagupta, it does not, in itself, prove
conclusively either the exact epoch of the era, or the scheme of the year; for the reason
that, being a date in a bright fortnight, this Jyêshṭha śukla 1 was the same tithi, and
ended on the same solar day, represented by the 28th April, all over India, in the southern
as well as the northern Śaka-Saṁvat 628, and in southern Vikrama-Saṁvat 762, as well
as in northern Vikrama-Saṁvat 763.
The Môrbi Grant of the year 586.
The last date that I have to notice from the present point of view, is contained in the
Môrbi grant of Jâiṅka, from Kâṭhiâwâḍ, published by Dr. R. G. Bhandarkar in the Ind.
Ant. Vol. II. p. 257 f. The record includes two dates. In line 16 f., in connection with
the making of the grant, we have (from the published lithograph)—pañch-âśîtyâ yutê=
tîtê samânâṁ śata-pañchakê l Gôptê dadâv=adô nṛipaḥ s-ôparâgê=rkka-maṇḍalê ll,—
"five centuries of years, together with eighty-five (years), having passed by, the king gave
this (charter) at (the village of) Gôpta, when the disc of the sun was eclipsed;" in which the
year is expressly coupled with a word meaning "expired." And in line 19 f., in connection
with the writing of the charter, we have—saṁvat 585 Phâlguna su(śu) di 5,—"the year
585; the month Phâlguna; the bright fortnight; the (solar) day 5;" without any indication
whether the year is an expired one, or current. The eclipse is also mentioned in line 3, in
the words — mârttaṇḍa-maṇḍal-âśrayiṇi Svavbhânô(read Svarbbhânau),—"while
Svarbhânu (i.e. Râhu, the personified ascending node) is resting on the disc of the sun."
There is some difficulty in disposing finally of the whole bearing of this record;
owing to the fact that the first plate was lost sight of, without being procured for
¹ Śaka-Saṁvat 628 current commenced rather early, on, approximately, Sunday, the 1st March,
A.D. 705. And hence the reason why the month Jyêshṭha, which ordinarily answers to May-June,
commenced on the 28th April, and of course ended before the end of May.
Nb

98 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.

examination at all; and now, even the second plate also, the published one, has been mis-
laid and is not forthcoming. And I have to point out that, in the second part of the
verse, Dr. R. G. Bhandarkar read Gauptê, instead of Gôptê; and translated "five
hundred and eighty-five years of the Guptas having elapsed." The lithograph, however,
shews distinctly that the original has Gôptê; and it is only by the correction of ô into au
that the name of the Guptas can be introduced into the passage.¹ But, even then, the
adjective Gauptê occupies an irregularly detached place, which any skilful composer
would have avoided, from the noun, śata-pañchakê, which it qualifies. While, on the
other hand, I have shewn, in my remarks on the nomenclature of the era, at page 19 ff. above,
that we have no reason at all to look for the use of such an adjective as Gaupta,
'belonging to the Guptas;' and, if we maintain the original reading of Gôptê, we have a
locative case, which we have every reason to expect in immediate connection with the
verb dadau, "he gave," and which will then give us the name of the village at which the
grant was made. And, until the original first plate of the grant is produced, to prove that
Gôpta was not the name of the village,² or otherwise to explain the passage, this is the
reading and interpretation that I adopt.
In accordance, however, with the palæography of the grant, I see no reason for
referring the date to any except the Gupta-Valabhî era; irrespective of the question
whether the era is mentioned by name, or not. This record, therefore, gives us, for
calculation, an eclipse of the sun, which took place on some unspecified date in Gupta-
Saṁvat 586 current, as the original text specifies that the year 585 had expired. And,
on the analogy of the Verâwal inscription of Valabhî-Saṁvat 945, the eclipse should be
found in Gupta-Saṁvat 586 + 242 = Śaka-Saṁvat 828 current, somewhere between ³
the 10th March, A.D. 905, and the 27th February, A.D. 906. Also, it should,
presumably, be visible at the place at which the grant was made on the occasion of it.
And, though there is nothing, in the existing remnant of the record, to indicate
with certainty the exact locality to which it belongs, still there is nothing against the
supposition that it really belongs to Môrbi itself, or to that neighbourhood. We have,
therefore, to look for a solar eclipse, occurring in Śaka-Saṁvat 828 current, and visible
at Môrbi, or near that town, in the north of Kâṭhiâwâḍ.

