146 Wisconsin Academy of Sciences , Arts and Letters . 
in German we have in the nominal inflections the frequent plural ending -er 
producing Umlaut of the root syllable, e. g., Thai, Thaler; Buck, Bucher; 
Grab, Grdber; Irrthum, Irrthumer; Kalb, Kalber, and very many others. 
All of these are modern except Kalb, Kalber, which must be considered as 
having furnished the starting point for the whole category. The plural 
Kalber therefore in the number of forms it has called into existence would 
furnish an exact parallel to SeSooxa. 
An interesting and instructive illustration is furnished also by the Italian. 
The Latin steti, stiti became in Italian stetti. This form was felt as 
consisting of st- as root and -etti as the ending. So by the analogy sto, 
stare, stetti, we got in Italian from do, dare, the perfect detti instead of 
diedi (Latin dedi). The ending -etti was then applied to other verbs, such 
as vendetti, fremetti, credetti, dovetti and many others. I have not been 
able to determine the exact extent of the application of this ending, from 
lack of the requisite books of reference, but am convinced that it is quite 
extensive. 
The example is at all events a striking one of how even one single form 
can furnish the model after which innumerable others may be constructed. 
Stetti, therefore, would furnish a complete analogy of what we have 
claimed for SsSGoxa. It is historically certain for stetti', ifc is therefore 
very possible for 8s 8 coxa. 
The, objection then to one form’s being the starting point for many 
would seem to be met by the above considerations. 
Hermann von der Pfordten in his Geschichte des griechischen Perfects 
calls attention further to the fact that the other perfects from stems in 
-x- as: 
8sSopx-a (root 8rpx-) 
jue-juvx-a (root juvx-) 
TE-TT/K-a (root T7/X-) 
/ isjUT/K-Go 5 (root ppx-) 
TC£-(ppiK-CX (root ppiK-) 
soix-a (root Feik-) 
may have contributed to the illusion that -xa was a tense suffix and so 
have assisted in its application to other roots. 
In fact von der Pfordten thinks that even the so-called -xa- aorists 
EdGoxa, E^pxa, i/xa may have contributed their share in the same direc¬ 
tion, especially when one considers the great frequency of these words 
(some 500 times) in the Homeric poems, the very stage of the language at 
which we find the -xa- formation just beginning to assume greater propor¬ 
tions. 
Brugmann’s theory then may be summed up in the following words: 
There was in Greek an early confusion of two roots of the same meaning, 
dca-and 8 cox-, da?-being used for the present and future, 8cox- for the aorist 
and perfect. By reason of this confusion the ending of SeScoxa was felt 
as -xa instead of -a, as it really was; and being once felt as a tense suffix, it 
