PREFACE

TO THE FIRST VOLUME OF THE FIRST EDITION.

AFTER five years spent in the collection of materials for an edition of the
Rig-veda and its Sanskrit commentary by S!yanAk!rya, the :first volume is
now completed, comprising the first Ashtaka (Ogdoad), and about the fourth
part of the whole 1•
When I :first entered on this undertaking, I saw but little chance that I
should ever succeed in carrying it out, and my only hope of success was derived
from the firm conviction that, in the present state of philological, historical,
and philosophical research, no literary work was of grei;i,ter importance and
interest to the philologer, the historian, and philosopher, than the Veda, the
oldest literary monument of the ludo-European world. There were many diffi-
culties to be overcome in carrying out this work. In the public libraries of
Germany no MSS. of the Rig-veda and its commentary were to be found, except
some old copies of the text and a small and worm-eaten fragment of S!yana's
commentary in the Royal Library at Berlin. It was necessary, therefore, to
spend several years in the libraries of Paris, London, and Oxford, in order to
copy and collate all the necessary Vaidik MSS. A complete apparatus criticus
having been brought together in this manner, it became possible to commence
a philological study of the Rig-veda, and . to prepare upon a safe basis a
critical edition of both its text and commentary. But a still greater difficulty
remained, the expense of publishing such a work. These obstacles have been
such, that although the wan-t of an edition of the Veda has been keenly felt b.y
all Sanskrit scholars, and although there were many fully qualified for such a
work, yet no one has been found to undertake it, since the first edition of the
Rig-veda by the late Dr. Rosen was interrupted by the early death of that
highly-gifted scholar. It is owing to a concurrence of many fortunate circum-
1 An introductory Memoir on the Veda. is in the Ancient Sanskrit Literature,' 1859 ; second edition,
Press, and will be published separately. [This was 1860.]
afterwards published under the title of ' History of
Vlll PREFACE TO THE

stances, and particularly to the kind encouragement and liberal assistance

which I have received from various quarters, that these difficulties have been
at length overcome. For several years I was able to advance but slowly, being
entirely left to my own resources, and having but few leisure hours to bestow
upon Vaidik studies. But the further I proceeded in my work, the more encou-
ragement I received. Amongst those who took an active interest in it, I have
to mPntion with sincere gratitude the names of Alexander von Humboldt
and Professor E. Burnouf in France, and of Chevalier Bunsen and Professor
H. H. \Vilson in England. The final success, however, of this undertaking is
owing to the well-known liberality of the Honourable the Court of Directors of
the East-India Company, whose enlightened views on this subject cannot be
better expressed than in their own words : 'The Court consider that the publi-
cation of so important and interesting a work as that to which your proposals
refer, is in a peculiar manner deserving of the patronage of the East-India
Company, connected as it is with the early religion, history, and language of
the great body of their Indian subjects.'
This first edition, however, of -the Rig-veda and its Sa;nskrit commentary is
not intended for the general scholar, but only for those who make Sanskrit
their special study, and for those among the natives of India who are still able
to read their O\vn Sacred Books in the language of the original. It would have
been more agreeable to myself to have kept for my own use the materials
which I had collected for the Veda, (I allude especially to the Sanskrit com-
mentary,) devoting all my time to their study, and communicating to the
public the last results only of my researches. But I felt that I should perform
a more useful work by at once making public those materials, without which
no philological study of the Veda was possible. A greater number of Sanskrit
scholars will thus be enabled to contribute their share towards the elucidation
of Vaidik antiquities, and we may now look forward to a more complete study
of Vai<lik literature than it is in the power of any single individual to bestow
upon so comprehensive a subject, and to a. better understanding of Vaidik
language, religion, and mythology, than can be expected from a scholastic
Indian commentator of the fourteenth century after Christ.
I determined therefore on publishing £rst a complete text of the Rig-ved.a-
samhita, (the SamhiU, and the Pada texts,) together with the only complete com-
mentary on the Rig-veda now existing, the M~dhaviya-ved~rtha-prakisa by
SA.yan~kfi.rya. As the limits of this publication were £xed, it became neces-
sary to sa.ve space as much as· possible, in order to get at least the whole of
 FIRST VOLUME OF THE FIRST EDITION. IX

the text and commentary into the prescribed compass of the edition. For this
reason, as well as because this edition was destined for the use of Indian as well
as European scholars, I had to exclude, and to reserve for a separate work, all
critical and explanatory notes of my own, together with the various readings
of the MSS.
My principal object in this present edition is therefore to give a correct text
of the Rig-veda, and to restore from the MSS. a readable and authentic text of
Sayana's commentary. The former was by far the easier task. The MSS. of
the Rig-veda have generally been written and corrected by the Brahmans with
so much care that there are no various readings in the proper sense of the
word, except those few which are found noticed as such in the commentaries
or in the Pd,tisitk.hyas. Even these are generally of small importanye, and
seldom affect the meaning of a sentence. For the most part they arise from
niceties of orthography and calligraphy, which by themselves are of little
importance to a European scholar, though they may become of interest if
considered with reference to the peculiarities of the old Slikhas or branches of
the V ed.a. The hymns of the Rig-veda are happily much more free from these
orthographic minutiae than the prayers of the Sama and Yagur-vedas. Of real
importance, however, for critical purposes, are the alterations which the verses
of the Rig-veda have undergone when incorporated into the ceremonial prayers
of the Sama, Yagur, and A.tharva-vedas. But neither are these alterations to
be considered in the light of variae lectiones, and, as they cannot be used for a
critical restoration of the received text of the Rig-veda, they will better be con-
sidered in a general critical account of the _whole Vaidik literature 1.
For the text of the Rig-veda I have made use of the following MSS. :-

I. Samhita Text.
S l. A manuscript in the collection of the Rev. Dr. Mill, now belonging to
the Bodleian Library at Oxford (Mill 147-150). It contains all the eight
A.shtakas (Ogdoads) of the Rig-veda. The first A.shtaka consists of 89 leaves
without a date, the last leaves having been replaced by a modern hand. The
second comprises 70 leaves, and has no date. The third, of 92 leaves, is dated
Samvat 1777· The fourth, of'100 leaves, is dated Samvat 1776. The fifth, of
102 leaves, is dated Samvat.1771. The sixth, of 104 leaves, has no date, the

The import.a.nee of these alterations ~s been

1 p. lvii. 'Die Hyi:nnen des Slma-veda herausgegeben,
pointed out by Profess01 Benfey, in his valuable iibersetzt und n:iit' Glossar versehn von Theodor
edition of the Slma-ve~a-samhitt ; Introduction, Benfey,' Leipzig, F. A. Brockhaus, xS48.
2 VOL. I. b
x PREF.A.CE TO THE

last leaves being of more modern origin. The seventh consists of 90 leaves,
and is dated Samvat I 777. .A.t the end of the sixth .A.dhyaya, Vargas 14-28 are
wanting in this MS., but have been added afterwards by the original writer on
two separate leaves. The eighth .A.shtaka consists of rn4 leaves, without date.
There are four different handwritings to be distinguished in this manuscript.
Ashtakas 3-7 are written by the same hand, about the year A. D. r 720, at
Benares. The name of the writer, however, is everywhere co.refully scratched
out with yellow ink. The last ten leaves of the sixth .A.shtaka are written by
the same person, who copied the second Ashtaka. The :first and last Ashtakas
again are copied by a third writer: while some few leaves on white paper
belong to a fourth and quite modern hand, and have probably been supplied
by the Pandit employed by Dr. Mill.
S 2. Another manuscript in the collection of Dr. Mill, now belonging to the
Bodleian Library (Mill 151-154). This also is a complete copy of all the eight
Asbtakas. The :first .A.shtaka consists of 103 leaves, without date. The second
has 93 leaves, and is dated Saka 1679 (A. D. I 757). The third fills 97 leaves,
and is dated Saka 1677. The fourth comprises 92 leaves, and is dated Sll.ka
1679. The :fifth consists of 62 leaves; the sixth of 80 leaves; and the seventh
of 76 leaves; all of them without dates. Leaves 12-37 in the seventh Ashtaka
have been supplied by a modern writer. The eighth .A.shtaka comprises l 30
leaves, and is dated S~a I 776. In this manuscript also four different writers
can be distinguished : to the :first belong .A.shtakas 1-4 ; to the second, .A.shtakas
5-7 ; to the third, the eighth .A.shtaka ; and to the fourth, the modern addi-
tions in the seventh Ashtaka.
S 3. The third manuscript belongs to Colebrooke's collection, deposited in the
library of the East-India House, where it forms Nos. 129-132 of the Cata-
logue. No. 129 contains the Gribya-s"fitras of Asvalayana and the first and
second Ashtak.as. The first .A.shtak.a contains 59 leaves, and is dated Sa:mvat
1802: the second contains 60 leaves, and is of the same date. No. x30 con-
tains the third and fourth Ashtak.as; the former of 53, the latter of 54 leaves;
both dated Samvat 1802. No. x31 contains Ashtakas 5 and 6; ~he former of
54 leaves, the latter of 56 leaves; both dated Samvat 1802. No. 132 contains
the SarvA.nukrama.and .A.shtak.as 7 and 8; the former of 56, the latter of 61
leaves; equally dated Samvat 1802. The whole manuscript was evidently
written by one person, about the year l 745 : his name is scratched out, but
seems to have been So~a.gopakAsinatha.
 FIRST VOLUME OF THE FIRST EDITION. Xl

II. Pada Text.

PI. A manuscript in Dr. Mill's collection, now belonging to the Bodleian

Library (Mill 155-158). It contains all the eight Asht~kas. The :first .A.shtaka
consists of 97 leaves, and is dated Sarnvat 1727, &\ka 1592 (A.D. 1670). The
second contains 129 leaves, and is dated Samvat 1728. The third fills 109
leav0s, and is not dated. The fourth has 107 leaves, and is dated Sam.vat
1727. The fifth contains 84 leaves, without a date; the last leaf having been
supplied by a modern band. The sixth .A.shtaka comprises 89 leaves, and is not
dated. The seventh consists of 95 leaves, and is dated Samvat 1672 (A.D. 1616).
It was difficult, however, to read the last page, which contains the date and
the name of the writer, but has been pasted over with yellow paper 1• The
eighth Ashtaka contains 86 leaves, but breaks off with the last Varga of the
seventh .A.dhyaya. The rest has been supplied by a modern manuscript, without
accents, dated Samvat 1857, Saka 1722 (A. D. 1800). In this manuscript also
four different hands may be traced. The oldest part contains the seventh and
eighth .A.shtakas, written in A. D. 1615; next come Ashtakas I-6, written in
A. D. I 760; thirdly, the supplement of the eighth Ashtaka, written in A. D. l 800 ;
and lastly, some few leaves of still more modern origin, probably copied by a
Pandit employed by Dr. Mill.
P 2. A complete copy of the Rig-veda-samhiM, bequeathed by John ·Taylor,
M. D., to the Hon. Court of Directors of the East-India Company, and entered
in the Catalogue under No. 2032. It has been copied at Bombay, and is bound
together in one large volume: its date is from. Saka 1736 to 1737: the name
of the ·writer Ri:\.mabhatta, called Sebenkara.
It was not necessary for an editor of the Rig-veda to collate a greater
number of MSS., or to classify them according to their age and origin. I have
seen nearly all the MSS. of the Rig-veda which exist in Europe, and I feel
convinced that no use can be derived from them as manuscripts, because
all of them are but transcripts, more or less carefully executed, of one and

1 All that can be read is ~ ~ C\-'e~ 1'tif and from the fragmentary passages which are still
tit- - - ~ - 'l(Cll4dq' I .. - - - iltiflCIClftctg- legible, I conclude the MS. to have been written by
" ' - -. Even this has been traced over with ink, Damodara-SadAsiva, who generally signs himself
by which it became still more illegible. Afterwards '1i1'fij'li"i\Cll4iiifi ~.C~u.iL.t.--:;:fdleciilCIS\­
another writer has given the date at which the accents ~ ~~ (or i(~l('~ or ~ti\i(«i( 1-
were 11tdded, but there also we can only read ~ ftf't;r) ~ I. He was still alive in Samvat
C\~ etc. To judge from the handwriting of the MS. x7o6.
b2
Xll PREFACE TO THE
the same text1. If there were, as in other Sanskrit works, corrupt passages,
on which doubts might exist, a comparison of the Samhitli text with the Pada
text, or a reference to the commentary, would have been sufficient to remove
such doubts. But so far from this being required, the reading of the Samhit:l
text, the Pn.da text, and the text which the commentator had before him, can
each be established with such certainty by the MSS., that it would be wrong
to correct even the smallest differences in the quantity or accent of vowels
which occur occasionally between these three texts, but which are always sup-
ported by the full testimony of each class of MSS. There are instances in
almost every hymn where a long vowel occurs in the Sam11ita text, while the
Pada text has a short one. The commentator considers these productions _of a
short vowel as Vaidik liberties. But in some cases where a long vowel seems
to be regular, and the Pada text has notwithstanding a short one instead, this
shortening is equally pronounced by Sayana as a Vaidik irregularity : for
instance, Rv. I. 37, r I. "Cl4cttifC't instead of "Ci4 rcttifif 2 ; I. 6 I, 14. ~instead of~ 3
Instances occur where the text followed by the commentator is different from
the text of' our MSS. Rv. I. II6, I. the Pada text has ~,while, according
to Sayana, the author of the Pada text (Sakalya) must have read ~'~
Rv. I. 61, 9. the Pada text ought to have, according to Sayana, '@'''U't' but all
the MSS. have '@.S~. Rv. I. 52, 10. ~has no accent in the Samhit:l and
Pada texts, while Sayana explains it as if it were a paroxytone 4 •
Notwithstanding the great accuracy with which the MSS. are written, occa-
sional mistakes occur. Letters, syllables, and words are sometimes left out,
sometimes misplaced in one or other of the MSS., owing to inevitable inadver-
tencies on the part or the copyist 5• They seldom occur, however, in more than
one MS. in each. instance.
1
The late Dr. 1F. Rosen, who had untlerta.ken a.n 1Qf"1ifit ~ ~; u, x. ~
while I.
edition of the Rig-veda.; pa.rt of which. wa.s pubhshed ~is explained by \\Q1(Cii(1§1Nt1Qi1..._i<4i
after lus lamented clea.th in 1837, has for the same ~I Of. PAniniVIII• .a, 17, 2.
reason give11 no various readings for the text of the 4
S!l.yana does not explain how ~ could be
Samhitl; and I may also quote Dr. A. Kuhn of Berlin,
without an accent in the text, though generally he
ou tbe saD1e subject, a.s a later but not less weighty
endeavours to account for irregular accents; that is to
a.uthority on questions connected with Va1dik literature,
say, to reconcile them with the rules of Plnmi. See,
:in J ahrbil.eher fur wissenscha.ftliche K11tik, Berhn,
for instance, I. 61, x, where he even admits of a
1844, p 131.
2~ • @\f~h•if\tt: •
""'"'"~ IEP=ttt~: a 5
This the writers of MSS. admit themselves in
3
"'<\ttt(q
• • ' •
qli(Cifit(: QI [See now Varietas several cases. For although they generally say at the
T"ectionis, p. 29.] Cf. I. 84, 4. ~; I. 84, 6. the end of a. MS. that they have copied it as it was in the
~ short 1 in ~ mstead of ~ is called -.if- original, and that it is not their fault if mistakes occur,
 FIRST VOLUME OF THE FIRST EDITION. xiii

