88 



species, do not jxisses.s ;inv siicli vnliii' tor classitication of I'aeitic cdust ti>lies. 

 I'^acli so-ealled si)e('ie.s seeins to he in a very unstable state of eciiiilibriuni, and not 

 to liave yet assumed or been able to retain, with any degree of pernianeuee. any set 

 of specific characters." " The crosswise series of scales [in At-osia niibiia ((Jrirard)] 

 varies from 47 to 7(» in nnnd)er ; the i)arbel [a generic character] is present or ab- 

 sent; the pliaryngeal teetli vaiv from 1, 4-4, to 2, 4-4, 1; and tiie dorsal fin 

 \aries mncli in position and somcwhal in si/e. These characters occur in various 

 combinations, and with some of these are often correlated i)ecnliarities of j)hysi- 

 ognoiny and general appearance, all of which may serve to put a cei'tain stamp 

 U|)i)n the iu'lividuals from a single stream, or even from one locality in a stream." 

 These observations, especially those contained in the last sentence, accord exactly 

 with the results obtained by me in another fish and contirm my statement which 

 will be further re-enforced by the jiresent paper, thai '"each locality has a variety 

 which in the aggregate is diti'erent from the variety of every other locality." 



The remarkable variation of the Pacific slope species, and more especially 

 the varia ion in the fin rays, was first noted in prejiaring my account of the speci- 

 mens coUecteil in the (ohnubin and Frazer basins.* This variation was most 

 lironounccd in the species of the late genus Kichardsonius. Of the species of this 

 uenus, I had about "J')!* specimens, coUecIed in the Frazer and Columbia systems, 

 from tide water to an elevation of i,?.*^!! feet. The later ex j)lorations of Gilbert and 

 l'".\ ermanii have increased this numl)er to S'2o, and these warrant a re-examination 

 4)1' llic points stated by me. \'\)V all (he ^\nt^^ concerning the lin rays of the s|)eci- 

 mens collected bv (Jilbert and Evermann, I am indebted to them. Their examina- 

 ii(rn of these specinu^ns was made to test certain conclusions reached l)y me, and 

 their data, therefore, join mine. In counting the anal rays, I counted the rndi- 

 nicuts at the beginning of the tin. These were not counted by (iilbert and Ever- 

 mann, and to bring their data in perfect accord with mine, it is necessary to add 

 two to the nnmln'r of anal rays. ^Vhile the number of rucrnuentai-y rays is not 

 alwavs two, it is so often that the exce|itions would probably not alter the gen- 

 (L'ral results. 



At the time 1 began my studies of these lorms, they weie regarded as two 

 species, forming a peculiar genus, Kichardsonius. 'I'hey were kiu)wn to inhabit 

 the ("olnnd)ia river and the streams about I'uget Sound. The compressi'd belly 

 behind tlu' ventral (ins was regarded as the cliaracler separating them generically 

 t'rgni the related fi>rms. It soon became exidcnt that, while some specimens 

 possessed this, if constant, un(pieslionai)le generic character, others did not show 



■•This variation in the same species does not seem to be confined to the fishe.*. Professor 

 Ritter Proc-. (',a\.\ Acad. Sci.. 2d ser. Vol. 1\'., p. ''7. finds tlie same in Prrnphrira nnnnctens 

 a new tunicate doscrilicd by hini. 



