22 BUREAU OF AMERICAN ETHNOLOGY [hull.g.j 



31. Rainbow. G, accade; Sk, akja'ielokl; 40. Yes. G, couam (or counns? or co- 

 Bo, dccaU'd-, dc-kioribek; Ir, hebnai. uausf), allous ( = ''oui certes"); Sk, 



32. B.ed. Sk, klru(-kwa'irR); Bo, chieoqvAr. ai'lo:; Fe, ailaou; H, alelkal-ou, tach,^ 



33. Run. Fi, ahl (=rush); H, ali; Bo, ouou;* Fi, o'o; Bo, da. 



dlcsc (=al80 run away!); Sk, a'ld§'; G, 41. Yesterday. G, area; Bo, dlcualdk. 



nl chy ( = "je vais partir"), alcherba 42. Thing, etc. "The suffix kwarrR is 



( = '"allons-nous-en"), yet lepert ^ (="il very common and seems to designate a 



s'enva"). condition or a quality possessed hy 



34. Sick, ill. Hjkoume or yaouil(=sick- something: hence dirk'sta, to sleep, 

 neBs);Fi,yau'hdl(=sickneaa);Bo,dl-ler, ci'rk's'ta-kwarrR, being asleep. The 

 dl-lel, dl-lelk, halen; Be, halen; Sk, colornamesalsoendin A- irarri?; a thing 

 d'l6l{k); Fe, alilki; Vy, deuf? is pa'lkwarrR, i. e., possesses a black 



35. Sit. H, choukouil; Fi, shUckd; Sk, color. Another example is a'rxkje:l, 

 H'J.drR; Bo, scidcdrk; G, houche. arrow; a'rxkje:lktvarrR, quiver='the 



36. Swim. H, kcl-i, lahnp-a'i; Fi, Um'pi; thing that holds the arrows' " (Skotts- 

 Sk, d'llpai:l. berg, d, 606). For examples of this 



37. Thin. Sk, d'jip; Bo, d-iepl. suffix in Bo, H, Fi, see the words al)ove 



38. To-morrow. Bo, terrudiacili, dl cud- forwhite,red,black, and green; the final 

 lak; G, colas. r or rrR is elided, as frequently, in H, 



39. Urinate. Sk, skarrS'; G, quesquer. Fi. Cf. also Group IV, 30, Bo, olacar. 



1 -lepert may be the same as Upper in Bo's sen- ^ In Yahgan the word da,s (Fitz-Roy) or tas 

 tence cial-chi'l-cdl (=they) jakscidrchier Upper (=go (Hyades, q, 270) is used for yes, but rarely. 



to the field). ^ A Yahgan word for assent is aoua'i (Hyades, 



2 Cf. 13, Bo, Sk. g, 270), auai (Bove), ow-wy (Despard, b, 718). 



DISCUSSION OF GLOSSARY 



0^ving to the paucity of available material and the total lack of 

 grammatical data, much in the foregoing comparative study is neces- 

 sarily tentative and provisional. Nevertheless, there appears to be 

 sufficient evidence on which to base certain dependable conclusions. 



The two most important lists. Bo and Sk, agree in from at least 

 80 to 90 per cent of the cases, so closely, in fact, that there can be no 

 reasonable doubt that they represent the same language. The same 

 is true of Fe. G agrees in quite the majority of cases with Bo, Sk, 

 and Fe, as do also the shorter lists Se, Lu, Sp, Li, Cy, Be, Si, Co, and 

 Ir, while some of the stray words from Duclos-Guyot, King, Mac- 

 douall, Meriais, and Wyse can also be identified. H and Fi closely 

 resemble each other and in the main show manifest afiuiity yviih the 

 other vocabularies, but on the other hand contain many words 

 peculiar to themselves. 



The 15 lists and other words tlierefore fall into two groups, one 

 represented by H and Fi, the other including the remaining material. 

 Do these two groups represent two distinct languages, or at least 

 two distinct dialects, or does the evidence call for some other ex- 

 planation ? 



That they represent one and the same language seems fairly, 

 reasonably clear, for in about 60 to 70 per cent of the 115 to 120 

 words for which comparison is possible, there appears to be either out- 