¹ The mistake of ô for au does occur in the word svavbhânô, for svarbbhânau, in line 3 of the
grant. But in line 9, in the word paurvva, the au is formed quite correctly and completely.
² We might easily find its present representative in the modern name of Gôp, which occurs in
the case of a village, in Kâṭhiâwâḍ, about seventy-five miles south-west of Môrbi; twenty-five miles
south of Nawânagar or Jâmnagar; and fifty miles east of Dhiniki, where there was found the copper-
plate grant of Jâikadêva, which purports to be dated in Vikrama-Saṁvat 794 (see page 62 above,
note 1.)
³ Of course, as it could only take place at a new-moon conjunction, it did not occur on the
given solar day for the writing of the charter.—The limits within which we must look for it, are the
first and the last days of Śaka-Saṁvat 828 current, as given in Indian Eras, p. 167.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 99
General Cunningham’s Table¹ mentions no solar eclipse, as having occurred during
the period defined above. But, by calculations from Prof. K. L. Chhatre’s Tables,
Mr. Sh. B. Dikshit finds² that there was an eclipse of the sun, on Tuesday, the 7th
May, A.D. 905, corresponding to the new-moon tithi of the Pûrṇimânta northern Jyesh-
ṭha of Śaka-Saṁvat 828 current, which fully answers the required conditions.³ It was visible
at Môrbî; over almost the whole of Southern India; and in Ceylon. The magnitude,
at Môrbî, was one ninth of the sun’s disc; and, in the southern parts of India, greater than
this. And the middle of the eclipse, at Môrbî, was at 12.9 midday of the Môrbî mean
civil time. This eclipse, accordingly, was very distinctly visible at Môrbî; even if it was
not known beforehand, from calculations.
This result, therefore, fully answers to the conditions of the record. It does not, in
itself, furnish conclusive proof, either as to the exact epoch of the era, or as to the scheme
of the year; for the reason that, if treated as a southern date, the day of the eclipse,
represented in that case by the new-moon tithi of the Amânta southern Vaiśâkha, belonged
to the southern Śaka-Saṁvat 828, and to the southern Vikrama-Saṁvat 962, as well as to
the northern Śaka-Saṁvat 828 and Vikrama-Saṁvat 963. But it answers, as is required,
to the running difference of two hundred and forty-two years between current Gupta-Valabhî
and current Saka years; and to the treatment of the Gupta year as a northern Śaka year.
And it gives Śaka-Saṁvat 828 current (A.D. 905-906), as the equivalent of the indicated
current Gupta year.
In respect of the given solar day for the writing of the charter, viz. the fifth solar day
in the bright fortnight of the month Phâlguna (February-March), as the name of the week-
day is not given, the only test that we can apply is the presumption⁴ that the running
number of the lunar tithi is the same as that of the solar day; i.e. that the fifth lunar tithi
ended on the fifth solar day of the fortnight. If the year 586, connected with this solar
day, is to be taken as expired, as in connection with the eclipse, this should be the case in
Śaka-Saṁvat 828. And, by calculations for that year from Prof. K. L. Chhatre’s Tables,
Mr. Sh. B. Dikshit finds that the preceding new-moon tithi, which was that of the
Pûrṇimânta northern Phâlguna, or the Amânta southern Mâgha, ended on Monday, the
27th January, A.D. 906; and that the fifth tithi of the bright fortnight of Phâlguna ended

¹ See Indian Eras, p. 213.
² For his calculations, which are based on the apparent longitudes of the sun and the moon, he
has taken the latitude and longitude of Môrbî, which I was not then able to supply to him, as ..2° 45′
N. and 70° 51′ E. I now find that, in Thornton’s Gazetteer of India, the figures are 22° 49′ N. and
70° 53′ E. Mr. Sh. B. Dikshit states, however, that the difference will not palpably affect his results.
³ So also Prof. K. L. Chhatre himself obtained the same eclipse; see Dr. R. G. Bhandarkar’s
Early History of the Dekkan, p. 99, where, with a slight difference of phraseology, the eclipse is
given as occurring "on the 30th of Vaiśâkha, Śaka 827;" the reference being to the Amânta
southern month and the expired Śaka year.
⁴ See page 85 above, note 1.