As to the spelling of words, I have endeavoured as much as possible to

preserve consistency, and never to deviate from the general laws of Sanskrit
orthography, except where Vaidik peculiarities were based on the unanimous
authority of all the MSS. Each MS. has its o\vn peculiar character, which
must be known and taken into account in order to make proper use of it 1 •
Some MSS., for instance, avoid certain groups of double letters, not only where
the reduplication arises from phonetic laws, but also where two independent
letters have been joined together. This shorter way of writing occurs not only
in V aidik, but also in other MSS., and cannot be considered as affecting the
pronunciation of words, because the simple letter makes the preceding vowel
long, as if a double letter had been written. I have seen, therefore, no reason
for adopting this way of spelling in a printed edition, because other Vaidik MSS.
frequently give the double letters where they ought to stand, according to the

yet they compla.m frequently of the hardships and d1f-

-m
added by a later hand m S 3 (Cf. Zur Lite1atur und
:ficult1es of their work I subjoin a few specimens of Gescbichte des Yed~ Drei Abhandhmgen von R.
their poetry Roth, p. 82.) Yet though S 3 is certainly a very
accurate copy, it could scarcely be expected to have
q~1titlif ~ preserved traces of an older ieading than Yaska. had
~l
before h.Im m the Nirukta, where this ve1se is quoted
ai\t~111M: ~ ~ 'if ri: q (Nirukta XII 22) with the long a. The fact is, as wtll

