BIHANG TILL K. SV. VET.-AKAD. HANDL. BAND 24. AFD. IV. N:0 9. d 
simultaneously the eggs were much smaller, not measuring 
more than 1 mm. in diameter. The ovaries in these were 
however so much swelled that it seems probable that the 
spawning might take place in the first part of the winter or 
perhaps in the later part of the autumn. The ovaries of both 
sides are broadly connected with each other. The colour is 
probably (reddish) yellow during life. The liver lies on the 
left side and covers the stomach and the anterior part of 
the ovary. It is not divided into lobes. The appendices 
pylorice are 5 in number. The intestine lies wholly on the 
right side. It makes a triple convolution between the pylo- 
rus and the anus. That is: it extends from the pylorus to 
just behind the anus from where it bends forward to the 
pylorus; from there it descends for the last time to anus. 
The last mentioned descending part lies most median, the 
ascending in the middle, and the first descending most lateral. 
The soft anatomy of this species thus resembles, as might be 
expected, that of Cottus scorpius, but the intestine is a little 
longer and the pyloric appendages are less numerous. 
CoLLETT ! found the food of this fish to be small crusta- 
ceans and annelids. LÖTKEN? only observed worms. I found 
in the stomach of one of the larger specimens fragments of 
Crangon boreas, in another smaller crustaceans (gammarids). 
It seems thus to be less voracious than its ally Cottus scor- 
pius. 
It is worthy of remark that only the largest specimen 
measuring 175 mm. in length is provided with granulated 
warts on the top of the head. The three others (not to speak 
of the yearlings) although mature and measuring resp. 165, 
145 and 136 mm. have the upper surface of the head smooth.? 
In this respect then these specimens agree with the north- 
eastern form Gymmnocanthus pistilliger PaALrraAs which F. A. 
SMITT in his standardwork »Scandinavian Fishes» (second edi- 
tion) has distinguished and characterized. But the other 
characteristics mentioned by SmMITtTt (1. ce. p. 161 swed. ed.) do 
not show such variations, so that the validity of G.-. pistilliger 
' The Norw. North-Atl. Exped. Fishes, p. 28. 
> Vidensk. Medd. Nat. Foren. Kjöbenhayvn 1876, p. 11. 
> A similar observation seems to have been made by LöÖTKEN (Il. c.), but 
as he unites G. ventralis and G. pistilliger into one species it is uncertain 
in which form he found the head smooth. (Conf. below.) 
