Vedic Metre in its Historical Development

E.V. Arnold

[p. 5]


13. Our single first-hand authority for the text of the Rigveda is the “Saṁhitā” recension, which has been preserved from an unknown antiquity without any variants of importance. The “Pada-pāṭha” text is now acknowledged to be an early commentary upon the Saṁhitā, and cannot be set in opposition to its authority. Other Vedic texts repeat the matter of many hymns of the Rigveda in slightly differing forms, but in such cases the Saṁhitā text generally appears to have the advantage, and it may be inferred with some confidence that the other texts were historically derived from it.

The Saṁhitā text can therefore only be revised upon internal evidence, that of the metre being the most important.

14. The following restorations are amongst those generally agreed upon:—

(i) where the text, in accordance with the classical rules of Sandhi, combines the final vowel of one word with the initial vowel of the next, final a ā must occasionally, and final i ī u ū generally, be read as separate syllables.

(ii) where the text omits initial a after final -aḥ or -e, the initial a must usually be restored as a separate syllable.

(iii) in numerous words and endings the value of a separate syllable is, either necessarily or optionally, to be given to y v of the text: e.g. *náriyaḥ is to be read for náryaḥ, *tuvám for tvám.

(Arnold's footnote: The technical spellings náriaḥ, tuám indicate the occasions on which such restorations are required. Though not in accordance with Indian practice, they have the advantage that they cannot be mistaken for Saṁhitā readings.)

(iv) in a few words long vowels or diphthongs are optionally to be read as equivalent to two syllables: thus śréṣṭhaḥ as * śráyiṣṭhaḥ, gop as gopáā.

(v) A few words are regularly misrepresented: thus for pāvaká we must always read pavāká, for suvānáḥ almost always svānáḥ, for mḷaya always mḷaya.

15. In other cases restoration is less certain, because the metre itself is open to doubt. The following are the most important:—

[p. 6]

(i) in numerous verses some part of the noun índra follows the cæsura, and the number of syllables in the verse appears to be ten, whereas most verses in the same hymn have eleven syllables. If we interpret índra metrically as a word of three syllables, e.g. índara, the verses in which the word occurs contain the normal number of syllables. But on the other hand índra is commonly a disyllabic, and verses of ten syllables are sufficiently common in the Rigveda to suggest a metrical solution of the difficulty.

(ii) in numerous verses ending with some genitive plural form in -ām there appear to be only seven syllables, but if we read instead -aam the regular number of eight is obtained. But verses of seven syllables are in many such hymns fairly common, and these verses may be, and are by some critics, so interpreted.

(iii) the final vowel of many flexional endings and of several adverbs is given by the text sometimes as short, sometimes as long. The distinction to a large extent follows the preferences of the metrical rhythm: and it is not easy to determine whether the quantity given represents the true value of the ending, or an artificial pronunciation by which a slight irregularity of metre is glozed [glossed] over. In these cases the Pada-pāṭha usually gives the short value, but this is only evidence of pronunciation of the word at the time this commentary was composed.

16. In the determination of the text of the Rigveda that of the metrical value of words and syllables is implicitly contained. The metrical value of a word depends solely upon the number and succession of the short and long syllables which it contains, without regard to the position of the accent. A syllable containing a long vowel or diphthong is necessarily long, the only doubt arising when the long vowel or diphthong is final and the next word in the same verse begins with a vowel. In such cases the quantity of a final diphthong is that of its prior element, as is shewn by a correct text in dáma á (for dáme á), tasmā adāt (for tāsmai adāt): but the quantity of a long final vowel when not combined by Sandhi is not always easy to determine.

A syllable the vowel of which is followed by two consonants is long. One or both of these consonants may belong to a word following, provided that it belongs to the same verse. The sounds represented by ṁ ḥ are for this purpose recognized as full [p. 7] consonants: and the Vedic lh (also variously represented ḷh or ḍh) is equivalent to a double consonant.