1 ~''
If I have written a mistake here, because I could not
appear from a more accurate collation of this MS.,
that S 3 dispenses most frequently with writmg the
see, or my mind was wandering, noble persons may long ''owel in cases of Anunasika. in words where there
correct it all, but let them not be angry with wliters.' can be no doubt that the long vowel is necessary, and
where all the other MSS. have it. Instances of this
~ ~ Jllli'l~lfil ~: l
occur continually, and have generally been corrected
~ lj'lii3j'if ~ qi ~~ 'if ~ " by the writer who added the accents; as 1n I 48, 14.
'A Muni even may err ; Bhi:ma even was vanquished
be it right or wrong, no fault must be given to me.' ~~; I.45,1.~~~; I.44,1.
~~~:; I.44,4.~~; I 44,7. ~Q;
lff11t ~ TIT wyt1f ~ iNf I I. 47, 5. . . . 'flfli, etc. Sometimes the long vowel
1tff( 1'1 iiljiif 'EfT ir1' ~ if fcrittl II is not written, but, according to the laws of Vaidik
'As I ha·n seen the book, so I have written it; be it
g1ammar, the quantity is marked by a particular sign.
right or wrong, it is not my fault.'
Rv I. 63, I. i{f ~-$~ ; I. 62, 12. ,it~-$~ ;
~q9t4ifl'.'4iq: ~' I. 59, 6. ~~-$~; here also another hand
cftir Nm ~ ~ Ji fdq lijt1'(( II has added the long vowel. Rv. I. 59, I, again, we
'My back, my hips, and my neck are broken; my eight 6.nd d~''atl instead of ~ ~. This does not
is stllf in looking down : keep this book with care at all exclude the possibility of an old mistake in this
which has been written with pain.' Others read verse (Rv. I. 50, 6), but it shows that, in order to
~: instead of~:. make proper use of a MS. it is not sufficient to collate
1 Rv. I. 50, 6, for instance, in order to support a
a few passages, but that the whole character of a. MS.
conjecture, great stress has been Ia.id on the fa.ct that must be studied by a careful collation, before it can be
in ~ ~ ~' the long a of '6l'1fi has been used as an authority for particular passages.
XIV PREFACE TO THE
laws of Sanskrit grammar, and because a deviation from these laws might lead
to confusion. Some MSS. write a double aspirate, where, according to the laws
of Sanskrit grammar, the first of the two letters ought not to be aspirated. I
mean forms like ~ mstead of ~ etc. As good MSS., however, reAtrict
:z
this peculiarity to the group 'f instead of 1, and as in this case also carefully
written MSS. preserve the regular form if, it would have been to no purpose
to give up a general phonetic law (on the incompatibility of two aspimtes) for
what may lJe after all a mere difference in writing 2• But although I have tried
to be as consistent as possible in the way of spelling, yet I have submitted to
the authority of the MSS. in cases where their testimony was quite unanimous,
particularly with regard to nasal letters, because in such cases there was
reason to suppose that certain peculiarities, if exhibited by all the MSS.,
might rest upon the authority of that Prll.tisakhya to which our MSS. belong.
Whether with the conflicting testimonies of old grammarians, quoted in the
PrAtis!khyas, it will be possible to restore the whole SanihiM. of the Rig-veda in
such a manner as. to include all the minute niceties of spelling prescribed by
different members of each SA.kha, is a question on which I should feel inclined
entirely to submit to Professor Roth's authority, who has devoted much time
and learning to this interesting branch of Vaidik literature.
There is only one case where I thought it better to deviate from the way of
spelling adopted by the Vaidik MSS. : this is with regard to the Avagraha.
The Vaidik: MSS. use the Avagraha where a hiatus arises from two vowels meet-
ing at the end and beginning of two words, while the common custom has been
to use this sign to mark the elision of an initial a, which has been dropped in
order to avoid a hiatus between it and a preceding vowel. If the Vaidik use
of the Avagraha had been adopted, it would have been necessary to introduce
a new sign for cases of real elision, which the Vaidik. MSS. do not mark at all
Instead of this I have prefeITed to retain the Avagraha where it is of real use
in marking the place where a letter has been dropped, and to exclude it where
it has no other purpose than that of marking a hiatus 8, which is quite as clear
to the eye without any such sign.

1
Other groups, which also occur occa.sionally, but often diftlcult to say which form is meant ; as words
never in a.11 the MSS. a.t t\>.e same time, are '\JS:'(,.,, like ~ also, where the~e can be no doubt as
and 'af. Of. Benfey, S&ma-veda, p. xuiv. to the double k and its pronunciation, are written
may be mea.n.t for I' in the same way a.a I\ is
2
f ~·
meant frequently for IT. Bukka's name is spelt 1af 8
The Avagraba is used in Vaidlk MSS., not only
and 11\; and in words like ~ and ~it ii where an elision ought to have taken place, according
 FIRST VOLUME OF THE FIRST EDITION. xv

I have now to state the principles which I have followed in editing the
commentary of Sayana. If the MSS. of the Rig-veda are generally the best,
the MSS. of the commentaries are nearly the worst to be met with in Sanskrit
libraries: they have generally been copied by men who did not understand
what they were writing, and the number of mistakes is at first sight quite
discouraging. No class of writings would have needed morP to be copied by
men who were masters of their subject than commentaries such as these, which
abound in short extracts, taken, without any further reference, from other books
on grammatical, etymological, ceremonial, theological, and philosophical subjects.
Most of these quotations are only detached fragments, full of technical expres-
sions, and often quite unintelligible by themselves. In order to understand,
nay frequently in order to read these passages, it was necessary to have
recourse to the works from which they were taken. Some of these works
were already published, but others existed in MS. only, and had :first to be
analysed, and furnished with alphabetical indices, before any use could be
made of them. By this process, however, a double advantage was gained. In
most cases a comparison with the work from which passages were quoted
served to correct the mistakes of the commentary ; while in other cases a
frequent recurrence of the same quotation in the commentary furnished also
the means of correcting false readings in the original works, or supplied, at all
events, a well-authenticated varietas lectionis. Sometimes, however, the same
passage is quoted differently in different places of the commentary. This may
be accounted for by the fact that Indian authors trust so much to their
memory as to quote generally by heart. Such slight differences, therefore, I
have left unaltered, whenever they were supported by the testimony of the
best MSS.
.A.13 to the other part of the commentary, which contains the original
explanations of M!dhava, as edited by S!yana, a similar advantage for a critical
restoration of corrupt passages was derived from the frequent repetition of
the same explanations in different hymns, which also made it easier to
become familiar with the style of the commentator, and his whole way of

to the general laws of Sanskrit grammar; (for instance, by the infiuence of one immediately following, afl ~
tctfl'd:, ~'1''41'fei, it-1'~, 1lfilcf\' '~' firir;q''~' ~.f"AA!; and even
.S'Wft' ;) but also after those vowels which Sanskrit 'vhere a :final consonant has been dropped on account
grammarians call pragrihya, and which are never af~ oh following vowel, as~'""~' f't1H 111 Nat Ifft,
fected by a following vowel,_ like ~'~, ~&'=ti fij:;q C1. By being employed for so many pur-
lijit iii af iJ..&'~ &'~, . . .&'""'!:· The same poses, and thlS not at a.11 consistently, the Avagra.ha,
sign is put also after a vowel which has been modifted as it stands in Vaidik MSS •• is of little use.
XVl PREFACE TO THE
thinking and interpreting the Veda. It was a further advantage that the
MSS. were most numerous for the first book of the commentary, and, as
SAyana says with regard to the first Adhyaya of his commentary, ,<'Ul\flA"ffl
~ 1ftlf' 1(ifiif?f ~Nlitl'( 'he who has got through this, can understand the rest,'
it migi1t, at all events, be said with some truth, that after having worked
through the £rst Ashtaka, an editor may go on to the rest with a smaller
number of MSS.
For the first Ashtaka I had twelve MSS. However, we have learnt from
Greek and Latin philology that a great number of MSS. is not at all desirable
for critical purposes. In most cases those numerous MSS. which have been
collated for classical authors have only served to spoil the text ; to make the
reading of doubtful passages still more doubtful; and to give rise to a mass of
conjectural readings, based either upon the authority of the transcriber of a
MS., or upon that of an ingenious editor. In this manner an immense deal of
labour has been wasted in classical philology; so that now, after the simple
rules for using MSS. have been laid down by a new school of critical philo-
logers, such as Bekker, Dindorf, Lachmann, and others, almost all the old
editions of classical authors have become useless for critical purposes, with the
exception of some of the editiones principes, which, as they simply reproduced
one MS., though generally a very bad one~ can claim for themselves at least a
certain degree of authenticity. Before MSS. can be used for critical purposes, it
is necessary that they should themselves be examined critically, in order to
determine their origin, their age, and their genealogical ramifications, and thus
to fix their relative value. If it were possible to recover the original MS. of a
work, as written by the author himself, there would be no need of criticism ; we
might dispense with all later MSS., and we should merely have to reproduce
the original text, pointing out at the same time such mistakes as the author
himself might have committed. But generally our MSS. are much later than
the composition of the works which they contain, and, if compared with one
another, they are found to differ from each other, partly in mistakes and omis-
sions, partly in corrections and additions, arising, in the course of centuries,
from the hands or heads of ignorant or learned transcribers. For the most
part these various readings are not peculiar' to one or th~ other MS. only, but
the same mistakes ?CC\lr generally in several MSS. at the same time. Now,
if there are, for instance, certain MSS. which omit a certain number of passages
tliat have been preserved in others, we may safely conclude that the MSS.
which coincide in omitting these passages flow from the same original source.
 FIRST VOLUME OF THE FIRST EDITION. XVll

But out of the number of MSS. which thus coincide in omitting certain sen..
tences, some may again differ in other characteristic passages, and thus form
new classes and subdivisions. By carefully collecting a large number of such
characteristic passages, all the MSS. of an author arrange themselves sponta-
neously, and form at last a kind of genealogical series, where each has its
proper place, and commands, according to its position, but not according to its
age, its proper share of authority. For a MS. may be of modern date, yet
if by a comparison of certain classical passages it can be shown to have been
copied immediately from an old MS., it inherits, so to say, a greater share of
authority than MSS. which, though of greater age, are of more distant relation-
ship. Here, however, a distinction must be made between the authenticity and
the correctness of a certain reading. As the date of the oldest MS. reaches but
seldom to the age of the author of the work, we can only expect by a critic~!,
and, so to say, genealogical arrangement of MSS., to arrive at the best authen-
ticated, not at the original and correct text of an author. It sometimes happens,
indeed, that all the MSS. of a work can be shovn1 to have originated from one
MS. which is still in existence, as is the case, for instance, with Sophocles. But
most frequently there remain in the end two or more different groups of~MSS.,
each with its own peculiar readings, and each group entirely independent of the
other. In the former case the best that can be done in a merely critical edition
is to reproduce the oldest and best authenticated MS. But it frequently
happens, that even in the oldest MS., upon which all the others depend, mis-
takes occur, which ha~e been corrected in more modem MSS., sometimes by
mere conjecture, sometimes by using quotations from an author occurring in
other works which have preserved a more ancient and more correct reading.
Such passages are open to philological discussions, and have to be treated in
notes. In the latter case, if there remain several independent branches of MSS.,
the task becomes more difficult; and as each class of MSS. may claim. for itself
the same degree of authenticity, it becomes the duty of an editor to choose in
each :earticular case the reading of that class of MSS. which may seem to him
most correct, and best in accordance with the general style of the author.
Frequently, however, even in this case one class of MSS. will be discovered,
which by its general character of correctness acquires a right to overrule the
testimony of the other classes in doubtful passages. .All this must be finally
settled before a critical edition of any author can be commenced; and it is
necessary, therefore, for an editor to collate most carefully even those passages
where the various readings of ¥SS. bear the evident character of mere mis-.
3 VOL. I. c
 xviii PREFACE TO THE
takes, but where, notwithstanding, the omission- of a single letter may often
serve to point out the connection of a certain class of MSS. Grave errors and
long omissions are generally much less charact~ristic as marking a family-
likeness between certain MSS. than small and insignificant mistakes, because
the former have often struck those who copied a MS., and have induced them
to correct erroneous readings on their own authority, or to supply important
,omissions from other MSS., in case they could be procured. The more insig..
nificant mistakes, on the contrary, were more likely to be overlooked and to
remain unaltered.
With regard to the twelve MSS. of the commentary to the first Ashtaka
of the Rig-veda, I have only succeeded in reducing them to three independent
classes. It is not very likely that MSS. should still be found in India contem-
poraneous with S~yana, though, if we could trust native authorities, copies of
SAyana's works have been buried in the ground near Vidyfinagara. Excluding
these MSS., the existence of which is extremely problematical, I am convinced
that there are no MSS. at present which have any claim to be considered as
exhibiting the commentary exactly such as it came from the hands of S~yana.
I shall proceed to give a list of those MSS. which I have made use of for
this edition. I shall call the three classes, to which all the MSS. belong, A, B,
and C, marking at the same time each particular MS. by its own number.

A.
A 2. A manuscript in four volumes, containing all the eight Ashtakas. It
was presented by Colebrooke to the library of the East-India House, where it is
entered under Nos. 2133-2136. It contains also the text of the hymns, but not
throughout. In some Ashtakas the accents also are marked. It is dated
between 1747 and I 760 1, and has been written by different hands at Benares.
It is on ~he whole the most valuable MS., and the only one which represents
1
'First Ashtaka. . Sixth Asbta.ka.: ~ C\1::0~=1750 A.D.
'tffifi fq'*!lq(l•i\l•i!fifi"ill ir1' ~
~ ~'C4'tiilfi ,r.111vn '4i'fl~~.\i\ n
:'bft+1i\lf4Qli(\Rinttt'l'ti\l-?t"ill,C(f ~ 1
Seventh Ashtaka. ~ qi::oC?, C!(if ~ ~
~ ~ift:c41qj f'lff.ft ~l(ql\C\ f<tNcft~(
~ ~ ~ . ~ 3Jtlttitl\l'4QJi(lf~Ml&*ICfi(\q1nil'1 '@lf
'ill.I '9'\ ctt¥lN41<4i JJ(\fictH:iht'ia n 'ClUf 'ii' H This gives r753 A.D.
This gives SAka 1673, or 1751 A.D.
Eighth Ashtaka.· ~ qt:q'e if lf.fttl'4 ii«QQ-
Fourth Ashtalta.:
'f4'ti(tfli ~Nt-ilf 1fll: ~: I ~
act4'•«tt4"fii~Q{ ~: 4tflf~nn ro, ~ ~ 'Jftifiltli ~ ll This gives
1'i43jWiliflC\lli ti~Jiii"if\tltaiil I 1761 A. D.
~This grves Baka 1669, or 1747 ,A..»,
 FIRST VOLUME OF THE FIRST EDITION. :Ex
this first class of MSS. throughout the whole of the commentary. though in
some of the later Ashtakas long passages are wanting in this MS. also.
For the first Ashtaka I have to include in this class-
A I. An old MS. of the National Library at Paris, containing the first
Ashtaka only. It is well written, and indeed gave me the first hope that a
critical edition of S~yana might still be possible. It is dated Samvat 1025 1
(1569 A. n.), and is in many respects more useful than A 2. But, though of
earlier date than A 2, it cannot be considered by any means as the original
from which A 2 was copied: for although the omission of passages which
stand in A I might be ascribed to the negligence of the transcriber of A 2, yet
there are also whole lines which are left out in A 1, but which are not wantmg
in A 2. Both MSS. flow from one original source, and their aut~ority is on
the whole equal; though A I, as being the earlier branch, has thereby some
advantages over A 2. The absence of the other seven Ashtak.as in A I is a
great loss for an edition of the commentary.
To the same class must also be referred A 3, the MS. of the first Ashtaka in
Sir R. Chambers' collection, now in the Royal Library at Berlin. Of this old
M.S., which is in a very bad state of preservation, I possess no complete colla-
tion, but only short notes and extracts which I made before I had seen the MSS.
at Paris and London, and before I was in a position to conceive the possibility
of a critical. edition of the commentary. A comparison of several characteristic
passages, however, shows the connection of this MS. with A I and A 2, with
which it coincides in several of its right as well as of its wrong readings. As I
was not able, however, to verify in each particular passage the reading of this
MS., it is not to be understood as included in the general designatioIJ. of class A~
unless especially mentioned.
B.
The second class, B, is represented by two MSS., both of them complete
copies of the commentary. I owe my first acquaintance with this class of
MSS. to the kindness and liberality of Professor E. Burnouf, who allowed me,
during my stay at Paris, to copy and collate the MS. of SAyana in his pos-
session. Besides several passages which are corrected or supplied by this MS.
in places where .mistakes or omissions occur in A or C, it contains also a number
of past'ages which evidently bear the character of later additions: they stand
frequently without any connection with the rest of the commentary, and I had
1 vrir: u ~ q~~'4 1ff afhtiitM•P:f-tt~cttd
rt.iiltti4(it(ifir S\wfl'flitl'ifinh I~
~ff('e~1-.ritt~~- ~,, . .TQ(~n
C2
xx PREFACE TO THE
no doubt that they owed their origin to marginal notes which had been added
by BrAhmans while studying the Veda, and which in later copies had been
incorporated into the text, though sometimes inserted in a wrong place. This
supposition I found fully proved l;>y another MS., which has lately been added
to the library of the East-India House, and which is evidently the very MS.
from which Professor Burnouf's copy was taken. In this MS. all those spurious
passages, which occur neither in .A nor C, have not yet been incorporated into
the text, but appear still as marginal notes. Nay, it is even easy to see how, by
mistaking the signs of reference, the transcriber was led to misplace some of
these additions. I call the MS. of the East-India House BI, and that of
Professor Burnouf B 2 ; though the latter is on the whole so carefully copied,
that both may be considered as one MS.
c.
The third class of MSS. is much more numerously represented, but consists
almost entirely of modern copies, executed, with more or less care, for the use
of European scholars. Yet this class of MSS. also was indispensable for restor-
ing a complete and correct text of S!yana: for though omissions and mistakes
are very frequent, yet some difficult passages are given more correctly in this
class of MSS. than in either A or B; while others, which are partly omitted in
.A. or B, receive occasionally great help from a comparison of C. Modern
additions occur, but very seldom, and their late origin is so evident that they
cannot be mistaken. The following is a list of this last class of MSS. :-
0 I. .A. complete copy of the commentary in the National Library at Paris.
It is advantageously distinguished from the rest, in so far as some very com~ider­
able omissions common to all the c MSS. have been supplied in I from a
another MS. Yet there can be no doubt that, with these exceptions, all the
rest of this MS. descends from the same original source as the other C MSS.
There is, for instance, a long omission at the erid of the fourtll. AdhyAya of the
first Ashtaka :· all the C MSS. break off in the third verse of the twenty-:6£th
Varga(p. 534[now 291] of my edition), with the words~' so that twenty
pages are altogether wanting. It is difficult to account for this omission, and I
suppose this loss to have happened very early, because in .A. and B also, where the
commentary goes on to the end of the fourth .A.dhy!ya, there is a peculiarity in
the style of the commentator not quite in accordance with the rest of his work.
That this omission has been supplied in C 1 from a different ·MS. is evident,
and c~ be traced even in the smallest particulars. Thus, for instance, through-
.out t~e whole of this supplement the merely grammatical part of the com-
 FIRST VOLUME OF THE FIBST EDITION. XXl

mentary is always divided by 11 lt from the rest, while in all the rest of this MS.
the beginning of the grammatical explanations is not marked at all.
C t b. A second MS. of the National Library at Paris, comprehending the
first Ashtaka only, and very negligently written.
C 2. The next MS. of this class is a copy which Dr. Mill brought over from
India, and which he kindly lent me for my edition. It will hereafter be
deposited in the Bodleian Library at Oxford, for which the whole collection of
Dr. Mill's Sanskrit MSS. has lately been purchased. It contains all the eight
.Ashtakas complete, with the usual omissions of the C MSS., and is written very
carefully and distinctly.
C 3. One volume, containing the first A.shtaka only. It belonged formerly
to the Sanskrit College at Calcutta, afterwards to Professor Wilson, and is now
deposited in the Bodleian Library.
C 4. A. complete copy of SAyana's commentary, forming Nos. 78-86 in Pro-
fessor Wilson's collection 9f Sanskrit MSS. in the Bodleian Library. It is dated
Samvat 1890 = 1834 A. D.
C 5. .Another copy belonging to the same collection, and entered in the
Catalogue under Nos. 5 7-60. It comprehends the first, third, and fourth A.sh-
takas only ; the second being supplied by another MS., No. 74, which contains
six only out of the eight .A.dhyayas.
C 6. A Bengali MS., containing the text and commentary of the first two
AdhyA.yas of the first Ashtaka. This likewise forms part of Professor Wilson's
collection, and is entered in the Catalogue of the Bodleian Library under No. 75.
That all these MSS. must be considered as separated from the MS. of
SAyana himself by at least one degree, I conclude from the existence of such
mistakes as are common to all the three classes of our MSS. I do not mean to
say that S~yana may not himself have committed mistakes in writing his com-
mentary. On the contrary, there are mistakes in all the MSS. which most
probably rest upon SA.yana's own responsibility. For instance, Rv. I. u7, 4, the
grammatical explanation of 1'\: evidently contains a mistake: yet all the MSS.
quote the same Sfitra of PA.nini (III. 3, 126), and there can be little doubt that
S!yana himself is the author of this wrong quotation 1• If mistakes of this
kind occur only in one class of MSS., or in two, but not in all three at the same
time: it must frequently remain uncertain whether they are to be laid to the
charge of S!yana or his transcribers. For instance, Rv. I. 97, 3, .A. has 4i1ii~Clfi41
'$~q411-~:tfialfll ~' 1'ftlftr ~ ~1(: I fit:;q h(i'~p(i=tt\ii I GfS'ift·
"''ti
1 Instead of
S&yan• ought to have wi·itten ~ ~ I iJ(tfht4'ft
""' a1 Mil I (cil qq~ I (QifiUfi fit@(E!i A
xxii PREFACE TO THE
~~~:, while B and C read ~~ttiGil~lfct(f;(~l\:rA~ct:. Sayana
might hav: written both, but I have retained the reading of A, as, in cases like
this, A generally represents the more authentic reading, B and C being more
liable to corrections. Rv. I. ro2, 3, however, I have adopted the reading of B
and C ~ ~~: instead of ~'Ui4i~: (A 2), because, as A I coincides here
with B and C, the reading of A 2 can only be considered as resting upon the
authority of the transcriber of A 2, and not upon the collective authority of A.
Sometimes old mistakes have been corrected in the more modern MSS. Hv. I.
66, 6, A and C read ~ ~~;rm ~: ~:, while in B the grammatical
fault, ~it,, has been corrected into 1l1J· Though -.~ may have originated
with Sayana himself I have of course adopted the reading of B.
I shall now quote, however, some passages where mistakes common to all
the MSS. cannot be ascribed to the author, but must have crept into the MSS.
before any of our present copies were written. There are evident traces of
corruption in the text of the commentary in explaining the grammatical forma-
tion of~' Rv. I. no, 2. All the MSS. omit (t'f\t.fil(.,.~, but have yet the '1'
after ~:, which leaves no doubt that {(OC6T<:@i'flf\" must have preceded it.
Again, Rv. I. I I 5, 5, where ~ is explained, the MSS. have only ~'
which explains merely the first part of ~' and necessarily requires the
addition of either Sf!° or ~lll, or some similar word, to explain the second
part. This, however, is omitted1 in all the MSS., and I was obliged to supply
it by conjecture. There is one passage towards the end' of the first Ashtaka
(Rv. I. I 20, 7) where the omission of several letters is marked in the .A. as
well as in the C MSS., and where the B MSS. also, though they do not mark
the omission, are of no use for restoring the text. In this case I was unable
to fill out this omission, and I have marked it in the same way as the MSS. do 1•
Sometimes old omissions have been supplied by the transcribers, but not always
successfully. Rv. I. 99, I, all the MSS. in explaining 'Ill (Uft'ftt: read t:ftt 1{({
'di(l'fl", thus making ~~: an accusative plural, while it ought to be the
genitive singular, and therefore ~ 'ft 'dC(r'fltcP. That 1(1! is indeed nothing
but a conjecture of the copyists, becomes clear when we see that A 2 has
Utt - - 'tC(l"tC\it, thus marking an old omission in the original MS. from which
it was copied, without any attempt to correct it. Such blanks occur most
frequently in A. 2, and in some cases evidently because the MS. from which

'l'h~re is probably also a corruption in. the words jecture Cftl'tt\ti, but this must remain uncertain tlll
1

1'if '1 I'If which precede the lacuna. ,One might con- new MSS oa.n be procured.
2
[Now found to be the reading of T.]
 FIRST VOLUME OF THE FIRST EDITION.
..
X.Xlll

it was copied was worn off at the margin, so that the blanks return always at
regular intervals, that is to say, always at the end of a line of the original MS.
Yet although in many other respects too, the A MSS. exhibit the best authenti-
cated reading, yet it is impossible to consider either B or C as descending from
A, on account of the omissions, additions, and mistakes which are peculiar to
each of the three classes of MSS., and have never found their way from one
class into another.
What I had to do therefore as an editor was first to find, by a collation
of the different copies of each class of MSS., the reading of each of the three
principal classes, and afterwards to choose that reading which, by weighing
the authority of the three classes, and by taking into account the whole style
of S~yana, seemed to be the most authentic. Considering, however, that this
edition of the commentary is not only a critical work, but at the same time
destined to be useful for studying the Veda, I have never carried these critical
principles so far as to leave a corruptiori in the text, which, though it might
rest upon the authority of the best MSS., was still so evident, that anybody,
if acquainted with the rules of the Sanskrit language, would have seen it, and,
if conversant with the style of Sayana, would have safely corrected it. I have
even added some few passages, which, though they belonged only to one class
of MSS., B or C 1, yet seemed to be useful where they stood. So that I may
safely assert, that whatever good was to be found in the MSS. will be found
in this edition, while much that was faulty in them has been corrected.
The laws of Sandhi and other euphonic laws I have endeavoured to observe
in the same way as they have been practically carried out in the best Sanskrit
MSS., considering it necessary, in a work like that of Sliyana, to avoid the inno-
vations of European, as well as the antiquated subtleties of Indian grammarians.
I have also followed the custom of the MSS., which sometimes suspend very

