Of the three later Vedas, the Samaveda is much the most closely connected with the Rigveda. Historically it is of little importance, for it contains hardly any independent matter, all its verses except seventy-five being taken directly from the Rigveda. Its contents are derived chiefly from the eighth and especially the ninth, the Soma book. The Samaveda resembles the Yajurveda in having been compiled exclusively for ritual application; for the verses of which it consists are all meant to be chanted at the ceremonies of the soma sacrifice. Removed from their context in the Rigveda, they are strung together without internal connection, their significance depending solely on their relation to particular rites. In form these stanzas appear in the text of the Samaveda as if they were to be spoken or recited, differing from those of the Rigveda only in the way of marking the accent. The Samaveda is, therefore, only the book of words employed by the special class of Ugat?i priests at the soma sacrifice. Its stanzas assume their proper character of musical Samans or chants only in the various song-books called ganas, which indicate the prolongation, the repetition, and the interpolation of syllables necessary in singing, just as is often done in European publications when the words are given below the musical notation. There are four of these songbooks in existence, two belonging to each division of the Veda. The number of Samans here given of course admitted of being indefinitely increased, as each verse could be sung to many melodies.

The Samaveda consists of 1549 stanzas, distributed in two books called archikas or collections of ?ich verses. The principle of arrangement in these two books is different. The first is divided into six lessons (prapa?haka), each of which contains ten decades (daÇat) of stanzas, except the sixth, which has only nine. The verses of the first twelve decades are addressed to Agni, those of the last eleven to Soma, while those of the intermediate thirty-six are chiefly invocations of Indra, the great soma-drinker. The second book contains nine lessons, each of which is divided into two, and sometimes three sections. It consists throughout of small groups of stanzas, which, generally three in number, are closely connected, the first in the group being usually found in the first book also. That the second book is both later in date and secondary in character is indicated by its repeating stanzas from the first book as well as by its deviating much less from the text of the Rigveda. It is also a significant fact in this connection that the verses of the first book which recur in the second agree more closely with the readings of the Rigveda than the other verses by which they are surrounded. This can only be accounted for by the supposition that they were consciously altered in order to accord with the same verses in the second book which were directly influenced by the Rigveda, while the readings of the first book had diverged more widely because that book had been handed down, since the original borrowing, by an independent tradition.

We know from statements of the Çatapatha Brahma?a that the divisions of the first book of the Samaveda existed at least as early as the period when the second part of that Brahmana was composed. There is, moreover, some reason to believe that the Samaveda as a collection is older than at least the Taittiriya and the Vajasaneyi recensions of the Yajurveda. For the latter contain verses, used also as Saman chants, in a form which shows the variations of the Samaveda in contrast with the Rigveda. This is all the more striking as the Vajasaneyi text has an undoubted tendency to adhere to the readings of the Rigveda. On the other hand, the view expressed by Professor Weber that numerous variants in verses of the Samaveda contain archaic forms as compared with the Rigveda, and were therefore borrowed at a time before the existing redaction of the Rigveda took place, has been shown to be untenable. The various readings of the Samaveda are really due in part to inferior tradition, and in part to arbitrary alterations made in order to adapt verses detached from their context to the ritual purpose to which they were applied.

Two schools of the Samaveda are known—the Kauthumas and the Ra?ayaniyas, the former of whom are said still to exist in Gujarat, while the latter, at one time settled mainly in the Mahratta country, are said to survive in Eastern Hyderabad. Their recensions of the text appear to have differed but little from each other. That of the Ra?ayanayas has been published more than once. The earliest edition, brought out by a missionary named Stevenson in 1842, was entirely superseded by the valuable work of Benfey, which, containing a German translation and glossary besides the text, came out in 1848. The Samaveda was thus the first of the Vedas to be edited in its entirety. The text of this Veda, according to the recension of the same school, together with the commentary of Saya?a, was subsequently edited in India. Of the Kauthuma recension nothing has been preserved excepting the seventh prapa?haka, which, in the Naigeya subdivision of this school, forms an addition to the first archika, and was edited in 1868. Two indices of the deities and composers of the Samaveda according to the Naigeya school have also been preserved, and indirectly supply information about the text of the Kauthuma recension.

