BOAS] HANDBOOK OF INDIAN LANGUAGES TAKELMA 265 



t.'eimi'^s one hundred (probably = one m.a\e[t!V-]) 



me^l t'gd°'^mi'^s crows earth-one ( = land packed full of crows) 



144.9, 11, 12, 13 

 de^ml'^s in-front-one ( = marching in single file) 

 almi"s all together 92.23, 24; 190.17 



Of the two forms for two, gd'plini'' seems to be the more frequently 

 used, though no difference of signification or usage can be traced. 

 gd'pHnV TWO and xi'hini'' three are evident compounds of the 

 simpler gd'^m and xi^n (seen in TiaHxVn eight) and an element -hinV 

 that is perhaps identical with -hini^ of Tia'-bini' in the middle. 

 gamga'ra four is evidently reduplicated from gd'^m two, the falling 

 accent of the second syllable being probably due to the former 

 presence of the catch of the simplex. An attempt has been made^ 

 to explain delial five as an adjectival form in -al derived from de^- 

 IN FRONT. The numerals six, seven, eight, and nine are best con- 

 sidered as morphologically verbs provided mth the compound prefix 

 Tia^i- IN THE HAND (scc § 35, 4), and thus strictly signifying one 



(finger) is in the hand; two, three, four (fingers) ARE IN THE 



HAND. No explanation can be given of -go^ in JiaHgo^ nine, except 

 that it may be an older stem for four, later replaced, for one reason 

 or another, by the composite gamga'nn two + two. i'xd/il ten is 

 best explained as compounded of l-x- hand (but why not lux- as in 

 lux-de'V MY hand?) and the dual -di^l, and as being thus equivalent 

 to two hands. 



It thus seems probable that there are only three simple numeral 

 stems in Takelma, ml'^^s one, gd'^m two, and xi'n three. All the 

 rest are either evident derivations from these, or else (dehal probably 

 and i'xdil certainly) descriptive of certain finger-positions. While the 

 origin of the Takelma system may be tertiary or quinary (if -go'' is 

 the original stem for four and dehal is a primary element), the 

 decimal feeling that runs through it is evidenced both by the break 

 at ten and by the arrangement of the numerals beyond ten. 



The teens are expressed by ten one above (i. e., ten over one), ten 

 TWO above ; and so on. ga^aH thereto may be used instead of gada'Y 

 over. Twenty is one man, i. e., both hands and feet. One hun- 

 dred can be plausibly explained as equivalent to one male person.^ 

 The other tens, i. e., thirty to ninety inclusive, are expressed by 



1 American Anthropologist, loc. cit., where five is explained as being in front, on the basis of the 

 method of fingering used by the Takelma in counting. 

 ' Loc. cit. 



§ 110 