1 Rv. r. I, I
(p. 44 [now 23]), the quotation 1'1'T because he followed drft:erent principles of criticism in
~ ~ etc., on the change of \I into ao' belongs his ed1t1on. Agam, Rv. I :z3, I, the explanation of
only to the B MSS. [hut see now Var. Leet. top. 23, ihfr: occurs only in the C MSS., and it is evident,
I. 39], yet I ha.ve not suppressed it, as it seemed from the quotation of the Manoramit, that this passage
to be useful. Dr. E Roer, who had begun an edition could not come from the hands of Sayana yet I
of SA.yana's commentary in Calcutta, (the first two thought it necessary not to suppress it on account of
Lectures of the Samhitt of tbe Rig-veda, in the Bib- the accent. If 'Cftif:were formed after Un. S. II. 29,
liotheca Ind1ca, fascioulus 1-4: Calcutta, 1848,) but we should expect it to be a pa1oxytone; but it is
who, on.hearing of my edition, has kindly given up his formed by ~' and not by "'-' m the same way as
own plan, and published instea.d his excellent ed1t1on ...-: and 1'1:!, wluch a.re also oxytones. Dr. Roer
of the Vnhadiranyaka, has left out this passage, either gives the explanation of cftln:, but with some slight
because B MSS. were not procutable at Calcutta, or .difl'erences, which must be peculia.r to his own MSS.
XXlV PREFACE TO THE
properly the laws of Sandhi in ora.er to avoid certain combinations ~f words, by
which either single words or the structure of whole sentences might become
obscure and doubtful. In this manner the Sandhi becomes for the Sanskrit
what punctuation is for other languages, only it is as difficult to lay down
general laws for the one as for the other.
I have now only to mention those works which I made use of for verifying
the quotations in S!yana's commentary. There is first of all PU.nini, whose
grammatical rules are most frequently quoted by Sil.yana, sometimes at full
length, sometimes only with a few words by way of reference 1• I have derived
great advantage for verifying and understanding these technical rules from
Professor Bohtlingk's edition of Panini, which, whatever may be said n,gainst
some parts of it, is a most excellent and meritorious work. In the quotn.tion of
rules I have seldom had occasion to differ from his edition, and where I have
done so, it has only been after mature consideration. In the quotation of the
Varttikas also I have followed Professor Bohtlingk's edition, though it is to be
regretted that he has left out many of them. These, however, could eaHily bP
found in the Calcutta edition of Pltnini, though for some of them I was obliged
to have recourse to the MahabMshya. In order to make this edition more
useful, I have been induced to add the references from Panini in the first
AdhyAyas, but afterwards I have done so only whenever a new rule was quoted
for the first time. Professor Bohtlingk (now Counsellor of State to His Imrerial
Majesty the Emperor of Russia) could not render a more valuable service, to
Sanskrit philology, than by publishing a second and complete edition of Pilnini
and his commentaries, a work for which he must possess at present the most
ample materials.
1
I must mention he1·e one expression of Sft.ya.na.'s, 1iiftftCfi and 'Cit !f"'Efi ca.11 thercforo, as far as I can
which occurs very frequently, but bas given rise to see, ha.ve no other meaning than 'occurring in tho
doubts, ancl is, as it seems, not yet understood rightly shth and occurring in the ejghth Ashta.lm:' yet I fllltl
There ai·e two i ules of PAn.mi,s, coJJs1sting of the words that, for some reason or othe1·, Dr. Wuber (in bis
~~ ~; the one (VI. r, 198) teaches that a edition of the Vdgasaneyi-samhitft, p. 7. l. 6) quotes
vocat1V& case has the accent on the first syllable; the Pan. VI. t, r98 as the .reference to '!llta:i~~· 'f:1',
other (Vlll. I, 19) restricts this rule, by saying that where the effect of this rule is distinctly said to be
if a. vocative be.pr~ceded byano~her word, and do not - 'iili!fitl40' firEmr:, SO.yana uses also 'Ell!]~!
atand a.t the begmnmg of a PA.da., 1t has no accent at all. in the same sense as applied to a rule occurring m the
In order to distinguish .between these two rules, S!yana fourth Ashtaka 0 ; PAnini: cf. Rv. I. s4, r 4, where the
ca.Us the 8.<lcent prescmb~d by the former 1 ule, which rule intended by the commentator is iu fa.ct to be
occurs in the ~ixth Ashtaka., 1'tftEfii4 C'§'4Y'ff \Ef; while found in the fourth Ashtaka of Paninit IV. .2, 86; so
the suppres~ion of the accent1 as Pl'escr1bed by the that I have no doubt that Dr.Weber's quotation is to
n
latter{irltle ln the eighth Ashtak:a., is called by him be consi'dered as a mispnn
1ttl .. .._ ~: or "ll!f1tcfi '~l~i(t'tt~ ;-
• • t.
 FIRST VOLUME OF THE FIRST EDITION. xxv

Two other collections of grammatical Sfttras which are quoted by Siyana

are the Unidi-sl1tras and the Phit-s-0.tras of Santanakll.rya. Both of them form
part of the Siddhanta-kaumudi, as published at Calcutta, I 8 I I, but they have
been edited with much less care than Panini's Slitras. They have been reprinted
in the Memoires de l'A.cademie Imperiale des Sciences de St Petersbourg, 1843
and 1844, by Professor Bohtlingk, but require, particularly the Un!di-sfttras, a
careful collation of MSS. and the help of commentaries. I have quoted the
Sutras after Professor Bohtlingk's text, as being more accessible than the
edition of the Siddh!nta-kaumudi; but I have been continually obliged to have
recourse to the MSS. and commentaries of the Un!di-sfitras 1•
A fourth grammatical work quoted by 81\yana is the Dh!tup~tha. Of this
work we have a most excellent edition by Professor W estergaa.t:d of Copenhagen,
at the end of his Radices Linguae Sanscritae. I have quoted it only a few
times, as it is very easy to find S!yana's quotations with the help of Professor
W estergaard's Radices. Sayana has himself written a commentary on the
Dhatupatha, before he wrote his commentary on the Veda, and has frequently
readings peculiar to himself, which he has defended in his commentary 2, and
which Professor Westergaard also has generally menti0ned in his edition.
Another work frequently used by Sayana for explaining the Veda is Y!ska's
Ni.rukta. This work existed only in manuscript when I began to print S!yana's
commentary, and as the greater part of the Nirukta is contained in S!yana's
works, I was obliged to copy and analyse it, in order to verify SAyanai's .quo-
tations. For though, with the help of the Sarv!nukrama, all the passages from
the Veda which are explained by YAska may be traced back to their places in
the text by referring to the commentary on the Nirukta, where the Devat! and
Rishi of each passage are given, yet it is vecy difficult, vice versa, to find always
the place in the Nirukta where a passage of the Veda has been explained by
Y!ska; still more so when only a few words out of Y!ska's explanations are
quoted by Styana. In the course of carrying this first volume through the
press, a very correct edition of the Nirukta has been published by my learned
1 The MS. from which I have derived the greatest mistaken. I merely men.tion this here to point out
use IS the Unld1vritti, by Uggvaladatta, a work which how \lnsafe it would be to. make use of our present
has been composed after a careful collation of old MSS. editions for lexicogra.pluo purposes; but I shall soon
and commentaries. It frequently points out words have an opportunity of returning to this subject, when
and Siltras as being of later origm, and as not occ11.l'- examining the b1stor1ca.l value of this and other works
ring in old commentaries. In our printed editions previous to PA.-nini.
2 Sayana. quotes his Dhituvritti, Rv. I. 42, 7;
some SO.tra.s are left out, others mixed with the com-
mentary ; some are incomplete, others incorrect; and I. 51, 8; I. 82, I, etc.
the meaning and formation of words a.re frequentl)'
4 VOL. I. d
XXVI PREFACE TO THE
friend Professor Roth in Germany. Professor Roth had kindly informed me
beforehand which of the two recensions of the Nirukta he would follow in his
edition, and I am glad to find that consequently the references which I have
always given, when the Nirukta is quoted by Sayana, coincide with his edition.
In some few places Sayana's quotations from Yaska do not exactly correspond
with the text of the Nirukta; but this is probably owing to Sayana's manner
of quoting, which, as I have mentioned before, is generally done from memory.
Although these diflerences were very slight, yet I could not, in accordance with
the principles of my edition, take it upon myself to correct them. I have not
added references to Sayana's quotations from the Nighantus, because these listA
of Vaidik words are already arranged systematically under different heads, and
thus require no further reference.
The same applies to Sayana's quotations from KA.tyayana's Sarvanukrama.
I have myself compared every passage quoted from this Index of the authors,
<leit~es, metres, etc. of each hymn. But as this Index follows exactly the same
order as the hymns of the Rig-veda, it would have been useless to add the
references. In those cases also where Sayana quotes from the Sarvanukrama
certain rules on metre and other subjects contained in the ParibhasM., I have
abstained from giving the references, because this Introduction to Katyayana's
Sarvanukrama is likewise so well arranged, and so short, that it is as easy to
find a reference as to find the quotation itsel£
Another author whom Sayana quotes most frequently with regard to the
Vaidik ceremonial is Asvalayana. There are twelve books of Srauta-st\tras,
and four books of Grihya-siltras, none of them as yet published. S!yan.a quotes
these Stl.tras continually, whenever a hymn or part of a hymn of the Rig-veda
occurs which is to be employed by the Hotri-priests at a certain act of a sacri..
:fice. Now if, like the Sfitras to the Yagur-veda, the Sfttras of .A.svalayana
followed the same order as the hymns, it would not have been difficult to find
S!yana's quotations in the MSS. of AsvaJ.ayana's Sfttras, and it would scarcely
have been necessary to give a reference to 01:\.Ch of Si\yana's quotations from
Asval~yana. But the Rig-veda has preserved its old arrangement and its
genuine form, and has not been supplanted by a Hotri-veda, or a prayer..book
for the Hotri-priests; such as the Yagur-veda is for the .A.dhvaryu-priests, and
the S~ma-veda for the Udga.t1"i-priests. If, like these two so-called· ceremonial
Vedas, the Rig-veda also consisted only of such passages as are requisite for
the Brahma~ic sacrifices, arranged in. the same order as they have to be recited
.l)y the Hotri-priests at difl't:'1-ent ceremonies, the order of the hymns and of the
 FIRST VOLUME OF THE FIRST EDITION. XXVll
..
Sutras, and probably also of the Brahmanas, would be the same. But, as it is,
the Rig-veda represents to us the old collection of sacred poetry, as it has been
handed down by tradition in different Vaidik families, each of which claimed a
certain number of ancient poets (Rishis) as their own. The poems therefore which
have been incorporated in the Rig-veda-sambiM. are arranged according to the
old families to which the poets of certain songs are said ·to have belonged,
and consequently those passages which in later times were selected as most
appropriate to be employed at the grand sacrifices by the Hot?i-priests, are
found scattered about in different parts of this old collection. Sayana, who of
course knew .A.sval:lyana's Sutras by heart, quotes these Sutras whenever one
of those verses occurs which Asvalayana has prescribed for any one of the dif-
ferent sacrifices. But all that Sayana adds, to enable one who has not learnt
by heart these sixteen book~ of ceremonial Sutras, to find their place in Asvala·
yana, consists in mentioning the name of the particular part of the ceremonial,
and sometimes in giving the beginning of the chapter where a certain Sfttra
occurs. By the help of Indices, however, I have succeeded in verifying these
passages also, and I have always added the book and chapter where Sayana's
quotations are to be found in Asvalayana's work. If, in the passages which
Sayana quotes from the Brahmanas, he had restricted himself to the Brahmanas
of the Rig-veda, I should have added references to these quotations also. But
as Sayana takes his quotations promiscuously from all the Brahmanas, whether
connected with the Rig-veda or the Sama-veda, Yagur-veda, and Atharva-veda,
I determined rather to give no references whatever for these Brahmana pas-
sages than to do it incompletely 1• Besides there was the difficulty that these
1 It 1s not only on account of the vastness of the ~ft itltllQJ4i~1: m sitl$&iiif<+i I
Eifft1rffl ~ 'tll'ft1f! ~II
BrAhma.na hterature that I found it impossible to
verify every quotation, but there a.re many Bi·ft.hmanas
of which there are not even MSS. to be procured in
....... ~ - ~ .
'1J11'11~ ~..,'H''""'"' ~q§)QMllC"fUt. I
4iftl1">tQMtt(tf\" 1f1rT iitlfitl'flf\Hi: ll
any of the European libraries. Some seem lost even
°"" "''1ecf1ttt@ltci4ff4tti~iilt\1iti1f\41i ~­
in India, and are only known by name, With regard
to the Brft.hmanas of the Sa.ma-veda, I had stated, in
a letter to my friend Professor Benfey at Gottingen,
~lftl11' ([Iii 1t I" i1 IQ~ I at Cifi I
~: I~ ~ ~ Sill,Qlci~
if'
I lfi lt t.U llfil-

that there are eight. Professor Benfey bas londly etc.

mentioned this in the Preface to his edition of the ' There a.re eight BrA.bmanas ; the Praudha 1s the
SA.ma-veda; and as Dr. Weber has lately published ftrst Bri.hman (this means the large :Brft.hma.na., the
some observations 'With regard to Professor Benfey's Panka'Vimsa, not prodha as Dr. W. writes); the one
and Mr. Oolebrooke's statements on the Brft.hmana called Shadvimsa. or Sha.clvimsad·Brft.hmana is the
literature of the S&ma.-veda, I owe it to Professor second ; then follows the Sbn1mdhi; then the Ar-
Benfey and to myself to make good my assertion. aheyt1o-BrAhma.nl!., the DevatadhyAya.-BrQ.hmana., and
Saya.na, m bis commentary on the S8maY1dMna- t'he. Upanishad. These with the Samhitopauisha.d and
Brlhmana, says : the Vamsa are called the eight books. In the great
d2
xxviii PREFACE TO THE

