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

THE GUPTA INSCRIPTIONS; INTRODUCTION. 19
Mahêndragupta, however, as pointed out by me on page 56, note 2, is due to nothing but
Dr. Mill's original mislection of the Bhitarî inscription, coupled with the real name of
Mahêndrâditya on Kumâragupta's coins. And the existence of the independent princess
‘Sanhârikâ,’ with her unnamed daughter, and of Mahâdaitya, and his daughter Dêvî, is
equally imaginary; being due, as indicated on page 1, only to early mislections of the
Allahâbâd pillar inscription. In his first list, Mr. Thomas entered the Mahârâja Gupta,
and his descendants, as belonging to the Solar lineage; and this mistake, though not
repeated in his second and third lists, has not even yet been completely eradicated; but,
as shewn on page 1, the statement rests on nothing but Dr. Mill's mislection of a passage
in the Allahâbâd inscription.
The Nomenclature of the Era.
Before proceeding further, however, it is desirable now to draw attention pointedly to
the fact that, though it is convenient to speak of the era in question as the Gupta era,
yet we have no ancient authority whatever for connecting the name of the
Guptas with it, as the establishers of it, and much less for accepting the early
existence of any such technical appellation for it, as Gupta-kâla.
It is true that this term is used by Albêrûnî, who speaks of the Gûbt-kâl or Gûbita-
kâl, i.e. Gupta-kâla. But, in just the same manner, he uses also the expression Śaka-
kâla for the Śaka era. In each instance, his use of the word kâla is in perfect consonance
with its meaning, ‘time, or a space of time,’ and with the application of it in the sense of
‘an era.’ And the Hindus who furnished him with the information that he recorded, but
who can have known nothing certain about the origin of the era, save only that it had
come down to them through the Gupta kings, would naturally speak of it as the Gupta-
kâla. But Albêrûnî's statements belong to the eleventh century A.D.; and furnish no
real authority, on such a point as this, for early times.
It is true, also, that the words Guptasya kâlâd, "from the era of Gupta," were supposed
by Dr. Bhau Daji¹ to exist in line 15 of the Junâgaḍh rock inscription of Skandagupta.
But the real reading there, as is quite clear even in the lithograph from which Dr. Bhau
Daji worked, is not Guptasya kâlâ[d]-gaṇanâṁ vidhâya, "counting from the era of Gupta,"
but Gupta-prakâlê gaṇanâṁ vidhâya, "making the calculation in the reckoning of the
Guptas." And there can be little doubt that Dr. Bhau Daji's reading, and his translation,
which was more or less unavoidable after the adoption of the reading, were suggested by
M. Reinaud's previously published translation of the extract from Albêrûnî, in which,
though translating Śaka-kâla by "the era of Śaka," he transliterated the original Arabic
by Gupta-kâla, with the words "era of the Guptas" added in brackets.² Dr. Bhau Daji's
¹ Jour. Bo. Br. R. As. Soc. Vol. VI. p. 207; Vol. VII. pp. 114, 123.
² Fragments Arabes et Persans, p. 143.

20 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
reading was unfortunately accepted without further investigation; and has stood up to the
present time. And, in particular, it was most emphatically endorsed by Mr. Thomas, who,
in 1876, and again in 1881, stated that he had previously had doubts as to the correctness
of the reading and translation, but had then become quite satisfied concerning it, from a
personal comparison of Dr. Bhau Daji's materials with others.¹ It is undoubtedly only to
this erroneous opinion, that we must attribute Mr. Thomas' later belief ² that on certain
coins of 'Syalapati,' he could read the syllables Gu and Gupta, with possible indications of
even Guptasya, shewing, it was supposed, that the dates on these coins referred to an era
of Gupta, or the era of the Guptas. These, however, are purely imaginary readings. And
the fact remains that, in the Junâgaḍh inscription, the expression Guptasya kâla does not
occur. And,—apart from the obvious meaning of the words; and from the fact, which
will be insisted on more strongly further on, that the Mahârâja Gupta, the founder of the
family, being only a feudatory Mahârâja, had not the power of establishing an era at all,—
that, in the real reading Gupta-prakâle, we are to interpret the first member of the com-
pound as representing the genitive plural, not singular, is expressly indicated by the use of
the full genitive plural, Guptânâm, governed by kâla, in another date in line 27 of the same
inscription. Of course, both these passages connect the era with the Guptas, as belong-
ing to them, and as being used by them; which had then been the case, so far as their epi-
graphical records go, for at least three generations. But not either of them suffices to
shew that it had been established by them; or even that it had, at that time, acquired the
technical name of the "Gupta era." And what the first expression really shews, is, simply
that the date was being recorded in an era which was new to Kâthiâwâḍ, and was not
the customary era of that part of the country.
In dated records, the genitive plural, Guptânâm, occurs again in line 2 of the
Kahâum pillar inscription of Skandagupta, No. 15, page 65. But there it is gov-
erned by vaṁśa, in vaṁśa-jasya · "of Skandagupta, who is born in the lineage of the
Guptas." So, also, we have Gupt-ânvayânâm in line 1 of the Udayagiri cave inscrip-
tion of the year 106, No. 61, page 258. But it is governed by kulasya; "in the augment-
ing sovereignty of the family of those who belong to the lineage of the Guptas." These
two passages, therefore, do not bear on the question at all.
Again, in the grants of the Parivrâjaka Mahârâjas Hastin and Saṁkshôbha,
No. 21, page 93, to No. 23, page 106, and No. 25, page 112, we have the expression Gupta-
nṛipa-râjya-bhuktau, "in the enjoyment of sovereignty by the Gupta kings." And this
¹ Archaeol. Surv. West. Ind. Vol. II. p. 22; and Jour. R. As. Soc. N. S. Vol. XIII. p. 538.
² See Sir E. Clive Bayley's "Remarks on certain Dates occurring on the Coins of the Hind
Kings of Kâbul," in the Numismatic Chronicle, Third Series, Vol. II. p. 128 ff.

expression is of importance, in shewing clearly that the Gupta dynasty and sway were
still continuing, and that the dates of these records are to be referred to the same era
with the dates of the Early Guptas themselves. But there is nothing at all in the expres-
sion, tending to give the era the name of the "Gupta era."
In the Môrbî grant of Jâïnka, again, as read by Dr. R. G. Bhandarkar,¹ we are sup-
posed to have a very clear intimation indeed, that the era was then known as the Gupta
era; his reading and translation being,—paṁch-âśîtyâ yutê=tîtê samânâṁ śata-paṁchakê ।
Gauptê dadâv=adô nṛipaḥ s-ôparâgê=rka-maṇḍalê,—of which a more literal translation than
that given by him, would be "the Gupta fifth century of years, together with eighty-five
(years), having elapsed, the king gave this, when the disc of the sun was eclipsed."
There is a difficulty in disposing finally of the whole bearing of this inscription; for the
reason that the first plate had been lost sight of, before the grant was obtained for
examination at all; as the result of which, the genealogy of Jâïnka is not known; and the
second plate contains no name of a place, unless we are to find it in the verse that I have
quoted. And now, even the second plate also, the published one, has been mislaid,
and is not forthcoming. That the date refers in all probability to the Gupta era, I do not
dispute. But the above reading of it, takes no notice of the fact that the real word, in
line 17, is not gauptê at all, but gôptê; the au being arrived at only by applying again,
as a component of the vowel, a perfectly distinct and separate sign, which is in reality
nothing but the single mark of punctuation after paṁchakê, at the end of the half-verse,
and which had already been properly interpreted as such. It is only by the deliberate
correction of ô into au,² that the name of the Guptas can be introduced into this passage;
but even then the adjective occupies a very inconveniently detached position as regards
the noun, pañchaka, which it qualifies. We might, with just as much reason, correct gôptê
into gôptrê, 'to the protector, i.e. the local governor;' and this would be even more sustain-
able; for the word stands immediately before dadau, 'he gave,' in connection with which
we have every reason to look for a dative, or some other case. Or, again, without any
correction at all, we might translate "the king gave this (charter) at (the village of)
Gôpta." And this, I expect, will be found to be the real meaning, if ever the first plate,
which is said to be now at Gaṅgâsâgar in Bengal, can be obtained for examination. As
matters stand, there is at least nothing in this passage that compels us to connect the
name of the Guptas with the date. And, at the best, even if a future discovery of the
first plate should force us to adopt the reading of Gauptê in the date, the grant belongs
to rather a late period, when, in Kâṭhiâwâḍ and Gujarât, any knowledge of the real origin
of the era must have been lost, and all that remained was that it was the Gupta kings
who had introduced the era into those parts.

