Mr. C. V. Buike 07i the Cyclogasteiidae. 509 



Arctic Ocean can bo identified by tlie followir)g key. All 

 of tliese species have not been carefully studied by the writer, 

 because of the lack of material, but they appear to be well 

 defined and easily recognized when we become familiar with 

 the specific characters and the amount of variation. In this 

 key the many described varieties of Cydogcifiter liparis are 

 not recognized. It is possible that more than one s[)ccies is 

 gron|)ed under this name. This species is one not thoroughly 

 studied by the writer because of the want of material. Jt is 

 likely that one or more species not recognized here will be 

 found represented in the museums. 



A. Peritoneum silvery or white, with few or no black dots. 

 B. Anal fin with less than '60 rays; dorsal fin usually 

 notched. 

 C. Pectoral fin with less than 30 rays. 



D. Anterior dorsal sometimes elongate ; prickles 

 sometimes present ; dorsal rays more than 30. 



(American.) atlanticiia. 



DD. Anterior dorsal rays never elongate; prickles 



never present ; dorsal 30 or less. (European.)., viontagvi. 

 CC. Pectoral fin with more than 30 rays ; pyloric caeca 



less than 20 liparis. 



BB, Anal fin with more than 30 rays (34-37) ; dorsal fin 



never notched ; rays 41 -41: tunicatus. 



AA. Peritoneum usually black, sometimes heavily pig- 

 mented with black ; dorsal fin with 48 rays; anal 40; 

 pectoral 34 major. 



The species of Careproctus are not so easily distinguished 

 as the sjjecies of Cijclogaster. This is because tiie s})ecific 

 characters of the species of the former genus are more difficult 

 to examine and we have a smaller amount of material to work 

 with. We find three species (not counting gelatinosus, 

 Pallas) of Careproctus recorded from the North Atlantic and 

 Arctic Ocean. The validity of all of these species has 

 been questioned. These species are C. reinhardi (Kiojer), 

 1862; C. ranida (Goode & Bean), 1879 ; and C. micropus 

 (Giinther), ]898. A number of writers have regarded 

 C. reinhardi AS, identical with C. gelatinosus''^. Garman (1892) 

 refers C. ranula to the synonymy of C. reinhardi. Liitken 



* The description oi Ci/clopterus ffelatitiosvs, Pallas (1769), is so un- 

 satisfactory that Ave are not justified in identifying any of our species 

 with that' described by Pallas. Furthermore, the species described by 

 Pallas was obtained in the Pacific Ocean, and we are hardly justified in 

 identifying any of the North Atlantic species with it. This is especially 

 true when we consider that, as recent inquiry has shown, none of the 

 species of the family can with certainty be said to exist on both sides of 

 the North American continent. 



