II. Euphonic Combination (Sandhi).
^^ 59
N. of a plant yielding honey, beside madhu-dugha- 'shedding sweetness'; sas[pa]-pinjara- (VS.) 'tawny like young grass.
2^. The second syllable is dropped in the d&'iwts pdumsyaja], ratnadheyj[ya], saMyd[y a], and abhikhyd beside abhikhyaya; also xi vrkdt[ati- 'destruction', beside vrkdtat- and devd-iati-; and at the beginning of the second member of a compound in sirsa-[saM- (AV.) 'headache' ^
a. A following syllable is sometimes dropped in spite of a different one intervening; thus in the dative marydda[ya] 'boundary'; and somewhat pecuHarly iri dvarrv[ar]ur, 3. pi. impf beside the 3. sing, d varivar and 3, sing. pres. a varTvarti.
II. EUPHONIC COMBINATION (SANDHI).
Benfey, Vollstandige Grammatik p. 21—70. — Whitney, Sanskrit Grammar p. 34 — 87. — Wackernagel, Altindische Grammatik i, 301 — 343. — Arnold, Vedic Metre p. 70 — 80.
65. The nature of Vedic Sandhi. — The sentence is naturally the unit of speech which forms an unbroken chain of syllables euphonically combined. It is, however, strictly so only in the prose portion of the AV.3 and the prose Mantras of the YV. As the great bulk of the Vedas is metrical, the RV. and the SV. being entirely so, the editors of the Sarnhita text treat the hemistich (consisting generally of two Padas or verses) t as the euphonic unit, applying the rules of Sandhi with special stringency between the Padas or metrical units which form the hemistich. The evidence of metre, however, shows that, in the original form of the text, Sandhi at the end of an internal Pada is all but unknowns. The verse, therefore, is the true euphonic unit®. The final of a word appears either at the end of this unit in pausa {avasane) 7, or within it as modified by contact with a following initial. The form which the final of a word assumes in pausa, being regarded as the normal ending, is generally the basis of the modification appearing within the verse. It will therefore conduce to clearness if the rules relating to absolute finals are first stated.
66. Finals in pausa. — a. Vowels' in this position undergo no change other than occasional nasalization.
1. In all the Sarnhitas a % u'^ are frequently nasalized when prolated; e. g. vindatiim = vindati (x. 146'); babhuvaiin = babhUva (AV. x. 2); vivesaim = vivesa (VS. xxiii. 49); mamdim = mama (TS. vii. 4^°).
2. In the Sarnhita text of the RV. there survive, at the end of a verse within a hemistich, from the period when such end also was accounted a pause, several instances of nasalized c, preserved to avoid hiatus and con-
■ This explanation is doubtful in uloka- for uhibka- (cp. Erugmann, Grundriss i, 624, p. 471); susmaya- (TS.) for iusma- mdya-; bhTmala- (VS.) 'terrible', for bhTma- mala-; improbable ia ne/anas for rujand-nas, 'with broken nose' (Bloomfield, JAOS. 16, xxxiv).
2 Cp. Whitney, Translation of AV. I, 123.
3 See Bloomfield, The Atharvaveda, in this Encyclopedia, §1 (beginning) and note i.
i In the Gayatrl metre, in which there are three Padas, the third alone constitutes the second hemistich; in the Pankti, which has five Padas, the last three constitute the second hemistich.
5 The only probable exception is RV. IX. 113, 7 c d; cp. Arnold, Vedic Metre 119.
6 The sentence within a Pada, as well as the Pada itself, is the unit of accent; cp. below, 83.
7 Cp. RPr. L 3; VI. 5; x. S; xi. 30.
8 Final r never occurs in the RV., its place being supplied by iir as nom. ace. s. n. e. g. sihitur 'standing' (cp. WaCKERNAGEL, KZ. 25, 287 f.); but the TS. already has
janayiif and bhartf (B ?), Wackernagel I, 259 a, note.
9 The vowels I and li, when dual termina- tions (pragrhya), cannot be nasalized.