¹ Ind. Ant. Vol. II. p. 258, line 16 f.
² 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.

22 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
And, finally, in a curious passage from the Achâraṭîkâ, which I give in full at page
32 below, note 3, we have the date "when there have gone by seven centuries of years of
the Guptas, increased by seventy-two ;" in which there is used again the genitive plural,
Guptânâm. But the passage contains, as will be seen, a remarkable confusion, which
cannot at present be cleared up, between the Gupta and the Śaka eras. And, like
the Môrbi grant, this record belongs to a late period, and furnishes no authority for
early times.
In the records of the Early Guptas themselves, the date is always expressed simply
by the word saṁvatsara, 'a year,' as in the Udayagiri cave inscription of Chandragupta
II., of the year 82, No. 3, page 25, line 2; or by the abbreviation saṁ as in the Sâñchi
inscription of the same king, of the year 93, No. 5, page 32, line 11; or by the fuller
abbreviation saṁvat, as in the Mankuwâr image inscription of Kumâragupta, No. 11, page
47, line 2. And such of their coins as are dated at all, give simply the numerical symbols,
without the accompaniment of any word meaning 'year,' and much less of any dynastic
qualification of it. For a reason which I indicate in Appendix I. below, these methods
of expression, which are exactly analogous to the usages of other dynasties also, do not
necessarily prove that the era was not established by the Guptas. At the same time, they
contain no proof that it was established by them; and nothing to connect their name
with it.
The fact remains, therefore, that in no early record can we find any indication
that the era was founded by the Guptas; and much less that there ever existed any
technical expressions analogous to Śaka-nṛipa-kâla, "the time or era of the Śaka king or
kings ;" Śaka-nṛipa-saṁvatsara, "the years of the Śaka king;" Śaka-kâla, "the Śaka era;"
Vikrama-kâla, "the Vikrama era;" Vikramâditya-ôtpâdita-saṁvatsara, "the years estab-
lished by Vikramâditya," &c. ;¹ or even analogous to the terms Valabhî-saṁ and Valabhî-
saṁvat, which are of actual occurrence in later times. This fact will become of more
importance, when we come to deal with the question of the probable origin of the
era. Meanwhile, I have drawn attention to it, because it is a point that must be borne in
mind throughout the whole discussion. It is necessary, however, in order to avoid
periphrasis, to have some convenient name for the era; and therefore, as a simple matter
¹ See my notes "On the Nomenclature of the principal Hindu Eras, and the use of the word
Saṁvatsara and its Abbreviations," in the Ind. Ant. Vol. XII. pp. 207 ff., 291 ff.; where, of course,
owing to our advance in knowledge, my remarks on the Kaira grants (p. 208), and on the Kâvî grant,
(p. 291 ff.) now require to be entirely recast.—I have shewn there that, as is now generally recognised,
the use of the abbreviation saṁvat is in no way restricted, as was formerly held to be the case, to the
Vikrama era; see also a pointed instance supplied by Dr. Bühler in the Ind. Ant. Vol. XIV. p. 63.
It is simply an abbreviation of the base saṁvatsara, 'a year,' or of any declensional case of it that
can be used in expressing a date (see No. 5, page 30, note 3). And, bearing this in mind, such
expressions as Gupta-Saṁvat, Śaka-Saṁvat, Vikrama-Saṁvat, &c., of which the two latter actually
occur in original records (see Ind. Ant. Vol. XII. pp. 213, 293), give a uniform, convenient, and
unobjectionable method of quoting the years of the different (

THE GUPTA INSCRIPTIONS; INTRODUCTION. 23
of convenience, I follow the custom of the last forty years, and speak of it as "the Gupta
era." Also, as the era did certainly in later times, in Kâṭhiâwâḍ, acquire the name of
"the Valabhî era," I shall, according to the convenience of the moment, speak of it
indifferently as "the Gupta era," "the Valabhî era," and "the Gupta-Valabhî era."
Subject to the remarks that I have made above, this nomenclature needs no further
apology, if I point out that, even by those who maintain that the Early Guptas were
themselves anterior to A.D. 319, it is admitted that the Gupta era which has come down
to us in the writings of Albêrûnî, had the same epoch with the Valabhî era; they only
maintain that the Early Guptas used a Gupta era which was not this Gupta era.

Albêrûnî's Statements, as rendered by M. Reinaud.

Until the discovery of the Mandasôr inscription, No. 18, page 79, the only direct
information available as to the epoch of the Gupta era, was the statement of Albêrûnî,
who, writing in the period¹ between the 30th April and the 30th September, A.D. 1030,
left on record the following notes, as taken from M. Reinaud's French translation in his
Fragments Arabes et Persans, page 138 ff:—
"People employ ordinarily the eras of Śrî-Harsha,² of Vikramâditya, of Śaka, of
"Ballaba,³ and of the Guptas. ............ Ballaba, who also has given his name
"to an era, was the prince of the town of Ballabha, to the south of Aṇhilwâra, at a
"distance of about thirty yôjanas.⁴ The era of Ballaba is subsequent to that of Śaka
"by 241 years. In order to make use of it, we set down the era of Śaka, and deduct at

¹ See Sachau's Albêrûnî's India, Preface, p. x.
² As is shewn by Albêrûnî's statement further on, this is not the era of Harshavardhana of Kanauj,
commencing A.D. 606 or 607, of which we have an instance in this volume, in No. 43, page 208; but
an earlier era, commencing B.C. 457, of which we have no epigraphical record, and, in fact, no inform-
ation of its existence beyond Albêrûnî's statement, which is accompanied by the remark that, in a
Kaśmîrî almanac, he had read that Śrî-Harsha was 664 years later than Vikramâditya, and that he
could obtain no explanation of the discrepancy (Sachau's Albêrûnî's India, Translation Vol. II. p. 5).
³ Reinaud's Arabic text contains no authority for his use of the double l throughout, and of the
aspirated bh in the name of the town. So, also, Sachau's text contains no authority for his use, once
of the form Ballabha, and once of Balabha. These forms can only be attributed to some fancied
connection between the name of the city of Valabhî and the Sanskrit word vallabha, 'lover, husband,
friend, favourite,' which was very often used as a proper name, but not in the case of any of the rulers
of Valabhî. The word valabhî itself means 'the wooden frame of a thatch, a roof; a turret or
temporary building on the roof of a house.' It occurs in No. 18, line 6, page 81; and, in the form
vaḍabhî, in No. 66, line 2, page 268.—Albêrûnî himself, in treating the name as that of a person, as
well as of a town, possibly wrote under some similar misconception. But, setting aside any such
mistake as this, which is rather like that of his treating Śaka as the name of an individual, instead of
a tribe, he is of course speaking of the era that was used by the well-known Kings of Valabhî.
⁴ The yôjana, a measure of distance, seems to have varied from 2½ to 9, or even 18, English
miles; but with an average length of between 4 and 5 miles.—Walâ, the modern representative of
the ancient Valabhî, is about one hundred and thirty-five miles almost due south of Aṇhilwâḍ.

