caste-system in India is very ancient, and existed already in the Vedic age. Some have sought to disprove its existence at that period, on the ground that, excepting this one, the old hymns contain no distinct reference to it. The hymn itself they assign to the very end of the Vedic era. In its present form, it is true, it is not older than the greater part of the hymns of the tenth book, and than those of the Atharvaveda. But the ideas which it contains are certainly of a primeval antiquity: in particular, the descriptive portion of it looks like a versified sacrificial formula. And in fact the hymn is found in the Yajurveda among the formulas connected with human sacrifices, which were formerly practised in India.

(b) "But even if it were to be admitted that this hymn, with its ideas, is but a very recent product of the Vedic age, that would afford no sufficient proof that castes did not exist at all in the earliest period of the Arian immigration into India. As we have seen above, p. 9, the Brahman already appears [R.V. iv. 50, 8, f., and i. 108, 7] in contrast to the Rājan, i.e. a member of the warrior caste (they are frequently called Rajanyas instead of Kshatriyas),—a circumstance which plainly points to a distinction of castes.

(c) "Besides, it is scarcely conceivable that the caste-system, to which reference is made in all the four Vedas, sometimes more frequently, sometimes more rarely, should have been suddenly formed in the later Vedic period.

(d) "The reason why the names of the castes are not mentioned in those hymns of the Rigveda which, rightly or wrongly (for a thorough investigation of this point with certain results is as yet wanting), are regarded as the oldest, may be a different one," [i.e. the reason why the castes are not there mentioned need not be that these castes did not then exist]. "The fact, namely, must not be overlooked, that by far the greater part of the Vedic hymns were composed for sacrificial purposes, and partly for quite definite ceremonies, are often merely the poetical variations of primeval sacrificial formulas, and further that the most of them are the compositions of Brahmans. As the ceremonial is not prescribed in the hymns, where almost everything turns upon the invocation of the most diverse gods, no opportunity was offered to make special mention of the castes. The sacrificers, i.e. those who caused the sacrifice to be offered (yajamānās), are called 'givers' in

general, also

the rich,' (maghavan) without any mention of their caste. . . . . Now as there were no hymns or sacrificial formulas composed specially for Brahmans, or Kshatriyas, or Vaisyas,—as there actually have been for the ceremonies of the Sudras, who are not allowed to hear verses of the Veda,-the Vedic poets had no opportunity to mention the castes in their hymns. Consequently the fact that the names of the individual castes do not appear there does not by any means prove their non-existence. This conclusion was in any case premature."

(e) "Besides the grounds already assigned, a further positive proof can be adduced that castes actually existed already in the most ancient period. In the religious records of the Iranians, who are so nearly allied [to the Indians], in the Zendavesta, the four castes are quite plainly to be found, only under other names, (1) Athrava, 'priest,' (Skr. Atharvan), (2) Rathaēstão, 'warrior,' (3) Västriyo fshuyās, 'cultivator,' (4) Huitis (Pehl. hutokhsh), 'workman' (Yasna 19, 17, Westerg.) No further data regarding the mutual relations of these castes are contained in the Zend writings; but we can conclude, from various circumstances, that the priests, the Athravas, already formed a caste. So, for example, Zarathustra is forbidden by Ahuramazda to communicate a sacred text to any one else but an Athrava, i.e. priest (Yasht 14, 46, West.)." [See Spiegel's Avesta, iii. 148.] "No one but the son of a priest may be a priest, and the daughters of members of the priestly caste may only be given in marriage within the caste,a custom which continues to this day. The distinction of the other castes has, however, become obliterated among the Zoroastrians, just in the same way as, among the Hindus, the Brahman caste alone has on the whole been maintained pure, though split up into innumerable subdivisions, whilst the other three castes have become dissolved into a great number of mixed castes, so that at the present day, properly speaking, four castes only exist in theory, but not in reality. This circumstance, now, that a remnant of the caste-system has still been preserved among the Zoroastrians, speaks strongly in favour of the assumption that that institution already existed among the Indians in the remotest times, in any case ever since their immigration into India. How close the connexion between the old Indians and the Iranians must have remained, even in the Vedic period, is shown by a distinct

allusion which I have lately discovered in the Zendavesta to the initial verse of the Atharvaveda."

(ƒ) In an earlier part of his Dissertation (p. 8, f.), Prof. Haug, after explaining that in the Vedic age Kshatriyas as well as Brahmans could take part in the performance of sacrifice, and were in some cases composers of Rik verses; and that even Kavasha Ailusha, the son of a slave, was the author of a hymn (see above, p. 397, f., note 85), adds: "Notwithstanding that the Brahmans in the Vedic age occasionally accorded to distinguished men of other classes a participation in the privileges they claimed, they nevertheless appear, even in the most remote period, to have formed a caste distinguished from the other classes, and already tolerably exclusive, into which no one who was not born in it could, without great difficulty, obtain an entrance." He then quotes the text R.V. iv. 50, 8, f., and refers to and compares i. 108, 7.