Brahmanas and Aranyakas, which as yet exist only m manuscript, are not
always divided in the same manner ; so that if I had adapted my refereuces to
the MSS., they might perhaps not have been found in accordance with the
editions of several of the Brahmanas which are now preparing for publication.
In many instances I have derived great help from the original MSS. of the Brah-
manas, particularly as Sayana's quotations from these works are generally full
of mistakes, arising from old V aidik. forms, which the transcribers did not
know and understand. Frequently, however, I found also that real differences
existed between a passage as quoted by Sayana and the text as exhibited in
the Br!hmanas, which can only be accounted for by the supposition that SAyana
used some BrAhmanas in a Sakha different from that which was accessible to me
in manuscript.
I have only to express, in conclusion, my sincere thanks for the instruc-
tion, the advice, the encouragement, and assistance which I have received, in
the course of my studies, from those distinguished Oriental scholars whose
lectures I have followed at the Universities of Leipzig, Berlin, and Paris, as
well as from those whom I met with there, and with whom I became con-
nected by the ties of kindred pursuits·and friendship. To mention the liberality
with which foreigners are admitted to the rich collections of the National
Library at Paris, the Library of the East-India House in London, and the
Bodleian Library at Oxford, would only be repeating what is known to all
who have had occasion to consult these Libraries. Yet this ought not to pre-
vent me from acknowledging the personal obligation under which I feel myself
towards M. Hase, M. Reinaud and M. Munk at the National Library at Paris,
and towards the Rev. Dr. Bandinel and the Rev. H. 0. Coxe at the Bodleian
Library at Oxford, for the kindness which I have received at their hands
during my studies at Paris and Oxford. Private collections also of Sanskrit
MSS. have been freely thrown open to me, in France by Professor Burnouf,
in England by the Rev. Dr. Mill ; and I avail myself of this opportunity to
return my thanks to both of these eminent Oriental scholars. I have aJ.sQ
thankfully ,to acknowledge the kind assistance of my learned :friend Dr. Ch. Rieu
at the British Museum, by whose careful corrections many misprints and mis-
Btahmana and the Shaclv1msa. the prm01paJ. s&erifices, dicate my assertion ; foi- nothing is more likely to
the Ekft.ha, Ahina, and Sattras, have been taught by bring Vaidik studies into discredit than mere auer-
which persons, fit for deiing sa.cnftces, may obtain twna and ingenious cou.jectures, of which I am sorry
life in heaven and other rewards. Now, m the third iio •a.y we ha.n had already a great number even in
Brft.hmana, called Slmavidbtna, other hymns will be print. [See now a letter of mine in the J.caclertly,
enjoined, etc.' Thia, I hope, will be suilicienHo vin- June 71 ;_c8.90, p. 390.)
 SECOND VOLUME OF THE FIRST EDITION. xxu:

takes have been removed, which, notwithstanding the great accuracy and ability
of the compositor employed on the present volume, could scarcely be avoided
in so extensive a work. .A.hove all, however, my thanks are due to Professor H. H.
Wilson. It would be presumptuous on my part were I to speak of his unequalled
achievements in different branches of Oriental philology. But it would be un·
grateful were I not to acknowledge the kindness with which he has assisted me
in my undertaking. To his recommendation I owe the liberal patronage which
the Honourable .the Court of Directors of the East.India Company have bestowed
upon this work, and without which its publication would scarcely have been
possible. While I was preparing this edition his books and manuscripts were at
my disposal; whenever I wanted advice, he was ready to give it; and he has
even given his valuable time to correct the press. The English translation of the
Rig-veda by Professor Wilson, which is soon to appear, will be a new proof of the
interest which he has taken in this work. To have been allowed to enjoy his
acquaintance, and to avail myself of his instruction, will always be to :me the
best compensation for what I have lost in living so 'long away from my own
country and my old friends.
M.M.
OXFORD, October 1849.

PREFACE
TO THE SECOND VOLUME OF THE FIRST EDITION.

THE principles followed in the :first volume in editing the hymns of the
Rig-veda, and in restoring the text of S!yana's commentary, have been strictly
adhered to in the second. My own conviction and the approval of those best
competent to judge have not allowed me to adopt a different course, although
others have thought S!yana's commentary undeserving a critical edition, and
the time that has been expended on it. Yet, perhaps, neither the age nor the
country of an author can justly be pleaded as an excuse for disregarding those
rules of critical scholarship which in classical philology have been ratified by
the experience of the last three centuries; and though I feel obliged to apologize
for the consequent delay in the publication of this second volume, I should be
more sorry if I had now to publish it without the conviction that all sources
available for the restoration of the text had been fully e•usted.
:xxx PREFACE TO THE
The MSS. of the Samhit~ and Pada texts of the hymns have been the same
for the second as for the :first volume, and therefore require no further notice.
The case differs with the MSS. of S~yana's commentary. For the :first
Ashtaka I was able almost always to determine the original reading of each
of the three families of MSS. by a comparison of the different memlJers belong-
ing to each ; and it was found that in all important passages the readings of the
three families, A, B, and C, could be so easily balanced that it rarely remained
doubtful which of the three had to give way before the others. On this account
it seemed to me unnecessary to publish the Varietas Lectioni.'l for the first
Ashtaka; and, with the exception of a few inaccuracies which really seem to be
unavoidable in the course of editing so voluminous a work, I have not met, in
the many excellent reviews of this edition of the Rig-veda, with any remarks
which could on this point have changed my opinion.
In the second Ashtaka, however, this course could be no longer followed.
When I began the printing of the second Ashtaka, to represent the A class
I had only one MS. instead of three, and this again was defective to such an
extent that it could hardly be used for more than half of the second book. The
MSS. of the B class were the same as for the first Ashtaka ; and I had even
received a new copy from Bombay, which I call B 3. This, however, proved of
little use, as it was but a new transcript of the same original from which Pro-
fessor Burnouf's MS. had been taken. It could not have been copied from B 1,
because this MS. was already in my hands when the new copy was made in
~ndia. There can be no doubt, however, that it was copied from a copy of B 1.
Most likely the Pandit who parted with his old MS. (B i) had a transcript
made before he sold the original, and this transcript was copied in B 3. With
regard to the C MSS. it will be seen that three only out of the seve11 contain
the second Ashtaka complete.
Under these circumstances it became impossible to determine with certainty
the original reading of each of the three families; nay, for many passages the
text of one or even two families was lost beyond recovery. Besides, almost
all the MSS. of the second and the following Ashtakas seem to have been less
frequently studied in India, and therefore less carefully corrected, than the
first Ashtaka. The principal difficulty, however, arose from the defective
state of A 2, because the text of the A class is throughout the whole com-
mentary by far the most authentic, and had been the principal authority for
my edition. Before the first A.shtaka was :finished I had therefore written to
India to obtain a new MS. for the second. After waiting, however, for
 SECOND VOLUME OF THE FIRST EDITION. xxxi
a considerable time, I received the news that the MS. which Dr. Roer had
procured for me at Calcutta had been lost by shipwreck. Not wishing to retard
the publication of the second volume still longer, I began to print with the MSS.
then at hand. Most fortunately, in the course of carrying the second Ashtaka
through the press, a new MS. was forwarded from India by Dr. Ballantyne,
the distinguished Principal of the Sanskrit College at Benares. It was sent to
Professor Wilson, and he most kindly presented it to me. It is a complete copy of
Sayana's commentary, and, as far as I know, the most ancient MS. of this author
in Europe. The first Ashtaka contains 648 leaves, but has no date. The second
has 260, the third 398, the fourth 312 leaves. At the end of this Ashtaka the
following date is given: ll ~ q~ it~ 11 'I it' q ~ ~ ~g lf( ll This is Samvat
1624 or A.. D. 1568. The £fth Ashtaka has 237, the sixth 264 leaves, the last
containing again a date; ~ q~~~ 'Cfif ih!iil' ~ {cUill@I 9_iCiil'1l' ~~:I This
is therefore a year earlier, 1623 Samvat or 1567 A. n. In the seventh Ashtaka,
which contains 274, and in the eighth, which contains 322 leaves, the last leaves
on which the date is usually marked, are lost. We may therefore ascribe this
copy to the years l 566 and 1567. The oldest MS. which I had used before was
that of the first Ashtaka at the National Library in Paris (A 1), which bears the
date Samvat 1625 or 1569 .A. n. Colebrooke's MS. (A 2) was dated between
I 74 7 and 1 760 A.. D.; and fragments of the commentary to the :first Ashtaka at
the Royal Library in Berlin (A 3) exhibit their date as Samvat 1664 and 1665,
which is 1608-1609 A. n. As all the other MSS. of S~yana's commentary are
still more modem, this last arrived MS., now in my possession. is therefore the
oldest in Europe.
But though it is the oldest MS., it does not, as I expected, belong to the A
class. On comparing it with the other MSS. I found that it belongs to class C,
and may in fact be considered as the prototype of all the C MSS. This would
by itself have made this copy extremely valuable. Its value,. however, was
further increased when I found that in the second Ashtaka considerable portions
had been most carefully compared and corrected after a MS. belonging to the A
class. If it had not been for the modern date of Colebrooke's copy, I should
have said that the collation had been taken from this very MS. (.A. 2). .As it is,
we must suppose that the Benares MS. was collated with an older .A. MS.,
whence afterwards Colebrooke's copy was taken. As these collations and cor-
rections, which are very numerous, were made in the usual Indian manner, by
covering the original text with yellow orpiment and writing the various reading
over it, or on the margin, it was possible in many cases to use this MS. as an
x:xxii PREFACE TO THE
authority not only for the C but also for the A class; a discovery which makes
this MS. invaluable, particularly for the second Ashtaka, where the text of
the A MSS. was frequently lost altogether. As to the C MSS. they are all,
though not immediately, derived from this copy, but before it had been corrected
by a co1lation with an A. MS This can be proved from accidental mistakes
and indistinct corrections which occur in this MS., and have afterwards found
their way into all the C MSS. I suppose therefore that this (my own MS.
from Benares) had been repeatedly transcribed there before it was collated
with an A MS. i that one of these transcripts was brought to Calcutta, and
that from it most, if not all, of the C MSS. were taken at Calcutta. Certain it is
that C 1, C 2, and C 3 came from Calcutta, and large portions, which were copied
for me from a MS. belonging to the Asiatic Society in Calcutta, contain literally
the same text as the other C MSS. There are readings, however, which are
peculiar to the C copies, and which do not occur in my Benares MS. Some
are found in one or two of the C copies only. These are either accidental mis-
takes .or occasional conjectures, which, after finding their way into one copy,
were afterwards repeated by successive transcribers. They are useful as indi-
cating the exact relationship of some of the C MSS., but they can never claim
any diplomatic authority. There is another class of various readings which
run through all the C MSS., but which have no guarantee in the old Benares C.
They deserve a more careful consideration, for they are evidently old errors, or
possibly old corrections. In passages where the C's agree among themselves,
but differ from the Benares C, this generally agrees with .A.. These C readings
therefore, as they do not proceed from the Benarea 0, are in reality only later
alterations of the original A text. Though I have hardly ever adopted the read-
ings 0£ the C's, where they differ from the A's and the Benares 0, I have
inserted them in the Varietas Lectionis, because occasionally they seem to con-
tain useful conjectures. Thus Rv. II. 6, 4 (p. 445. 1. 10 [now vol. ii. p. 19. 1. 14])
Ca has~ lff; the same mistake occurs in A, while all the C's have the
correct reading~ 11t. It also happens, as has been shown in the preface
to the first volume, that passages originally wanting in the 0 family have
been supplied from other sources. In this manner C and .A. readings have
been mixed up together in one and the same copy; particularly in Dr. Mill's,
so as to cause considerable confusion in some parts of Sayana's work. In later
portions of the commentary some of the C's cease to be C's, and become A's,
owing to the writer having changed the original from which he copied.
~ By this new MS., which in its original text I count with the C, in its
 SECOND VOLUME OF THE FIRST EDITION. XXXlll

corrections with the A. MSS., while if it has to be quoted by itself, it will be

marked Ca, the apparatus criticus became much simplified. The various read-
ings of the C MSS. have no longer any authority by themselves, but only so
far as they agree with the original C before it was collated. The text, of the
A MSS. is supplied by Ca, where the A's are wanting, and confirmed where
they and the corrections in Ca agree.
It might be asked, why, if the A MSS. agree with the corrected text of Ca,
the whole family of the A's should not rather be considered as branching
off from Ca, after the general text of the C family was corrected as we find it in
Ca, and why the alterations in Ca should not be taken for independent correc-
tions made by the writer of Ca, instead of being gleaned from a collation with an
A MS. ~ On all important points the A MSS. agree with the con-ected Ca; and
on minor points it is impossible to deny that slight differences could as well be
ascribed to the several copyists of the A MSS. There are passages, however,
which make it impossible to entertain this supposition : Rv. II. 9, 3, for instance,
ca(\fV~ in Ca has been corrected into 'dC(lf(f.1. Here the C MSS. have :+J(lf\tq,
because they were copied from Ca before it was corrected. .But A also has 'iC(1 f~'4.
This by itself might be a mistake of the copyist of A ; but this peculiar mistake
is not likely to have happened if A had been copied from Ca, where this very
mistake had already been very distinctly emendated. I have selected this
passage for another purpose also. It shows that occasionally corrections are
made in Ca which cannot be traced back to A.. They generally refer to such
palpable errors, as for instance·~ instead of 'di(lf('4; and in these cases it is
clear that the intelligent writer of Ca was guided by his own judgment without
looking to the authority of any MS. In order to prove that A was copied from
Ca, it would be necessary to point out passages where corrections in Ca had been
misread by the writer of .A.. This happens not unfrequently with the C MSS.
Alterations made by the original writer of Ca were mistaken by the later
copyists, and these mistakes are occasionally of such a charact'er that they could
only have happened with the very ~eaf of Ca before the eyes of the copyist.
Now I have not been able to find any evidence of this kind against the A MSS.
Quite at the beginning of the second A.shtaka, for instance (p. I. 1. 9 [now vol. i.
p. 551. l 7]), Ca reads •fic4ltf4•iW(~fd l(f°~ittlllf41 1 The writer seeing that be
had left out ft: between' ftl, made a mark at the top of 1', and wrote 'R: on. the
margin. The copyist of the first C MS. mistook this mark for a vowel, and
without look:iug to the margin wrote o~. .A. however has 0 t1' f\fit. The same
happened Rv. II. 6, 4, where Ca had 1:!1""("1\1~ This the original writer himself
5 VOL.-I. e
XXXlV PREFACE TO THE