The Yajurveda introduces us not only to a geographical area different from that of the Rigveda, but also to a new epoch of religious and social life in India. The centre of Vedic civilisation is now found to lie farther to the east. We hear no more of the Indus and its tributaries; for the geographical data of all the recensions of the Yajurveda point to the territory in the middle of Northern India occupied by the neighbouring peoples of the Kurus and Panchalas. The country of the former, called Kurukshetra, is specifically the holy land of the Yajurvedas and of the Brahma?as attached to them. It lay in the plain between the Sutlej and the Jumna, beginning with the tract bounded by the two small rivers D?ishadvati and Sarasvati, and extending south-eastwards to the Jumna. It corresponds to the modern district of Sirhind. Closely connected with, and eastward of this region, was situated the land of the Panchalas, which, running south-east from the Meerut district to Allahabad, embraces the territory between the Jumna and the Ganges called the Doab (“Two Waters”). Kurukshetra was the country in which the Brahmanic religious and social system was developed, and from which it spread over the rest of India. It claims a further historical interest as being in later times the scene of the conflict, described in the Mahabharata, between the Panchalas and Matsyas on the one hand, and the Kurus, including the ancient Bharatas, on the other. In the famous lawbook of Manu the land of the Kurus is still regarded with veneration as the special home of Brahmanism, and as such is designated Brahmavarta. Together with the country of the Panchalas, and that of their neighbours to the south of the Jumna, the Matsyas (with Mathura, now Muttra, as their capital) and the Çurasenas, it is spoken of as the land of Brahman sages, where the bravest warriors and the most pious priests live, and the customs and usages of which are authoritative.

Here the adherents of the Yajurveda split up into several schools, which gradually spread over other parts of India, the Ka?has, with their subdivision the Kapish?halas, being in the time of the Greeks located in the Panjab, and later in Kashmir also. The Ka?has are now to be found in Kashmir only, while the Kapish?halas have entirely disappeared. The Maitraya?iyas, originally called Kalapas, appear at one time to have occupied the region around the lower course of the Narmada for a distance of some two hundred miles from the sea, extending to the south of its mouth more than a hundred miles, as far as Naasik, and northwards beyond the modern city of Baroda. There are now only a few remnants of this school to the north of the Narmada in Gujarat, chiefly at Ahmedabad, and farther west at Morvi. Before the beginning of our era these two ancient schools must have been very widely diffused in India. For the grammarian Patanjali speaks of the Ka?has and Kalapas as the universally known schools of the Yajurveda, whose doctrines were proclaimed in every village. From the Ramaya?a, moreover, we learn that these two schools were highly honoured in Ayodhya (Oudh) also. They were, however, gradually ousted by the two younger schools of the Yajurveda. Of these, the Taittiriyas have been found only to the south of the Narmada, where they can be traced as far back as the fourth century A.D. Their most important subdivision, that of the Apastambas, still survives in the territory of the Godavari, while another, the Hira?yakeÇins, are found still farther south. The school of the Vajasaneyins spread towards the south-east, down the Ganges Valley. At the present day they occupy a wide area, embracing North-East and Central India.

Each of these four schools has preserved one or two recensions of the Yajurveda. The text of the Maitraya?i Sam?hita, which consists of four books (ka??a), subdivided into fifty-four lessons (prapa?haka), has been edited by Professor L. v. Schroeder (1881–86). The same scholar is preparing an edition of the Ka?haka Sam?hita, the recension of the Ka?ha school. These two recensions are nearly related in language, having many forms in common which are not found elsewhere. Of the Kapish?hala-Ka?ha Sam?hita only somewhat corrupt fragments have hitherto come to light, and it is very doubtful whether sufficient manuscript material will ever be discovered to render an edition of this text possible. The Taittiriya Sam?hita, which comprises seven books, and is subdivided into forty-four lessons, is somewhat later in origin than the above-mentioned recensions. It was edited by Professor A. Weber in 1871–72. These texts of the Yajurveda form a closely connected group, for they are essentially the same in character. Their agreement is often even verbal, especially in the verses and formulas for recitation which they contain. They also agree in arranging their matter according to a similar principle, which is different from that of the Vajasaneyi recension.

The Sam?hita of the latter consists entirely of the verses and formulas to be recited at the sacrifice, and is therefore clear (Çukla), that is to say, separated from the explanatory matter which is collected in the Brahma?a. Hence it is called the White (Çukla) Yajurveda, while the others, under the general name of Black (k?ish?a) Yajurveda, are contrasted with it, as containing both kinds of matter mixed up in the Sam?hita. The text of the Vajasaneyins has been preserved in two recensions, that of the Madhyam?dinas and of the Ka?vas. These are almost identical in their subject-matter as well as its arrangement. Their divergences hardly go beyond varieties of reading, which, moreover, appear only in their prose formulas, not in their verses. Agreeing thus closely, they cannot be separated in their origin by any wide interval of time. Their discrepancies probably arose rather from geographical separation, since each has its own peculiarities of spelling. The White Yajurveda in both these recensions has been edited by Professor Weber (1849–52).

It is divided into forty chapters, called adhyayas. That it originally consisted of the first eighteen alone is indicated by external as well as internal evidence. This is the only portion containing verses and prose formulas (both having the common name of mantras) which recur in the Taittiriya Sam?hita, the sole exceptions being a few passages relating to the horse-sacrifice in chapters 22–25. Otherwise the contents of the last twenty-two chapters are found again only in the Brahma?a and the Ara?yaka belonging to the Taittiriya Sam?hita. Moreover, it is only the mantras of the first eighteen chapters of the Vajasaneyi Sam?hita which are quoted and explained word by word in the first nine books of its own Brahma?a, while merely a few mantras from the following seventeen chapters are mentioned in that work. According to the further testimony of an ancient index of the White Yajurveda, attributed to Katyayana, the ten chapters 26–35 form a supplement (khila).

