212 BUKEAU OF AMERICAN ETHNOLOGY [bull. 40 



two nouns that we have chosen as types da^-nxde^V shows no forma- 

 tive element in the proper sense of the word, while the -gw- of t.'iha- 

 gwa'nt'Y is such an element (cf . from stem liu- look liu-gw-ax-de'V 



MY face) . 



More characteristic of the Takelma noun than derivational suffixes 

 is a group of elements that are never found in the absolute form of 

 the noun, but attach themselves to it on the addition of a pronominal 

 suffix or local pre-positive. The -n- and -{a)n- of da°-nxde^Y and 

 t!ihagwa^nf¥ , respectively, are elements of this kind (cf. lia-da-n-de 

 IN MY ear; ha-t.'ihagw-an-de in my pancreas), also the -a- of danaH'¥ 

 MY ROCK (cf. Tia-dan-a' in the rock [from da'n rock]), and the -u of 

 Jia-fgdu in the earth 33.7 (from fgd earth). The function of these 

 elements, if they have any and are not merely older formative suffixes 

 that have become crystallized in definite forms of the noun, is not 

 at all clear. They are certainly not mere connective elements serv- 

 ing as supports for the grammatical suffixes following, as in that 

 event it would be difficult to understand their occurrence as absolute 

 finals in nouns provided with pre-positives ; nor can they be plausibly 

 explained as old case-endings whose former existence as such was 

 conditioned by the preceding pre-positive, but which now have 

 entirely lost their original significance, for they are never dependent 

 on the pre-positive itself, but vary solely with the noun-stem: 



ha-dan-a^ in the rock; dd'^-dan-a^ beside the rock; dal-dan-a^ 

 among the rocks; dan-a^-fk' my rock; da¥-dan-a-de over my 

 rock (with constant -a- from da^n rock 16.12) 



Jia^gwdH-a'm in the road 62.6; dd'^-gwd'^l-a^m along the road; 

 gwdH-a^m-f¥ my road (96.8) ; da¥-gwdH-am-de over my road 

 (48.6, 8) (with constant -am- from gwdn road 148.7) 



For want of a better term to describe them, these apparently non- 

 significant elements will be referred to as noun-characteristics. 

 Not all nouns have such characteristics : 



Jia-gela^m in the river (from gela^m river 21.14) as opposed to xd'^- 

 gulm-a'n among oaks (from gulu'm ook. 22.10, 11) 



Whether such nouns were always without them, or really preserve 

 them, but in a phonetically amalgamated form, it is, of course, 

 impossible to decide without other than internal evidence. 



A fourth nommal element, the pre-pronominal -x-, is found in a 

 large number of nouns, including such as possess also a characteristic 



