^{page 170 note 1} In the case of Sk. ā, a + two conss. the development was as follows: It became a + double cons, in the pre–second century A.D. period; then, when the double conss. became single, i.e. second century A.D., this preceding a was compensatorily lengthened [cf. § 50, 3]. This a when followed by i became ā, just as original Sk. ā – i > d: cf. Sk. vāpi– > vav– likewise, Sk. hastin. > pre–second century A.D. hatti– (cf. Mid. Ind. hatthi–) > second century A.D. to fourth century A.D. hāti– (āti–) > fourth century A.D. onwards āt–.

^{page 172 note 1} The distribution of use between the two gerundial suffixes, –tvā (with uncmpd. roots) and –ya was entirely ignored in Mid. Ind. [cf. Chatterji, ODBLang. § 738].

^{page 175 note 1} In the case of Sk. u + two conss. the development was as follows: it survived as u –fdouble cons, in the pre–second century A.D. period'; then, when the double cons, became single, ie. second century A.D., this preceding u was compensatorily lengthened [cf. § 50, 3]. This i > i –i, just as original Sk.

^{page 181 note 1} In the case of short vowels (or short followed by long), contraction has taken place in loanwords which entered the language after the time of assimilation of the vowels concerned, but before the loss of the consonants [cf. Introduction].