24 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
" the same time the cube of 6 (216) and the square of 5 (25). The remainder is the era
" of Ballaba. This era will be discussed in its place. As to the Gupta-kâla (the era
" of the Guptas), we understand by the word gupta certain people who, it is said, were
" wicked and powerful ; and the era which bears their name, is the epoch of their exter-
" mination. Apparently Ballaba followed the Guptas immediately ; for, the era of the
" Guptas also commences the year 241 of the era of Śaka. The era of the astronomers
" commences the year 587 of the era of Śaka. It is to this era that the Kandakhâtaka
" Tables of Brahmagupta are referred. This work has among us the title of Arkand.
" Accordingly, placing ourselves in the year 400 of the era of Yazdajird,¹ we find ourselves
" in the year 1488 of the era of Śrî-Harsha ; the year 1088 of the era of Vikramâditya ; the
" year 953 of the era of Śaka ; the year 712 of the era of Ballaba and of that of the
" Guptas."
Albêrûnî's Statements, as rendered by Prof. Sachau.
And to this we have now to add Prof. Sachau's English rendering of the same
passages, in his Albêrûnî's India, Translation,² Vol. II. page 5 ff. :—
" For this reason, people have given up using them,³ and have adopted instead the
" eras of— (1) Śrî Harsha ; (2) Vikramâditya ; (3) Śaka ; (4) Ballabha ;⁴ and (5)
" Gupta. . . . . . . The era of Balaba is called so from Balaba,
" the ruler of the town Balabha,⁵ nearly 30 yôjanas south of Anhilvâra. The epoch
" of this era falls 241 years later than the epoch of the Śaka era. People use it in
" this way. They first put down the year of the Śakakâla, and then subtract from it
" the cube of 6 and the square of 5 (216+25=241). The remainder is the year of the
" Balaba era. The history of Balaba is given in its proper place.⁶ As regards the
" Guptakâla, people say that the Guptas were wicked powerful people, and that, when
" they ceased to exist, this date was used as the epoch of an era. It seems that Balaba
" was the last of them, because the epoch of the era of the Guptas falls, like that of the
" Balaba era, 241 years later than the Śakakâla. The era of the astronomers begins 587
¹ This era dates from the accession of Yazdajird III., a Sassanian king of Persia, in A.D. 632 (see
Prinsep's Essays, Vol. II. Useful Tables, p. 302 and note). The "gauge-year," 400, selected
by Albêrûnî for the comparison of dates, is one year ahead of that in which he was writing.
² It is not certain whether Prof. Sachau's Translation will have been actually published before
the issue of this volume. But, as it will be the standard translation, he has kindly allowed me to
make the necessary references to it, from advanced proofs.
³ i.e. the eras of the Bhârata war and of the Kaliyuga, and certain other methods of reckoning
time, just previously detailed by Albêrûnî, who states that they had been abandoned because of the
very large numbers involved in the use of them.
⁴ and ⁵ See page 23 above, note 3.
⁶ The reference appears to be to the story of the fruit-seller Raṅka and the king Vallabha ;
Translation, Vol. I. p. 192 f., in Chapter XVII., "on Hindu Sciences which prey on the Ignorance of
People."

THE GUPTA INSCRIPTIONS; INTRODUCTION. 25
“years later than the Śakakâla. On this era is based the canon Khaṇḍakhâdyaka by
“Brahmagupta, which among Muhammadans is known as Al-arkand. Now, the year 400
“of Yazdajird, which we have chosen as a gauge, corresponds to the following years of
“the Indian eras:—(1) to the year 1488 of the era of Śrî Harsha; (2) to the year 1088
“of the era of Vikramâditya; (3) to the year 953 of the Śakakâla; (4) to the year 712
“of the Balaba era, which is identical with the Guptakâla.”
The Results of the above Renderings.
The passages given above mention an era, connected with the names both of the
Guptas and of the Valabhî family, in respect of which we have to notice the following
leading points.
In the first place, Albêrûnî calls this era, both “the Gupta era,” and “the
Valabhî era.” As regards its connection with the Guptas, he tells us, according to
the above renderings, that it commenced at the time of the cessation of the Gupta sway,
and seems to imply that, as a matter of fact, it was that event which led to the establishment
of it; and, of course, the result of this would be, that the dates recorded in the inscriptions
of the Guptas themselves, could not be referred to this Gupta era, but must belong to
some era with a considerably earlier starting-point. And, as regards its connection with
the Valabhî family, from the fact that their era had the same starting-point with this
Gupta era, he infers that the members of this family came after the Guptas in chrono-
logical sequence; but he does not indicate any connection between them and the estab-
lishment of the era.
And secondly, in respect of the initial point, he seems first to state, very clearly,
that the epoch of the era, under either name, was when Śaka-Saṁvat 216+25=241 had
expired; which, taking the expired Śaka year in the way in which it has to be used for the
purposes of the present Tables,¹ gives us A.D. 319-20 current² for the epoch or year 0 of
the era, and A.D. 320-21 current for its commencement or first current year. This is
shewn, first by the rule for subtracting 241 complete Śaka years; and secondly by Gupta-
Valabhî-Saṁvat 712 being quoted as equivalent to Śaka-Saṁvat 953, which is known to be
intended as an expired year, since Albêrûnî gives it as the equivalent of the year 400 of
Yazdajird, which is A.D. 1031-32 current. In his next mention, however, as rendered by
M. Reinaud, he apparently speaks of the era as commencing with, i.e. as having for its first
current year, Śaka-Saṁvat 241, which must be understood as an expired year; according
¹ e.g. those in Gen. Cunningham's Book of Indian Eras, and Prof. K. L. Chhatre's Graha-
sâdhanâchîṁ Kôshṭakêṁ, or “Tables for calculating the Places of the Planets.”
² Or, more accurately, the period from the 9th March, A.D. 319, to the 25th February, A.D. 320.
—It is, perhaps, desirable to point out, for general readers, that the quotation of two years of the
Christian era in this way, “A.D. 319-20,” does not denote the whole period of the two years, but only
the period from the commencement of the Śaka year in the first of them, to the end of it in the second.
d,

26 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.

to which its epoch would be when 240 years had expired. While, in a third passage, a
little further on in the book,¹ in mentioning the "roundabout way" in which the Hindus
computed the date (in January, A.D. 1026) of the destruction of Sômnâthpâṭan by Mah-
mûd of Ghaznî, which event took place "in the year of the Hijra 416, or 947 Śakakâla,"
he tells us that they first wrote down 242, then 606 under it, and then, again, 99; with
the result, by addition of the figures, of Śaka-Saṁvat 947, which, as an expired year, brings
us to the period A.D. 1025-26 current, inclusive of the month of January, A.D. 1026.
And here the first figures would seem, at first sight, to indicate that, in this calculation,
the epoch of the era was treated as being when Śaka-Saṁvat 242 had expired.²

¹ Albêrûnî's India, Translation, Vol. II. p. 9.
² This last passage is introduced by Albêrûnî in connection with his remarks on the Lôkakâla
or popular reckoning by cycles of a hundred years. And he expresses himself as being inclined to
think that the number 242 indicates the years which preceded the time when the Hindus commenced
to use the cycle of a hundred years, and that they adopted this cycle together with the Gupta era;
also that the number 606 represents the completed cycles, viz. six, "each of which they must reckon
as 101 years;" and that the number 99 gives the expired years of the current cycle. He goes on to
say that the rule, as found by him in the writings of Durlabha of Multân, was, to write down 848, and
add the Lôkakâla; the sum of which would give the Śaka year. But, in proceeding to apply this
rule to Śaka-Saṁvat 953 (expired), as corresponding to the year of Yazdajird which he had already used
as a "gauge-year," he points out that, subtracting 848, there remained 105 for the Lôkakâla, while the
destruction of Sômnâthpâṭan would fall in the 98th year of the cycle.—There are subsidiary difficulties
here, which cannot at present be fully cleared up. One of them is, the reference of the destruction of
Sômnâthpâṭan to both the ninety-eighth and the ninety-ninth years of a Lôkakâla cycle; with the
addition, moreover, that the ninety-eighth year is indicated as current, and the ninety-ninth is
inferred to be expired.—Another is, that, according to the only Lôkakâla reckoning the nature of
which has been fully explained (Indian Eras, p. 6 ff.), viz. that used in Kaśmîr,—which, Albêrûnî tells
us, had been adopted by the people of Multân a few years before his own time,—the event in question
would fall in the first current year of a cycle. Thus, Kalhaṇa, in the Râjataraṁgiṇî, i. 52
(Calcutta edition, p. 3), makes a very explicit statement regarding the equation between the Śaka era
and the Lôkakâla of Kaśmîr. His words are — laukikê-bdê chatur-viṁśê Śaka-kâlasya saṁprataṁ
saptaty-âtyadhikaṁ yâtaṁ sahasraṁ parivatsarâḥ,—"at this present moment, in the twenty-fourth
laukika (or popular) year, there have gone by one thousand years, increased by seventy, of the
Śaka era." In this passage, he quotes the Śaka year as expired, in accordance with the practice of
astronomers; but the Lôkakâla year as current, as would naturally be the case in using a popular reckon-
ing of that kind. He was writing, therefore, in Lôkakâla 24 current, and when Śaka-Saṁvat
1070 had expired; which is equivalent to A.D. 1148-49 current. And this gives Lôkakâla 1 current
of the same cycle, as corresponding to Śaka-Saṁvat 1047 expired; which is equivalent to A.D. 1025-26
current. Now, that the scheme of each year of the Kaśmîrî Lôkakâla cycle was identical with the
scheme of the Śaka years of Northern India, commencing with the first day of the month Chaitra
(March-April), is shewn by Albêrûnî's statements, and by the notes put together by Gen. Cunningham
in his exposition of this reckoning. And it follows that the first year, current, of each Kaśmîrî Lôkakâla
cycle coincides exactly with the forty-seventh year expired, and the forty-eighth current, of each century
of the Śaka era, and with part of the twenty-fifth and part of the twenty-sixth years, current, of each
century of the Christian era. The month of January, A.D. 1026, therefore, fell in Lôkakâla 1
current, of Kaśmîr, which coincided with Śaka-Saṁvat 947 expired, and extended (see Indian Eras,
p. 171) from the 3rd March, A.D. 1025, to the 21st March, A.D. 1026. And it is difficult to see how, in the