17. The following stanza will illustrate the general character of metrical restoration in the Rigveda:—

(i) Saṁhitā text (following Aufrecht's edition):—

vayám adyéndrasya préṣṭhā vayáṁ śvó vocemahi samaryé ¦ vayám pur máhi ca no ánu dyn tán na bhukṣ narm ánu ṣyāt ¦¦

(ii) Text metrically restored:

vayám ady índarasya práyiṣṭhāḥ

vayáṁ śua vocemahi samaryé ¦

vayám pur máhi ca na ánu dyn

tán na bhukṣ naráam ánu ṣyāt ¦¦

i 167 10.



18. The units of Vedic metre are the ‘ verse,’ the ‘stanza,’ and the hymn.’

19. A verse consists most commonly either of eight syllables, when we distinguish it as a dimeter verse: or of eleven or twelve syllables, both of which varieties are included under the name trimeter verse.

20. The number of syllables in a verse is not quite rigidly prescribed. Thus many dimeter verses contain seven syllables only: such verses, if they correspond in rhythm to an ordinary dimeter verse with loss, of the last syllable, we term catalectic dimeter verses: if otherwise, the more general name heptasyllabic may be used.

21. Similarly, trimeter verses which contain only ten syllables are not uncommon, such verses being usually equivalent to the verse of eleven syllables with the loss of a syllable before or after the cæsura. These shorter verses may be termed decasyllbic.

22. The most typical forms of the stanza are

(i) the Anuṣṭubh, which consists of four dimeter verses, and

(ii) the Triṣṭubh, which consists of four trimeter verses, each of eleven syllables.

Four trimeter verses, each of twelve syllables, form a Jagatī stanza.

[p. 8]

23. Stanzas may contain more or fewer verses than four. Thus the Gāyatrī consists of three dimeter verses, the Paṅkti of five, the Mahāpaṅkti of six: whilst three Triṣṭubh verses form the metre known as Virāj, and two decasyllabic verses that known as Dvipadā Virāj.

24. Stanzas may also consist of combinations of dimeter and trimeter verses, the latter being then usually of twelve syllables: all metres of this type we group as lyric metres. Lyric metres may also include verses of four syllables, and even of sixteen, but these are comparatively rare. The most important lyric metres are Uṣṇiḥ (8. 8. 12 or 8. 8. 8. 4), Kakubh (8. 12. 8). Bhatī (8. 8. 12. 8), Satobhatī (12. 8. 12. 8), and Atyaṣṭi (12. 12. 8, 8. 8, 12. 8).

25. The strophe (pragātha) is comparatively rare in the Rigveda, and consists of the combination of a Kakubh or Bhatī stanza with a Satobhatī stanza.

26. A few hymns are composed of lyric stanzas of very various structures: these we call hymns in mixed lyric metres.

27. The rare combination of verses of eight and eleven syllables in one stanza gives uneven lyric metre.

28. A hymn may contain any number of stanzas, but usually it consists of not less than three or more than fifteen stanzas, generally uniform in character, except in the case of strophic and mixed lyric hymns. It is also not uncommon for the last stanza of a hymn to contain, one or even two additional verses.

29. Where the number of stanzas in a hymn is very large, or the metre suddenly changes, it becomes probable that we have a composite hymn, that is, two or more hymns combined in the Saṁhitā text.

The number of such composite hymns in the Rigveda is very large: but the hymns thus combined are usually homogeneous in character, and, if so, it is not necessary for the present purpose to separate them. If however, the suspicion arises that one of the hymns thus combined differs in date or character from another to which it is joined, it becomes at once expedient to treat them as separate hymns.

30. The definitions just given may be said to determine the external form of the different metres, and describe their more [p. 9] obvious characteristics. In accordance with them the metres may be grouped as belonging respectively to (a) the lyric, (b) the Anuṣṭubh, (c) the Triṣṭubh-Jagatī families respectively. The lyric metres being practically unknown in later literature, may be presumed to belong to the more distinctively early parts of the Rigveda.

