Below is added a statement of the usage, as regards the auxiliary vowel, of all the roots found quotable — for the most part, in the form of a specification of those which add the tense-sign directly to the root; in brackets are further mentioned the other roots which according to the grammarians also refuse the auxiliary vowel.

a. Of roots ending in vowels, the great majority (excepting those in ṛ) take no i. Thus, all in ā (numerous, and unnecessary to specify: but compare c below); those in i, as kṣi possess,ci gather, ci note, mi, si or sā bind (siṣya), hi; from i, kṣi destroy, and ji occur forms of both classes; çri [and çvi] has i; — those in ī, as krī, bhī, mī, vlī; but çī lieand nī have both forms [and ḍī takes i]; — those in u, as cyu, dru, plu, çru, hu; but su press out and stu have both forms [and kṣu, kṣṇu, nu, yu, ru, snu take i]; — of those in ū, dhū and bhū take i; sū has both forms. But all in ṛ (numerous, and unnecessary to specify) take i [those in changeable ṛ, for so-called ṝ-roots (242), are said by the grammarians to take either i or ī; no ī-forms, however, are quotable].

b. Of roots ending in mutes, about half add the tense-sign directly. Thus, of roots ending in gutturals, çak; — in palatals: in c, pac, muc, ric, vac, vic, vraçc, sic (but yāctakes i); in ch, prach; in j, bhañj, mṛj (mārkṣya and mrakṣya), yaj, bhuj, yuj, vṛj, sṛj [also bhrajj, rañj, sañj, svañj, nij, ruj], while tyaj, bhaj, andmajj (man̄kṣya and majjiṣya) have both forms, and vij (vijiṣya and vejiṣya) and vraj take i; — in dentals: in t, kṛt cut and vṛt [also cṛt and nṛt] make both forms; in d, ad, pad, çad fall, skand, syand, chid, bhid, vid find, nud [also had, khid, svid, kṣud, tud]; while sad (satsya and sīdiṣya) and vid know make both forms [also chṛd and tṛd], and vad has i; in dh, vyadh (vetsya), rādh, sidh succeed, budh, yudh, rudh, vṛdh. [also sādh, krudh, kṣudh, çudh], and bandh and sidh repelhave both forms; in n, tan, while man and han have both forms; — in labials: in p, āp, kṣip, gup, tṛp, sṛp (srapsya and sarpsya) [also çap, lip, lup], while tap, vap, svap, dṛp, and kḷp have both forms; in bh, yabh and rabh, labh having both forms; in m, ram, while kram, kṣam, nam, and yam make both forms.

c. Of the roots reckoned by the grammarians as ending in semivowels (761 d–g) all take i. And vā or vi weave, vyā or vī envelop, and hvā or hū call take a y-form, as in their present-system, to which then i is added: thus, vayiṣya, vyayiṣya, hvayiṣya (but also hvāsya).

d. Of roots ending in spirants, the minority (about a third) are without the auxiliary vowel. They are: roots in ç, diç, viç, dṛç (drakṣya), spṛç (sprakṣya) [also daṅç, riç, liç, kruç, mṛç], while naç be lost has both forms (nan̄kṣya and naçiṣya); in ṣ, piṣ, viṣ, çiṣ [also tviṣ, dviṣ, çliṣ, tuṣ, duṣ, puṣ, çuṣ], while kṛṣ has both forms (krakṣya and karṣiṣya); — in s, vas shine, vas clothe [also ghas], while vas dwell has both forms; — in h, mih, duh, druh [also nah, dih, lih], while dah, vah, sah, andruh have both forms.

e. In the older language, a majority (about five ninths) of simple roots add the sya without auxiliary i; of the futures occurring in the later language only, nearly three quarters have the i, this being generally taken by any root of late origin and derivative character — as it is also uniformly taken in secondary conjugation (1019, 1036, 1050, 1068).