THE GUPTA INSCRIPTIONS; INTRODUCTION. 27
Albêrûnî's Statements, as rendered by Prof. Wright.
The leading historical item of Albêrûnî's information, as rendered by M. Reinaud,—
vis. that the Gupta era commemorated the epoch of the extermination of the Gupta
dynasty,—naturally at once attracted special attention; because of the primâ-facie impro-
bability of such a fact. And, as will be seen further on, various attempts were made, to
find an explanation for it; and with very conflicting results.
application of a Lôkakâla reckoning introduced from Kaśmîr, an event occurring in that month can be
correctly referred even to Lôkakâla 99 expired; and, much more so, to Lôkakâla 98 current. To suit
the former case, we require a cycle commencing one year later than the Kaśmîrî cycle; and to suit
the latter case, a cycle commencing three years later than the same. And, that there were varying
starting-points of this kind, as well as a want of uniformity in respect of the scheme of the years, is
indicated by Albêrûnî's remark that the totally different accounts of the Lôkakâla reckonings, given
to him, rendered him unable to make out the truth about it.—One point, however, seems clear. As
to the number 606, given by Albêrûnî in his first illustration, it is impossible that a centenary cycle
can consist of a hundred and one years. And Albêrûnî himself had previously said distinctly
(Albêrûnî's India, Translation, Vol. II. p. 8) "if a centennium is finished, they drop it, and
simply begin to date by a new one." It is plain, in fact, that the odd six years do not belong to the
cycles of the Lôkakâla. To that reckoning, only the six even centuries belong. I shall shew hereafter
that the epoch of the Gupta-Valabhî era really was A.D. 319-20 current; which might be quoted
either as Śaka-Saṁvat 241 expired or 242 current. If we add the odd six years to Śaka-Saṁvat
241 expired, we obtain Śaka-Saṁvat 247 expired, or 248 current, equivalent to A.D. 325-26 current ;
and this, as we have just seen, would coincide with the first current year of a Lôkakâla cycle, as
reckoned in Kaśmîr. But it is one year too early for the first current year of any cycle of a
reckoning, in a subsequent cycle of which the event in question belonged to the ninety-ninth year
expired and the hundredth year current. Now, I suppose that we must assume that Albêrûnî has
quoted Durlabha correctly. And, if so, then the first current year of each cycle in the Multân
reckoning really was one year later than in the Kaśmîrî reckoning; and coincided with the forty-
eighth year expired, and the forty-ninth current, of each century of the Śaka era, and with part of
the twenty-sixth and part of the twenty-seventh years, current, of each century of the Christian
era. And, if the introduction of the reckoning, at Multân, could be carried back so far, it commenced
with Śaka-Saṁvat 248 expired, and 249 current. This year might have been obtained by adding
seven to Śaka-Saṁvat 241 expired. But, as will be seen hereafter, the real use of Śaka-Saṁvat 241
expired is only for obtaining the basis with which Gupta-Valabhî dates have to be calculated; and it
only brings us to the beginning of the Gupta epoch. The difference between Gupta-Valabhî and Śaka
years, both treated as current for purposes of comparison, is 242. And Śaka-Saṁvat 242 expired
brings us to the beginning of the first current Gupta year. This is the starting-point that was really
wanted for a process of the kind shewn to Albêrûnî. And this is why that year was selected as the
apparent basis of the computation; the true basis being Śaka-Saṁvat 848 expired.—It is evident,
therefore, that the process illustrated by the figures given to Albêrûnî does involve a method of adapting
the Lôkakâla reckoning to the Gupta era; or, more properly, of converting Lôkakâla dates into Śaka
dates through the Gupta reckoning. So far, however, from the figures tending to support any inference
that the Lôkakâla reckoning was introduced by, or in the time of, the Early Guptas, the fact that
Durlabha of Multân would deduct 848, with a remainder, in the particular instance, of 105, or one
complete Lôkakâla cycle and five years over, seems to indicate very clearly that the use of this
reckoning in that part of the country commenced with Śaka-Saṁvat 848 expired, equivalent to A.D.
926-27 current; had it been otherwise, Durlabha's rule would surely have been worded in such a way
d 2