31. We proceed to consider the internal form, or rhythm, of the different kinds of verse. The internal form is even less rigidly defined than the external, and is treated by native commentators as of comparatively small importance. There are few parts of the verse in which the poets do not consider themselves free at times to depart from the usual rhythms, so that it may perhaps be said that there are no ‘rules’ of rhythm in the Rigveda. On the other hand, there is no considerable part of the verse in which certain rhythms are not steadily favoured, and others avoided: everywhere there exist metrical preferences. In consequence of the greater elasticity of the internal form, its study is more difficult, but at the same time yields results of greater historical importance than the study of external form.

32. In all metres in the Rigveda the quantities of the first and last syllables of each verse are indifferent, and (with some exceptions) each verse is independent in structure.

33. In almost all metres a general iambic rhythm may be noticed, in the sense that the even syllables, namely the second, fourth, and so on, are more often long than short. Hence it has been supposed that Vedic metres have arisen historically from some combination of iambic ‘feet,’ such as is found in so many Greek metres. This supposition meets with many difficulties, and the use of the terms ‘dimeter’ and ‘trimeter’ verse in this book is not intended to imply acceptance of any such historical theory.

34. In all metres the rhythm of the latter part of the verse is much more rigidly defined than that of the earlier part. Hence it has often, but incorrectly, been supposed that the earlier part of the verse is non-rhythmical.

35. In the early part of the verse there is a general preference for long syllables, in the latter part of the verse for short syllables. These preferences modify considerably the general iambic rhythm prevailing in both parts.

[p. 10]

36. The dimeter verse falls naturally into two parts, or members, each consisting of four syllables. There is no cæsura. We distinguish the two parts as the opening and the cadence.

37. In the opening the first syllable is indifferent, according to § 32. The second and fourth syllables are preferably long (§ 33), but often short (§ 34). As to the third syllable, the principles laid down in § 33, 35 appear each to neutralise the other, so that the quantity of this syllable is indifferent. But if the second syllable is short, the third is almost invariably long. This fact seems to imply some principle of transfer of metrical weight, which may be described by the term syncopation, used in a somewhat similar way in modern music.

38. In the cadence the fifth and seventh syllables are almost always short (§ 33-35), but it is not unusual to find a short syllable in the sixth place also (§ 35).

39. The following stanza, in which the division into parts or members is marked by upright lines, is therefore fairly typical of Vedic dimeter verse:—

tuáṁ valá¦sya g mataḥ ¦

ápāvar ad¦rivo bílam ¦

tuṁ dev ¦ ábibhyuṣas ¦

tujyámānā¦sa āviṣuḥ ¦¦

i 11 5.

[Ed: here and elsewhere I mark the end-syllable according to its natural weight.]

4O. In Gāyatrī stanzas preference is sometimes shewn for a trochaic rhythm in the cadence. Hymns in which such preference is fairly consistent are said to be in the Trochaic Gāyatrī metre. In the following stanza all the verses have a strict trochaic cadence:—

uktháṁ caná ¦ śasyámānam ¦

ágor arír ¦  ciketa ¦

ná gāyatráṁ ¦ gīyámānam ¦¦

viii 2 14.

41. An important variation of the Anuṣṭubh stanza is found in those parts of the Rigveda which have been already referred to (§ 10) as similar in character to the Atharvaveda, and is regular in the Atharvaveda itself.

In this variation the verses are grouped in pairs, so that the stanza consists of two distichs, and each distich of two dimeter [p. 11] verses. In each distich the cadence of the first verse and the opening of the second are modified in the direction of lengthening the third syllable.

In order to make clear the nature of this change, we give to the last part of the first and third verses of the stanza the special title semi-cadence, and to the opening of the second and fourth verses the title re-opening. The variation may then be defined as follows:—

The seventh syllable, which in the cadence is regularly short, becomes in ‘semi-cadence’ frequently long.