• 2h

100 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
on Saturday, the 1st February, which was the fifth successive solar day. If this date
is accepted, then the charter was written nine months after the making of the grant.¹ On
the other hand, if the year 585 is here to be taken as current, there should be the same
agreement of the lunar tithi and the solar day in Śaka-Saṁvat 827. And, for this year,
Mr. Sh. B Dikshit finds that the preceding new-moon tithi ended on Thursday, the 7th
February, A.D. 925; and that the fifth tithi of the bright fortnight of Phâlguna ended
on Tuesday, the 12th February, which was again the fifth successive solar day. If this
date is accepted, then the charter was prepared two months before the actual making of
the grant.
It may perhaps be argued, hereafter, in opposition to my results for the exact epoch
of the era, that all the Gupta-Valabhî dates are recorded in expired years, whether the
fact is distinctly stated or not; and, consequently, that it is as an expired year, not cur-
rent, that the year 165 of the Êraṇ pillar inscription is equivalent to A.D. 484-85 current,
and that the year 585 expired, of the present record, is equivalent to A.D. 904-905 cur-
rent. In that case, the solar eclipse would have to be found in Gupta-Saṁvat 585+242=
Śaka-Saṁvat 827 current, somewhere between ² the 21st March, A.D. 904, and the 9th
March, A.D. 905. During this period, there were two eclipses of the sun;³ on Saturday,
the 16th June, A.D. 904, corresponding to the new-moon tithi of the Pûrṇimânta northern
Âshâḍha of Śaka-Saṁvat 827 current; and on Saturday, the 10th November, A.D. 904,
corresponding to the new-moon tithi of the Pûrṇimânta northern Mârgaśîrsha of the
same Śaka year. In respect of the first of them, Mr. Sh. B. Dikshit finds that it was
not visible anywhere in India; but only in the more northern parts of the earth. This,
therefore, cannot be the eclipse intended. In respect of the second of them, he finds
that it was visible at Môrbi; over more than half the northern part of Kâṭhiâwâḍ; and,
to the south, along the coast, as far as Surat, one hundred and seventy miles south-east
of Môrbi, and in the interior, a little further still. And, at Môrbi, the middle of the eclipse
was at 11.54 A.M. of the Môrbi mean civil time. At Ahmadâbâd, one hundred and twenty
miles east by north from Môrbi, one twelfth of the sun's disc was eclipsed; and, in the
more northern parts of India, a considerably greater surface. But, at Môrbi itself, the
magnitude of the eclipse was very small; extending there to only one twenty-fifth part
¹ There is nothing in the record itself, to indicate whether the writing of the charter preceded,
or followed, the making of the grant. The Râjim grant of Tîvaradêva, No. 81, page 291, furnishes
another similar instance. In that instance, the grant was made on the eleventh tithi of Jyêshṭha
(May-June); while the charter was written, or assigned, on the eighth solar day of Kârttika (October-
November); and there is nothing to shew specifically whether it was the following, or the preceding,
Kârttika. That charter may have been written, or assigned, either five months after, or seven months
before, the making of the grant recorded in it.
² See Indian Eras, p. 167. ³ See Indian Eras p. 213.