corrected into ,..n::~«ll(J only that the space did not allow him to do more than
to make a mark at the top of 'e{, and to draw a slight line between ' and C{.
'fhe copyist of C mistook this for t_f\q«1..._, which henceforth became the reading
of the C MSS. Indeed it looks very much like !f<1fg;1,in Ca. A however reads
'{f<:fb\tl(. Rv. I. r8r, 8 (p. 366. 1. 24 [now vol. i. p. 765. l. 18]) Ca had originally
~: Jh(iiffctitit 'efT Accordingly all the C's read ~:. A however has ~~:,
and from it the correction is made in Ca, where, subsequent to the taking of
the first C copy, the' was covered with yellow, and changed into~· ~:is
evidently an old mistake, because it occurs in B also. Rv. III. l, 3 (p. 626. 1. 14
[now vol. ii. p. 127. 1. 28]) the C MSS. ha~e ~ otiif ~m, thus leaving out
two lines. The omission was originally made by the writer of Ca, but afterwards
supplied on the margin from A. That it was not supplied from B, but from A,
can be proved by the words '@'if 1(1!1' ~ ~' which are not in B, but in. A
only. At the beginning of the third Ashtaka (Rv. III. 7, 1) the invocation in
Ca is simply 'illatiilill"I 'l'f: 11 if ~ etc. It is the same in the C's. .A. 2 has
:1flatfflttlll if1l: ll l!tl\t\@fa 'ifil: H'!ft~ 1{i{! II II ~ etc. The B's have ~1ff1.T
if1f: ~~ ifil: ll ~; only B (Taylor) has a slight correction 3!f\a1iQt(l"I etc.
Coincidences of this kind though they may seem irrelevant are really the most·
convincing. Again, Rv. III. 7, l (p. 661.1. 7 [now vol. ii. p. r 49. 1. s]) A 2 has 'ff 1f
'I' '9Aflt:, while Ca and the C's have the right reading 'ff":!f 1f 'I' ~:. In the same
verse (p. 662.1. 3 [now p. 149.1. 18]) A 2 has fttfo9tnl~4l; Ca and the C's coincide
in giving f'J:M~l111 I 9Htii(\tfl. Rv. III. 7, 2 (p. 663. 1. I l [now vol. ii. p. I 50. 1. 14])
Ca has f~<fttll'l'li instead of the more primitive reading (\di-Cfif4ifiiht41, which
is in A 2 and B. The C's have again (\m"lttti like Ca. As no case of this
kind can be made out against A, it is necessary to regard the corrections
in Ca as collations from a MS. which was connected with the A class. How
A and Ca came originally to differ, is hard to say. But as it is impossible
to account for the text of the one simply as a corruption of the other, A and Ca
must be considered as two branches diverging from the original stock of S!yana's
work which is lost to us. , .A. and Ca are coordinate, not subordinate, though A
stands high~r than Ca. B also stands independent from .A. and Ca. The C
MSS. on the contrary are subordinate to -Ca, or, as in later portions of the
commentary, to .A., and must therefore be discarded where the evidence for a
reading has to be. weighed by the authority of MSS. Even in cases where the
C MSS., particularly C 2 and C 5, contain large portions of the commentary
whieh are taken from an A. MS., they are of very little use for restoring the
original text of A., except occasionally in the second Ashtak.a where A. 2 is
 SECOND VOLUME OF THE FIRST EDITION. xxxv

defective. Here C 5 (Wilson 74) has sometimes supplied the deficiency, but
the modern date and general incorrectness of this MS. hardly allow us even
then to ascribe much weight to its readings.
If we compare the three MSS., A, Ca, and the C's at the beginning of the
second Ashtaka, it will be seen that they always confirm our view of their
mutual relationship even in the most minute points. In 1i'iflriltl'f: (p. r. 1. 2 [ no"\v
vol. i. p. 55 I 1. 3]) Ca had left out the Visarga, and it is wanting in the O's. It
was afterwards added in Ca, above the line, i!iiU!i~l:, and it stands above the line
in A also. P. I. 1. 6 [now p. 55I. I. s] Ca had~ instead of '1ft. ~"ft became
the reading of the C's, while in Ca it was altered into itf,n, which is the reading
of A. P. I.1. 7 [now p. 55r. l. 6] Ca had 1fffiiliflit, which was corrected by the
writer himself (not secunda manu or with yellow orpiment) into ttaf.,fittill'"·
This correction was therefore adopted by the C's, and it is also the reading of A.
P. I. 1. 8 [now p. SSL 1. 7] Ca had 0~; one of the vowels was afterwards
covered with yellow. The C MSS. therefore read ilttC!f" ; A. has ilttct tli. P. I.
1. 9 [now p. 551. 1. 7] we read in Ca ciih(tl:\f. The C MSS. adopt this as ti~i(tif,
while in Ca it is corrected into ci'Ef<"lif as it stands in .A.. P. I. 1. Io [now p. 551.
1. 8] ~in Ca was left out, and added on the margin. It is wanting in the C's.
P. I. 1. I 1 [now p. S5 I. 1. 8] Ca had Jr lf1' ~· Here the copyist of the :first
C MS. must have corrected the mistake, which was indeed very palpable, for the
C's read~' which is also in Ca, but as a correction. P. 1. I. I 1 [now p. 551.
1. 9] Ca had-.. M'EfT Rifirc1:. Thus the C's have 'Qr M'EfT 1 ~:I Ca was
afterwards corrected into f?(itlll"'filq1an ~:, but in such a manner that none
but a learned copyist could have made out the correction. A however has the
right reading. P. 1.1. 17 [now p. 551.1. 14] Ca had Qlntiliti4i, which became the
reading of the C's. In Ca it was altered into qf@i'*fl1wi, by put~ing q over 11,
and ~ over w, and covering part of 1' with yellow. A has distinctly Ql41•nftlli.
Again (1. 18 [now 15]) aft~r ~' 1{C{ is left out in Ca, and added from A in the
margin. It is wanting in the C's. .Afterwards Ca had originally "'tJ'Ld ~
~~ lfi S(iiiiif~ t This was, not at all distinctly, corrected into~ ll'f ?11!1
j(lctM, which is the reading of A. But the old mistake remained in the C's.
After oMlltt41~ (1. 19 [now 15]), ~ is left out in Ca and C; it was added in
Ca from A. P. 2. 1. I (now p. 55x. 1. 16], QWn'J'..1(1. Ca and C; 'Q'11T 1ft'l. Ca
sec. man. and A. P. 2. 1. 3 [now p. 551. I. 17], '1ft ~ Ca and C; ~ ff(1ft Ca
sec. man. and A. P. 2. 1. 5 [now p. 55r. l. 19], tijfqtgf'tail~tir Ca. The i in fihit
is covered with yellow in Ca, so that one should expect fitit, in the C's, instead
of which they have the correct reading itit. On looking more closely, however,
e2
XXXVl PREFACE TO THE
we find that the i, before it was covered with yellow, was already struck out
with ink p1"ima manu, which accounts for the adoption of the con,ected reading
in the C's. P. 2. I. 5 [now p. SSI. I. 19], ~Ca and C; ~Ca sec. man.
and A. P. 2. I. 6 [now p. 55 1. I. I 9], lfcil: Ca and C ; U'ff: Ca sec. man. and A.
Ibidem [now I 20], f~ ~ 'I! Ca; firo'(t ~ ~: C. The text of Ca is
corrected into fifcfit"1 'flc111f~ail 'I:, but in such a confused way, that the writer
of A could hardly have got his correct text~~ ~~ ": from Ca. P. 2.
I. 8 [now p. 551. 1. 21], f~ ~ Ca and C; ~: ~~ Ca sec. man. and A.
P. 2. 1. 9 [now p. 55I. I. 22], ti\RP'•i~ Ca and C; 4/l'il:sc:tti~ Ca and A; both
sec. man. P. 2. I. Io [now p. 55 1. I. 2 2 ], ~: I fcfi: Ca. C ; ~: 'fcfi: I Ca sec man. and
A. P. 2. I. IO [now p. 551.1. 2 2 ], ~'"\:llrt I ~: Ca. C; el f@'\411 hfr~T=tr: Ca sec. man.
and A. P. 2. 1. 1 o [now p. 55I • 1 2 3], '1(1 iitcf Ca. C ; 'ai( t"ih.c:i Ca sec. man. and A.
These are the various readings for one verse only. I have given them in
full, partly to show that they confirm my view of the relation of the MSS. in the
minutest detail, partly to convince those, who are so eager for a complete Vari-
etas Lectionis, of the entire uselessness of such a compilation.
There is one more MS. of SA.yana (Aa) which has come to hand since the publica-
tion of the first volume, and which, though rarely quoted in the Varietas Lectionis,
deserves to be mentioned here. It was sent to me as a present by Mr. Fitz-
Edward Hall of Benares, to whose zealous exertion~ Sanskrit literature is already
greatly indebted. Unfortunately it is but a fragment, beginning with the twenty-
seventh Varga of the third Adhy:\ya of the third Ashtaka, and ending at the end
of the third Ashtaka. But it is of great importance in so far as it, served to
confirm those principles of diplomatic criticism on which all the other MSS. of
SA.yana had been previously classified, and according to which their respective
weight and authority had been determined. Its title is~ q~~g 'efit 11 II 111tiflCClff(
qq ~ fi«fiflt U~II II Cifi48tt41it1J II Ufi{q444Q \t~dlant' q(i!fflfil(ttt ~ l ~ ~
~ 1fT1ft °'~ ~ 1l"'fU ~: II II ~ ~ n This shows that its age is
exactly the same as that of Ca; Samvat 1624 or .A. D. I 568. Its readings
would therefore have had considerable weight whether they coincided with any
of the three classes or not. But on a close examination it was found that not a
single reading, not even a mistake, occurred in this MS. which was not known
already from the .A. MSS. Nay, I feel almost convinced that Colebrooke's .A. MS.
was copied directly from the fragment which is now in my possession. Rv. IV.
2,'13 (vol. iii. p. 18 [now vol. ii. p. 351]) the whole explanation of the thirteenth
verse had been left out in Aa, but was afterwards added so as to cover the
"whole margin, not without giving rise to many ambiguities by the indistinctness
 SECOND VOLUME OF THE FIRST EDITION. xxxvii

of the letters. All these ambiguous letters are reproduced with the greatest
exactness in Colebrooke's MS., where the explanation stands no longer in the
margin, but forms part of the commentary. .A. still clearer case occurs Rv. IV.
4, 4 (vol. iii. p. 35 [now vol ii. p. 360] ). Here also the commentary on verse 4
had been left out in Aa, and been supplied as far as the space allowed on
the margin. We have fust two long lines at the top of the page. .After this
the text is continued on the margin of the right side, then on the margin of the
left, and one line is placed under the last line of the text. I give as exact a
copy of it as possible:
~ ~lif II~ frfiffr I fcHf(fit fitatt 1~'(4Q I~~~~~: WT'f1(t * ~ (1'!lCiiii tif ~
ift ~I tl'{'11fff '9;(('~ I *11"114~ ~'11(¥1 I~:~: •fil'4lii 1fl.'l. ac'tiidl(( f'Rltje(t t ~ ~
~~ ~~
if I lf'.fl 33 flt VMr
*"" .. '!(it ~ 1': "Sir
ff~'( "'-'"~
tN ~ 1":lfli an
~'ifl~ ~r.r
~ '@l'lft' ~: ~
*Qilfl ""' f'lfllt ~
'ffnH• ii(il! I ;(11 ~
'\lA' I ~ '1\' 11 (I flttt
tt...n~ ~1'1'~
itftt 'i(1f\' ' 1' " ~ '
• fr v; (ffr • """1' 1'fn1
"" ~: ..m-: iii: :Q\fir ~ fimcf: " ii """ '(

This is reproduced in Colebrooke,s MS. (A 2) in the following manner:

~'""'
11~11 ~ fttlftr II Mittl'ft fd•ift4'11-~ ~ 11'11 'V ~: 'ftTlll l ~
~ 1::'tftti1•ttfi ~~I il\titfft '8' ~ .. 1414\l fit'dl('* ~:. .: ilf'"lllii 1l\'t
wO'lf\ti( fll<ri<i ~l ~ ~: ~ 1f '""'" 1'4""'K"ri ~ 1llW ~ ~ ~eumu
'S1il ~ ~ 'l4914ifdif <thi'i'i~il! ~I '(t ~ Vti ·i(~M 1t1fr '111'frW~1'i
~: 'fir: 1fi! ~ """ fimcf: II
"ftlllfft °(f"lttl'IN "f! '911T't.1"° 4114ii(IM ~ ~: •<tfitif 1(l1(1i ..- ~Ii
~ 1fm ~ \lf°;r ~CAT
If we compare the readings of the two MSS., and. remember that both came
from Benares, thete can hardly be a doubt that the latter (A 2) watt copied
directly from Aa. It is clear that the writer of Colebrooke's MS. was misled by
XXXVlll PREFACE TO THE
the marginal notes of Aa, that he left out all that was written on the right
margin, and went on from the end of the second line to the third line on the left
margin. After having proceeded to the end of the verse he became aware of his
mistake, and then added the whole omitted passage on the margin. This is
nearly convincing. But there are other points to show that A 2 is a direct
descendant of Aa. The ~ in ~ is very indistinct in Aa ; a blank is left in
A. 2. The lf in 5:11"1 l'liEi looks, according to Aa's style of writing, very much like
\l ; A. 2 has \l. In ~ 1 the Yid.ma under Cf is covered in Aa by the ft of the
:next line; the writer of A 2 did not perceive this, and thus altered <l:.l into ?fl.
In Aa the marked fil in the fourth line is meant as a correction of the marked
frf in the third. The writer of A 2 began his marginal notes with the fourth
line, mistook the third as an addition to come in after the marked fitl, and thus
wrote flrfi{ftr ~instead of f'1fit(f. tpPi in .A.a would look exactly like 1{Tt(;f to
one not familiar with the peculiar style of writing used in these marginal notes ;
in A 2 we find 1tl1l1f. Other coincidences between Aa and A 2 require no expla-
nation. The only real difference between the two is f.,.(<i<i in A 2 instead of
firatj in A.a ; a mistake evidently produced by the distraction of the writer, and
the omission of the passage between ~ and ~· This will be sufficient to
show that Aa is really the prototype of the A. MSS., and that therefore, if the
whole of this MS. could have been recovered, it would have rendered to the A
class the same service which was rendered to the C class by Ca.
I cannot conclude this second volume of the Rig-veda without acknowledging
my obligation for much useful ad.vice and kind assistance which from many
quarters I have continued to receive. Professor Wilson has taken the same
active interest in this as in the first part, and there is not a sheet that has not
received the benefit of his careful perusal. The present volume has not indeed
had the advantage of Dr. Rieu's revision, to whom much of the praise bestowed
on the correctness of the press in the first is due. But in the latter portion of
this volume I have been able to avail myself of the assistance and active coopera-
tion of my learned friend Dr. Aufrecht of Berlin, and the benefit hence derived
cannot be too highly valued. I may hope that by his continued assistance
I shall be able to bring this edition to an end in a much shorter time than
I at first expected. ·Dr. Ballantyne of Benares, Dr. Roer of Calcutta, Mr. Fitz-
Edward Hall of Benares, and lately Mr.Walter Elliott of Vizagapatam, have all
in the most obliging manner assisted me in the execution of my work by the
transmission of several most valuable MSS., without which, I fear, I should
never have succeeded in accomplishing a work which I commenced without
 SECOND VOLUME OF THE FIRST EDITION. XXXlX