The third syllable, of which the quantity in the opening is indifferent, becomes almost invariably long in the ‘re-opening.’

To this metre the name of ‘later Anuṣṭubh’ has been given, on the ground that it is only found in the later additions to the Rigveda. Without questioning the truth of this view, it may be better to choose a more colourless epithet, and the name epic Anuṣṭubh is justified by the resemblance between this metre and that of the epic śloka.

The following stanza shews a fully developed ‘epic’ rhythm: but the Rigveda shews every gradation from the normal Anuṣṭubh rhythm to this.

vāyúr asmā ¦ úpāmanthat ¦

pináṣṭi smā ¦ kunannamá ¦

kes viṣá¦sya ptreṇa ¦

yád rudréṇá ¦ pibat sahá ¦¦

x 136 7.

42. Trimeter verse may be analyzed in two ways:—

(i) into two parts, as separated by the cæsura which regularly follows either the fourth or the fifth syllable: or

(ii) into three members, namely (a) the opening, which consists of the first four syllables; (b) the break, consisting of the fifth, sixth, and seventh syllables; and (c) the cadence, which includes the remaining syllables, beginning with the eighth.

43. In the case of Triṣṭubh verse, the two parts consist either of four and seven syllables, or of five and six syllables each, according as the cæsura is early (that is, after the fourth syllable), or late (that is, after the fifth syllable). If the cæsura is early, we have a short first part and a long second part: if late, we have a long first part and a short second part.

[p. 12]

In the case of Jagatī verse, the second part is in each case longer by one syllable.

In all cases the second part regularly begins with two short syllables.

44. The ‘opening’ of trimeter verse, if the cæsura is early, is identical in rhythm with that of dimeter verse. But if the cæsura is late, the iambic rhythm is much more strictly observed, and we have a further example of syncopation in the principle that if the fourth syllable is short, the fifth must be long.

4. The middle member of trimeter verse is termed the ‘break’ because the general iambic rhythm is broken at this point by the preference for two short syllables after the cæsura. The remaining syllable is usually long: so that the most regular forms are //⏑−⏑ or //⏑⏑, where the symbol // marks the position of the cæsura.

46. Sometimes the general iambic rhythm is not broken in the middle part of the verse, which therefore takes one of the forms //⏑−⏑ or //−⏑. Hymns in which such rhythms are prevalent, as is the case in particular in the seventh Maṇḍala, may be said to be in Iambic Triṣṭubh (or Jagatī) metre.

On the other hand, a still stronger contrast to the iambic rhythm is not uncommon, after an early cæsura, when the break takes the form of a cretic //−⏑−). Hymns in which this rhythm is favoured maybe said to be in Cretic Triṣṭubh (or Jagatī) metre.

The iambic and cretic variations are occasionally found in all parts of the Rigveda, but they are only to a small extent combined in the same hymns. Hymns in the lyric metres incline to the employment of iambic variations, which therefore appear to be characteristic of the more distinctively early parts of the Rigveda.

47. The ‘cadence’ of trimeter verse shews the same general rhythm as that of dimeter verse. Thus the ninth syllable, and in Jagatī verse the eleventh, are regularly short: the eighth and the tenth are regularly long, but either or both are occasionally short.

We are, however, able to trace a progressive tendency to eliminate the employment of short syllables in the eighth and tenth places. Thus it is only in a small part of the Rigveda that [p. 13] a short tenth syllable is at all common: whilst the short eighth syllable is chiefly found in the shape of a syllable containing a short final vowel. The employment of these final syllables in a position from which other short syllables are excluded requires careful investigation, and has evidently some connexion with the varying quantities which, as has been already noticed (§ 15 iii.), the Saṁhitā text assigns in many cases to final vowels.

