a. Atm. to go,' Atm. 'to be aroused,' Atm.' to be born,' may form their 3d sing. as if they were passive verbs (see 475): thus, wife, Du. 3. WTत्सातां; अबोधि (or optionally सबुद्ध), Du. 3. सभुत्सातां; सजनि (or optionally अजनिष्ट).

b. Roots ending in and must change these letters to Anusvára before s, and म् becomes न् before ध्व: thus, मन् makes अमंसि, अमंस्थास्, अमंस्त (or if in class 8. अमनिष्ट, or by c. below समत); क्षम् makes अशंसि &c., Du. 2. अक्षन्ध्वं .

हन् (generally Parasmai) drops its nasal before the Atmane terminations (सहसि , &c.; initial s being rejected according to 419).

गम् does so optionally (सगंसि or अगसि, अगंस्थास् or अगथाम्, &c.).

c. Roots in and of the 8th class, which properly take the B terminations, are allowed an option of dropping the nasal in the 2d and 3d sing. Atmane, in which case initial s is rejected according to 419: thus, makes सतनिष्ट or wd (Páṇ. II. 4, 79).

e.

3.

d. Similarly, क्षण् makes 3. अक्षणिष्ट or अक्षत; and ऋण, वार्णिष्ट or जाते. to give' is allowed the option of lengthening the a, when ʼn is dropped: thus, Sing. 2. अमाथास् or असनिष्ठाम्, 3. असात or असनिष्ट. Compare 354. a, 339 (Pán. II. 4, 79).

ƒ. The nasal of 'to bite' becomes before सदानं, Du. 2. सदष्टं ; Atm 1. सदह्नि, Du 2. अदण्ड

and before thus, See 303.

425. to carry' (see 611) changes its radical vowel too before those terminations which reject an initial s by 305. a: thus, aváksham, avákshís, avákshít (Lat. vexit), avákshva, aroḍham, &c.; Átm. avakshi (Lat. rexi), aroḍhás, avoḍha. a. Atm., to bear,' generally takes the B terminations (asahishi, &c.), though the form is also given for the 3d sing.

426.to tie,' 'to fasten,' makes anátsam, anátsís, anátsít, anátsva, anáddham, &c.; and Atm. anatsi, anaddhás, &c., by 306. b (compare 183).

a.

to dwell' (see 607) makes avátsam, &c., by 304. a.

427. Verbs which assume i, and so take the B terminations at 418, require a different rule for the formation of their base, as follows:— a. If a root end in the vowelsi, kí, cu, Hú, Hri, rí, vṛiddhi those vowels in the Parasmai before all the terminations, and gunate them in the Atmane.

Thus, from 'to purify' come the two bases apau for Parasmai and apo for Atmane (apau + i + sam = fa by 37, apau+i+ís = अपावीम्, apau + i + it = अपावीत्, &c. ; Atmapo + i + i = अपविधि, &c., by 36), see 583; and from 1st c., to cross,' comes the base atár for Parasmai (atár +i+sam = fi &c.).

So, from शो ‘to lie down ' comes अशयिषि, अशयिष्ठास, &c. ; but roots ending in any other vowel than ú and long rí more frequently take the A terminations, as they generally reject i.

b. If a root end in a single consonant, gunate the radical vowel in both Parasmai and Atmane (except as debarred at 28. b, and except in the roots enumerated at 390. a).

Thus, बुध् budh, 1st c., 'to know,' makes its base abodh (abodhisham, &c.), see 583; and vṛit, 'to be,' avart (avartishi, &c.); and edh, 'to increase,' aidh (aidhishi, &c., 260. b), see 600.

T ल्

428. A medial a in roots ending in and is lengthened in the Parasmai, but not in Atmane : thus, चर् ' to go' makes अचारिषं; ज्वल् 'to blaze,' wf. The roots to speak' and a 'to go' also lengthen the a in Parasmai (अवादिषं ; but not in Atmane अवदिषि &c.).

a. But those in 7, 4, never lengthen the a in Parasmai: thus, to sound' makes eyfay. The following roots also are debarred from lengthening the a : कख्, कग्, रग्, लग्, सग्, स्थग्, हुग्, द्वग्, कट्, क्षण्, चत्, क्वथ्, पथ्, मथ्, चद्, बध्, श्वस्, हस्. One or two do so optionally ; as, कण् and नद् 'to sound.’