THE GUPTA INSCRIPTIONS ; INTRODUCTION. 101
of the disc.¹ This eclipse, therefore, setting aside all other considerations, is not in any way
as satisfactory as that of the 7th May, A.D. 905.
The use of the Twelve-Year Cycle of Jupiter in Records of the Early
Gupta period.
I now come to a still more interesting and important part of my inquiry ; viz. to the
use of the Twelve-Year Cycle of Jupiter in the dates of some of the records of the
Early Gupta period.
These dates are found in the inscriptions of the Parivrâjaka Mahârâjas Hastin and
Saṁkshôbha, No. 21, page 93, to No. 25, page 112. And the extreme value of the
records, from the present point of view, is due to the fact that in each instance, except
in No. 24, page 110,² the date is directly connected with an expression which shews expli-
citly that, at the time mentioned, the Gupta sovereignty was still enduring ; and conse-
quently,—since the figures of the years are naturally referable to the same uniform
series with the years quoted in the records of the Early Guptas themselves ; and since the
palæography of the inscriptions is entirely in favour of such a reference,—which shews
also that the dates are recorded in the identical era that was used by the Early Gupta
kings.
The evidence derivable from these records has hitherto been completely misapplied ;
in consequence of the adoption of the view, that the duration of any saṁvatsara or year
of this cycle, is the same with that of the years of the Śaka era, from Chaitra śukla 1 to
Chaitra kṛishṇa 15; and that the means of exactly determining the saṁvatsaras of this
cycle, are provided by the last remainder obtained from certain rules given by Varâhamihira
and others, which in reality only shews what saṁvatsara of the Sixty-Year Cycle of the
same planet according to the Northern System, and of the Twelve-Year Cycle according to
the mean-sign system, is current at the commencement of any given Śaka or Kaliyuga year,
and which does not provide for the determination of the saṁvatsaras on any other given
date in the year.³
¹ Mr. Sh. B. Dikshit has not made actual calculations for the village of Gôp (see page 98 above,
note 2) ; but is able to state that both the eclipses, of the 7th May, A.D. 905, and of the 10th Nov-
ember, A.D. 904, were visible there ; the circumstances of the former eclipse, in respect of visibility,
being more favourable, and those of the latter being less so, at Gôp than at Môrbî.
² For the possible explanation of the omission in this instance, see page 8 above.
³ See, for instance, Indian Eras, p. 26 ff.—This interpretation of the rules in question leaves un-
utilized, and unexplained, the first remainder, obtained from the division by 3750 according to Varâha-
mihira's rule, and by 1875 according to the rule of the Jyôtishṭava. In connection with the Jyôtish-
ṭava rule, however, Warren has shewn (Kala-Sankalita, p. 202) how this remainder gives the means
of determining the actual commencement of each saṁvatsara.—From the use of some Tables drawn
up by Mr. Sh. B. Dikshit, for determining the actual commencement of each saṁvatsara of the Sixty-
Year Cycle according to the Northern System, and of each saṁvatsara of the Twelve-Year Cycle

The correct theory of the cycle, according to the requirements of the system actually
applied in the records now under consideration, with the proper method of determining each
of the saṁvatsaras, has now been demonstrated by Mr. Sh. B. Dikshit, in his paper which is
published as Appendix III. below. And, by calculations based on the Sūrya-Siddhānta,
he has worked out all the results required for a full treatment of the dates in question ; giv-
ing the full English and Hindu dates throughout, in order that both European and Hindu
astronomers may be in a position to easily check his results. In publishing his results, I am
confident that no essential errors can be established in them ; even though it should be
shewn, by more exhaustive calculations, that his longitudes for the heliacal risings of
Jupiter are capable of slight corrections.¹ And, as will be seen, his results corroborate, in
the most emphatic manner, all that has been established in the preceding pages ; first,
approximately, from the statements of Albêrûnî, the Mandasôr inscription of Mâlava-Saṁ-
vat 529 expired, and the Eraṇ pillar inscription of Budhagupta of Gupta-Saṁvat 165 ; and
then, exactly, by the Verâwal inscription of Valabhî-Saṁvat 945.
It is not essential, for the most part, to enquire whether any of the dates, which give
correct results with the running difference of two hundred and forty-two years between
current Gupta and current Śaka years, also give, by coincidence, correct results with any
of the suggested earlier epochs. And, indeed, to make any such inquiry systematic and
thorough, it would be obligatory to make the calculations, throughout, for at least a dozen
years or so before and after the proposed epochs. Such inquiries are certainly unneces-
sary in respect of any epoch as early as that proposed by Mr. Thomas. It seemed worth
while, however, to calculate the dates now under consideration, and to give the results,
for the epochs proposed by General Cunningham and Sir E. Clive Bayley ; both of whom
have dealt with the whole subject on the clear understanding that these dates are recorded
in the era that was actually used by the Early Guptas themselves. And it will be seen