48. The following stanzas illustrate the most important forms of Trimeter verse, the division of the ‘members’ being again shewn by perpendicular lines:—

Normal Triṣṭubh stanza:—

bhaspátiḥ // ¦ prathamáṁ ¦ jyamānaḥ ¦

mahó jyóti¦ṣaḥ // para¦mé víoman ¦

saptsias // ¦ tuvijā¦tó ráveṇa ¦

ví saptáraś¦mir // adha¦mat támāsi ¦¦

iv. 50 4.

Normal Jagatī stanza:—

ádadā ár¦bhām // maha¦té vacasyáve ¦

kakṣvate // ¦ vcaym ¦ indra sunvaté ¦

ménābhavo // ¦ vṣaṇaś¦vásya sukrato ¦

víśvét tá te // ¦ sávane¦ṣu pravciā ¦¦

i 51 13.

Triṣṭubh stanza with iambic variations:—

agníṁ vo de¦vám // agní¦bhiḥ sajóṣāḥ ¦

yájiṣṭhaṁ dū¦tám // adhva¦ré kṇudhvam ¦

yó mártie¦ṣu // nídhru¦vir tvā ¦

tapurmūrdhā // ¦ ghtá-an¦naḥ pavākáḥ ¦¦

vii 3 1.

Jagatī stanza with cretic variations:—

imáṁ stómam // ¦ árhate ¦ jātávedase

rátham iva // ¦ sám mahe¦mā manīṣáyā ¦

bhadrá hí naḥ // ¦ prámatir ¦ asya saṁsádi

ágne sakhyé // ¦ má riṣā¦mā vayáṁ táva ¦¦

i 94 i.

49. ‘Decasyllabic’ verses correspond generally to Triṣṭubh verses with the omission of a syllable immediately before or after the cæsura: this omission we term a rest. Two types are used with sufficient regularity to have given special names to the [p. 14] metres of the hymns which contain them: but these, as well as others, are frequently found in combination with Triṣṭubh verse.

50. The Virāṭsthānā verse has a rest at the fifth place, followed by a trochee, as:—

[−−⏑−¦ · −⏑¦−⏑−−]
rsi kṣáyaṁ // ¦ · rsi ¦ mitrám asmé ¦

[−⏑−⏑¦ · −⏑¦−⏑−⏑]
rsi śārdha // ¦ · indra ¦ mrutaṁ naḥ ¦

[⏑−⏑−¦ · −⏑¦−⏑−−]
sajóṣaso // ¦ · yé ca ¦ mandasānḥ ¦

[⏑−⏑−¦ · −⏑¦−⏑−⏑]
prá vāyávaḥ // ¦ · pānti ¦ ágraṇītim ¦¦

ii 11 14.

51. The Dvipadā Virāj verse has a rest at the sixth place, the quantity of the seventh, syllable being indifferent: the two parts are then so similar in form that the stanza of two decasyllabic verses may also be described as containing four verses of five syllables, or pentads.


rayír ná cit¦r // · ś¦ro ná saṁdk ¦

áyur ná prā¦ṇó // · nít¦yo ná sūnúḥ ¦¦

i 66 1.

[alternatively if we follow Arnold's suggestion in the last sentence we could (and probably should) scan the lines like this:]

rayír ná citr ¦ śro ná saṁdk ¦

áyur ná prāṇó ¦ nítyo ná sūnúḥ ¦¦

Here the division into members is only of historic interest, shewing the relation of this metre to the Triṣṭubh.

52. In two hymns of the Rigveda a similar metre is based on the Jagatī verse, the ‘rest’ occupying the position of the syllable first following the cæsura, whether early or late. to this metre, following the traditional name of the author of the hymns, we may give the name Bhārgavā. If the rest be denoted by a point · ; the most common forms of the break in Bhārgavā verse are // · ⏑−¦ and // · −¦ .

