84 BUREAU OF AMERICAN ETHNOLOGY [Bull. 191 



in the charts are morphologically defined, i.e., their occurrence cannot 

 be predicted in terms of phonological environment. Phonologically 

 defined allomorphs (e.g., giiz derived by morphophonemic vowel 

 assimilation from gaz) are not listed in the charts but may be predicted 

 from the morphophonemic processes described under "Morpho- 

 phonemics." 



The same upper case letters have been used to identify allomorphs 

 which tend to be affiliated with the same group of verbs. This 

 consistency is especially evident in transitive verbs and within each 

 decade class: 



cf-zuwA (lOlB + verb stem) I -paid him 



sd^i-zuwA (102BH-verb stem) I {didn't) pay him 



tl-zuwA (103B + verb stem) maybe I paid him 

 etc. 



This tendency also is manifested to a lesser degree from one decade 

 class to another within each mode: 



ci-zuwA (lOlB + verb stem) I paid him 



si-ziiwA (lllB + verb stem) you paid him 



zl-ziiwA (12lB + verb stem) he paid him 

 etc. 



The fact that not all affixes have the same number of allomorphs, 

 makes it obvious that this tendency cannot hold true in the case 

 of every verb. Intransitive verbs, especially, often show a shift from 

 one series of allomorphs to another even within the same decade class : 



S9za-'i*Abe (lOlG+verb stem) we two ate 



sga-'i'Abe'i* (102A+verb stem) we two (didn't) eat 



tazd-'i'Abe (103E + verb stem) maybe we two ate 

 etc. 



The phonetic shape and distribution of many affixes of Century 

 Class 100 is suggestive of certain historical processes. It is probable 

 that many of the B-allomorphs were derived from the corresponding 

 A-aUomorphs either through glottalization (compare 141B with 141A, 

 for example) or through palatalization (compare lllB with 111 A). 

 B-allomorphs that appear to be a palatalized form of the corresponding 

 A-aUomorphs nearly always occur with verb stems beginning in -i-. 

 The few exceptions prevent the phenomenon from being described as a 

 morphophonemic process. 



Another type of historical development is suggested by the 

 overlap in phonetic form of certain allomorphs of affixes belonging to 

 different decade classes. This overlap is evident when one compares 

 affixes of Decade Class lOOl with those of 160, affixes of Decade Class 

 1101 with those of 170, and affixes of Decade Class 120 with those of 130. 

 This suggests that the system of pronominal affixation may have 