28 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
That the true solution was to be found, not in wrong information given to Albêrûnî
or in a mistake made by him in reporting correct information, but in an erroneous inter-
pretation of his meaning, was first indicated to me by Mr. Rehatsek, who, in December,
1886, gave me, from M. Reinaud's published text, the following literal rendering of the crucial
passage,—"and (as regards) the Gupta era, it was, as is said, a nation wicked (and)
strong; and when they perished, dating was made according to them."¹ Such a rendering
as this would enable us to give to Albêrûnî's words a meaning perfectly clear and consis-
tent with the usual order of things; vis. that the Guptas, though wicked and inferentially
unpopular, yet had exercised so powerful a sway that, even when their dynasty came to
an end, the era, that had been used by them, still continued to be used.
So also, somewhat later, Mr. H. C. Kay, translating the words by "dating was made
by (or, according to) them," added the following remarks on the interpretation of them .—
"The author's meaning is not clear. But, taking the words as they stand, I think they
"can most consistently be understood as signifying an adoption or continuation of the
"method of dating that had been used by the Guptas. The preceding words 'when they
"came to an end' suggest the possible meaning that the dating ran from that event. But
"it seems to me that this construction can be properly preferred, only if there be something
"else in the context, or in the known facts of the case, that would make it obligatory; or,
"at least, that clearly points to it."
I had the hope that, before it came to my own turn to write finally on the subject,
Prof. Sachau's translation of the Arabic text would have removed all doubt as to Albêrûnî's
real meaning on this point. This, however, has not been the case; because of the intro-
duction in his translation as in M. Reinaud's, of the word 'epoch,' which does not exist in
the original text, and the use of which gives the translation a fixed obligatory meaning
that, at any rate, a literal rendering of the original does not compel us to adopt.
that, in the particular instance, 948 must be deducted, with a remainder of only 5 years over.—Gen.
Cunningham, in giving a somewhat similar explanation of the figures 606 (Indian Eras, p. 16), has
assumed that in this process Śaka-Saṁvat 242 is a mistake for 241. But this, as I have now shewn, is
not the case. And I do not agree with the manner in which he deals with the figures as a whole.
Śaka-Saṁvat 241 expired or completed brings us to the end of A.D. 318-19 and the beginning of A.D.
319-20. If to this we add 6 and 600 and 99 complete years, it only brings us to Śaka-Saṁvat 946
expired, or to the end of A.D. 1023-24 and the beginning of A.D. 1024-25; and leaves us still one
year short of that in which the event in question occurred.
¹ Similarly, some thirteen years ago, Mr. Blochmann (see the Jour. Beng. As. Soc. Vol. XLIII.
Part I. p. 368) proposed to translate—"as regards the Guptakâl, they were, as is related, a people
wicked and powerful; and when they were cut off, it was dated in them (the era commenced?)."—
This translatior, however, is spoilt by the bracketed words "(the era commenced?) ;' the use of which
shews why Mr. Blochmann, though giving a translation capable of a totally different meaning,
expressed himself as not able to see any fault in Reinaud's translation.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 29
I am, therefore, glad to be able now to publish the following transliteration, with in-
terlinear word-for-word rendering, and translation, which Prof. William Wright, of
Cambridge, has recently been kind enough to give me, from Prof. Sachau's published
text, of the original of the whole passages of which M. Reinaud's and Prof. Sachau's
renderings have been printed above:—
Text and Literal Rendering.
{ Wa-li-dhâlika 'a'radû 'an-hâ wa-jâ'û 'ilâ
{ And for this they have turned away from them and have come to
tawârîkh Shrî-Hrish wa-Bigarmâdita wa-Shaka wa-Bilaba
the eras of [Śrî-Harsha] and [Vikramâditya] and [Śaka] and [Valabhî]
{ wa-Kûbita. ........... Wa-'ammâ ta'rîkh Balba
{ and [Gupta]. ........... And as regards the era of [Valabhî]
{ wa-huwa ṣâhib madînat Balbah wa-hiya janûbîyah 'an
{ and he lord of the city of [Valabhî] and it south from
{ madînat 'Anhlwârah bi-ḳarîb min thalâthîn jozhan fa-'inna
{ the city of [Aṇhilvâḍa] by near to thirty [yôjana] behold
'auwalu-hu muta'akhkhir 'an ta'rîkh Shk bi-mi'atain
the first of it posterior to the era of [Śaka] by two hundred
{ wa-ihdâ wa-'arba'în sanah. Wa-musta'milû-hu yada'ûn Shg-kâl
{ and one and forty years. And the users of it put down [Śaka]-era
{ wa-yankuṣûn min-hu majmû' muka"ab 'as-sittah wa-murabba'
{ and lessen from it the sum of the cube of six and the square of
{ 'al-khamsah fa-yabkâ ta'rîkh Balba. Wa-khabaru-hu 'âtin
{ five and remains the era of [Valabhî]. And his history is coming
fî mauḍi'i-hi. Wa-'ammâ Gûbt-kâl fa-kânû kamâ kîla
in its place. And as regards [Gupta]-era they were as is said
{ ḳauman 'ashrâran 'aḳwiyâ'a fa-lammâ 'inḳaraḍû 'urrikha
{ a people wicked strong and so after they perished it was dated
bi-him. Wa-ka-'anna Blb kân 'akhîru-hum. Fa'inna
by them. And as if that [Valabhî] was the last of them. And behold
'auwala ta'rîkhi-him 'aîdan muta'akhkhir 'an Shg-kâl
the first of their era also posterior to [Śaka]-era

80 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
{ 241. Wa-ta'rîkh al-munajjimîn yata'akhkhar 'an Shg-kâl
{ 241. And the era of the astronomers is posterior to [Śaka]-era
{ 587 wa-'alai-hi buniya zîj Kndkâtk li-Brhmgûpt
{ 587 and on it is built the canon [Khaṇḍakâṭaka] by [Brahmagupta]
{ wa-huwa 'al-ma'rûf 'inda-nâ bi'l-Arknd.
{ and it the known with us (chez nous) by (the name of) Al-Arkand.
{ Fa-'idhan sinû ta'rîkh Shrî-Harish li-sanati-nâ
{ And so then the years of the era of [Śrî-Harsha] to our year
{ 'al-mumaththal bi-hâ 1488 wa-ta'-rîkh Bkrmâdt 1088
{ that is used as an example 1488 and the era of [Vikramâditya] 1088
{ wa-Shg-kâl 953 wa-ta'rîkh Balba 'alladhî huwa 'aidan
{ and [Śaka]-era 953 and the era of [Valabhî] which it also
{ Gûbita-kâl 712.
{ [Gupta]-era 712.
Translation.
“And for this reason they have given them up, and have adopted the eras of Śrî-
Harsha, Vikramâditya, Śaka, Valabhî, and the Guptas. ............ And as regards
the era of Valabhî,—who was the ruler of the city of Valabhî, which was south of
the city of Aṇhilvâḍa by nearly thirty yôjanas,—its beginning was later than the Śaka
era by two hundred and forty-one years. Those who use it put down (the year of)
the Śaka era, and subtract from it the sum of the cube of six and the square of five;
and there remains (the year of) the era of Valabhî. His history is coming in its
proper place.¹ And as regards the Gupta era,—(the members of this dynasty) were, it is
said, a race wicked (and) strong; and so, after they became extinct, people dated by
them. And it seems as if Valabhî was the last of them. And so the beginning of their
era also is later than the Śaka era (by) 241 (years). And the era of the astronomers is
later than the Śaka era (by) 587 (years); and on it is based the astronomical canon
(named) Khaṇḍakhâṭaka, by Brahmagupta, which among us is known by (the name of)
Al-Arkand. So, then, 1488 years of the era of Śrî-Harsha are in correspondence with the
year (of Yasdajird) that we have taken as a gauge; and 1088 of the era of Vikramâditya;
and 953 of the Śaka era; and 712 of the era of Valabhî, which is also the Gupta era.”
The essence of the whole matter, of course, lies in the precise meaning that is to be
given to the words which follow the statement that the Guptas were wicked and powerful.
¹ See, however, page 24 above, note 6.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 31
Prof. Wright states that, in the original, we have a vague impersonal passive, meaning
"it was dated by them," "there was a dating by them," or "people dated by them;" but
that this certainly does not expressly imply that this dating took place from the year of
the extinction of the Gupta power, and in consequence of that event. That such an
interpretation might, if found on other grounds to be justifiable, be given to this expression
may be admitted. But it is at the least equally open to us to interpret the expression as
meaning that the Guptas had been so powerful that, even when they were dead and gone,
people still used their era to date by. And we shall have to determine hereafter, from an
examination of the details of the recorded Gupta and Valabhî dates, which of these two
possible interpretations is the one that must be adopted.
There is also one other point in the revised translations, to which special attention
should be paid. Prof. Wright's rendering, "(the year) 712 of the era of Valabhî, which
is also the Gupta era,"—and Prof. Sachau's, "the year 712 of the Balaba era, which is
identical with the Guptakâla,"—are essentially different, in their ultimate bearing, from
M. Reinaud's, "the year 712 of the era of Ballaba and of that of the Guptas." They
shew very clearly that Albêrûnî was speaking of absolutely one and the same era, under
two names; not of two different eras, with the same, or almost the same, epoch.
The Theories based on M. Reinaud's rendering of Albêrûnî's Statements.
Up to the present time, however, the only rendering of Albêrûnî's statements that has
been available, is that given by M. Reinaud. His translation has been taken as the
basis of argument by all writers on the subject. And, for a full comprehension of the
whole question, it will be necessary now to consider the matter from the light that was
given by him.
According to his rendering, we had three years to choose between, for the epoch of
the era,—Śaka-Saṁvat 240, 241, or 242, expired; involving a question, especially in res-
pect of the exact determination of the corresponding date in the Christian era, that could
only be settled by accurate calculations of the data available from the inscriptions, ex-
plained in detail, so that general readers might see that the processes were satisfactory.
But, whatever might be the final settlement of this point, the fact remained that
Albêrûnî had information given to him of the existence of an era, coupled with the name
of the Guptas and of the city of Valabhî, which began at some point in A.D. 319, or within
a year on either side of that date, and which it is convenient to speak of as the Gupta,
the Valabhî, or the Gupta-Valabhî era. And, that this era was actually used in connec-
tion with the name of Valabhî, at any rate, was proved by the Verâwal inscription of the
Chaulukya king Arjunadêva of Aṇhilwâḍ,¹ in which the leading record of the year is

¹ Originally brought to notice, many years ago, by Tod, in his Annals of Râjasthân; but
critically edited, for the first time, by Hultzsch, in 1882, in the Ind. Ant. Vol. XI, p. 241 ff.

