THE HAMILTON ASSOCIATION. 71 
be remarked that the apparent irregularities in Greek are merely the 
maintenance of the primitive type which in Latin has been wiped 
eut by the influence of analogy. We may add that all appearances 
go to make it probable that originally the oblique cases, except the 
acc. and in some degree the Locative, accented their suffixes in all 
words and not merely in monosyllables. 
In the light of this explanation some irregularites so-called ap- 
pear in their true light as preserving a preethnic relation lost in the 
bulk of words through the working of analogy. Thus the difference 
between zez of zeus and dir of dios, are simply like that between dyau 
of dyavs and div of div-ds, etc., the weak form being caused by its 
immediately preceding the accented suffix. Even the accent of zez7 
is explained by the fact that vocatives when accented always ac- 
cented the first syllable. A similar recessive accent, by the way, in 
the zeut. comp. -ion isa retention of preethnic accentuation. Compare 
svad-syaus with hedion and hédion and it will be seen that the latter 
preserves the primitive accent. 
We see, therefore, that the principle of gradation or ablaut, 
resting as it does upon the shifting accent of preethnic times, far 
from being confined to Teutonic conjugation, is a guiding principle 
in Greek and Sanskrit conjugation and declension, and in the word 
formation of all these tongues. Had time permitted we might 
have shown its traces in Latin (e.g. as in fido ( fezdo), foedus, fides), in 
Trish (as den gen mna cf Dor. dana, and mna-omat), and in Avestan . 
but enough has been adduced to show its wide-reaching importance 
in moulding the primitive “‘ speech-material,” and how its recogni- 
‘tion throws light upon many obscurities and combines apparently 
unconnected phenomena into one harmonious whole. 