429. Observe, that as the majority of Sanskrit verbs assume i, it follows that rule 427. a, b, will be more universally applicable than rule 420, especially as the former applies to the aorist of intensives, desideratives, and nominals, as well as to that of simple verbs.

(अकुचिषं

430. The special rules for the two futures at 390. a-o will of course hold good for the aorist: thus the roots enumerated at 390 and 390. a ( &c.) forbid Guna; and,,, generally change their finals to uv (fi &c., fa &c.); but when is written it makes &c., see 421.f, and may also make अधाविषं, and नू,

अनाविषं.

a. ऊर्णु makes और्णविषं or और्णविषं or और्णुविषं &c, and in Atmane और्णविधि or wtyfafa.

b. According to 390. C. दीधी, घेवी, and दरिद्रा drop their finals (अदीधिपि, weffe, &c.; see also 433).

rí, such

431. In the Atmane, 'to choose,' 'to cover,' and all roots in long as स्तृ ‘to spread,’ may optionally lengthen the inserted i: thus, अवरिषि or अवरीषि &c., अस्तरिषि or अस्तरीषि ; but in Parasmai only अवारिषं, अस्तारिषं.

432. f'to swell' and

&c., see also 440. a;

'to awake' take Guna instead of Vriddhi (f

fi &c.).

a. ग्रह् according to 399. a. makes अग्रहीषं, and by 390. m. गुह् makes अगूहिषं. The latter also conforms to 439 and 439. b. See 609.

433. Many roots iná, e, o, and ai, with three in m, viz. 44 yam, ram, nam, assume i, but in the Parasmai insert s before it; final e, o, and á: thus, from 'to go' comes fa, &c. (see 644); अशासिषं, &c. ; from यम् ' to restrain, ' अयंसिषं, &c.

ai, being changed to

from शो — to sharpen

f'to be poor' makes adaridrisham or adaridrásisham, &c.

434. In the Atmane these roots reject the i and the s which precedes it, and follow 418: thus, from 'to measure' comes f, &c. (see 664. a); from 'to cover,' (see 421. c); from to sport,' wife, wiera, vim, &c.

436. Observe No confusion arises from the similarity which this form bears to the imperfect or 1st preterite, as in all cases where the above terminations are used for the aorist, the imperfect presents some difference in the form of its base: thus, 'to go' make sagattham for its imperfect, agamam for its aorist (see 602); भिद् fre to break' makes abhinadam for its imperfect, abhidam for its aorist (see 583). So again, the sixth conjugation, which alone can show a perfect identity of root and base, never makes use of this form for its aorist, unless by some special rule the base of its imperfect is made to differ from the root: thus, foto smear' (cf. aλeipw), which makes alipam in its aorist, is alimpam in its imperfect (281). So in Greek, compare the imperfect eλerov with the 2d aor. ἔλιπον; ἐλάμβανον with ἔλαβον; ἐδάμνην with ἔδαμον, &c.

a. Note―This form of the Sanskrit aorist corresponds to the Greek 2d aorist (compare asthám, asthás, asthát, with corny, eσTns, eoT), and the first form is more or less analogous to the 1st aorist. The substitution of i for e, and áthám, átám, for ethám, etám, in the Atmane of form II, is confined to a class of roots mentioned at 439.

437. Rule for forming the base in verbs of the first nine classes. In general the terminations are attached directly to the root; as in agamam, &c., abhidam, &c., at 436. So also, to perish' makes अनशं (also सनेशं, see 441, 424 ) .

a. Observe, however, that most of the roots which follow this form in the Parasmai, follow form I at 418 in the Atmane: thus, fr 'to break' makes abhitsi, &c., in Atmane; see the table at 583: similarly, f'to cut,' see 667. And a few roots, which are properly restricted to the Atmane, have a Parasmai aorist of this 2d form:

thus, Atm. to shine,' 'to be pleasing,' makes Parasmai arućam, रुच्

as well as Atmane aroćishi.

b. One or two roots in á, i, and T e reject their finals; and one or two in ri andri gunate these vowels before the above terminations: thus,

makes a &c.,

'to go,' war; 'to go,'

c.