perhaps a full consciousness of all its difficulties. I have been also honoured by
that distinguished scholar and noble patron of Sanskrit literature, Raja Radha-
kanta Deva of Calcutta, with a most valuable present, the Sabda Kalpa Druma,
a work which by its comprehensive range and its exoellent arrangement stands
unrivalled in Indian philology. At Oxford, again, I have had the advantage of
the Bodleian Library, where, under the liberal management of Dr. Bandinel,
scholars of all countries may avail themselves of the treasures of one of the finest
collections in Europe with the same facility as if at work in their own private
library. '
Nor can any one, perhaps, acknowledge more thankfully than myself the
valuable additions lately made to our knowledge of Vaidik literature and anti-
quities by the combined labours of so many distinguished scholars in India,
England, France, and Germany. Many difficulties, against which I had to con-
tend singlehanded in the first volume, have been removed by their publications.
On commencing this edition I had first to copy and collate many works con-
tinually quoted by SAyana, or otherwise essential to a full comprehension of the
Rig-veda. There were then no editions of the other Vedas, of the Nirukta, the
Aitareya-br!hmana, Asval!yana, and similar works. Yet it would have been
impossible to print even the :first pages of SAyana's commentary without having
to a certain extent established a critical text of these writings. Several have
since been published, and their text has been settled with an accuracy greater
than the limits of my time· allowed me even to aim at in these supplementary
treatises. Our comprehension of these works has been considerably facilitated
either by translations -0r by notes and indices. The only works from which assist-
ance could be derived when I cotnmenced this edition were Rosen's :first book of
the Rig-veda and some valuable essays by Professor Neve and Dr. Kuhn. At
present we have Professor Roth's edition of the Niruk.ta, Professor Benfey's
SAma-veda, Dr. Weber's Yagur-veda, valuable treatises on the Gnbya-sO.tras by
Professor Stenzler, useful indices to the Vedas by Messrs. Pertsch and Whitney,
and last, not least, the first specimen of Vaidik lexicography by Roth. Many
obscure points in the earliest literature of India have 1\eceived new light in the
first and second volumes of Lassen's classical work on Indian Antiquities.
Names formerly known to few have become familiar to all through the inde-
fatigable industry of Dr. Weber. On the whole an entirely new direction has
been given to Sanskrit philology, and during the last six years the Vaidik has
received greater attention than any other period of Sanskrit literature.
But since the publication of the first volume of the Rig-veda we have suffered
:xl PREFACE TO THE
one irreparable loss. The death of Eugene Burnouf has deprived Sanskrit phi-
lology of one of its chief supporters, of one of its greatest ornaments. His loss
will be long felt in different departments of Oriental learning, where his name
is associated with some of the most brilliant discoveries of our age ;-nowhere
longer and more keenly than among the friends and students of Sanskrit
literature. Of Burnouf's works I need not here speak. As the first scientific
decipherer of the Cuneiform inscriptions, he has erected to himself a monument
more lasting than the mountain-records of Persia. As the first P:tli scholar
and the historian of Buddhism, his fame will not easily be surpassed by
future researches. As the first editor and interpreter of the Zend-avesta, his
memory will endure so long as the human race values the traditions of its
early childhood. But Burnouf's key to all these discoveries was Sanskrit ; and
in Sanskrit philology, where his influence was most beneficial, his loss is now
felt most severely. I do not here allude to the Bhagavat-pur~na and other
monuments of his persevering industry now left unfinished, nor to the works he
contemplated, nor to the treasures he had collected. In losing Burnouf, we have
lost not only an indefatigable fellow-labourer, not only a disinterested teacher,
but a most respected judge ; in his approval valued by all, in his censure feared,
in his verdict distinguished unfailing!y by fairness and by truth. Though he
published but little on the Veda, yet I may safe~y assert-and those with whom
I had the benefit of attending his lectures at the College de France, Barthelemy
St. Hilaire, Gorresio, Neve, Pavie, Foucaux, Roth, Goldstucker, Bardelli, and
many others, will bear me out-that there was not then a scholar in Europe
more conversant with the language and the traditions of the Veda than
Burnouf. The intimate friend of Rosen, he alone kept up after Rosen's death
the tradition of Vaidik studies. He impressed their importance on all who
came to study under him, and he proved that for a true appreciation of the
early history of mankind, and for a comparative study of the religions of the
East, a knowledge of the Veda was indispensable. The new impulse given to
Sanskrit philology in that direction, and the subsequent publication of numerous
Vaidi.k works, were due to him ; and for my own part I can only repeat, that
without Burnouf's advice, encouragement, and assistance, I should never have
been able to undertake this edition of the Rig-veda. When I heard of his death
I felt-and I believe that many engaged in similar studies shared the feeling-
as if our work had lost much of its charm and its purpose. 'What will Burnouf
say 1' was my earliest thought, on completing the first volume of the Rig-veda.
And now as I am finishing the second, in its turn submitted to the judgment
 THIRD VOLUME OF THE FIRST EDITION. xli
of so many scholars whose friendship I value, and whose learning I admire,
my thoughts turn again to him who is no longer among us, and I think:, not
without sadness, of what his judgment would have been.
M.M.
Sm RoBEBT TAYLoR's INSTITUTION, OxF011.n,
Christmas 1853.

PREFACE
TO THE THffiD VOLUME OF THE FffiST EDITION.

IN laying before the public the third volume of my edition of the Rig-veda
and its commentary by SA.yan.A.kA.rya,, it gives me much pleasure to acknow-
ledge the increasing interest which of late years has been evinced by the most
eminent scholars in England, in India, and on the continent, with regard to
these ancient remnants of the sacred poetry of the Bcl.hmans. Their importance
for Sanskrit literature had been felt 'ever since Sir William Jones, Colebrooke,
and F. Rosen informed us of their existence, and gave us the first specimens of
their contents ; and no one acquainted with the later literature of India, the
epic poems, the law-books, the systems of philosophy, could fail to see that our
knowledge of the historical growth of the Indian mind must remain incomplete
until we had gained an insight into that period of literature which precedes
Vyba and Valmtlri, Manu and Gaimini, and to which the poets, the lawgivers~
and philosophers of India point with common consent as the highest authority
for their inspirations, their belief, and their institutions. Sanskrit literature
without the Veda was like Greek literature without Homer, like Arabic litera-
ture without the Koran, like English without Shakespear.
But as the study of Sanskrit owes its permanent interest chiefly to the fact
that the ancient language of India has been proved to be most intimately
connected with the classical languages of Europe, and that in it has been found
the key to the most secret archives of the history of language in general, the
Veda would never h~ve engaged the serious attention of a large class of
scholars, if this ancient literary relic had not been found to shed the most
unexpected light on the darkest periods in the history of the most prominent
6 VOL. I.
xlii PREFACE TO THE
nations of antiquity. The religious traditions of the Persians or the Zoroastrians
have been traced back to their source in the Veda. Many of the most obscure
grammatical forms in the arrow-headed inscriptions of Darius and Xerxes have
been df:'ciphered by means of the Veda. The mythology of Greece and Italy,
nay of Germany and Iceland, has suddenly assumed a new aspect and an intel-
ligible expression by being confronted with the poetical language of the Veda.
Even civil institutions, local customs, and proverbial expressions, which we
meet with in the later history of the Aryan nations, have received an unexpected
explanation in the simple poetry of the Veda. In this manner the Veda, though
not yet known in its completeness, has assumed an importance which no other
literary production of India could ever have claimed; and w~ may rest con-
vinced, that as long as man cherishes the records of his family, in the widest
sense of the word, these simple songs will maintain their place among the most
valued annals of ancient history. There is one class of readers that may have
been disappointed-men who study ancient literature less on account of its
historical than its poetical value. Those who expected in the Veda, strains like
the elaborate odes of Pindar, or the vague and misty exhalations of Ossian, will
have found but very little answering their expectations. But the true historian
values facts, ancient and genuine ; and a corroded copper As of the Roman
republic is of greater value to him than an imperial gold medal of the most
exquisite workmanship. What Schelling says with regard to the deities of the
later Hindu pantheon, such as they are represented to us in the MahtbM.rata,
the poems of KUidasa, and the Puranas, applies to all facts of history: 'Hideous
or not, they stand before us, and so require a rational explanation 1.'
But it has been a still greater pleasure to me, while engaged for so many
years in preparing a critical edition, not only of the text of the Rig-veda, but
also of its commentary by SayanAkarya, to observe how the conviction seems
to be growing more and more general, that without this commentary an accurate
and scholarlike knowledge of the Veda could never have been obtained. There
was at first much controversy as to the value of Sayana, and as to the necessity
of an edition, and particularly of a critical edition, of his commentary. Now it
seemed to me, that his strong and his weak points must have been so apparent
to all who entered honestly into the study of Sayana, that I hardly thought it
incumbent on me to defend him against his enemies, or to save him from his
friends. ·For though we all admired the quick perception and the brilliant

1
Schelling's Philosophy of Mythology, p. 24..
 THIRD VOLUME OF THE FIRST EDITION. xliii
divination displayed by some distinguished scholars in their attempts to guess
the meaning of these ancient hymns without the help of that tradition which
SAyana embodied in his commentary-and though a work like that of the late
M. Langlois, who actually published a complete translation of the Veda at a
time when most scholars were content with deciphering a few lines, will always
excite our admiration by the boldness, the perseverance, and the undoubted
ingenuity which it displays-yet before the tribunal of a more severe scholar-
ship such works could not be approved ; and it begins to be recognised that the
errors which they propagated have proved so mischievous as to outweigh the
many right guesses which no doubt they contained.
It would have been equally wrong, however, to consider SA.yana's com-
mentary as an infallible authority with regard to the interpretation of the
Veda. SAyana gives the traditional, but not the original, sense of the Vaidik
hymns. These hymns-originally popular songs, short prayers and thanks-
givings, sometimes true, genuine, and even sublime, but frequently childish,
vulgar, and obscure-were invested by the BrAhmans with the character of an
inspired revelation, and made the basis of a complete system of dogmatic
theology. If therefore we wish to know how the BrAhmans, from the time of
the composition of the first BrAhmana to the present day, understood and
interpreted the hymns of their ancient Rishis, we ought to translate them in
strict accordance with SAyana's gloss. This is the object which Professor
Wilson has always kept in view in his translation of the Veda ; and for the
history of religion, which in India, as elsewhere, represents the gradual
corruption of simple truth into hierarchical dogmatism and philosophical
hallucination, his work will always remain the most trustworthy guide. Nor
could it he said, that the tradition of the ·Br!hmans, which SAyana embodied
in his work, after the lapse of at least three thousand years, had changed
the character of the whole of the Rig-veda. By far the greater part of these
hymns is so simple and straightforward, that there can be no doubt that
their original meaning was exactly the same as their traditional interpre-
tation. But no religion, no poetry, no law, no language, can resist the wear
and tear of thirty centuries; and in the Veda, as in other works, handed down
to us from a very remote antiquity, the sharp edges of primitive thought, the
delicate features of a young language, the fresh hue of unconscious poetry,
have been washed away by the successive waves of what we call tradition,
whether we look upon it as a principle of growth or decay. To restore the
primitive outlines of the V aidik period of thought will be a work of great
f 2
xliv PREFACE TO THE
difficulty. i 'We may collect all the passages where an obscure word occurs,
we may compare them and look for a meaning which would be appropriate to
all; but the difficulty lies in finding a sense which we can appropriate and
transfer by analogy into our own language and thought. We must be able to
translate our feelings and ideas into their language at the same time that
we translate their poems and prayers into our 01\-'11. We must not despair
even where their words seem meaningless and their ideas ban-en or wild.
What seems at first childish may at a happier :moment disclose a sublime
simplicity, and even in helpless expressions we may recognise aspirations after
some high and noble idea. When the scholar has done his work, the poet and
philosopher must take it up and :finish it. Let the scholar collect, collate, sift,
and reject-let him say what is possible or not according to the laws of the
Va.idik language-let him study the commentaries, the Stltras, the Bri.hmanas,
and even later works, in order to exhaust all the sources from which informa-
tion can be derived. He must not despise the tradition of the Brahmans, even
where their misconceptions and the causes of their misconceptions are palpable.
To know what a passage cannot mean is frequently the key to its real meaning ;
and whatever reasons may be pleaded for declining a careful perusal of the
traditional interpretations of YA.ska or S!yana, they can all be traced back to
an ill-concealed "argumentum paupertatis." Not a corner in the Brihmanas,
the S1ltras, Yaska, and SAyana should be lea un~xplored before we venture to
propose a, rendering of our own. S!yana, though the most modern, is on the
whole the most sober interpreter. Most of his etymological absurdities must
be placed to YAska's account, and the optional renderings which he allows for
meta.physical, theological, or ceremonial purposes, are mostly due to his regard
for the Brihmanas. These BrA.hmanas, though nearest in time to the hymns of
the Rig-veda, indulge in the most frivolous and ill-judged interpretations. When
the ancient Rishi exclaims with a troubled heart, " Who is the greatest of the
gods? Who shall first be praised by our songs ?"-the author of the Bri.hmana
sees in the interrogative pronoun "Who " some divine name, a place is allotted
in the sacrificial invoc~tions to a god" Who," and hymns ad.dressed to him are
ealled "Whoish " hymns. To make such misunderstandings possible, we must
assume a considerable interval between the composition of the hymns and the
Brahmanas. As the authors of the Brthmanas were blinded by theology, the
authors of the still later Niruktas were deceived by etymological :fictions, and

