causals end in â, these forms are of frequent occurrence, and sometimes even an is inserted, as . In the old ha future, the â of the stem and the initial i of the termination frequently coalesce into, as— निशिचर मारि तुमहिं ले हें ॥ fas ye arceufe an ÎÊ | "For a few days, mother, sustain thy courage, Slaying the demons, will carry thee off; The three worlds, Nârada and all shall sing his praise." -Tulsi, Ram., S.-k. 36. where ऐहें = आइहें, जैहें = जाइहें, and गेहें = गाइहें ; so also we find पैहें “ they will find,” for पाइहें, mod. पावेंगे (Tulsi, Ram., S.-k. 10). In other places, however, we find the junction-letter व, as पावहि, आवहि, etc. In Panjabi the junction-letter for the aorist, imperative, and future is regularly व, as जावे “ he goes," जावेगा “ he will go,” but before o it is omitted, as जाओ " go ye,” जाओगे “ ye will go." For the past participle it is, as "been," masc. sing., and is omitted before e, as "been," masc. pl. In the three first-named tenses the is regularly inserted in pure P., but in speaking it is now sometimes, under the influence of Hindi, omitted, and is heard instead of the more characteristic देवेगा. In Sindhi all verbal stems end in a vowel, those stems, which in other languages end in a consonant, having in that language a short a or i. In this class of stems, before the neuter infinitive in y, a a is inserted, as "fall," infin. qay; fa “bow," infin. f. Before the active infinitive in no junctionletter is employed, as "measure,” infin. Ħy. Stems, whether active or neuter, ending in í and û, and sometimes those ending in o, shorten those vowels in the infini stems in â, including causals, retain the long vowel, as— The aorist follows generally the type of the infinitive, re taining the short vowel. In the persons is inserted as in P., except before û or o; thus at "I fall," is declined PI. 1. पऊं 2 पत्र 3. पवनि. "thou Sing. 1. पवां 2. पवें 3. पवे. The may be dropped before 2 and 3 sing., as au sayest," or "he says." The common verb fy "to give," undergoes contraction in this, as in all the other languages; thus 2 sing. "thou givest," not fi, 3 pl. fa, not feufa. 66 Verbs ending in a insert before â, ú, and o, as “I grow old,” “we grow old,” i “ye grow old.” बुढायूं The imperative and other tenses follow the general rule, which may, for Sindhi, and, to a great extent, for the other languages also, be thus stated; the junction-letters are ☎ and , is omitted before vowels of its own organ, as u and o, and is omitted before i and e; before â both are employed, but preferentially after short vowels, and after long ones. Thus, in the present participle, which is used as a future, a is either inserted or omitted, as Contraction also occurs as चवणु “ to speak, " चंदो (चवंदो ) अणु “ to be," हंदो (हुअंदो) डिअणु “ to give,” डींदो (डिअंदो). The past participle regularly ends in or, and the inserted is naturally dropped before it, thus If the stem ends in a palatal vowel or consonant, the i of the termination is dropped, as― थिअणु “ to become,” थिओ, not थीओ = थि + इओ. कुछणु “ to speak,” कुछो. सोझणु “to inquire,” सोझो . Passives naturally drop the euphonic of the active infinitive before their palatal junction-vowel, as— चवणु “ speak,” imperfect चउ, Passive चजणु “ to be spoken.” The stems quoted above, as shortening their radical long vowel before the termination of the infinitive active, naturally retain the long vowel in the passive, as There is very little to notice, in this respect, about G. The orthography of that language is still in so unfixed a state, that it is impossible to seize upon any principles as to junctionletters. One writer will insert them, while another omits them, or the same writer will insert them on one page and omit them on the next. Thus we find होए, होये, होअ, होय written in differently, also and aru. Until the natives of that province make up their minds as to how their language ought to be spelt, it is impossible for foreigners to evolve any laws or rules on the subject. § 78. Marathi is slightly more sensitive to hiatus, and has a greater fondness for the -sound than the other languages. There exists, consequently, in some persons of certain tenses, a system of Sandhi for Tadbhava words and forms, which differs in its general principles from that prevailing in Sanskrit. The grammar-writers, unfortunately, either omit entirely or only casually note these important combinations. The following remarks are offered as a contribution to the subject.1 The tenses of the M. verb, whose terminations begin with a vowel, are the aorist, imperative, future, and subjunctive, also the participles present and past, the conjunctive and infinitive. These are for the neuter verb; in the active verb the दू, which is inserted between the stem and termination, causes a collision of two vowels in the other tenses also. Verbal stems ending in all the vowels except a have here to be considered (khâ "eat," pî "drink," ghe “take”). Aorist, (in modern usage past habitual) and in the 3 pl. the final is elided. In 2 sing. both and (not are changed to the palatal semivowel before â, though not before û, so that we have, side by side with प्यूं), घेऊं (not यूं). For the illustrations to this section I have to thank Captain G. A. Jacob, Inspector of Schools, Punâ, who kindly furnished me with details which are wanting from most of the grammars. In the future there is similarly in 2 pl. खाल, प्याल, घ्याल, and so also in the imperative 2 pl. घ्या, प्या; ये “come,” makes aorist 2 pl. या, future याल, and occasionally one hears ये aorist 3 sing. for the more regular येई. In the subjunctive the semivowel occurs again, as घ्यावा, घ्यावा, but यावा simply from ये. This last verb should, by analogy, form, but the double y in such a position would be unpronounceable, and a single y is therefore exhibited. It must not be supposed that the e of has simply been dropped. For the potential the termination of the present a might have been expected to be simply added to the verbal stem, as no hiatus would thus be caused. But the origin of this form from the Skr. part. in renders this course impossible. The of having suffered elision, there naturally results an hiatus. Thus from खादितव्यं, Pr. खाइअव्वं gives खावव, the first supplying the place of the lost, so that in the potential we get not खावतें, but खाववतें. Similarly पीववतें, घेववतें, येववतें, and even in stems ending in ह, as लिह " write,” लिहववतें. Stems ending in u preserve the hiatus almost throughout, thus dhu "wash ". In the present participle only त is added, not ईत or अत, as खात, पीत, घेत, येत, धुत, and in the past participle the semivowel is generally used, as |