236 BtlEEAU OF AMERICAN ETHNOLOGY Fedll. 40 



MY FOREHEAD [from di^%l-]) . The third person is, at least super- 

 ficially, without ending in all nouns of this group whose pre-pro- 

 nominal form is not monosyllabic. The third personal form is 

 characterized by a falling accent on the final syllable, -a- and -i- 

 being lengthened to -a'" and -l'^ respectively. Other forms are : 



ts-!ele'i his eye 27,8; 86.7, 9; (cf. 54.6) 



do'^ma'l his testicles 130.8; 136.5 



xaHa'm his urine 



gwit'.i'n his wrist 



There is no doubt, however, that these forms without ending origi- 

 nally had a final -V, as indicated by the analogy of third personal 

 forms in -da in Scheme II, and as proved by the preservation of the 

 -t- before the reflexive sufiix -gwa and in monosyllabic forms : 



p!a''^nf his liver 120.2, 15 



nl'H' her teats 30.14; 32.7 



m'H' her husband ( 17.13) 



sd'H'^ his discharge of wind 166.8 



Though the conditions for the loss of a final -f are not fully under- 

 stood, purely phonetic processes having been evidently largely inter- 

 crossed by analogic leveling, it is evident that the proportion ^y^Z^'* 

 HIS house: nl'H' her teats = s'as-inl he stands: wif he travels 

 ABOUT represents a by no means accidental phonetic and morphologic 

 correspondence between noun and verb (Class II intransitives) . 

 The falling pitch is peculiar to the noun as contrasted with the verb- 

 form (cf. lie'^^l SONG, but liel sing!). Monosyllabic stems of Scheme 

 III seem to have a rising accent before -fgwa as well as in the first 

 person. Thus: 



lat'gwa his own excrement 77.1 



tnt'gwa her own husband (despite t.'i'H") 45.14; (59.16; 60.2); 

 128.22 



Nouns with characteristic -i- prefer the parallel form in -i'-Xndagwa 

 to that in -i'-fgwa. Thus : 



hu^hini'xdagwa his own arm, rather than hii^hini'fgwa, despite 

 hu^hiniYlc' my arm 

 The limitation of each of the two schemes to certain definite pho- 

 netically determined groups of nouns (though some probably merely 

 apparent contradictions, such as ga'l-H'ek' my bow and dl^^a%fV 



1 -t'¥ always requires preceding rising or raised accent. As gal- bow seems to be inseparalily connected 

 witli a falling accent (very likely because of the catch in its absolute form), it is, after all, probably a phonetic 

 reason that causes it to follow Scheme U rather than ni. 



§ 92 