The internal evidence of the Vajasaneyi Sam?hita leads to similar conclusions. The fact that chapters 26–29 contain mantras relating to ceremonies dealt with in previous chapters and requiring to be applied to those ceremonies, is a clear indication of their supplementary character. The next ten chapters (30–39) are concerned with altogether new ceremonies, such as the human sacrifice, the universal sacrifice, and the sacrifice to the Manes. Lastly, the 40th chapter must be a late addition, for it stands in no direct relation to the ritual and bears the character of an Upanishad. Different parts of the Sam?hita, moreover, furnish some data pointing to different periods of religious and social development. In the 16th chapter the god Rudra is described by a large number of epithets which are subsequently peculiar to Çiva. Two, however, which are particularly significant, IÇana, “Ruler,” and Mahadeva, “Great God,” are absent here, but are added in the 39th chapter. These, as indicating a special worship of the god, represent a later development. Again, the 30th chapter specifies most of the Indian mixed castes, while the 16th mentions only a few of them. Hence, it is likely that at least some which are known to the former chapter did not as yet exist when the latter was composed.

On these grounds four chronological strata may be distinguished in the White Yajurveda. To the fundamental portion, comprising chapters 1–18, the next seven must first have been added, for these two parts deal with the general sacrificial ceremonial. The development of the ritual led to the compilation of the next fourteen chapters, which are concerned with ceremonies already treated (26–29) or entirely new (30–39). The last chapter apparently dates from a period when the excessive growth of ritual practices led to a reaction. It does not supply sacrificial mantras, but aims at establishing a mean between exclusive devotion to and total neglect of the sacrificial ceremonies.

Even the original portion of the White Yajurveda must have assumed shape somewhat later than any of the recensions of the Black. For the systematic and orderly distribution of matter by which the mantras are collected in the Sam?hita, while their dogmatic explanation is entirely relegated to a Brahma?a, can hardly be as old as the confused arrangement in which both parts are largely mixed up.

The two most important portions of the Yajurvedas deal with the new and full moon sacrifices, as well as the soma sacrifice, on the one hand, and with the construction of the fire-altar on the other. Chapters 1–10 of the White Yajurveda contain the mantras for the former, chapters 11–18 those for the latter part of the ceremonial. The corresponding ritual explanations are to be found in books 1–5 and 6–9 respectively of the Çatapatha Brahma?a. In these fundamental portions even the Black Yajurveda does not intermingle the mantras with their explanations. The first book of the Taittiriya Sam?hita contains in its first four lessons nothing but the verses and formulas to be recited at the fortnightly and the soma sacrifices; the fourth book, nothing but those employed in the fire-altar ritual. These books follow the same order as, and in fact furnish a parallel recension of, the corresponding parts of the Vajasaneyi Sam?hita. On the other hand, the Taittiriya Sam?hita contains within itself, but in a different part, the two corresponding Brahma?as, which, on the whole, are free from admixture with mantras. The fifth book is the Brahma?a of the fire ritual, and the sixth is that of the soma sacrifice; but the dogmatic explanation of the new and full moon sacrifice is altogether omitted here, being found in the third book of the Taittiriya Brahma?a. In the Maitraya?i Sam?hita the distribution of the corresponding material is similar. The first three lessons of the first book contain the mantras only for the fortnightly and the soma sacrifices; the latter half of the second book (lessons 7–13), the mantras only for the fire ritual. The corresponding Brahma?as begin with the sixth and the first lesson respectively of the third book. It is only in the additions to these fundamental parts of the Black Yajurveda that the separation of Mantra and Brahma?a is not carried out. The main difference, then, between the Black and the White consists in the former combining within the same collection Brahma?a as well as Mantra matter. As to its chief and fundamental parts, there is no reason to suppose that these two kinds of matter, which are kept separate and unmixed, are either chronologically or essentially more nearly related than are the Vajasaneyi Sam?hita and the Çatapatha Brahma?a.

The Yajurveda resembles the Samaveda in having been compiled for application to sacrificial rites only. But while the Samaveda deals solely with one part of the ritual, the soma sacrifice, the Yajurveda supplies the formulas for the whole sacrificial ceremonial. Like the Samaveda, it is also connected with the Rigveda; but while the former is practically altogether extracted from the Rigveda, the Yajurveda, though borrowing many of its verses from the same source, is largely an original production. Thus somewhat more than one-fourth only of the Vajasaneyi Sam?hita is derived from the Rigveda, One half of this collection consists of verses (?ich) most of which (upwards of 700) are found in the Rigveda; the other half is made up of prose formulas (yajus). The latter, as well as the verses not borrowed from the Rigveda, are the independent creation of the composers of the Yajurveda. This partial originality was indeed a necessary result of the growth of entirely new ceremonies and the extraordinary development of ritual detail. It became impossible to obtain from the Rigveda even approximately suitable verses for these novel requirements.

