14 ANDERSSON, COMPARISON OF COTTUS POECILOPUS WITH COTTUS GOBIO. 
My measurements given in the table above show that 
the length of the pectoral fins is somewhat greater in the 
smallest specimens of C. gobio than in specimens of C. poeci- 
lopus of the same size, but that in others the fins are of 
about the same length in the two forms, generally reaching 
to the beginning of the second dorsal fin in both. 
The length of the innermost ray in the ventral fins. 
HECKEL says in his first work that the innermost ray in 
the ventrals of OC. poecilopus is very short, only !/a4 of the 
length of the fin. In C. gobio it is, he says, considerably 
longer. Nevertheless, he seems to attach no great weight to 
this character as he does not mention it in his later work. 
This must be regarded as very singular, for this character 
is, as Professor Smitt remarks, the only one of those given 
by HEcKEL which is proved to be constant. This character 
is not stated by GÖNTHER either in his work, »Die Fische des 
Neckars» nor in »the Catalogue of Fishes»>; yet he in the latter 
distinguishes the two species. NILSSON, LILLJEBORG, and SMITT 
all employ this character in their diagnoses; likewise JEITTELES. 
To judge from my measurements, this ray is in OC. poeci- 
lopus about 2/3, in OC. gobio about ?/3 of the length of the 
ventral fin and in not a single specimen of C. poecilopus was 
it half as long as the fin, while in all the specimens of CO. 
gobio it exceeded this measurement. The greatest length in 
C. poecilopus was 46,9 2 of that of the fin and the least in 
C. gobio 56,8 Zz. The following table shows the average length 
of the ray in proportion to the length of the fin. 
C. poecilopus. | C. gobio. 
: EN [| 45,2| 57,5] 76,1) 42,51 57,91 75,5 
Length of the body in millimetres (average) 12 43 .4| 62.71 76.1] 42.4l Bs 
4 ,” ,! , 5, , 
Length of the innermost ray in 4 of the | 31 36,0) 36,7| 33,8) 63,4 | 72,3) 70,7 
TERO OH OL RE MITT SES RA hare dh AE | ; ; fd i 
12 | 39,71 32,6| 28,81 67,2| 68,7) 69,3 
Consequently, by this character we are always able to 
distinguish the forms, at least the Scandinavian ones. An- 
other question still remaining is what systematic value we 