according to the mean-sign system, I find that, for the purposes of such rules as that given by
Varâhamihira in the Bṛihat-Saṁhitâ, viii. 20, 21, the Śaka years must be treated as commencing with
the Mêsha-Saṁkrânti or entrance of the Sun into Aries, and not with Chaitra śukla 1 ; though
the latter is the initial day that is required for the notation of tithis. For instance, by Varâhamihira's
rule, the Viśvâvasu saṁvatsara of the Sixty-Year Cycle was current at the commencement of Śaka-
Saṁvat 748 current (A.D. 825-26); and, by Gen. Cunningham's extension of the rule (Indian Eras,
p. 27), the Mahâ-Âśvayuja saṁvatsara of the Twelve-Year Cycle was current on the same date. In
Śaka-Saṁvat 748, the Mêsha-Saṁkrânti occurred on the 21st March, A.D. 825; and Chaitra śukla 1
ended on the 22nd February. The Viśvâvasu saṁvatsara of the Sixty-Year Cycle according to the
Northern System, and, with it, the Mahâ-Âśvayuja saṁvatsara of the Twelve-Year Cycle by the mean-
sign system, actually commenced on the 15th March; and thus they were current at the Mêsha-Saṁ-
krânti, but not on Chaitra śukla 1. And the same will often happen when the saṁvatsaras are com-
mencing shortly before the Mêsha-Saṁkrânti. This application of such rules as that given by Varâ-
hamihira, is, in fact, the natural one ; though it may not be apparent at first sight. For, the Mêsha-
Saṁkrânti is a very definite point in the year ; whereas, Chaitra śukla 1 is always shifting backwards
and forwards by about eleven and nineteen days at a time, and its circumstances could not be met by
any such hard-and-fast rules.
¹ See, e.g., page 110 below, note 1.

that, with the same heliacal-rising system that is applied to prove the epoch of
A.D. 319-20, the results, as a whole, emphatically fail. In calculating the details with
these two epochs, as with the epoch that I am proving, the Gupta year has been treated
as a Śaka year, commencing with Chaitra śukla 1, and with the Pūrṇimānta northern
arrangement of the lunar fortnights. In some of the cases in which, by this treatment,
and with these two epochs, the results fail, more favourable results might be obtained by
adopting a year commencing with, either the preceding, or the following, Kārttika śukla 1 ;
and by slightly varying the proposed epochs accordingly. But a full consideration of the
details will shew that, with neither of these two epochs, as they stand, and probably with
no epochs in any way approximating to them, is there any method of bringing out perfectly
uniform results throughout.
It might, however, be claimed that General Cunningham's epoch, or Sir E. Clive
Bayley's, or even both of them, could be proved, for the present records, by the other sys-
tem of the Twelve-Year Cycle ; according to which the saṁvatsaras are determined by the
passage of Jupiter among the signs of the zodiac ; and which is, in fact, the system that
they have sought to apply in support of their theories. The results, therefore, according
to this system also, will be given. It will be seen that it is not the system which applies
to the epoch that I am proving ; inasmuch as it gives correct results in only two cases
out of the four by which any absolute proof can be established ;¹ viz. in the case of the
grant B. dated in Gupta-Saṁvat 163, and in the case of the grant C. dated in Gupta-
Saṁvat 191. As regards General Cunningham's and Sir E. Clive Bayley's epochs, it has
always been recognised that this system fails in the case of the grant B. dated in Gupta-
Saṁvat 163, unless the given year is deliberately altered from 163 to 173. As there is
really no justification for this alteration of the original text,² the failure of the system
even in this instance alone, is enough to shew that the proposed epochs are not sustain-
able by means of it. But further, though the fact has not hitherto been recognised, the
system fails also in respect of another of the records. Thus, in the case of the grant C.
dated in Gupta-Saṁvat 191, the given saṁvatsara really was current at the commence-
ment of the given year, in accordance with the rules applied by General Cunningham, and
accepted by Sir E. Clive Bayley, from which, however, it was not properly to be inferred
that it was current through the whole of the same year ; but it came to an end, and was
followed by the next saṁvatsara, with General Cunningham's epoch, nearly three and
a half months, and with Sir E. Clive Bayley's epoch, nearly seven months, before the sub-

