750 BUREAU OF AMERICAN ETHNOLOGY [bull. 40 



found after wl- as well as a-. Now, it should be observed that we have 

 liu after ii- m some stems regularly; in others it never occurs. As li 

 is unquestionably used as a glide, we are at once tempted to regard 

 the u as a morphological element. But a direct proof is wsiwiginitcin^ 

 HE WHO DWELLED THERE 80.9, 12, 20; 82.10, 22; 84.10, 21; 86.2, 20. 

 This form is a participial (§ 33), showing the characteristic change 

 of u to wd (§ 11). Hence the wd points to an initial u, which can 

 not be a glide, as nothing precedes; and li is absent. Now, this u is 

 found in d^Mwr^'wlgewatc^ when they went to live somewhere 

 66.15 (a — wdtd, § 29; Jclwi is an extended form of Jcl, an initial stem 

 denoting indefinite motion, § 16 ; 'Ic for Ic regularly after a). — T. M.] 



Other additions, like h, w, y, are clearly glides : 



d'hutciHc^ whence he came, the independent form of which is 



U'tclw"' HE CAME FROM SOME PLACE 



d'hundpdmiHc^ when she took a husband, a temporal form 



for una' 'pdmi'^w"- she took a husband 

 o'wlwA^n^ his wife (from owi-Ani) 

 owl'^tdwA^n^ his brother-in-law (from owi''td-Ani) 

 Jcetdsi'yutd^w"^ he crawls up hill (from Jcetdsi-utdwa) 

 M'jdwd''w"' he is jealous (from M-dwdwa) 



Consonantic Accretion 



A frequent type of accretion is w or y with k, forming a cluster: 



tca'^kwiwind^w'^ he is short-horned (from tCAgi-windwa) 

 tca'^kwdpyd'^w^ it is short (from tcAgi-dpydwi) 

 sdsl'ga ky d^w"' he scattered it (this is just the same in meaning 

 as sdslgd'kdw"') 



Intervocalic Consonants 



The most common accretion is t.^ It falls in between two vowels, 

 each of which is part of a different iiiember in a word-group. 



Examples : 



Between i and e: a^Jcwitepyd^g^ top of the water 

 Between e and a: neiA'^pAud'^n^ I laugh 

 Between a and o: d'wAtb^w"- he carries it away 

 Between d and u: py?i' insd'^w^ he came walking 

 Between o and d: p%'tbi')k''w°' he crawls in 



1 t serves as a connective in an inanimate relation, and will be mentioned again. 



§8 