The language of the Mantra portion of the Yajurveda, though distinctly representing a later stage, yet on the whole agrees with that of the Rigveda, while separated from that of classical Sanskrit by a considerable interval.

On its mythological side the religion of the Yajurveda does not differ essentially from that of the older Veda; for the pantheon is still the same. Some important modifications in detail are, however, apparent. The figure of Prajapati, only foreshadowed in the latest hymns of the Rigveda, comes more and more into the foreground as the chief of the gods. The Rudra of the Rigveda has begun to appear on the scene as Çiva, being several times mentioned by that name as well as by other epithets later peculiar to Çiva, such as Çankara and Mahadeva. Vish?u now occupies a somewhat more prominent position than in the Rigveda. A new feature is his constant identification with the sacrifice. The demons, now regularly called Asuras, perpetually appear as a group of evil beings opposed to the good gods. Their conflicts with the latter play a considerable part in the myths of the Yajurveda. The Apsarases, who, as a class of celestial nymphs endowed with all the seductive charms of female beauty, occupy so important a place in post-Vedic mythology, but are very rarely mentioned in the Rigveda, begin to be more prominent in the Yajurveda, in which many of them are referred to by individual names.

Certain religious conceptions have, moreover, been modified and new rites introduced. Thus the word brahma, which in the Rigveda meant simply “devotion,” has come to signify the essence of prayer and holiness, an advance towards its ultimate sense in the Upanishads. Again, snake-worship, which is unknown to the Rigveda, now appears as an element in Indian religion. That, however, which impresses on the Yajurveda the stamp of a new epoch is the character of the worship which it represents. The relative importance of the gods and of the sacrifice in the older religion has now become inverted. In the Rigveda the object of devotion was the gods, for the power of bestowing benefits on mankind was believed to lie in their hands alone, while the sacrifice was only a means of influencing their will in favour of the offerer. In the Yajurveda the sacrifice itself has become the centre of thought and desire, its correct performance in every detail being all-important. Its power is now so great that it not merely influences, but compels the gods to do the will of the officiating priest. By means of it the Brahmans may, in fact, be said to hold the gods in their hands.

The religion of the Yajurveda may be described as a kind of mechanical sacerdotalism. A crowd of priests conducts a vast and complicated system of external ceremonies, to which symbolical significance is attributed, and to the smallest minutiÆ of which the greatest weight is attached. In this stifling atmosphere of perpetual sacrifice and ritual, the truly religious spirit of the Rigveda could not possibly survive. Adoration of the power and beneficence of the gods, as well as the consciousness of guilt, is entirely lacking, every prayer being coupled with some particular rite and aiming solely at securing material advantages. As a natural result, the formulas of the Yajurveda are full of dreary repetitions or variations of the same idea, and abound with half or wholly unintelligible interjections, particularly the syllable om. The following quotation from the Maitraya?i Sam?hita is a good example: Nidhayo va nidhayo va om? va om? va om? va e ai om? svar?ajyotih?. Here only the last word, which means “golden light,” is translatable.

Thus the ritual could not fail to become more and more of a mystery to all who did not belong to the Brahman caste. To its formulas, no less than to the sacrifice itself, control over Nature as well as the supernatural powers is attributed. Thus there are certain formulas for the obtainment of victory; by means of these, it is said, Indra constantly vanquished the demons. Again, we learn that, if the priest pronounces a formula for rain while mixing a certain offering, he causes the rain to stream down. Hence the formulas are regarded as having a kind of magical effect by exercising compulsion. Similar miraculous powers later came to be attached to penance and asceticism among the Brahmans, and to holiness among the Buddhists. The formulas of the Yajurveda have not, as a rule, the form of prayers addressed to the gods, but on the whole and characteristically consist of statements about the result of employing particular rites and mantras. Together with the corresponding ritual they furnish a complex mass of appliances ready to hand for the obtainment of material welfare in general as well as all sorts of special objects, such as cattle or a village. The presence of a priest capable of using the necessary forms correctly is of course always presupposed. The desires which several rites are meant to fulfil amount to nothing more than childish absurdity. Thus some of them aim at the obtainment of the year. Formulas to secure possession of the moon would have had equal practical value.

Hand in hand with the elaboration of the sacrificial ceremonial went the growth and consolidation of the caste system, in which the Brahmans secured the social as well as the religious supremacy, and which has held India enchained for more than two thousand five hundred years. Not only do we find the four castes firmly established as the main divisions of Indian society in the Yajurveda, but, as one of the later books of the Vajasaneyi Sam?hita shows, most of the mixed castes known in later times are already found to exist. The social as well as the religious conditions of the Indian people, therefore, now wear an aspect essentially differing from those revealed to us in the hymns of the Rigveda.