32 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
Valabhî-Saṁvat 945, together with Vikrama-Saṁvat 1320, which is equivalent to A.D.
1263-64, and with the Hijra year 662, which ran¹ from the 4th November, A.D. 1263, to
the 23rd October, A.D. 1264.
So much was certain. But it was felt to be highly improbable that the Gupta era
should date from the extermination of the Guptas. And students of the subject divided
themselves, almost at once, into two schools.
The first, represented most publicly and with undeviating tenacity up to the last by
the late Mr. J. Fergusson, accepted Albêrûnî's statement as to the period of the com-
mencement of the era, but,—on the analogy of the statement which, on the authority of
Hindu tradition, he also makes, that the Śaka era, too, dated from the overthrow of the
Śakas; a statement which is certainly wrong,²—rejected the addition that it dated from the
downfall of the Guptas; and took A.D. 318 for the date of the rise of the dynasty, as well
as the establishment of the era; selecting this particular year on the theory that the era
did not date from the accession of a king, or from any particular historical event, but, for
convenience of comparison, was simply regulated by the completion of four of the Sixty-
Year Cycles of Jupiter from the commencement of the Śaka era.
The other school accepted A.D. 318, or thereabouts, for the downfall of the Guptas,
and took the Valabhî era of Arjunadêva's inscription, which indisputably began at that
period, as being separate altogether from the real Gupta era, and as having been established
in commemoration of the overthrow of the Gupta power; and began then to look about for
an earlier date for the rise of the Gupta dynasty, and for the starting-point of the era used
by the Gupta kings, and also,—as was maintained, with some inconsistency, by the
majority,—by the Kings of Valabhî, in preference to their own Valabhî era. The chief
exponents of this school have been—the late Mr. E. Thomas, who held that the era was
identical with that of the Śakas,³ with the epoch of A.D. 77-78;—General Sir Alexander
¹ See Indian Eras, p. 126. ² See Appendix I. below.
³ A curious instance of confusion between the Gupta and the Śaka eras, on the part of a Hindu
writer, is furnished by the following two passages from the commentary named Âchâraṭîkâ, by Śîlâ-
chârya, on the Jain work called Âchârâṅgasûtra; I have taken them from a manuscript, supposed to
be about three hundred years old, shewn to me in the early part of 1883 by Dr. Bhagwanlal Indraji.—
This first passage, on pp. 207b and 208a, is in metre, and runs—dvâsaptaty-adhikêshu hi śatêshu saptasu
gatêshu Guptânâṁ | saṁvatsarêshu mâsi cha Bh[â]drapadê śuklâ(kla)-paṁchamyâṁ || Śîlâchâryêṇa*
kṛitâ Gambhûtâyâṁ sthitêna ṭîk=aishâ | samyag-upayujya śôdhyâ mâtsarya-vimukritair=âryê (ryai)ḥ ||.
This passage gives Gupta-Saṁvat 772, expired, the fifth day of the bright fortnight of the month
Bhâdrapada, as the date on which this portion of the Commentary was completed by Śîlâchârya, at
Gambhûtâ (? Cambay).—The second passage, on p. 256b, at the end of the whole book, is in prose,
and runs — Śaka-nṛipa-kâl-âtîta-saṁvatsara-śatêsashu(read śatêshu) saptasu | ashṭânavaty-adhikêshu
Vaiśâkha-śuddha-paṁchamyâṁ Âchâraṭîkâ kṛit=êti || Ba || Saṁvat (page 256b ends here; and the
next page, containing the repetition of the date in figures, and the last final words of the author, is
lost). This passage gives Śaka-Saṁvat 798, expired, the fifth day of the bright fortnight of the month
Vaiśâkha, as the date of the completion of the whole Commentary.—The two passages, indicating, as

THE GUPTA INSCRIPTIONS; INTRODUCTION. 33
Cunningham, who finally fixed on A.D. 166-67 as the epoch ;—and the late Sir E.
Clive Bayley, who selected A.D. 190-91.
Examination of the previous Theories.
Mr. Fergusson’s theory of A.D. 318-19 for the epoch of the era, and A.D. 319-20
for its commencement, is easily disposed of. It was wrong ; but by one year only. The
reason is, that, as has been said, his assumption,—based apparently on a sugges-
tion thrown out by Dr. Bhau Daji in 1864,—was, that the commencement of it was regu-
lated only by the completion of four of Jupiter’s Sixty-Year Cycles from the commence-
ment of the Śaka era, in order that there might be always an even and convenient
difference of two hundred and forty years between the Śaka and Gupta dates. This how-
ever, could be arranged only by applying the Sixty-Year Cycle as used now in Southern
India, where it is not in reality an astronomical cycle at all ; since there the cyclic years run
on in regular succession, without any adjustment of them to the motion of the planet, with
reference to his sign-passing or his heliacal rising, by the omission of a year on certain
occasions, and are taken as commencing and ending with the luni-solar years. According
to the present southern system, Śaka-Saṁvat 1 current (A.D. 78-79) was the Bahudhânya
saṁvatsara ; and Śaka-Saṁvat 241 current (A.D. 318-19) was again the same cyclic year,
Bahudhânya ; and, by this means, some justification might be found for Mr. Fergusson’s view.
But I shall shew hereafter that the real epoch of the Gupta era was A.D. 319-20, which
does not correspond to Saka-Saṁvat 241 current ; so that there was, in reality, a com-
pletion of four cycles and one year, even by the southern system ; and this, alone, is fatal to
they stand, that Śîlâchârya treated the Gupta and Śaka eras as identical, contain a mistake
of some kind or another, which must be attributed to a pedantic desire on his part to introduce a
mention of an era,—whether the Gupta or the Śaka, as the case may be,—with which he was only
imperfectly acquainted. And the mistake cannot be cleared away, unless we can obtain some inde-
pendent record of the real date of Śîlâchârya, sufficient to shew whether the Âchârâṅka was written
during Gupta-Saṁvat 772 to 798 expired (A.D. 1092 to 1118), or during Śaka-Saṁvat 772 to 798 expired
(A.D. 850 to 876). As to this, I can only say that the Śaka era was used so rarely, if at all, in Gujarât
and Kâṭhiâwâḍ, except by the Gujarât branch of the Râshṭrakûṭa family, that probably the Gupta era
gives the real date for Śîlâchârya. And I would throw out a suggestion that possibly some mistake of
the same kind may account for the dates, Śaka-Saṁvat 400 and 417, of the apparently spurious Umêtâ
and Ilâô grants of Dadda II. (Ind. Ant. Vol. VII. p. 61 ff. ; and Vol. XIII. p. 115 ff.).—The passages,
however, are of some interest, in shewing that, in Śîlâchârya's time, there was still a recollection of the
fact that the era,—which must have been known best from its use by the rulers of Valabhî, and which
came eventually, in Kâṭhiâwâḍ, to be called the Valabhî era,—was connected originally and specially
with the Gupta kings, by whom it was introduced into Kâṭhiâwâḍ a. d the neighbouring parts.—After
writing this note, which appeared originally in the Ind. Ant. Vol. XV. p. 188, I found an allusion by
Dr. Bhau Daji, in 1864, to what is evidently the same manuscript, though he quoted only the Gupta
date. He wrote (Jour. Bo. Br. R. As. Soc. Vol. VIII. p. 246) “I have a Jaina manuscript which is
dated in the 772nd year of the Guptakâla ; but unfortunately the corresponding Vikrama or Śâli-
vāhana’s year is not given ; nor is it possible at present to ascertain the exact date of the author
from other sources.”
2

36 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.

