In 9, hi is dropped after ána, substituted for the characteristic ní of the 2d sing. imperative, Parasmai, in the case of roots ending in consonants. A form tát (cf. Latin to, Greek Tw) may be substituted for hi and tu, and even for ta, to imply benediction, chiefly used in the Vedas.

Only eight roots, viz. śru, stu, dru, sru, kṛi, bhṛi, sṛi, vṛi, reject the initial i from the terminations marked with *; and of these eight all but vṛi (meaning 'to cover') necessarily reject it also in the 2d sing. Parasmai. These eight roots also take dhve for dhve in the 2d pl. Atmane. The option of iḍhve for idhve is allowed in other roots when a semivowel or h immediately precedes.

Many roots prefix i to the above terminations: thus, 1. itásmi, 2. itási, &c.
E lengthens this i;vri and all roots in long rí optionally do so.

Many roots prefix i to the above terminations: thus, 1. ishyámi (r. 70), 2. ishyasi, &c. lengthens this i; and all roots in long rí optionally do so.

Aorist or third preterite (requiring the augment a).

FORM I.-Terminations of the memorial scheme.

dhvam is used for dhvam after any other vowel but a or á, or after 3d immediately preceding.

The same terminations with i prefixed, except in 2d and 3d sing., where

iḍhvam may be used for idhvam when a semivowel or h immediately precedes. E lengthens the i throughout; and all roots in long rí optionally do so in Atm. FORM II.-Terminations resembling those of the imperfect or 1st preterite.

Many roots prefix i to the Atmane, but not to the Parasmai, of the above: thus, 1. ishíya, &c. lengthens the i in this tense also, but no other root can do so. is used for sídhvam after any other vowel but á, and optionally after the prefixed i, when immediately preceded by a semivowel or h.

Many roots prefix i to the above terminations throughout: thus, 1. ishyam, 2. ishyas, &c. lengthens this i; and all roots in long rí optionally do so.

Observe-We shall in future speak of the 1st, 2d, and 3d preterites under the name of imperfect, perfect, and aorist, respectively.

a. If we examine the terminations exhibited above, we shall find that they are composed of two distinct elements, one marking person, number, and voice; the other, mood and tense. The terminations in which the former element prevails may be called simple, and belong to the present, imperfect, imperative, perfect, and 2d form of the aorist; those which include the second may be called compound, and are peculiar to the other tenses. Thus the terminations of the potential consist of i or í or yá as characterising the mood, and of am, s, t, va, tam, tám, &c., as marking person, number, and voice. So, also, in the 2d future the syllable sya prefixed to all the terminations, characterises the future tense, while the mi, si, ti, vas, thas, tas, &c., mark person, number, and voice. If, then, such initial parts of every termination as mark mood or tense were left out, an examination of the remaining parts would show that the present and imperfect are the prototypes of the terminations of all the other tenses, that is to say, that the formation of the terminations of every other tense may be referred back to one or other of these two. The present tense may in this way be connected with the two futures. These three tenses agree in showing a certain fulness of form, which is wanting in most of those connected with the imperfect. The terminations of the perfect, however, partake of the character of both the present and imperfect. In the Atmane-pada they very closely resemble the present. Many of them exhibit the same fulness as that tense, while some of the other terminations of the perfect show even more lightness than those of the imperfect*. It should be observed, too, that the terminations of the imperative, though evidently connected with the imperfect, are in some instances even more full than those of the present.

b. Although comparative grammarians have bestowed much labour on the elucidation of the origin of Sanskrit verbal terminations, the only point which may be asserted with probability is, that they stand in a certain relationship to the pronominal bases ma, tva, ta. The m of the first persons is related to the base ma ; the t, th, sr, s, of the second persons, to the base tra of the second personal pronoun ; and the t, of the third person, to the base ta. We may also observe a community of character between the termination nti of the 3d plur. and the plural of neuter nouns like dhanavat (dhanavanti). But whether the v in the dual stands for m or relates to a pronominal base va occurring in á-rám, va-yam; whether the terminations of the dual and plural are formed from those of the singular by adding s as a mark of the plural, or by the composition of several pronominal bases; whether the terminations of the Atmane-pada are formed from those of the Parasmai-pada by guṇation or by composition of the latter with other bases, these and others are questions which cannot be determined with any certainty.

c. As an aid, however, in committing the terminations to memory, the student

* Comparative grammar, however, has established that these terminations were originally as full as those of the present.