¹ The date in the Bhumarā pillar inscription, E. below, furnishes no definite proof in itself ;
because the current Gupta year is not mentioned in it; and consequently the given saṁvatsara
could be proved equally well for epochs differing by a year or more, on either side, from the exact
epochs that are being considered. We can only test it, in so far as to see whether, under any
particular circumstances, the system fails through an omission of the given saṁvatsara.
² See Texts and Translations, page 102, note 1.

104 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
sequent given date in that year. In fact, out of the four leading dates on which any
arguments can be based, with these two epochs this system gives correct results only in
respect of the grant A. dated in Gupta-Saṁvat 156, and the grant D. dated in Gupta-
Saṁvat 209. And, as with the heliacal-rising system, so with this system also, a full
consideration of the details will shew that, even by adopting a scheme for the Gupta year
other than the scheme of the northern Śaka year, with neither of these two epochs is
there any method of bringing out uniform results throughout.
A.—The Khôh Grant of the year 156.
The first inscription is one of the Khôh grants of the Mahârâja Hastin,
No. 21, page 93; in which the date (line 1 ff.) is—shaṭpañchâś-ôttarê=bda-śatê Gupta-
nṛipa-râjya-bhuktau Mahâ-Vaiśâkha-saṁvatsarê Kârttika-mâsa-śukla-paksha-tṛitîyâyâm,
—“in a century of years, increased by the fifty-sixth (year); in the enjoyment of
sovereignty by the Gupta kings; in the Mahâ-Vaiśâkha saṁvatsara; on the third lunar
day of the bright fortnight of the month Kârttika.”
This gives us, for calculation, the Mahâ-Vaiśâkha saṁvatsara, as current on the
third tithi or lunar day of the bright fortnight of the month Kârttika (October-November)
in Gupta-Saṁvat 156 current. And, on the analogy of the Verâwal inscription of
Valabhî-Saṁvat 945, this should be the case in Gupta-Saṁvat 156 + 242 = Śaka-
Saṁvat 398 current;¹ in which year the given tithi corresponds to Sunday, the 19th Octo-
ber, A.D. 475.
Mr. Sh. B. Dikshit finds (see Table IV. page 105 below, Col. A.) that Jupiter’s
rising,² next before the given date, took place on Kârttika śukla 1 of the same year,
Śaka-Saṁvat 398 current, corresponding to Friday, the 17th October, A.D. 475; or,
by the English calendar, on Saturday, the 18th October.³ His longitude then was

¹ Here, and throughout, the year is treated as a northern year. But the details of these dates do
not furnish any actual proof as to the Pûrṇimânta or Amânta arrangement of the lunar fortnights.
² i.e., throughout, his heliacal rising. But the actual calculation is for his first daily rising after
his becoming capable of rising heliacally.
³ Jupiter’s daily rising, next after his becoming capable of rising heliacally, takes place about
forty-four minutes before sunrise, and therefore in the period during which the Hindu and the English
week-days are not identical (see a note in Appendix II. below). In the present case, it took place
at the time in question before sunrise on the English Saturday, the 18th October. Kârttika śukla 2
did not end till after sunrise on that day. Consequently, as current tithis are not quoted, unless
under certain very exceptional conditions not applicable to such occurrences as this, the tithi
on which he rose was Kârttika śukla 1. And this tithi, ending after sunrise on the Friday (and
before sunrise on the Saturday), has to be coupled with Friday, the 17th October, as its week-day.
Hence the apparent, but not actual, difference of a day, according as we take the Hindu or the
English calendar. And a similar difference runs through all the dates of the heliacal risings given
below.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 105
TABLE IV.
The Saṁvatsaras of the Twelve-Year Cycle of Jupiter.