death of Skandagupta. All else, serviceable for our present inquiry, that I can find in
Mr. Prinsep's writings, is his statement in the same volume (p. 354), that the dates record-
ed in the Valabhî charters must be referred to the Vikrama era.
In 1845, M. Reinaud republished collectively, under the title of Fragments Arabes
et Persans, certain extracts, with French translations, from works relating to India, which
he had previously published separately in the Journal Asiatique, in September and
October, 1844, and February-March, 1845. In this book, as we have already seen, he
renders Albêrûnî (id. p. 143) as stating that the Gupta era dated from the extermination
of the Guptas. I do not find that he makes any reference to Prinsep on this specific
point. But he shews, throughout, so good an acquaintance with Prinsep's writings, as also
with those of other English scholars, that he must certainly have read Prinsep's transla-
tion of, and comments on, the Kahâum inscription. And, though he may not have inten-
tionally allowed himself to be guided by Prinsep's views, it can hardly be doubted that he
had a reminiscence of the purport of them, when he was translating Albêrûnî's remarks.
In fact, in the face of Mr. Blochmann's, Mr. Rehatsek's, and Prof. Wright's versions, it is
difficult to see how M. Reinaud can have arrived at the exact words used in his trans-
lation, except under some such predisposing influence.
After this, in a paper read on the 15th April, 1848, and published in 1850 in the
Jour. R. As. Soc. F. S. Vol. XII. p. 1 ff., Mr. Thomas entered into an extensive
disquisition on the history connected with the dynasty of the so-called 'Sâh' kings¹
of Saurâshṭra or Kâṭhiâwâḍ; in the course of which, it became incumbent on him to
consider the Early Gupta dates. Accepting it as proved by M. Reinaud's rendering

¹ As I have had occasion to remark elsewhere (Ind. Ant. Vol. XIV. pp. 65, 325), this name of
'Sâh,'—and, with it, also an idea that these princes were Śakas or Indo-Scythians,—owes its origin to
nothing but the fact that, on the silver coins of this series, as also on the silver coins, and even some
of the gold coins, of the Early Guptas, it was the custom, as a rule, not to cut on the dies such vowels
as, if engraved, would fall on or above the top line of the writing of the marginal legend. The
reason, of course, was, want of sufficient space. This custom was observed uniformly in the Gupta
silver coins; and hence such legends as (id. p. 65 f.) paramabhagavata-maharajadharaja-śra-
Chandragupta-Vakramadatya, which represents paramabhâgavata-mahârâjâdhirâja-śrî-Chandra-
gupta-Vikramâditya; "the most devout worshipper of the Divine One, the Mahârâjâdhirâja, the
glorious Chandragupta-Vikramâditya." On the Saurâshṭra coins, the custom was observed almost
uniformly, but not quite so; and hence, to select an instance of its partial observance, such legends
as (id. p. 325) rajñá mahakshatrapasa Rudradamna putrasa rajñá mahakshatrapasa Rudrasîhasa,
in which the superscript vowel î is exceptionally introduced in the last word, and which represents
râjñô mahâkshatrapasya Rudradâmnah putrasya râjñô mahâkshatrapasya Rudrasîhasya; "of the
Râja, the Mahâkshatrapa Rudrasîha, the son of the Râja, the Mahâkshatrapa Rudradâman." The
word sîha, i.e. simha, 'a lion,' is the termination of the names of several of the Kshatrapas or
Mahâkshatrapas. And, because the long vowel î, or the short vowel i with a following anusvâra, is
usually omitted in the coins, thus giving the readings of saha and sahasya, these princes came to be
invested with the imaginary family or dynastic name of 'Sah' or 'Sâh.' Owing to the same fact,
also, the termination of some of the names in the list has occasionally been read as sêna.

THE GUPTA INSCRIPTIONS; INTRODUCTION. 37
of Albêrûnî's statements, and by the Verâwal inscription of Valabhî-Saṁvat 945, that
the Valabhî era commenced A.D. 319 (id. p. 4) or A.D. 318-19 (id. p. 4, note 1),
and, by Albêrûnî's statements, that a family of Guptas reigned as paramount sovereigns
in Gujarât shortly before that date, and were then exterminated, he arrived at the con-
clusions—(1) that this Valabhî era, commencing A.D. 319, may have been established
by the Mahârâja Guhasêna of Valabhî, and may have dated from his accession, or
from some striking event of his reign ; (2) that there could scarcely be a doubt that these
Guptas, who had reigned shortly before A.D. 319, were identical with the Guptas of the
Allâhâbâd, Junâgaḍh, and Bhitarî inscriptions; (3) that the Guptas were the immediate
successors of the Indo-Scythians in Saurâshtra, though traces of the latter might be found,
west of the Indus, down to the end of the fourth century A.D.; and (4) that the so-called
Sâh kings preceded the Indo-Scythians. His chronological results are tabulated on
p. 48 of the same volume. Anterior to B.C. 157 he placed “one or more Sâh kings,”
who are represented on p. 49 by “Îśvaradatta, son of Varsha.”¹ He is followed by thir-
teen Sâh kings, whose coin-dates are interpreted as being all dated in the fourth century
of an era which was assumed to be Albêrûnî's era of Harsha commencing B.C. 457; and
the result for the thirteen kings was from B.C. 157 to B.C. 57. Then comes the Indo-
Scythic conquest; to which is allotted the date of B.C. 26. Then the Guptas. And
then the Valabhî era, commencing A.D. 319. In this Table, curiously enough, no date is
entered opposite the Guptas. But this was probably a printer's omission. For, referring
to the note commencing on p. 4, we find it distinctly stated (id. p. 5) that the dates in the
Gupta inscriptions, and in the Valabhî charters, are to be referred to the Śaka era.
These results contained the first intimation that the rulers of Valabhî acted with such in-
consistency as to invent an era of their own, dating from the extermination of the Guptas,
and yet to continue to use the Gupta era in preference to their own; and also the first
intimation of the exact epoch, anterior to A.D. 319, to which the Gupta dates were to be
referred. One point on which Mr. Thomas was evidently inclined to lay some stress (id.
p. 13 ff.), is Albêrûnî's statement,—based on, but in its application different from, Hindu
tradition,—that the Śaka era commemorated the defeat and slaughter of a Śaka or
Scythian king by a king named Vikramâditya, who, as Albêrûnî points out, is evidently
not identical with the supposed founder of the Vikrama era ; coupled with the occurrence
of this last name, Vikramâditya, as a secondary title, on some of the Early Gupta coins.²
¹ i.e., literally, “Îśvaradatta, son of a year”!! The legend on his coin is given, on id. p. 50, as end-
ing with Varsha putha, in which putha is taken (id. p. 51) as a Zend substitute for the Sanskrit putra,
'a son'!! The two aksharas are in reality the first two syllables of the Sanskrit word prathame,—
varshe prathame, “in the first year ;” see Newton, in Jour. Bo. Br. R. As. Soc. Vol. VII. p. 8, and
Plate, No. 8 ; see also Plate i. No. 1, accompanying Thomas' paper, where the legend is almost equally
clear.
² I myself think it not at all improbable that, hereafter, it may be shewn that the name of
Vikrama, or Vikramâditya, came to be connected with the Mâlava era of B C. 57, in consequence of
some confused reminiscence of a conquest of the Indo-Scythians by Chandragupta I. or II. But the
establishment of the Śaka era is not concerned in this question at all.

38 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
And, in general corroboration of his views, he (id. p. 12, note 4) quoted, some remarks by
Major Kittoe to Colonel Sykes, that the grant of the Mahârâja Hastin of the year 163,
No. 22, page 100, shewed,—on the supposed identity of this Mahârâja with the Hastivar-
man, king of Vengî in the south, who is mentioned in line 20 of the Allahâbâd pillar
inscription,—that one hundred and sixty-three years of the Gupta dynasty had passed in
the time of Samudragupta, and thus proved that the Guptas reigned from the second to the
fifth century A.D. This, however, hardly fits in with Mr. Thomas' reference to A.D.
172 (id. p. 5, note) of a Chandragupta whose date in the Gupta era is the year 93, and
who could only be the father or the son of Samudragupta.
In 1854, General Cunningham published his Bhilsa Topes, in which, on p. 138 ff.,
he drew particular attention to the fact that Albêrûnî three times mentions the Gupta and
Valabhî eras, and identifies them with each other; and on each occasion specifies them
as starting from the same date, A.D. 319. He proceeded to remark—" but it appears to
" me that the most important of these passages must be either corrupt or obscure; for
" the translation given by M. Reinaud makes the epoch of the Guptas commence from
" the date of their extermination! If this is a correct translation, there can be little
" doubt that the text of Abu Rihân must be erroneous; for we know positively that the
" Guptas were reigning during the fifth and sixth centuries of our era. ............
" The statement made in M. Reinaud's version is so extraordinary, that, even without
" any direct proofs of its inaccuracy, I would have set it aside as erroneous. The era of
" the Seleukidæ began with the foundation of the Syrian empire by Seleukos; the Chris-
" tian era is dated from the establishment of Christianity; and the era of the Guptas
" without doubt commenced with the settlement of their own dynasty. For the Guptas,
" as I have mentioned before, date their inscriptions in an era of their own; which, though
" not so named by them, was actually a Gupta-Kâl, and must, therefore, have been called
" such by the people." And he suggested a different translation of Albêrûnî's statement;
to the effect that the Gupta era became extinct with the Gupta dynasty, instead of
dating from the extermination of it; and throughout the rest of his book, he applied the
epoch of A.D. 319 to the Gupta dates. If General Cunningham, our leader in Indian
archæology since the death of Prinsep, had continued to hold these views, and had
worked out his other researches in accordance with them, we should probably have heard
but little more of any theory assigning to the Guptas an earlier period. But, as we
shall see, he soon adopted other opinions
In 1855, in his paper "On the Epoch of the Gupta Dynasty," published in the
Jour. Beng. As. Soc. Vol. XXIV. p. 371 ff., Mr. Thomas made a special reply to
General Cunningham's views and arguments as set forth in Bhilsa Topes. This paper,
however, contains nothing at all worth quoting. And we pass on to 1858, when, under
his own editing, he published a collection of the works of Mr. James Prinsep, then
deceased, under the title of Essays on Indian Antiquities. Prinsep, in his account of

