308 BULLETIN OF THE BUREAU OF FISHERIES 



seas, states that it is very seldom taken shoaler than 75 fathoms there and chiefly 

 below 90 fathoms. That is to say, they are living at about 37° to 39° 68 — not in the 

 coldest layer — and in west Greenland, says Jensen, numbers of them sometimes 

 come to the surface dead in winter, apparently having succumbed to cold. The 

 fishery experiments of the Norwegian fisheries steamer Michael Sars 6 " have proven 

 that the rosefish is no more characteristic of Arctic temperatures on the European 

 than on the American side of the Atlantic, for while its geographic range extends 

 far to the north, indeed right up to Nova Zembla and Spitzbergen, it is caught 

 there only in the overlying layer of Atlantic water at temperatures of 39° to 43°, 

 never in the colder Polar water deeper down, though the latter supports typically 

 Arctic fishes in abundance. It is worth emphasizing that in thus avoiding Polar 

 temperatures the rosefish occupies very different bathymetric zones in the Gulf of 

 St. Lawrence and in west Greenland waters on the one hand, where it is confined 

 to the bottom stratum, than off northern Europe on the other. 



This is perhaps an appropriate place to note that there is no positive record of 

 the rosefish from the east coast of Labrador, north of the Straits of Belle Isle, 

 Packard's young specimens (dredged in 15 fathoms) being as likely from the north 

 shore of the Gulf of St. Lawrence. It therefore remains to be seen whether it is 

 endemic in the undiluted Polar water of the Labrador current, and for this same 

 reason it is most desirable to establish its status on the east coast of Newfoundland 

 more definitely than can be done from data yet available. 



Food. — The diet of the rosefish includes a great variety of crustaceans, espe- 

 cially mysids and euphausiids, shrimps, small mollusks, and various other inverte- 

 brates as well as small fish, and it bites on any bait. In turn it is itself the prey 

 of all larger predaceous fish, its fry being devoured in quantity by cod, older 

 rosefish, and halibut. 



Breeding habits. — It has long been known that this fish is viviparous, the eggs 

 developing and hatching within the oviduct of the mother. It is not likely that 

 any rosefish are born in the Guff of Maine before late May, for we have found no 

 gravid females or larvae prior to the end of that month. Breeding is evidently 

 well under way in June, for not only did Welsh see several mother fish containing 

 well-developed yoimg taken on Georges Bank from the 20th to 26th in 1912, but 

 we have towed a few newborn fish (7 to 10 mm.) off Boothbay and Mount Desert 

 on May 31 and June 14. July 8, however, is the earliest that we have taken them 

 in any numbers in our tow nets (57 larvae off Cape Cod in 1913), with July and 

 August covering the height of the breeding season and with very few young pro- 

 duced after the first week of September, while our latest seasonal record for young 

 rosefish (a single specimen from Southwest Harbor) is for the 14th of that month. 

 Previous authors have similarly described it as breeding from June to September 

 off Massachusetts, and most of the Canadian records of rosefish larvae, both within 

 and without the Gulf of St. Lawrence, are likewise for late June, July, and the 

 first half of August. In north European waters young rosefish are produced over 

 a longer period — from mid-April through August, according to locality. 



« See The Danish Ingolf-Expedition, Vol. I, Part 2, 1S99, for temperatures in this general region. 



» Captured in abundance on line trawls suspended at 50 to 100 fathoms over much greater depths in the northern part of 

 the Norwegian Sea. Murray and Hjort, The Depths of the Ocean, 1912, pp. 435, 648. London. 