The Rig-, Sama-, and Yajur-vedas alone were originally recognised as canonical collections. For they only were concerned with the great sacrificial ceremonial. The Atharva-veda, with the exception of the last book, which was obviously added in order to connect it with that ceremonial, is essentially unconnected with it. The ceremonial to which its hymns were practically applied is, with few exceptions, that with which the G?ihya Sutras deal, being domestic rites such as those of birth, marriage, and death, or the political rites relating to the inauguration of kings. Taken as a whole, it is a heterogeneous collection of spells. Its most salient teaching is sorcery, for it is mainly directed against hostile agencies, such as diseases, noxious animals, demons, wizards, foes, oppressors of Brahmans. But it also contains many spells of an auspicious character, such as charms to secure harmony in family and village life, reconciliation of enemies, long life, health, and prosperity, besides prayers for protection on journeys, and for luck in gambling. Thus it has a double aspect, being meant to appease and bless as well as to curse.

In its main contents the Atharva-veda is more superstitious than the Rigveda. For it does not represent the more advanced religious beliefs of the priestly class, but is a collection of the most popular spells current among the masses, who always preserve more primitive notions with regard to demoniac powers. The spirit which breathes in it is that of a prehistoric age. A few of its actual charms probably date with little modification from the Indo-European period; for, as Adalbert Kuhn has shown, some of its spells for curing bodily ailments agree in purpose and content, as well as to some extent even in form, with certain old German, Lettic, and Russian charms. But with regard to the higher religious ideas relating to the gods, it represents a more recent and advanced stage than the Rigveda. It contains, indeed, more theosophic matter than any of the other Sam?hitas. For the history of civilisation it is on the whole more interesting and important than the Rigveda itself.

The Atharva-veda is extant in the recensions of two different schools. That of the Paippaladas is, however, known in a single birch-bark manuscript, which is ancient but inaccurate and mostly unaccented. It was discovered by Professor BÜhler in Kashmir, and has been described by Professor Roth in his tract Der Atharvaveda in Kaschmir (1875). It will probably soon be accessible to scholars in the form of a photographic reproduction published by Professor Bloomfield. This recension is doubtless meant by the “Paippalada Mantras” mentioned in one of the PariÇish?as or supplementary writings of the Atharva-veda.

The printed text, edited by Roth and Whitney in 1856, gives the recension of the Çaunaka school. Nearly the whole of Saya?a’s commentary to the Atharva-veda has been edited in India. Its chief interest lies in the large number of readings supplied by it which differ from those of the printed edition of this Veda.

This Sam?hita is divided into twenty books, containing 730 hymns and about 6000 stanzas. Some 1200 of the latter are derived from the Rigveda, chiefly from the tenth, first, and eighth books, a few also from each of the other books. Of the 143 hymns of Book XX., all but twelve are taken bodily from the established text of the Rigveda without any change. The matter borrowed from the Rigveda in the other books shows considerable varieties of reading, but these, as in the other Sam?hitas, are of inferior value compared with the text of the Rigveda. As is the case in the Yajurveda, a considerable part of the Atharva (about one-sixth) consists of prose. Upwards of fifty hymns, comprising the whole of the fifteenth and sixteenth, besides some thirty hymns scattered in the other books, are entirely unmetrical. Parts or single stanzas of over a hundred other hymns are of a similar character.

That the Atharva-veda originally consisted of its first thirteen books only is shown both by its arrangement and by its subject-matter. The contents of Books I.–VII. are distributed according to the number of stanzas contained in the hymns. In Book I. they have on the average four stanzas, in II. five, in III. six, in IV. seven, in V. eight to eighteen, in VI. three; and in VII. about half the hymns have only one stanza each. Books VIII.–XIII. contain longer pieces. The contents of all these thirteen books are indiscriminately intermingled.

The following five books, on the contrary, are arranged according to uniformity of subject-matter. Book XIV. contains the stanzas relating to the wedding rite, which consist largely of mantras from the tenth book of the Rigveda. Book XV. is a glorification of the Supreme Being under the name of Vratya, while XVI. and XVII. contain certain conjurations. The whole of XV. and nearly the whole of XVI., moreover, are composed in prose of the type found in the Brahma?as. Both XVI. and XVII. are very short, the former containing nine hymns occupying four printed pages, the latter consisting of only a single hymn, which extends to little more than two pages. Book XVIII. deals with burial and the Manes. Like XIV., it derives most of its stanzas from the tenth book of the Rigveda. Both these books are, therefore, not specifically Atharvan in character.