S

may observe that the letter m generally enters into the 1st sing. Pasasmai; into the 2d sing. Parasmai and Atmane; and t into the 3d sing. du. and pl. Parasmai and Atmane of all the tenses. Moreover, that the letter v occurs in the 1st dual, m in the 1st plural of all the tenses, and dhe in every 2d plural Atmane-pada. In the imperfect and potential Atmane, and in the perfect Parasmai, th is admitted, instead of s, into the 2d sing.; and in the 2d pl. of the last tense, th has been dropped, owing to the influence of the heavy reduplication. For the same reason the m and t are dropped in the 1st and 3d sing. perfect. Observe also - When the 1st dual Parasmai is vas, the 2d and 3d end in as (except the 3d du. 1st future), and the 1st plural is mas. When the 1st dual Parasmai is va, the 2d and 3d end in tam, tám (except in the perfect), and the 1st plural in ma. When the 1st dual Atmane is vahe, the 1st plural is mahe, and the last letter of the remaining terminations is generally e. When the 1st dual Atmane is vahi, the 2d and 3d end in ám ; the 1st plural is mahi, and the 2d plural is dhvam.

d. The frequent occurrence of m in the 1st sing., of s in the 2d, of t in the 3d, of mas and ma in the 1st pl., of ta in the 2d pl., and of ant in the 3d pl., suggests a comparison with the Greek and Latin verb. We may remark, that m, the characteristic of the 1st person sing., is suppressed in the present tense active of all Greek verbs except those in μι (asmi = εἰμί, Dor. ἐμμί for εσμι, dadámi=δίδωμι), and also in Latin verbs (except sum and inquam); but w and o answer to the Sanskrit á of bharámi = pépw, fero. In the Greek middle and passive, the μ, which originally belonged to all active verbs, becomes μa; while the Sanskrit, on the other hand, here suppresses the m, and has e for at; bhare (for bhara-me) = pépoμas. In the imperfect, Greek has for the Sanskrit and Latin mute m; atarpam = ἕτερπον, adadám=èdídwv, astṛiņavamèoтópvvy, avaham vehebam. Greek has μ in the 1st sing. optative or potential; and in verbs in μ, v takes the place of the mute m of Sanskrit and Latin: thus, bhareyam=pépoμ, feram; dadyám=dıdvíny, dem ; tishtheyamiotaíny, stem. In the first Greek aorist, m is suppressed, so that Sanskrit adiksham (3d pret.)="deığa; but not in the 2d aorist, so that adám= wv. In the perfect the Sanskrit a=Greek α, tutopa=TÉτvpa.

mas=

=

In the Greek

middle and passive futures, m is retained, but not in the active; dásyámi=dwow, dekshyamideíkow, dásyedwooμa. As to the 1st person plural, the Sanskrit mas of the pres. is μev (for μes) in Greek, and mus in Latin; tarpá-mas=тépπOμεv; sarpá-mas="pño-μev, serpi-mus; dad-mas=dido-μev, da-mus; tishțháloτa-μev, sta-mus. The Atmane-pada mahe answers to the Greek μeba; dad-mahe=didó-μela. As to the other tenses, in the imperfect 1st pl. abharáépépo-μev, fereba-mus; avahá-ma-veheba-mus; adad-ma= édído-μer; abhará-mahi=èpepóuela. In the potential 1st pl. bhare-ma = pépoi-μes (-μev), fera-mus; dadyáma=didoínμes (-μer), demus; dadí-mahi—didoí-μeba. In the 2d future, dásyá-mas – dwσo-μev, dekshyá-mas-deikoo-μev. In the 2d pers. sing. active, the characteristic s has been preserved in all three languages: thus, in the present, the Sanskrit asi (for original assi) – éσoí, es ; dadá-si = dídws, das; bharasi = pépets, fers; vahasi = vehis. In the Atmane, the Sanskrit se (for sai, by 32)