&c.; f'to swell,' ; to call' makes

to tell' (see 595);

to stop' makes ni;

to see' gunates its vowel (e, see 604).

d. Penultimate nasals are generally dropped: thus,

स्यन्द् ‘to distil,' अस्यदं; स्कन्द 'to mount,' अस्कदं; भ्रंश 'to fall,' अभ्रशं.

e. A form occurs in the Veda, from to eat,' the medial a being dropped. 438. In the Parasmai certain roots ending in long á and e conform still more closely to the terminations of the imperfect, rejecting the initial vowel, as indicated by the brackets in the table at 435. In the 3d plur. they take us for an: thus, 3d c., 'to give,' makes adám, adás, adát, adáva, &c.; 3d pl. adus, see 663. So also, 3d c., 'to place,' makes adhám, &c., 664; and 1st c., 'to stand,' makes asthám, &c., 587.

a. Similarly, 1st c., 'to be,' except 1st sing. and 3d pl. (wri, TH, wm, wa, &c. ; but 3d pl. 7, see 585).

b. Observe, however, that some roots in á, like yá, 'to go,' follow 433

c. And some roots ine ando, which follow 433, optionally follow 438; in which case e and o are changed as before to á: thus, dhe, 1st c., 'to drink,' makes either adhásisham &c., or adhám &c., also adadham, see 440. a; so, 4th c., *to come to an end,' makes either asásisham or asám, see 613.

d. In the Atmane-pada, roots like, vi, TM, à, è, ¿ì, follow 421. d.

e. 'to go' makes its aorist from a root : thus, agám, agás, &c. Note-Adadám, the imperfect of the root dá, 'to give,' bears the same relation to its aorist adám that ediowy does to dwv. So also the relation of adhám (aorist of dhá, 'to place') to adadhám (imperfect or 1st pret.) corresponds to that of en to ἐτίθην. Compare also abhavas and abhus with ἔφνες and ἔφυς.

ह्

439. Certain roots ending in έ, sh, h, enclosing a medial i, u, or ri, form their aorists according to form II at 435; but whenever confusion is likely to arise between the imperfect and aorist, s is prefixed to the terminations, before which sibilant the final of the root becomes k by 302 and 306.

Thus, f to point out,' the imperfect of which is fasi, makes f&c. in aorist (compare the Greek 1st aorist edea). Similarly, faq 2d c., 'to hate,' makes adviksham &c. 657; 'to milk,' makes y adhuksham &c. by 306. a. See 660.

Bb

2d c.,

a. This class of roots substitutes i for e, and áthám, átám, for ethám, etám, in the Atmane terminations: thus, adikshi, adikshathás, adikshata, adikshávahi, adiksháthám, &c.; 3d pl. adikshanta.

b. A few roots inh (viz. fe, fee, E,

optionally in the Atmane re

ject the initial a from the terminations of the 2d and 3d sing., 1st du., and 2d pl.: thus, fe may make afmfa, aciera, uole;

661 and to milk,' f, g, &c.

:

अली,

Du. 1. afsafe; Pl. 2. wote,

See 661, 659, 609, 660.

c. According to some authorities, a few roots (e. g. 7, 74, 74) which generally follow form I, A, in Atmane, may optionally conform to form II, taking the terminations i, áthám, átám, rejecting initial a and á from the other terminations, and taking ata for anta: thus, atṛipi, atṛipthás, atṛipta, atṛipvahi, &c.

440. Causal verbs and verbs of the 10th class make use of form II, but the base assumes both reduplication and augment (as in the Greek pluperfect): thus, 1st c., 'to know,' makes in the causal aorist अबूबुधं, &c. This will be explained at 492.

a. A few primitive verbs besides those of the 10th class take a reduplicated base, analogous to causals (see 492): thus, f'to have recourse' makes aff &c.; fto swell' makes ff (also and fq, see 432, 437.6); 1st c., ' to run,'; 'to flow,' असुस्रुवं; धे ‘to drink,' सदधं; कम् 'to love,' अचकमे, &c. This last is defective when it belongs to the 1st c., having no conjugational tenses; but when it belongs to the roth c. (Pres. कामये, &c.) its aorist is अचीकमे.

441. The following primitive verbs take a contracted form of reduplicated base: 2d c., to speak,' makes a avoćam (from ८ -- अवउचं for अववचं 650 ) ; पत् ist c., 'to fall,' अपनं (from अपपतं ; compare Greek ETTTOV ) ; शास् 2d c., ' to rule,' अशिषं (from अशिशसं, but the Atmane follows 427; see 658); 4th c., to throw,' (from आससं, contracted into आथ्मं for सत्सं 304. a, whence by transposition आस्थं) ; न‍ 4th c., ' to perish,' चनेशं (from अनइयं for चननिशं 620, 436).

Precative or benedictive.

6

Terminations of precative or benedictive repeated from page 128.

L

Sidhvam is used for sídhvam when immediately preceded by any other