The last two books are manifestly late additions. Book XIX. consists of a mixture of supplementary pieces, part of the text of which is rather corrupt. Book XX., with a slight exception, contains only complete hymns addressed to Indra, which are borrowed directly and without any variation from the Rigveda. The fact that its readings are identical with those of the Rigveda would alone suffice to show that it is of later date than the original books, the readings of which show considerable divergences from those of the older Veda. There is, however, more convincing proof of the lateness of this book. Its matter relates to the Soma ritual, and is entirely foreign to the spirit of the Atharva-veda. It was undoubtedly added to establish the claim of the Atharva to the position of a fourth Veda, by bringing it into connection with the recognised sacrificial ceremonial of the three old Vedas. This book, again, as well as the nineteenth, is not noticed in the PratiÇakhya of the Atharva-veda. Both of them must, therefore, have been added after that work was composed. Excepting two prose pieces (48 and 49) the only original part of Book XX. is the so-called kuntapa hymns (127–136). These are allied to the danastutis of the Rigveda, those panegyrics of liberal kings or sacrificers which were the forerunners of epic narratives in praise of warlike princes and heroes.

The existence of the Atharva, as a collection of some kind, when the last books of the Çatapatha Brahma?a (xi., xiii., xiv.), the Taittiriya Brahma?a, and the Chhandogya Upanishad were composed, is proved by the references to it in those works. In Patanjali’s Mahabhashya the Atharva had already attained to such an assured position that it is even cited at the head of the Vedas, and occasionally as their only representative.

The oldest name of this Veda is Atharvangirasah?, a designation occurring in the text of the Atharva-veda, and found at the beginning of its MSS. themselves. This word is a compound formed of the names of two ancient families of priests, the Atharvans and Angirases. In the opinion of Professor Bloomfield the former term is here synonymous with “holy charms,” as referring to auspicious practices, while the latter is an equivalent of “witchcraft charms.” The term atharvan and its derivatives, though representing only its benevolent side, would thus have come to designate the fourth Veda as a whole. In its plural form (atharva?ah?) the word in this sense is found several times in the Brahma?as, but in the singular it seems first to occur in an Upanishad. The adjective atharva?a, first found as a neuter plural with the sense of “Atharvan hymns” in the Atharva-veda itself (Book XIX.), is common from that time onwards. The name atharva-veda first appears in Sutras about as early as ?igveda and similar designations of the other Sam?hitas. There are besides two other names of the Atharva-veda, the use of which is practically limited to the ritual texts of this Veda. In one of these, Bh?igu-angirasah?, the name of another ancient family of fire-priests, the Bh?igus, takes the place of that of the Angirases. The other, brahma-veda, has outside the Atharvan literature only been found once, and that in a G?ihya Sutra of the Rigveda.

A considerable time elapsed before the Atharva-veda, owing to the general character of its contents, attained to the rank of a canonical book. There is no evidence that even at the latest period of the Rigveda the charms constituting the Atharva-veda were formally recognised as a separate literary category. For the Purusha hymn, while mentioning the three sacrificial Vedas by the names of Rik, Saman, and Yajus, makes no reference to the spells of the Atharva-veda. Yet the Rigveda, though it is mainly concerned with praises of the gods in connection with the sacrifice, contains hymns showing that sorcery was bound up with domestic practices from the earliest times in India. The only reference to the spells of the Atharva-veda as a class in the Yajurvedas is found in the Taittiriya Sam?hita, where they are alluded to under the name of angirasah? by the side of Rik, Saman, and Yajus, which it elsewhere mentions alone. Yet the formulas of the Yajur-veda are often pervaded by the spirit of the Atharva-veda, and are sometimes Atharvan even in their wording. In fact, the difference between the Rigveda and Yajurveda on the one hand, and the Atharva on the other, as regards sorcery, lies solely in the degree of its applicability and prominence.

The Atharva-veda itself only once mentions its own literary type directly (as atharvangirasah?) and once indirectly (as bheshaja or “auspicious spells”), by the side of the other three Vedas, while the latter in a considerable number of passages are referred to alone. This shows that as yet there was no feeling of antagonism between the adherents of this Veda and those of the older ones.

Turning to the Brahma?as, we find that those of the Rigveda do not mention the Atharva-veda at all, while the Taittiriya Brahma?a (like the Taittiriya Ara?yaka) refers to it twice. In the Çatapatha Brahma?a it appears more frequently, occupying a more defined position, though not that of a Veda. This work very often mentions the three old Vedas alone, either explicitly as Rik, Saman, Yajus, or as trayi vidya, “the threefold knowledge.” In several passages they are also mentioned along with other literary types, such as itihasa (story), pura?a (ancient legend), gatha (song), sutra, and upanishad. In these enumerations the Atharva-veda regularly occupies the fourth place, coming immediately after the three Vedas, while the rest follow in varying order. The Upanishads in general treat the Atharva-veda in the same way; the Upanishads of the Atharva itself, however, sometimes tacitly add its name after the three Vedas, even without mentioning other literary types. With regard to the Çrauta or sacrificial Sutras, we find no reference to the Atharva in those of Katyayana (White Yajurveda) or La?yayana (Samaveda), and only one each in those of Çankhayana and AÇvalayana (Rigveda).