I add a few remarks on some portions of Prof. Haug's argument. (a) In regard to the Purusha-sūkta, I refer to what has been said above, in the first volume of this work, p. 11. Prof. Haug now states his opinion that the ideas of the hymn are very old, but not its diction.

(b) See the first volume of this work, pp. 246, and 247, note 15 in the latter page, and p. 263, f., also Böhtlingk and Roth's Lexicon, 8.v. "kshatra." Does not Prof. Haug found too much on the passages to which he refers (R.V. iv. 50, 8, f., and i. 108, 7)? See, however, the verse of Manu, iii. 13, quoted below, where the word "rājan" is employed for Rajanya. In R.V. iv. 50, 8, f., the word seems evidently to denote a "king." In R.V. i. 108, 7, however, it may mean a man of the ruling tribe or class. Comp. Böhtlingk and Roth's Lexicon, 8.v. "rajan."

(c) The Vedic period was of considerable duration. Professor Haug himself estimates the period during which "the bulk of the Sanhitā " was composed to have been from 1400-2000 before our era, and "thinks "the oldest hymns and sacrificial formulas may be a few hundred years more ancient still, so that we would fix the very commencement of Vedic literature between 2400 and 2000." Ait. Br. i. 47, f. The entire Vedic period would thus be a thousand years, which however, is, perhaps, too large an estimate.

(d) The reasons here assigned in explanation of the non-occurrence

in the hymns of other references to the castes, supposing them to have then existed, and to have borne the same names as afterwards, seem scarcely sufficient. The hymns do not appear to be so exclusively sacrificial in their character as is here assumed; and might in many passages have admitted of allusions to the existence of castes.

(e) The observations already made upon Professor Kern's Dissertation are applicable here.

(ƒ) In reference to these remarks, see the first volume of this work, p. 265, ff. Prof. Kern, in his Dissertation, p. 18, cites a passage from the Mahābhārata, xiii. 2505, ff., in regard to the intermarriages of Brahmans with the other two next classes, one verse of which (2515) is to the following effect: Abrāhmaṇam tu manyante S'ūdra-putram anaipuṇāt | trishu varneshu jāto hi brāhmaṇād brāhmaṇo bhavet | "They regard from want of skill as not a Brahman the son of a Sūdrā woman [by a Brahman father]. A son begotten by a Brahman in the three castes [i.e. on a woman of either of the upper three classes] will be a Brāhman." And Manu says, iii. 13, S'ūdraiva bhāryā Sūdrasya să cha svā cha viśah smṛite | te cha svā chaiva rājnaś cha tāś cha svā ch' āgrajanmanaḥ "A Sūdrā female only can be the wife of a Sūdra. She and a woman of his own caste may be the wives of a Viś, i.e. Vaiśya. These two and a woman of his own caste may be the wives of a Rājan, i.e. Rājanya; these three and a woman of his own caste may be the wives of a Brahman." From these texts it would appear that purity of caste blood was not much regarded among the Hindus in early ages.

NOTE C.-Page 258.

"The conformities [between the languages of the same family] are astonishing; and especially so, because they enter into the minutest details, and even into the anomalies. It is a curious phenomenon to discover such an inconceivable tenacity in idioms which might appear to be nothing more than passing caprices. The most volatile portion of languages, I mean their pronunciation, has evinced its stability: in the midst of mutations of letters, which are, nevertheless, subject to certain rules, vowels, long or short, have often preserved their quantity." "On the other hand, the disparity is great: the distances which

the languages have traversed in their individual development are immense. After we have exhausted all the analogies, even the most secret, there remains in each of these languages a portion which is no longer susceptible of comparison with the other languages of the same family. We must therefore admit as the causes of that partial incommensurableness, two opposite principles, viz., oblivion and invention. The oblivion of forms and words formerly in use is but too manifest in the languages with whose history we are most intimately acquainted; and it has frequently injured their richness and beauty. Such oblivion must always follow a retrograde movement in civilization in proportion as the intellectual sphere is contracted, a generation which has relapsed into ignorance and barbarism, abandons expressions which have now become superfluous. And as regards invention, I find no difficulty in that either, since in order to comprehend the absolute origin of language, we have no choice between having recourse to a miracle, and conceding to mankind an instinctive power of inventing language."-A. W. von Schlegel, de l'origine des Hindous, Essais; and in the Transactions of the Royal Society of Literature of the United Kingdom, vol. ii. p. 433.

NOTE D.-Page 277.

Mr. Geldart argues the question both from a negative and a positive point of view. Under the first head, he remarks that "language is too uncertain an ethnological test to be of any practical value," and instances the complete discrepancy which exists between the races and the languages of the British Isles. Cumberland and Cornwall, for example, in language agree with London and disagree with Wales, while as to race, it is directly the reverse." The same thing is shown, he observes, "by many similar examples: the accumulative evidence of all amounted to this, that since in so many cases where the ethnological indications of language can be compared with the actual testimony of history, the latter completely contradicts the former," a common language is "not even primâ facie evidence in favour of a common lineage." "Secondly, in a positive point of view, it was shown that in all the instances above cited, there had taken place between the races a close assimilation of (1) political, (2) religious, (3)