Of the four sibilants which the cuneiform symbols are capable of expressing, Hittite evidently had no use for two; s and ṣ occur so rarely in Hittite words that we are justified in explaining away such instances as are found (e.g. one should read zé rather than ṣe). The commonest Hittite sibilant is ordinarily written by the symbols which most Assyriologists transcribe áš, šá, iš, ši, uš, šú, etc. It corresponds etymologically to Indo-European s (šagg- ‘see’ = Gothic saíhwan, šipand- ‘pour a libation’ = Greek σπέvδω, šittar ‘star’ = Sanskrit star-, ešun ‘I was’ = Skt. āsam, ešhar ‘blood’ = Skt asr̥k, kwiš ‘who’ = Latin guis, etc.). The simplest interpretation of the facts would lead us to pronounce these words with s, and there is nothing in the history of the cuneiform writing to conflict with such an interpretapretation.

The Classic Latin words for ‘reed’ were harundo, f. and calamus, m.; the first, apparently, has not come down in any Romance tongue, the second is Ital. calmo, Fr. chaume ‘thatch’ and chalumeau ‘rustic pipe’. The Roman people also used for ‘reed’ canna, which was a borrowing from the Greek, and this word has been a flowing source of derivatives in all the Romance languages. The present paper aims to show that the list of derivatives is even more extensive than has been supposed.

The history of English since 1050 may be regarded as consisting of two periods with the year 1300 as a very rough terminus ad quem for the first and terminus a quo for the second. The period from 1300 to the present has been characterised by very extensive changes in the phonetic form of English and by important syntactic changes but not by correspondingly great changes in the morphological pattern of English speech. The morphological changes in nouns and adjectives have been almost wholly the result of the loss of final e and to a very small extent the result of analogical processes. Analogical processes have resulted in important changes in the inflectional pattern of verbs but in verbs also changes in morphological structure have been to a much greater extent the result of other causes, especially the displacement of forms ending in -en by forms ending in e and the subsequent loss of final e.

The Greek κ-Perfect presents the most favorable conditions possible for testing the validity of a widely prevalent theory as well as for detecting the steps by means of which a formation grows from its first beginnings to an extensive category. Since it is formed from only twenty or twenty one verbs in vowel stems in the language of Homer, it is reasonable to believe that these represent, in approximate fashion, the earliest strata of κ-Perfects in the Greek language. We are thus able, as it were, to observe the category in the very process of formation, and a close scrutiny of the Homeric forms ought to give an answer to several questions. Are the verbs such as to give support to the common derivation of the κ-Perfect from the ‘root determinative’ κ found in the singular aorists ἔδωκα, ἔθηκα, and ἕηκα, κα as opposed to the plurals ἔδoμεν, ἔθεμεv, and -εἷμεν?

The origin and signification of the prefixes of Tibetan have in the past been the centres of a variety of theories, and among them m-has naturally excited speculation. The light thrown upon the matter by the native grammarians does not seem to greatly clarify the issue, although in the case of m- their main statement has, perhaps, a significance so far unnoticed. The object of this note, accordingly, is to bring out this fact, and to illumine the origin of at least some of the substantives in m-. To this end let us first examine the main contention of the Tibetan grammarians with regard to this prefix.