514 BULLETIN OF THE BUREAU OF FISHEBIES 



Food. — This is an invertebrate eater like other small-mouthed flounders, 

 European experience pointing to small crustaceans, starfish, small moUusks, and 

 worms as its chief diet.'* It is not known to eat fish and seldom takes a bait. 



Breeding habits. — In European waters this flatfish spawns from late April until 

 September, and while so far as we know ripe fish have not been reported from the 

 Gulf of Maine our captures of eggs, almost certainly of this species, in July and 

 August are evidence that it is equally a summer spawner there. The fact that we 

 have taken larvse as long as 20 to 23 mm. by the first week in July (p. 515) and 

 others as small as 9 or 10 mm. as late as mid-October also indicates that spawning 

 commences as early in the Gulf and endures as late there as on the other side of the 

 Atlantic. Probably July and August see the height of production. Thus its breed- 

 ing season overlaps that of the haddock (p. 442). At present our only positive egg 

 records are at the mouth of Massachusetts Bay and off Penobscot Bay, but probably 

 it breeds all along the coastal zone from Cape Cod to Mount Desert and off the 

 west coast of Nova Scotia as well. Apparently it does not breed successfully in 

 the Bay of Fundy, for neither its eggs nor its larvse have been found there. No 

 definite evidence has yet been obtained by capture either of ripe fish or of eggs or 

 larvse in the tow net that it spawns on Georges Bank or on Browns, though it 

 probably does so. Both the Gulf of St. Lawrence and the waters off the south coast 

 of Newfoundland likewise serve as breeding grounds, for larvae less than 10 mm. 

 long were taken in both these regions by the Canadian Fisheries Expedition in 1915, 

 but there is no evidence that the witch spawns west of Cape Cod. 



The witch necessarily spawns through a wide range of both temperature and 

 salinity, breeding as it does over so protracted a period and over so many degrees 

 of latitude. In the GuK of Maine its eggs are shed in temperatures ranging from 

 39° to 41° at the beginning of the season to from 43° to 48° in midsummer, but, 

 being buoyant (p. 515), the temperature in which their development takes place and 

 which thus governs the success of reproduction may be considerably higher than that 

 of the deep water in which the spawning fish lie. In fact it is doubtful if any eggs 

 develop m water as cold as 42° or 43° in the Gulf, though they may be spawned in 

 lower temperatures, nor is there any reason to suppose that witch eggs spawned in 

 the icy cold bottom water off Newfoundland or in the Gulf of St. Lawrence actually 

 develop at lower temperatures than those produced in the Gulf of Maine, for the 

 surface stratum to which they rise immediately after they are shed is comparatively 

 warm (upward of 45°) during the spawning season. Experiment has shown that 

 incubation proceeds normally and rapidly at 46° to 49°, hence, this is evidently a 

 favorable figure. It appears from this that no part of the GuK of Maine is cold 

 enough in summer to hinder the successful reproduction of the witch, hence its 

 failure to breed in the Bay of Fundy is due to some other cause. Our captures of 

 eggs and of newly hatched larvce near the surface in July prove that incubation 

 can take place successfully in water at least as warm as 50° to 55°, but the upper 

 limit to normal development can not be stated from the evidence yet in hand, for 

 with a temperature gradient as steep as it is over most of the Gulf of Maine in 



9" No witch stomachs have been examined in the Gulf of Maine. 



