79 JOURNAL AND PROCEEDINGS. 
tion. These examples, which might be multiplied indefinitely, 
show how important is gradation in derivation as in conjugation, 
and that not in one or two branches of the Indo-European family, 
but throughout its whole range. 
A question naturally arises, are these nouns and adjectives 
- which show root forms similar to those of certain tenses, derived 
therefrom? This would involve great difficulties, as the laying aside 
of verbal and the addition of nominal suffixes. Rather we should 
say that the vocalization of the noun is called out by the same 
accentual conditions as is the vocalization of the similar verb-form. 
This is plain in the case of those which, like drunkard, p1s-tos, 
bhit-ti, ducem, have weak forms. And even in such words as £/opoés, 
tomé, the requisite condition may have been furnished by certain 
collocations of words For, as Sweet says, ‘* At first all sound- 
changes are carried out consistently throughout each breath-group 
without regard to word division” (whence he adds the initial muta- 
tions of Celtic and the Sandhi of Sanskrit, comparing even our 
English ¢ie man, the-erth.) And the words which have acute 
accent on the o or the weak vowel-form may not possess their 
original accentuation, but may have for some reason changed it. 
Thus there is an important rule that with the suffix o (Skr. a) adjec- 
tives andnomina agentis are mostly oxytone, as fords, /a-ra, nomina 
actionis barytone as gdz-0-s, Skr. gan-a, birth. 
As these remarks have swelled to a tedious extent, it would be 
well to pass to declension. Here we see that Sanskrit and Greek 
monosyllables have a similar rule being oxytone in the oblique 
cases other than the acc. (in Skr. also the Locative, but not acc. 
pl.) thus : 
podos, podi, podes, pois, poda, podas,  podon, 
pat, pad-am, padas, padé, padas, padas, padam. 
But in Sanskrit we find that while the nom. and acc. sing. and nom, pl+ 
have stem pad with @ long the other cases have fad with ashort. Here 
the principle is plain. The forms that accent the stem have the strong, 
those that accent the suffix have the weak form of the stem. This isa 
fundamental rule in Sanskrit declension, every noun of certain wide 
spread classes having its strong and its weak stem. Thus f7/a, patér 
has strong stem pitar (pater) and its weak stem Aéér (like pair) or 
before a consonant fitr, (like patra for patr in patrast.) Here it may 