In all this sacrificial literature there is no evidence of repugnance to the Atharva, or of exclusiveness towards it on the part of followers of the other Vedas. Such an attitude could indeed hardly be expected. For though the sphere of the Vedic sacrificial ritual was different from that of regular magical rites, it is impossible to draw a distinct line of demarcation between sacrifice and sorcery in the Vedic religion, of which witchcraft is, in fact, an essential element. The adherents of the three sacrificial Vedas would thus naturally recognise a work which was a repository of witchcraft. Thus the Çatapatha Brahma?a, though characterising yatu or sorcery as devilish—doubtless because it may be dangerous to those who practise it—places yatuvidah? or sorcerers by the side of bahv?ichas or men skilled in Rigvedic verses. Just as the Rigveda contains very few hymns directly connected with the practice of sorcery, so the Atharva originally included only matters incidental and subsidiary to the sacrificial ritual. Thus it contains a series of formulas (vi. 47–48) which have no meaning except in connection with the three daily pressings (savana) of soma. We also find in it hymns (e. g. vi. 114) which evidently consist of formulas of expiation for faults committed at the sacrifice. We must therefore conclude that the followers of the Atharva to some extent knew and practised the sacrificial ceremonial before the conclusion of the present redaction of their hymns. The relation of the Atharva to the Çrauta rites was, however, originally so slight, that it became necessary, in order to establish a direct connection with it, to add the twentieth book, which was compiled from the Rigveda for the purposes of the sacrificial ceremonial.

The conspicuous way in which Çrauta works ignore the Atharva is therefore due to its being almost entirely unconnected with the subject-matter of the sacrifice, not to any pronounced disapproval or refusal to recognise its value in its own sphere. With the G?ihya or Domestic Sutras, which contain many elements of sorcery practice (vidhana), we should expect the Atharva to betray a closer connection. This is, indeed, to some extent the case; for many verses quoted in these Sutras are identical with or variants of those contained in the Atharva, even though the Domestic, like the Sacrificial, Sutras endeavoured to borrow their verses as far as possible from the particular Veda to which they were attached. Otherwise, however, their references to the Atharva betray no greater regard for it than those in the Sacrificial Sutras do. Such references to the fourth Veda are here, it is true, more frequent and formulaic; but this appears to mean nothing more than that the G?ihya Sutras belong to a later date.

In the sphere, too, of law (dharma), as dealing with popular usage and custom, the practices of the Atharva maintained a certain place; for the indispensable sciences of medicine and astrology were distinctively Atharvan, and the king’s domestic chaplain (purohita), believed capable of rendering great services in the injury and overthrow of enemies by sorcery, seems usually to have been an Atharvan priest. At the same time it is only natural that we should first meet with censures of the practices of the Atharva in the legal literature, because such practices were thought to enable one man to harm another. The verdict of the law treatises on the whole is, that as incantations of various kinds are injurious, the Atharva-veda is inferior and its practices impure. This inferiority is directly expressed in the Dharma Sutra of Apastamba; and the later legal treatise (sm?iti) of Vish?u classes the reciter of a deadly incantation from the Atharva among the seven kinds of assassins. Physicians and astrologers are pronounced impure; practices with roots are prohibited; sorceries and imprecations are punished with severe penances. In certain cases, however, the Atharva-veda is stated to be useful. Thus the Lawbook of Manu recommends it as the natural weapon of the Brahman against his enemies.

In the Mahabharata we find the importance and the canonical character of the Atharva fully recognised. The four Vedas are often mentioned, the gods Brahma and Vish?u being in several passages described as having created them. The Atharva is here often also referred to alone, and spoken of with approbation. Its practices are well known and seldom criticised adversely, magic and sorcery being, as a rule, regarded as good.

Finally, the Pura?as not only regularly speak of the fourfold Veda, but assign to the Atharva the advanced position claimed for it by its own ritual literature. Thus the Vish?u Pura?a connects the Atharva with the fourth priest (the brahman) of the sacrificial ritual.

Nevertheless a certain prejudice has prevailed against the Atharva from the time of the Dharma Sutras. This appears from the fact that, even at the present day, according to Burnell, the most influential Brahmans of Southern India still refuse to accept the authority of the fourth Veda, and deny its genuineness. A similar conclusion may be drawn from occasional statements in classical texts, and especially from the efforts of the later Atharvan writings themselves to vindicate the character of their Veda. These ritual texts not only never enumerate the Vedas without including the Atharva, but even sometimes place it at the head of the four Vedas. Under a sense of the exclusion of their Veda from the sphere of the sacrificial ritual, they lay claim to the fourth priest (the brahman), who in the Vedic religion was not attached to any of the three Vedas, but being required to have a knowledge of all three and of their sacrificial application, acted as superintendent or director of the sacrificial ceremonial. Ingeniously availing themselves of the fact that he was unconnected with any of the three Vedas, they put forward the claim of the fourth Veda as the special sphere of the fourth priest. That priest, moreover, was the most important as possessing a universal knowledge of religious lore (brahma), the comprehensive esoteric understanding of the nature of the gods and of the mystery of the sacrifice. Hence the Gopatha Brahma?a exalts the Atharva as the highest religious lore (brahma), and calls it the Brahmaveda. The claim to the latter designation was doubtless helped by the word brahma often occurring in the Atharva-veda itself with the sense of “charm,” and by the fact that the Veda contains a larger amount of theosophic matter (brahmavidya) than any other Sam?hita. The texts belonging to the other Vedas never suggest that the Atharva is the sphere of the fourth priest, some Brahma?a passages expressly declaring that any one equipped with the requisite knowledge maybe a brahman. The ritual texts of the Atharva further energetically urged that the Purohita, or domestic chaplain, should be a follower of the Atharva-veda. They appear to have finally succeeded in their claim to this office, doubtless because kings attached great value to a special knowledge of witchcraft.

