III. COTTOIDS OF NOETH AMERICA. 13 



they are only accidentally taken either with the hook or the net, and must be, as 

 already remarked, sought for under stones. In the United States where the 

 markets are abundantly supplied with larger fishes, the Miller's Thumbs pass 

 entirely unnoticed. 



Some species spawn in the spring, others in summer, from the month of March 

 to July. When only one species was admitted in both continents, the difference of 

 latitude in America was called on to explain the difference in spawning. Thus, to 

 see the Cottus of the State of Connecticut spawn in March and April, and those in 

 the Great Bear Lake only in May, appeared quite natural. But how explain the 

 fact, that in the Seine, at Paris, the C. gobio spawns in May, June, and July, 

 whilst in the State of Connecticut it spawns in March and April ? The isothermal of 

 Paris, it is well known, passes below that of any part of Connecticut. The Cotti 

 of the Danube again, spawn in March and April. 



It has for a long time been known that certain fishes construct nests, such as 

 Sticklebacks. Others take care of their spawn and young; so do the lump-fish, the 

 cat-fishes, and, according to Fabricius, the male of his C. gobio does the same. We 

 know nothing of the behaving of other Cottoids in this respect. 



The food of Cottoids consists generally of crustaceans, 4nsects and of their larvae. 

 Fabricius says that his C. gobio feeds on ammodytes (Sand eel) Holoilmria priapus 

 (Sea cucumber), worms and small fishes. Sir John Richardson tells us that frag- 

 ments of small fish were found in the stomachs of C. cognatus. We have examined 

 the contents of the stomach of many marine and fresh water Cottoids, and in the 

 Acanthocotti of the Boston Bay found them to consist of crabs (Pilwmnus Harrisii) ; 

 in Cottus viscosus and (jracilis, of insects and larvaa, and in Triglopsis Thompsonii 

 of shrimps. We have never found any indications of fishes. 