ma=

=

=

=

=

=

ἵστα-σαι). In other

answers exactly to the Greek as of verbs in μ (tishṭha-se = Greek verbs, s has been rejected, and ea contracted into 7, something in the way of Sanskrit (TUT) for TÚTTE-σα). In the 2d dual, thas Greek Toy, and in the 2d plur. tha = 7€ and tis; bhara-thas = pépe-rov; tishtha-tha ==""σTa-Te, statis ; bhara-tha = pépe-TE, fer-tis. In the 2d pl. Átmane, bhara-dhve = pépeole. As to the other tenses, in the 2d sing. imperfect, atarpas eTeρπes, avahas = vehebas, &c. So also, tam = Tov, adat-tam = édídc-Tov, ta = TE, adat-ta = ¿dido-Te. In the Atmane, thás is found for sás in the 2d sing. of the imperfect, as well as of the potential; hence abhara-thás épépe-co, adat-thás édido-oo, dadí-thás = dido(). In the 2d sing. potential, tishṭhes = iσrains, stes; dadyás = didoíns, des ; vahes vehas; bhares pépois, feras: in 2d du. bhare-tam = pépos-Tov: in 2d pl. tishtheta = ioTainte, stetis; dadyáta = didointe, detis; bhareta = pépoite, φέροιτε, feratis. In the 2d sing. imperative, hi and dhi answer to Greek . Dhi was originally universal in Sanskrit (see 291), as in Greek verbs in μ; e-dhi=””σ-01, viddhi = "o-01, de-hi dido-01. Many verbs drop the termination hi both in Greek and Sanskrit; as, = pépe, and compare deíkvu with ćinu, &c. In the 2d du. imp: tam = Tov, and ta = TE. In the imperative Atmane, sva = the old form σo; bhara-sva = pépe-00 (old form of pépov); dat-sva== dído-00; áthám = €50ov, &c. In the perfect, the tha of the 2d sing. = Latin sti; dad-itha = dedi-sti, tasthitha steti-sti, tutodi-tha tutudi-sti. In the aorist, adás = edws, avákshís = veristi. In the 3d pers. sing. active, Greek has dropped the characteristic t (except in Tí Sansk. asti, Lat. est); bharati pépe (T), fert; vahati vehit. Verbs in have changed t to s; dadáti = δίδωσι (for δίδωτι). In the Atmane, bharate = pépera. In the imperfect, avahat = vehebat, abharata=èpépero. In the potential, bharet = pépoi, dadyát = didcin. In the imperative, bhara-tu or bhara-tát pepé-Tw, fer-to. In the perfect, tutopa TéTUDE. In the aorist, arákshít = vexit, adikshata=deíkoαтo. As to the 3d pl., in the above tenses, bharanti =pépovo, ferunt; vahanti = vehunt; bharante = pépovτai; dadati = didoñoi; tishṭanti =stant; bhareyus = pépoiev; bharantu=ferunto; abharan=epepov; abharanta = ἐφέροντος ásan = ἦσαν; atarpishus = ἔτερψαν; dasyante = δώσιν

ται.

=

248. The above terminations are supposed to be applicable to all verbs, whether primitive or derivative: and as in nouns, so in verbs, the theory of Indian grammarians is, that before these terminations can be affixed to roots, an inflective base must be formed out of the root. Ten different rules, therefore, are propounded for forming verbal bases out of roots in the first four tenses; while all verbs are arranged under ten classes, according to the form of the base required by one or other of these rules. In the other tenses there is one general rule for forming the base, applicable to all verbs of whatever class.