THE GUPTA INSCRIPTIONS INTRODUCTION. 39
Hindu eras, had mentioned the Valabhî era (id. Vol. II. Useful Tables, p. 158), to which,
on the authority of the Sômnâthpâṭan or Verâwal inscription of Valabhî-Samvat 945, he
allotted the epoch of A.D. 318. But he had made no mention of a Gupta era. Mr.
Thomas, however (id. Vol. I. p. 270 ff.), introduced his own previously expressed opinion
that the Gupta dates had to be referred to the Śaka era; and supplied some further facts
which seemed to give additional corroboration to this view. And, on this occasion, he
wound up with the general conclusions (id. Vol. I. p. 276) that the dates in the Valabhî
grants, if applied to the Valabhî era of A D. 318-19, would give far too modern a period;
that these dates did not appear to belong to the same consecutive series with the numbers
employed by the Guptas themselves; and that, while still applying the Śaka era to the
Gupta dates, the Vikrama era was probably the one which, "in spite of any apparent
inconsistency involved," should apply preferentially to the Valabhî grants. Curiously
enough he quoted (id. Vol. I. p. 271, note 1) a modification of meaning which one portion
of Albêrûnî's original words would bear; viz. "again, the Kûbat-Kâl (Gupta era),
"that was, as is said, a wicked and powerful family; when it ceased, it was dated from;
"and, as it were (it would seem that), Balab was the last of them, for the first of their era
"also is 241 years after the Śaka-Kâl." But, he still had not arrived at the absolutely
literal translation of the words which he rendered by "when it ceased, it was dated from;"
and, partly through this, partly through having his attention directed principally to
the connection between the Kings of Valabhî and the Guptas, as described in the words
italicised by him, he still failed to see the real bearing of this passage on the epoch of
both the Gupta and the Valabhî eras. In this treatment of the question, he quoted Prof.
Lassen's views (Indische Alterthumskunde, Vol. II.), to the effect that the rise of the
Guptas took place between A.D. 150 to 160; but I have not had the opportunity of
examining that theory.
Meanwhile, in 1853, 1857, and 1858, there had been published M. Stanislas Julien's
French translation of the Life and Travels of the Chinese pilgrim Hiuen T'siang, in which
there is contained the important statement that, when the pilgrim visited Valabhî,—which
was in or about A.D. 640,—the reigning king there, a nephew of Śîlâditya of Mâlava, a
son-in-law of Śîlâditya of Kanauj, and a member of the Kshatriya caste, was named Tou-
lou-p'o-po-t'o (id. Vol. I. p. 206), Tou-lou-po-pa-tch'a (id. Vol. I. p. 254), or T'ou-lou-p'o-po-
tou (id. Vol. III. p. 163). Such were M. Julien's renderings of the Chinese transliter-
ation of the original Sanskṛit name; which latter he restored as 'Dhrouvapaṭou,' i.e. Dhru-
vapaṭu. And it had already been suggested that this name represented that of one of the
Dhruvasênas of the Valabhî family. Mr. Thomas himself (Prinsep's Essays, Vol. I.
p. 267, note 4) had attached no value to this supposed identification. But Hîuen Tsiang's
statement had begun to be looked on, and very properly, as an important factor in the

40 CORPUS INSCRIPTIONUM INDICARUM, VOL. III.
general inquiry. And I, therefore, now give, for ready reference, in Table II. on
page 41 below, a complete genealogy of the Valabhî family, with the official titles of the
members of it, and their dates as far as I have been able to verify them. Here, however,
in connection with Hiuen Tsiang's statements, I have to draw attention to one or two points
which still remain to be cleared up. In his general account of the Life and Travels of the
Chinese pilgrim, M. Julien (id. Vol. I. p. 206) represents him as telling us, in respect of
the kingdom of Valabhî, that "the present king is of the race of the Kshatriyas (Tsa-
"ti-li); he is the son-in-law of Śîlâditya (Chi-lo-'o-tie-to), king of Kanyâkubja (Kie-jo-
"kio-che); his name is Dhruvapaṭu (Tou-lou-p'o-po-t'o)." Whereas, in the more detailed
account of the Travels, M. Julien renders the pilgrim as speaking, in the same connection,
of not one king only, but more than one, and as saying (id. Vol. III. p. 163) "the kings of
"the present period are of the race of the Kshatriyas (T'sa-ti-li); they all are nephews of
"king Śîlâditya (Chi-lo-'o-t'ie-to), of the kingdom of Mâlava (Mo-la-p'o). Now the son of
"king Śîlâditya (Chi-lo-'o-t'ie-to), of the kingdom of Kanyâkubja (Kie-jo-ko-che), has
"a son-in-law named Dhruvapaṭu (T'ou-lou-p'o-po-tou)." While, in Mr. Beal's Buddhist
Records of the Western World, published in 1884, Vol. II. p. 267, this latter passage
appears in the singular again,—"the present king is of the Kshatriya caste, as they all
"are. He is the nephew of Śîlâdityarâja of Mâlava, and son-in-law of the son of Śîlâditya,
"the present king of Kanyâkubja. His name is Dhruvapaṭa (T'u-lu-h'o-po-tu),"—without
any comment explaining the important difference between his rendering and that given
by M. Julien. And again, according to M. Julien (id. Vol. I. pp. 254 f., 260), the pil-
grim speaks of a Dhruvapaṭu (Tou-lou-po-pa-tch'a, and also simply Pa-tch'a), king of
Southern India; but the kingdom of Valabhî can hardly be included in, and much less can
it include, Southern India; and the statement is inconsistent with the fact that, at that time,
the king of the greater part, if not of the whole, of Southern India, was Pulikêśin II., of the
Western Chalukya dynasty, for whom we have on record no title that at all resembles the
Chinese transliteration, unless it can be found in the second component of the name of
Satyâśraya-Dhruvarâja-Indravarman, the supreme lord or governor of four vishayas and
maṇḍalas, stationed or resident at Rêvatîdvîpa, who is mentioned in line 4 f. of the Goa
grant of Maṅgalîśa,¹ the uncle of Pulikêśin II., and the wielder of sovereignty during the
minority of the latter. These passages present points which must be carefully consider-
ed, before any final opinion is arrived at in respect of the identity of the person, or persons,
intended by Hiuen Tsiang : the more especially because the dates render it impossible
that he should be Sîlâditya VII. of Valabhî, the only one in the family for whom as yet
we have obtained the second name of Dhruvabhaṭa; and because M. Julien tells us (id.
Vol. III. p. 163, note) that the Chinese translation of the name of Dhruvapaṭu of Valabhî
was Tch'ang-jouï, 'constantly intelligent,' which of course supports the supposition that
¹ Jour. Bo. Br. R. As. Soc. Vol. X. p. 365.—This person, however, may be a son of Maṅgalîśa;
see my Dynasties of the Kanarese Districts, p. 22.