The following stanza is entirely composed of Bhārgavā verses, the two types alternating:—

[⏑−−−¦⏑, · −¦−⏑−⏑⏑]
śriyé máryā¦sa // · añ¦jr akṇvata ¦

[⏑−⏑−,¦ · ⏑−¦−⏑−⏑⏑]
sumrutaṁ // ¦ · ná púr¦vr áti kṣápaḥ ¦

[⏑−−−¦⏑, · −¦−⏑−⏑−]
divás putr¦sa // · é¦tā ná yetire ¦

[−−−−,¦ · ⏑−¦−⏑−⏑⏑]
āditysas // ¦ · tá ak¦r ná vāvdhuḥ ¦¦

x 77 2.

53. It seems evident that the decasyllabic metres (amongst which we include the Bhārgavā) presuppose and are derived from the more common trimeter metres. Their precise relation [p. 15] to them is an important subject of investigation. As none of these metres are found in the Atharvaveda or later literature, there is a presumption that, like the lyric metres, they belong to the distinctively early parts of the Rigveda.

54. An important feature in trimeter verse is contamination, or the combination in one verse, stanza, or hymn of parts that do not correspond. Contamination may be external or internal.

55. External contamination is the combination of Triṣṭubh and Jagatī verses in the same stanza, and of Triṣṭubh and Jagatī stanzas in the same hymn. In hymns of which the general character is Jagatī it is quite usual for one or two Triṣṭubh stanzas or verses to appear towards the end of the hymn; consequently it is rather the absence of such variations that calls for notice, at least in hymns of any length. But the contamination of a Triṣṭubh hymn by Jagatī is only exceptional in the Rigveda, though very common in the Atharvaveda and later verse, down to and including the epic period. It forms therefore a connecting link between the Rigveda and this later literature.

56. Internal contamination is the combination of two parts of a Triṣṭubh verse which correspond to different positions of the cæsura. The combination of two parts consisting of four and six syllables respectively gives the equivalent of a decasyllabic verse with a rest at the cæsura; verses of this type, whatever their origin may be, are fairly common in all parts of the Rigveda, as for example

[⏑−−−,¦ · ⏑⏑¦−⏑−⏑]
inó rājann // ¦ · ara¦tíḥ sámiddhaḥ ¦

x 3 1a.

The combination of two parts consisting of five and seven syllables respectively may be termed the hybrid verse. It is often found in the same hymn with external contamination the two features are exemplified in the first and third verses respectively of the following stanza

uruvyácā ¦ no // mahiṣáḥ ¦ śárma yasat ¦

asmín háve // ¦ puruhū¦táḥ purukṣúḥ ¦

sá naḥ praj¦yai // hari¦aśva mḷayā ¦

índra m no // ¦ rīriṣo ¦ m párā dāḥ ¦¦

x 128 8.

The ‘hybrid’ verse is again a connecting link with the Atharvaveda and later literature.





[p. 19]

64. The time has perhaps hardly come when a fair estimate can he made of the literary and aesthetic powers of the Vedic bards: but it is important to examine the conditions of the mechanical problems presented to them, and to form some estimate of the skill shewn by them in handling their material. The sketch already given of the metrical contents of the Rigveda has shewn that each period, except the ‘normal’ period, was inventive. The bards were occupied in constructing fresh metrical schemes, as well as in producing verse following established models. They frequently compare their craft to the highest kind of workmanship known to them, that displayed in the construction of a war-chariot: and they shew confidence that a ‘new’ song will be more pleasing to the gods than one which is old-fashioned. It is a common-place of Western criticism that in many of the mechanical arts the Hindu workman follows too submissively ancient rules and models. But no such statement can justly be made with regard to the poetic literature of India, either ancient or modern: rather the faculty of inventing and appreciating new and delicate variations of rhythm seems to be a special gift of the race.

