The present volume of translations comprises about one third of the entire material of the Atharva-veda in the text of the Saunaka-school. But it represents the contents and spirit of the fourth Veda in a far greater measure than is indicated by this numerical statement. The twentieth book of the Samhitâ, with the exception of the so-called kuntâpasûktini (hymns 127-136), seems to be a verbatim repetition of mantras contained in the Rig-veda, being employed in the Vaitâna-sûra at the sastras and stotras of the soma-sacrifice: it is altogether foreign to the spirit of the original Atharvan. The nineteenth book is a late addendum, in general very corrupt; its omission (with the exception of hymns 26, 34, 35, 38, 39, 53, and 54) does not detract much from the general impression left by the body of the collection. The seventeenth book consists of a single hymn of inferior interest. Again, books XV and XVI, the former entirely Brahmanical prose, the latter almost entirely so, are of doubtful quality and chronology. Finally, books XIV and XVIII contain respectively the wedding and funeral stanzas of the Atharvan, and are largely coincident with corresponding Mantras of the tenth book of the Rig-veda: they are, granted their intrinsic interest, not specifically Atharvanic. Of the rest of the Atharvan (books I-XIII) there is presented here about one half, naturally that half which seemed to the translator the most interesting and characteristic. Since not a little of the collection rises scarcely above the level of mere verbiage, the process of exclusion has not called for any great degree of abstemiousness.

These successive acts of exclusion have made it possible to present a fairly complete history of each of the hymns translated. The employment of the hymns in the Atharvanic practices is in closer touch with the original purpose of the composition or compilation of the hymns than is true in the case of the other collections of Vedic hymns. Many times, though by no means at all times, the practices connected with a given hymn present the key to the correct interpretation of the hymn itself. In any case it is instructive to see what the Atharvan priests did with the hymns of their own school, even if we must judge their performances to be secondary.

I do not consider any translation of the AV. at this time as final. The most difficult problem, hardly as yet ripe for final solution, is the original function of many mantras, after they have been stripped of certain adaptive modifications, imparted to them to meet the immediate purpose of the Atharvavedin. Not infrequently a stanza has to be rendered in some measure of harmony with its connection, when, in fact, a more original meaning, not at all applicable to its present environment, is but scantily covered up by the 'secondary modifications of the text. This garbled tradition of the ancient texts partakes of the character of popular etymology in the course of the transmission of wofds. New meaning is read into the mantras, and any little stubbornness on their part is met with modifications of their wording. The critic encounters here a very difficult situation: searching investigation of the remaining Vedic collections is necessary before a bridge can be built from the more original meaning to the meaning implied and required by the situation in a given Atharvan hymn. Needless to say the only correct and useful way to translate a mantra in the Atharvan, is to reproduce it with the bent which it has received in the Atharvan. The other Vedic collections are by no means free from the same taint. The entire Vedic tradition, the Rig-veda not excepted, presents rather the conclusion than the beginning of a long period of literary activity. Conventionality of subject-matter, style, form (metre), &c., betray themselves at every step: the 'earliest' books of the RV. are not exempt from the same processes of secondary grouping and adaptation of their mantras, though these are less frequent and less obvious than is the case in the Atharva-veda.

Obligations to previous translators: Weber, Muir, Ludwig, Zimmer, Grill', Henry, &c., are acknowledged in the introduction to each hymn. I regret that the work was in the hands of the printer prior to the appearance of Professor Henry's excellent version of books X-XII. The late lamented Professor Whitney kindly furnished me with the advance sheets of the late Shankar Pandurang Pandit's scholarly edition of the AV. with Sâyana's commentary, as also with many of the readings of the Cashmir text (the so-called Paippalâda-sâkhâ) of the AV. Neither the Paippalâda nor Sâyana sensibly relieves the task of its difficulty and responsibility.

MAURICE BLOOMFIELD.

JOHNS HOPKINS UNIVERSITY,

BALTIMORE: April, 1896.