106 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
195° 24'. By both the systems of unequal spaces for the longitudes of the ending-points
of the nakshatras (see Appendix III. Table IX.), he was then in Viśâkhâ; and the
saṁvatsara which then began (see Appendix III. Table VIII.), must have been named
**Mahâ-Vaiśâkha.**¹ Jupiter's next following rising took place on Mârgaśîrsha śukla 13
of Śaka-Saṁvat 399, corresponding to Monday, the 15th November, A.D. 476; or, by the
English calendar, on Tuesday, the 16th November. His longitude then was 225° 35'.
By the Brahma-Siddhânta system of unequal spaces, he was then in Jyeshṭhâ; and the
saṁvatsara which then began, must have been named Mahâ-Jyeshṭha. While, by the
Garga system of unequal spaces, he was then in Anurâdhâ; and the saṁvatsara which then
began, must have again been named Mahâ-Vaiśâkha; which shews that, by this system,
there was at this period a repetition of a saṁvatsara. This difference as to the following
saṁvatsara, however, does not affect the given date. By both the systems of unequal
spaces, with the running difference of two hundred and forty-two years between current
Gupta and current Śaka years, the Mahâ-Vaiśâkha saṁvatsara was current on the
given date. And the result gives Śaka-Saṁvat 398 current (A.D. 475-76), as the equi-
valent of the given current Gupta year.
In connection with the results for this record, the following points have to be noted.
In the first place, Mr. Sh. B. Dikshit has described three systems of fixing the ending-
points of the nakshatras; one of equal spaces; and two of unequal spaces. An exami-
nation of Jupiter's longitudes, as given in Table IV., page 105 above, for each rising
next before the given dates, will shew that, in all the remaining instances, the current
saṁvatsara is proved by all three systems; the only variation is that, in the case of E1,
Jupiter's position, at his rising next before the given date, was, by the system of equal spaces,
in Âślêshâ; but, even then, the current saṁvatsara would be named Mahâ-Mâgha, as also
by the two systems of unequal spaces. So, also, it may be seen that, by all three systems,
we have almost the same results in respect of the following saṁvatsaras; the only variation
is that, in the case of D., Jupiter's position, at his rising next after the given date, was,
by the system of equal spaces, in Bharaṇî, and, accordingly, the saṁvatsara which then
began, would again be named Mahâ-Âśvayuja; which shews that, by this system, there was
a repetition of a saṁvatsara at this period; but this does not affect the saṁvatsara current
on the given date. So far, therefore, as those dates are concerned, the correctness of the
records might be proved by any of the three systems. The same, however, is not the case
in respect of the present record. By the system of equal spaces, Jupiter's position, at his
rising next before the given date, was in Svâti; the saṁvatsara which then began, would
¹I have not been able to obtain the original authority for the use of the prefix mahâ (mahat),
'great.' And it does not occur in connection with the two saṁvatsaras mentioned in the Halsî grants
of the Kâdamba chieftain Mṛigêśavarman; viz. the Pausha saṁvatsara, in line 8 of the grant dated
in his third year (Ind. Ant. Vol. VII. p. 35). and the Vaiśâkha saṁvatsara, in line 10 of the grant
dated in his eighth year (id. Vol. VI. p. 24). I use the prefix, however, throughout, in accordance with
the custom of the original records now under examination.