65. The comparison of the earlier parts of the Avesta indicates that the first Vedic poets were not far from the period when verse was measured solely by the number of syllables, without any regard to their quantity. We have already noticed (§ 32) that the quantity of initial and final syllables is always indifferent in [p. 20] the Rigveda, and we have every reason to suppose that this feature is inherited from the earlier period of purely syllabic measurement. But in all other parts of the verse we find rhythmical tendencies, which determine with varying regularity the quantity required in each position.

66. The principle of alternation of short and long syllables is simple, and we may well believe that the Vedic bards could have carried it to mechanical perfection, had they so desired. But in addition they were guided by other metrical motives, some of which have been described above (§ 34, 35). Above all, they were actuated by the desire to combine these motives, and thus give variety to their metre. We are not perhaps entitled to ascribe the ‘break’ of trimeter verse to a deliberate design of interrupting the iambic rhythm, because it is quite possible that the cæsura and the two short syllables that follow it may have been earlier in date than any such rhythm. But the ‘cretic’ rhythm certainly suggests a deliberate aiming at variety of form: and if it be granted that the ‘epic’ Anuṣṭubh is based upon the Anuṣṭubh of the Rigveda proper, the motive of the change seems evidently to be the same.

67. ‘Contamination’ readily suggests careless workmanship, and may in some cases have been due to this cause. But it is also a ready instrument of the inventive temper. The historical relation of Triṣṭubh and Jagatī verse has often been discussed: and if one of the two is historically the earlier, the Triṣṭubh has certainly the better claim to this position. Either then Jagatī has been derived from Triṣṭubh by contamination with dimeter cadence, or a dimeter verse added to a similar half-verse has been assimilated to the rhythm of the Triṣṭubh. By one method or the other a metre not without a beauty of its own has been constructed. The contamination at a later period of Triṣṭubh with Jagatī did not lead to the construction of any harmonious metre: but perhaps in principle it is not to be distinguished from that mingling of the lyric metres from which grew the Bhatī-Satobhatī strophe, which is probably generally recognized as the most beautiful metrical form of the Rigveda.

68. The origin of the decasyllabic verse is a matter as to which a conjecture will be hazarded later (Ch. viii), but it is only [p. 21] through this stage that we can account at all for the existence of the Dvipadā Virāj or Pentad metre, which was actually brought to as polished a perfection as any metre of the Rigveda.

69. In all these cases it is not hard to recognize that apparent irregularity is the result of the inventive spirit, sometimes leading to the construction of new and harmonious forms, sometimes apparently ending in failure. This observation cannot fail to suggest reserve and care before bringing forward any suggestions of incompetence against the versifiers or of carelessness against the editors of our text. It is difficult to think that a professional bard should without motive have left his verse with an irregular rhythm, when any European scholar, without serious practice of the art of versification, can put it into order for him with hardly a perceptible alteration in the meaning. It is also difficult to think that professional reciters and their instructors could by mere accident have left stanzas in a shape which must make them a perpetual burden to the memory. In these ‘irregularities’ there may be meanings not easily recognized, and for this reason they deserve to be carefully studied.

70. To whatever conclusions we may further be led in detail, it must be plain that as works of mechanical art the metres of the Rigveda stand high above those of modern Europe in variety of motive and in flexibility of form. They seem indeed to bear the same relation to them as the rich harmonies of classical music to the simple melodies of the peasant. And in proportion as modern students come to appreciate the skill displayed by the Vedic poets, they will be glad to abandon the easy but untenable theory that the variety of form employed by them is due to chance, or the purely personal bias of individuals: and to recognize instead that we find all the signs of a genuine historical developement, that is of united efforts in which a whole society of men have taken part, creating an inheritance which has passed through the generations from father to son, and holding up an ideal which has led each in turn to seek rather to enrich his successors than to grasp at his own immediate enjoyment. If this was so, then the Vedic bards also are to be counted amongst ‘great men, and.......such as sought out musical tunes, and set forth verses in writing.’