1
See the Author's Essay on the Veda and Zendavesta, p. 13.
 THIRD VOLUME OF THE FIRST EDITION. xlv

both conspired to mislead by their authority later and more sensible com-
mentators, such as S~yana. Where Sayana has no authority to mislead him,
his commentary is at all events rational; but still his scholastic notions would
never allow him to accept the free interpretation which a comparative study of
these venerable documents forces upon the unprejudiced scholar. We must
therefore discover ourselves the real vestiges of these ancient poets ; and if W• ·
follow them cautiously, we shall find that with some effort we are still able to
walk in their footsteps. We shall feel that we are brought face to face and
mind to mind with men yet intelligible to us, after we have freed ourselves
from our modern conceits. We shall not succeed always : words, verses, nay,
whole hymns in the Rig-veda, will and musli remain to us a. dead letter. But
where we can inspire those early relics of thought and devotion with new life,
we shall have before us more real antiquity than in all the inscriptions of Egypt
or Nineveh ; not only old names and dates, and kingdoms and battles, but old
thoughts, old hopes, old faith, and old enors, the old " Man " altogether-old
now, but then young and fresh, and simple and real in his prayers a.nd in hiA
praises.'
How the Veda should be interpreted, and how S~yana's commentary should
be made use of for that purpose, has lately been shown in a work by M. Ad.
Regnier, ' Etude sur l'idiome des Vedas et les origines de la Langue Sanscrite,
Premiere Partie, Paris 1855.' I may be allowed to quote from this excellent
essay the following passage, which lays down with fairness and exactness the
principles which ought to be followed by every student of the Veda. ' J e joins
au texte des hymnes celui du commentaire de Sayana Aoharya, que je suivrai.
dans son interpretation, partout ou il me semblera que la logique et la gram-
maire le permettent; toutes les fois que ,fadopterai un autre avis que le sien,
j'en donnerai les raisons: d'abord pour bien etablir le sens, parce que, daus
une matiere souvent aussi obscure, il iaut toujours savoir d'abord l'avis des
Indians eux-mAmes; puis, parce que ces scolies nous donneront !'occasion de
faire eonnaissance avec quelques-unes des habitudes les plus ordinaires d'inter-
prettttion des glossateurs. Tous ceux qui ont eu le bonheur de suivre le eOUl'S
de M. Eugene Burnouf savent quelle importance il attacha.it a !'explication des
comm.entaires. Le meilleur moyen, selon lui, d'assurer et de hAter les progres
et de se rompre aux difficultes de la langue, o' etait de se familiariser de bonne
heure avec la methode et le style des grammairiens, style souvent tres-abstrait et
OU les procedes d'expression synthetique sont pousses frequemment al'exces.'
Even if the author had not paid this tribute to the memory of E. Burnouf,
xlvi PREFACE TO THE

the accuracy and painstaking minuteness of his work would have shown that
he belonged to Burnouf's school ; and it is pleasing to see how the spirit of that
eminent scholar seems still to be alive in that brilliant senate of learning of
which he once formed so illustrious a member, when we read that the French
Academy has proposed as one of its last prizes-
' Un commentaire particulierement exegetique et grammatical, soit sur une
partie suivie, soit sur un choix d'hymnes du Rig-veda, ou l'on aura soin d'exposer
toujours et de discuter, s'il y a lieu, meme quand on ne l'adoptera pas, !'opinion
du commentateur Saya:ga Acharya.'
Such a prize, while it gives a sanction to my work, for which I cannot
sufficiently express my gratitude, will, it is to be hoped, act as an encouragement,
and bring some of our young Sanskrit scholars toward that line of study which
Burnouf pointed out to all of his pupils, as the most sure to lead to real and
lasting results.
After what has been stated in the prefaces to the first and second volumes, I
have little to add with regard to the MSS. which I used for the third volume.
There is one notice which I lately received from the Rev. Dr. Stevenson, the
distinguished editor and translator of the S!ma-veda, and which I subjoin here
in confirmation of my views on the local origin of the three families of the MSS.
of Sayana's commentary.
' As I see you have formed a particular family of the two MSS. B I and
B 2, I may as well mention to you all I have learnt of their history. BI was
procured at Puna from a Wakil, who procured it from the family of the Guru
of the late Peshwah; at least, so he said; and, as the family was poor, and no
one else likely to have such a work, there is no reason to discredit his story. It
is, as you will see, written by two different scribes, the greater part in what we
at Bombay call Kayasthi lipi, the hand writing of KAyasths from the province of
Guzerat. The letters in this portion are very deep. The rest is written by a
Deccani Brahman in what we call the Dakshant lipi, and not so deep as the other.
This difference is discernible even in the Cave inscriptions in the old character.
'B 2 was copied for M. Burnouf from that MS. by a Puna Brahman, whom
I got to transcribe it for him.
'I have also an imperfect copy of another MS. of the BhA.shya. The whole
of the seventh Ashtaka is wanting, and I have only two .Adhyayas of the eighth.
If you would like to see them, I shall be happy to send them; and indeed the
whole MS. is at your service. There is a complete copy of the Bhtshya of the
.Rig-veda SamhiU. in the library of the R. .A. Society, Bombay branch. The :first
 THIRD VOLUME OF THE FIRST EDITION. xlvii
Ashtaka is copied from B I, and would be of' no use to you ; but the rest was
taken, I am told, at Mr. Elphinstone's expense, or at that of the Bombay
government, and deposited in the library, from a copy belonging to Dr. Taylor,
which was caITied to England, as I understood, to be deposited in the library of
the India House. As, however, you take no notice of such a work there 1,
I must have been under a mistake about that. However, there is no mistake
about Dr. Taylor's having had such a work, and the Bombay copy having been
taken, sometime about 1820, from it, with the exception of the first .A.shtaka,
which was omitted, why I cannot say. The imperfect copy I have is partly
t~en from this, and partly made up of portions of a MS. received from
the Wakil.'
.Another communication on the MSS. of Sayana's commentary was kindly
sent to me by my learned friend Mr. Fitz-Edward Hall, while yet at Benares.
I had applied to him for help with regard to some extremely difficult passages
in the fifth Ashtaka, and in the hope that some more fragments of the MS.
mentioned in my preface to the second volume, p. xii [above, p. xxxvi], might
still be discovered in India. He wrote to me, Benares, Jan. I 5, I 85S:
' I was much afraid that I should have to send off this letter without being
able to furnish you with the means of perhaps supplying the lacunae you have
encountered in the fifth Ashtaka. In fact, but a few hours have elapsed since
I was able to put together the extracts from MS. ~' which I enclose. We have
in the College library but one MS. of SA.yana's commentary on the R.V. It
was copied in the Samvat year 1851, and is, as you will see under the heading
of MS. a., of little value. Notwithstanding repeated enquiries, I was unable to
get sight of any other MS. until ~ was brought to me yesterday. This MS. is
without date of transcription, and has no external indications of any antiquity.
I think the passages from it, which I now send, fully justify me in ordering a
copy to be made forthwith of the whole of the fifth .A.shtaka 2• If you request
it, I will have the remainder of the MS. copied; but, as there is a possibility
that its character will differ in different Ogdoads, you had better send some
test-passages by which to decide its value in the subsequent Ashtakas.
'As for the sheet which I send you, it has reference to the last passage or

1 of the B class, though, particularly in the later books,

Dr. Taylor's copy was not mentioned in the
preface to the first volume, because it only begins it has peculiar readings, and is sometimes evidently an
with the third Ashtaka. It will be seen from my a.bb1eviation of the or1gmal text of the B MSS.
2 This copy bas since been received. It begins
preface to the second volume, which Dr. Stevenson
had not 1&1ce1ved, that I recognISed this MS. as one with the third Adhylya. of the fifth Asbtaka.
xlviii PREFACE TO THE
passages noted in the paper you sent. I was not sure what it was you required.
Accordingly, if I have erred, it has been on the right side The first copy was
made from a., which I afterwards changed by interlineations and erasures to
coITespond to /3. I shall be happy if I hear that I have been the instrument of
rescuing your work from even a single imperfection.
'.A.re you acquainted with an abridgement of S~yana's commentary by
Mudge.la 1 The grammatical explanations are omitted altogether, and the
remainder of the comment so abridged that the whole takes up about a fourth
part the space of the original. Our copy professes to be taken in the Sanivat
year I 47-. The last figure is unsupplied. Strange to say, it does not break off
in the fourth Adhy!ya of the first Ashtaka, where all your MSS. of the .A. and
B classes terminate; but it runs on to the words fll'4WC<:<(, p. 538. 1. 5 [now
p. 294. 1. 8]. .Afterwards there occurs the same appearance of supplial by a later
hand, to whioh you call attention ; a peculiarity which I observed also in the
MS. /l If this epitome may be trusted, the mutilated passage at the bottom of
page 969 [now p. 542. 1. 3] should run thus: ~ Tff 'f.lffiircrat~ ~C'I'{. 11•
'Did you ever hear of a Rig-bhAshya by R&vana 1 Sfirya Pandit, in his
Paramartha-prabhl, a commentary on the Bhagavadglta, professes to have seen
it. I am also told, that a, commentary by Ravana on one of the Sa,khAs of the
Yagur-veda is still extant.
1
This MS. has since been sent to the hbrary of 'iiltVlo 'I( fimo lf\"'qo ~~
the East-India. House. It contains Aehtakas I, II,
~: I~ 3'\tt4tl=tfCfittl ~ ~\WMCtit6ltif
a.nd III, and the last three Adhy!yas of .Ashtaka IV.
Some fragments of the fust and seventh Ashtak&H
@tllCSl'tfi~ltrf f'*4i1"!1Mif~ l N'i11'1f1
were p1esented to me by Mr. Fitz-Edward Ha.11. I 111<44'11 tlf ~f't•f<t I d*it H4 I(411i;<:leit! lif-
subjoin the begmning of the first Ashtaka. so far as ~ ~ ll •1"C(4iPt4tltii11 41i4'41Elif\
it 1s legible, and without attempting to correct all 1tliEOT ~ I iifi<*> I iri~ ~ftt tf;:r 'l{NT! I "
mistakes . 1( 1400\ift d Ielf C{1[ af\I& Ifir I ~ 1"fil' 1i&
"ifl•lflttl+.t 111f: I ~ofiht 1uct tCfi1: 1?{'I J4tt44 •uctiil -=.ftf1lr ~ ,
"ff lJM~l.f ~ fi41t'lltf.i 'lif 1 ~I ti?C(4'f'1Jf ~~~T'f.t 1'illftr?TTf1r I ~
JiQt111~ If.I ~•i(Qttt ~ 1!1f: h
~ ~ Mf~11'11(Qltf.i ~ I
iiNi\& Ult
.C('lli1Cfi
1''" 1"{1f ~ I flP&tfit,.Szetl
-.A: I ~ ~ 11 "if q :
~ ~ '111f6'1~ ftlu ~ H -.1:a1 QI qifit II iff -.fl;rifta; '(flf I ;grfq ~ ~
44ll"'1"1l 'i.~ 'lif ~~: (~:i) I ~ I 14lit1t4lN I ~ ~t'ff ~ ~­
q1[ ltif41tt~ .. ~ ~" f\lfil i~ ~~1Q"l"f~ ff 1911: cit11{ I
~~ t41\tfN4<il+llt4 ~I
~ ~~~ lTitR 'ii~ I~ 'frit!
~ ~ ~ lTifi!rtrt ~II ~~ ';flf~ ~ I ~: cftt1f l 1'Ji{-
if'i••nll1Q' 'lif ~If~ \f\i(f~il I \11lffi l ~(i(fi't'Ql(Qf <:Cit1iflilfit1('" ~­
~'4t~lf ~ qlilt
4
mf1{q mw II f'-lm~ II
~~· ~ 1\qtt4'4 '1{iit f1IU '6M 1
 THIRD VOLUME OF THE FIRST EDITION. xlix
'I have failed entirely in all attempts to trace the history of the MS. of
Sayana, of which I sent you some fragments. But I have since procured a
fragment of the eighth Ashtaka, very like it in outward appearance.'
In another letter, dated Ajmere, 24th May, 1856, the same distinguished
scholar wrote to me :
'It must be perplexing to be obliged to illuminate Sayana's text with stars.
I am not going to say that the publication of this work was undertaken
prematurely: but I am pretty well assured, from what I have observed, that
there still lie hidden, in the libraries of rigid Brahmans, scores of hereditary
copies of this commentary; and it is scarcely unreasonable to suppose that, if
procurable, they might supply your lacunae.'
There is one more MS. which ought to be mentioned here. It belongs to the
library of the East-India House (No. 26 r 2 ), and bears the title~·
It is, however, not the sixth, but the fifth Ashtaka of the Rig-veda. As it bears
the date~ q~~g, it might have been expected to yield some help towards the
restoration of Sliyana's text; but on closer inspection it turned out an exact
and literal reproduction of my own MS. Ca.
When I began this edition, I thought the whole of it would be completed in
three or four volumes, and I now find that the first three volumes contain hardly
more than half of the whole work. I must confess that I could have wished
that the ancient poets of the Veda and their Indian commentators had been less
diffuse ; for though I believe that no edition of any author in Sanskrit or
any other language, for which manuscripts had first to be copied, others to be
collated, innumerable references to be verified, and an index to be made of every
word, has ever been brought out so rapidly as this edition of the Rig-veda, yet I
feel that ten years of my life are gone, and I know not whether I shall have
sufficient time left to finish a work which I once undertook perhaps with too
much confidence. Yet even if I should not see the completion of this work, I
should not be sorry for the time that I have spent on it ; and nothing will ever
induce me to change the principles which I have hitherto followed, and to give a
hasty copy of a MS., instead of a critical edition of the text and commentary of
the Rig-veda. I have had again for this volume the valuable assistance of my
learned friend Dr. AufrAcht, and I sincerely regret that I shall no longer enj~y
this advantage, as much of the correctness and accuracy of the last volumes was
due to his conscientious cooperation, joined to the kind assistance which I ha~e
never failed to receive from my honoured friend Professor H. H.Wilson.
I have to express my deep obligation to the Court of Directors of the Honour- .
7 VOL, L g
I PREFACE TO THE THIRD VOLUME, FIRST EDITION.
able the East-India Company, under the chairmanship of Colonel Sykes, and to
the Board of Control, under the presidency of the Right Honourable Vernon
Smith, for having sanctioned the continuation of this work, and granted funds
necessary for its completion-an act of enlightened liberality, which will be
applauded by all persons interested in the history of India and in the history of
mankind, and by which one of the most important monuments of antiquity will
be rescued from oblivion and restored in its integrity.
M.M.
OXFORD, June 5, 1856.