be named Mahâ-Chaitra; and the Mahâ-Vaiśâkha saṁvatsara would not begin till Jupiter's
rising next after the given date, when, by the same system, his position was in Anurâdhâ.
Accordingly, the system of equal spaces could be applied to the present record, only with a
running difference of two hundred and forty-three years between current Gupta and current
Śaka years; which would be in contradiction with the fact that, to prove the saṁvatsaras of
all the remaining records, it must be applied with the running difference of two hundred and
forty-two years. It is evident, therefore, that, in dealing with these records, the system of
equal spaces is not the correct one; and that we have to apply one or other of the
systems of unequal spaces. This, however, is only natural; for they are both more
ancient than the system of equal spaces; and, the older the system, the greater the cer-
tainty that it is the one in use in the Early Gupta period. Also, the Dêôgaḍh inscription
of king Bhôjadêva of Kanauj, dated Śaka-Saṁvat 784, indicates very plainly that one or
other of the systems of unequal spaces, if not both of them, continued in use, in what had
formed a part of the Early Gupta territory, down to at least the last half of the ninth cen-
tury A.D.¹ Of the two systems of unequal spaces, whether we are to apply the Brahma-
Siddhânta system, or the still more ancient Garga system, cannot at present be decided;
¹ In this inscription, which was brought to notice by Gen. Cunningham in the Archæol. Surv. Ind.
Vol. X. p. 101, and Plate xxxiii. No. 2, and is on a pillar in a detached portico in front of the hall
of a temple at Dêôgaḍh, about sixty miles to the south-west of 'Jhansi' in Scindia's Dominions in
Central India, the date (from an ink-impression; lines 6 ff., 10) runs—saṁvat 919 Asva(śva)yuja-
śukla-paksha-chaturddaśyâṁ Vṛihaspati-dinêna Uttar[â*]-Bhâdrapad[â*]-nakshattrê idaṁ
stambhaṁ samâptam=iti ......... Śakakâl-âbda-sapta-śatâni chatur-aśîty-adhikâni
784,—"the year 919; on the fourteenth tithi or lunar day in the bright fortnight of Aśvayuja; on the
day of Vṛihaspati; under the Uttarâ-Bhâdrapadâ nakshatra, this pillar was finished; ......
...... seven centuries of years, increased by eighty-four, of the Śaka era; (or in figures) 784."
This gives us, for calculation, Vikrama-Saṁvat 919 and Śaka-Saṁvat 784; the month Âśvayuja
(September-October); the day of Vṛihaspati or Jupiter, i.e. Thursday; and the Uttarâ-Bhâdrapadâ
nakshatra. The date, of course, is a northern date; but, as the given tithi is in a bright fortnight,
this point is immaterial.—Taking the Śaka year in question as an expired year, Mr. Sh. B. Dikshit
has obtained as the English equivalent, by both Prof. K. L. Chhatre's Tables and the Sûrya-Siddhânta,
Thursday, the 10th September, A.D. 862. The tithi ended at 56 ghaṭîs, 34 palas, or 22 hrs. 37 min.
36 sec., after sunrise on that day. By the equal-space system of the nakshatras, the moon was in the
Pûrvâ-Bhâdrapadâ nakshatra up to 53 ghaṭîs, 31 palas, or 21 hrs. 24 min. 24 sec., after sunrise on
the Thursday; and she then entered the Uttarâ-Bhâdrapadâ nakshatra; i.e. assuming sunrise at 6
A.M., only 2 hrs. 35 min. 36 sec. before sunrise on the Friday (the time, all through, is for Ujjain,
as I have not the exact longitude of Dêôgaḍh; taking it, approximately, as 78° 15′ E., the time in each
case, will be a little less than ten minutes later). This, however, is a highly improbable hour for the
completion, as stated in the record, of the pillar on which the inscription is engraved. But, by the
unequal-space systems of the nakshatras, the Purvâ-Bhâdrapadâ nakshatra ended at about 23
ghaṭîs, 40 palas, or 9 hrs. 28 min., after sunrise on the Thursday; and the moon then entered the
Uttarâ-Bhâdrapadâ nakshatra; i.e., roughly, about half past three in the afternoon, leaving ample
time, in ordinary working hours, for the completion of the pillar before sunset.—It is evident, there-
fore, that for the determination of the nakshatra in this record, we have to apply one or other of
the systems of unequal spaces.
O 2