The geographical data contained in the Atharva are but few, and furnish no certain evidence as to the region in which its hymns were composed. One hymn of its older portion (v. 22) makes mention of the Gandharis, Mujavats, Mahav?ishas, and Balhikas (in the north-west), and the Magadhas and Angas (in the east); but they are referred to in such a way that no safe conclusions can be drawn as to the country in which the composer of the hymn in question lived.

The Atharva also contains a few astronomical data, the lunar mansions being enumerated in the nineteenth book. The names here given deviate considerably from those mentioned in the Taittiriya Sam?hita, appearing mostly in a later form. The passage in which this list is found is, however, a late addition.

The language of the Atharva is, from a grammatical point of view, decidedly later than that of the Rigveda, but earlier than that of the Brahma?as. In vocabulary it is chiefly remarkable for the large number of popular words which it contains, and which from lack of opportunity do not appear elsewhere.

It seems probable that the hymns of the Atharva, though some of them must be very old, were not edited till after the Brahma?as of the Rigveda were composed.

On examining the contents of the Atharva-veda more in detail, we find that the hostile charms it contains are directed largely against various diseases or the demons which are supposed to cause them. There are spells to cure fever (takman), leprosy, jaundice, dropsy, scrofula, cough, ophthalmia, baldness, lack of vital power; fractures and wounds; the bite of snakes or injurious insects, and poison in general; mania and other ailments. These charms are accompanied by the employment of appropriate herbs. Hence the Atharva is the oldest literary monument of Indian medicine.

The following is a specimen of a charm against cough (vi. 105):—

Here is a spell for the cure of leprosy by means of a dark-coloured plant:—

A large number of imprecations are directed against demons, sorcerers, and enemies. The following two stanzas deal with the latter two classes respectively:—

A considerable group of spells consists of imprecations directed against the oppressors of Brahmans and those who withhold from them their rightful rewards. The following is one of the threats held out against such evil-doers:—

Another group of charms is concerned with women, being intended to secure their love with the aid of various potent herbs. Some of them are of a hostile character, being meant to injure rivals. The following two stanzas belong to the former class:—

Among the auspicious charms of the Atharva there are many prayers for long life and health, for exemption from disease and death:—

Another class of hymns includes prayers for protection from dangers and calamities, or for prosperity in the house or field, in cattle, trade, and even gambling. Here are two spells meant to secure luck at play:—

A certain number of hymns contain charms to secure harmony, to allay anger, strife, and discord, or to procure ascendency in the assembly. The following one is intended for the latter purpose:—

A few hymns consist of formulas for the expiation of sins, such as offering imperfect sacrifices and marrying before an elder brother, or contain charms for removing the defilement caused by ominous birds, and for banishing evil dreams.

A short hymn (vi. 120), praying for the remission of sins, concludes with this stanza:—

Another group of hymns has the person of the king as its centre. They contain charms to be used at a royal election or consecration, for the restoration of an exiled king, for the attainment of lustre and glory, and in particular for victory in battle. The following is a specimen of spells intended to strike terror into the enemy:—

Here is a stanza from a hymn (v. 21, 6) to the battle-drum meant to serve the same purpose:—

Among the cosmogonic and theosophic hymns the finest is a long one of sixty-three stanzas addressed to the earth (xii. 1). I translate a few lines to give some idea of its style and contents:—

The four hymns of Book XIII. are devoted to the praise of Rohita, the “Red” Sun, as a cosmogonic power. In another (xi. 5) the sun is glorified as a primeval principle under the guise of a Brahman disciple (brahmacharin). In others Pra?a or Breath (xi. 4), Kama or Love (ix. 2), and Kala or Time (xix. 53–54), are personified as primordial powers. There is one hymn (xi. 7) in which even Ucchish?a (the remnant of the sacrifice) is deified as the Supreme Being; except for its metrical form it belongs to the Brahma?a type of literature.

In concluding this survey of the Atharva-veda, I would draw attention to a hymn to Varu?a (iv. 16); which, though its last two stanzas are ordinary Atharvan spells for binding enemies with the fetters of that deity, in its remaining verses exalts divine omniscience in a strain unequalled in any other Vedic poem. The following three stanzas are perhaps the best:—