which the द has entirely dropped, and instances हसि for हसितं "laughed," fe for ufa “recited," and this form has become the type of most modern languages. In Old-Hindi this participle regularly ends in sing. ì m., † ƒ., pl. u m., †ƒ., as arent m., at ƒ., etc., "burnt." Here the represents the जल्यौ जली of the Prakrit, hardened into a semivowel vowel. In the feminine it is merged in the before the final of the affix, and in the plural lost altogether, for easily passes into . Chand uses this form throughout, as तन रह्यौ जोति गय देव "his body remained bright, he went to the abode of the gods" (i. 299); "done," "gone," etc. it It is, however, more frequently used as a tense than as a participle, and further illustrations will be given in a following section. The form in lasts all through the medieval poets, and is still in use in the dialects of Rajputana and in Braj. In the former a slight change has occurred, sing. , pl. while in Kumaon the form is sing. मारियो, pl. मारिया . Modern classical Hindi has sing. मारा m., मारी / ; pl. मारे m., मारीं./., “ struck.” Panjabi retains the of the Prakrit, and has sing. ĦIFTAI M., मारी./.; pl. मारे ., मारीआ/., “ struck ;” so also does Sindhi, sing. हलिओ or हल्यो m., हली / ; pl. हल्या "., हलिऊं. /. Trumpp seems to be here in error in saying that the has been inserted to fill the hiatus caused by the elision of the . It is rather theof hardened to a semivowel, as in Old-H. and P. Oriya has rejected the final syllable, just as it has in its present participle, and has an indeclinable past participle in i, as dekhi. This is never used alone, but only in composition, with an auxiliary forming a tense. The past participle used to form the passive ends in â, like H., as dekhâ jîbâ “to be seen. The same form is found in Gujarati, as sing. (chhoḍyo) m., छोडी./., छोड्णुं (chhodyü) n.; pl. छोड्या ., छोडी /, छोड्यां n. G., however, in common with M. B. and O., has another form of this participle ending in an affix, whose special type is B. देखिल, in Old-B. देखिला (only used as a tense combined with as), The Bhojpuri dialect of Hindi has also an indeclinable past participle मारल, in some districts also मारिला, from which it forms a tense. Here the junction vowel varies much. In B. O. and the active of M. it is . In G., on the other hand, it is , while in the neuter of M. it is a. M. has a long string of verbs, both active and neuter, with the junction vowel â; some of these are causals by origin, as ч "flee,” p.p., for palâ-ilâ (as in B. and O. ). Others, again, owe the long vowel to a Skr. ay, as उड “fly,” p. p. उडाला, Skr. उड्डयित_ / उद + डी. Others are denominatives, as दीप “ be dazzled,” दीपाला, Skr. tuf; there are, however, some which I am not able to explain on any of the above grounds. The list comprises about thirty verbs only, and in twenty-five of them participles, with the junction vowel a, are also in use. The usual explanation of this form in 7 is that it is derived from the Skr. p.p.p. in ita, through Pr. ido, by change of to , and thence to. The change is undoubtedly possible as far as and are concerned, or as far as andare concerned; but the change from to is a great stumbling-block. द ड The great authority of Lassen (p. 363) is usually quoted in support of this view, but even he cannot avoid being struck by the coincidence between this and the Slavonic preterites in l. As regards the change from to, it is observable that it only occurs in those Skr. preterites which contain a cerebral. Thus becomes in Magadhi (Mr. 270). Here, however, there was evidently a form कर्त = कट = कड, so that there is no question of a द at all. So also in वाबड for व्यापृत = व्यापर्त = = (Mr. 227). The only other instance known to me is for = (Mr. 276), but here we may fairly assume a false analogy with a. So widespread a form as the modern participle in / must rest upon some firmer proof than the rare examples given above. I am disposed to think that we have in this participle the survival of an ancient form which has not been preserved in classical Sanskrit, nor in the written Prakrits. Perhaps (but here I tread on ground somewhat beyond my own domain) that type of the passive past in Skr. which ends in or may be the classical representative of this ancient form; thus we have from √लू “ cut,” लून:, from / भिदु “ cleave,” भिन्नः, and in some roots both forms, that in a and that in, stand side by side, thus √पॄ “fill,” makes पूर्ण: and पूर्त:, √ नुद् “push,” नुन्नः and नुत्तः Even in the Slavonic languages, however, the characteristic / of the preterite is thought to have arisen from an original d, and that again from t. If this be so, we have here an ancient change which took place before the separation of the various members of the Indo-European family, and not a mere local corruption confined to Indian ground. In Russian the preterite is a participle with forms for gender, thus from dielať “to make," pret. sing. dielal' m., dielala, dielalo n., pl. dielali mfn.2 In Servian the same form occurs, trés "to shake," has― Sing. trésao m., trésla f., tréslo n. Pl. trésli m., trésle f., trésla n. 1 Rapp, Verbal-organismus, vol. i. p. 99. from "to trouble." The similarity is striking, and seems to be more than a mere accidental coincidence. Moreover, the connection between this Slavonic and n is shown in more than one instance. Thus, the Russian verb has from nes "to drag" a pret. past sing. nesén m., neséna ƒ., neseno n., pl. neseny. The same form occurs in the Czech. But we are getting beyond bounds. The comparison is attractive, and, if there were time to study the Slavonic languages as well as the Indian, might perhaps be worked out to some conclusive result. All that can be said at present is that two groups of the same family have a preterite in 7, and that there may be some connection between the two; while, on the other hand, the derivation of this preterite from a past participle in t seems strained and ill-supported as regards the Indian group, and if true for the Slavonic group, must have occurred a long while ago, before the separation of the families, and has strangely failed to leave any traces of itself in the most important language of the Indian group in its most cultivated stage. § 46. Let us turn to matters more within our scope. The passive past participle is the only part of the modern verb which affords an exception to the general rule of the unchangeableness of the stem-syllable. Each one of the modern languages has a few such participles, which, being derived from the Prakrit developments of the Skr. p.p.p., differ from their respective verbal stems, which latter are derived generally from the form of the root used in the present tense. These early Tadbhava participles, as they may justly be called, are most numerous in Sindhi. Trumpp gives (p. 273) a list of no less than 140 of them, a number which far exceeds that to be found in any other of the languages. They owe their existence chiefly to the omission in Skr. of the intermediate, so that the affix of the p.p.p. is added directly to the root, and when this root ends in a consonant, there arises a strong or mixed nexus, which in Prakrit has to be dealt with according to the ordinary phonetic laws. Sometimes, as we saw in $ 14, the stem of the verb itself is entirely borrowed from the p.p.p., and in that case the modern participle does not differ from the rest of the verb; but when the ordinary stem is derived from the older present, and only the participle from the old p.p.p., the two differ so much that it is difficult at first sight to recognize the connection between them. The verbs given in § 19 have mostly old Tadbhava participles, and it is through these participles that the clue is found to the derivation of the verb. Thus— |