530 BUIiLETIW OF THE BUREAU OF FISHERIES 



fishermen of southern New England have long been familiar with them, though it 

 was not untU about 1871 that Alexander Agassiz demonstrated their true parentage." 

 The eggs occasionally become isolated, perhaps by some storm shredding the mucous 

 veil to pieces, and when this occurs they float like any ordinary buoyant fish eggs. 

 We have not actually foimd them in this condition in the Gulf of Maine, though 

 Agassiz and Whitman saw isolated eggs at Newport. 



The eggs themselves, large nmnbers of which have now been examined, are 

 spherical or slightly oval, 2.13 to 2.5 mm. in diameter and averaging about 2.3 mm. 

 as they lie in their mucus cells,'^ but isolated eggs as large as 3. 11 mm. and others 

 as small as 1.67 mm. have been reported. The yolk is homogeneous, straw-colored, 

 and with either one large oil globule of 0.4 to 0.56 mm. or several smaller ones. 



The duration of incubation is not known. The larvae " are about 4.5 nam. 

 long at hatching and float at first with the yolk uppermost. Within 4 days or 

 so the first dorsal fin ray (which is to form the second head spine of the adult) appears 

 as a lobe at the margin of the embryonic finfold on the nape, while at about 7 days, 

 when the larva is 5.5 mm. long, the pectorals are formed and the ventral fins have 

 appeared as two long conical processes below and behind the pectorals (fig. 275). 



In summer temperatures the absorption of the yolk and the formation of the 

 mouth are complete and the larva rights itseK in the water in about two weeks, 

 while either just before or shortly after the. disappearance of the yolk (North 

 American and North Sea specimens differ in this) and at a length of 8 to 10 mm. 

 a second dorsal ray appears behind the first, the ventrals elongate and become 

 two-rayed, and the pigment congregates in three masses behind the vent. From 

 this point on larval goosefish described from different seas have shown considerable 

 differences at different sizes, depending on the rapidity of development as com- 

 pared with the rate of growth in waters of different temperatures; also in the 

 detailed structure of the fins and in the general outlines and proportions of head 

 to body, but the successive stages have been essentially simdar in all. Thus North 

 Sea specimens of about 10 mm. show a third dorsal ray on the nape behind the two 

 previously formed, while the first traces of the rays of the second dorsal fin and 

 of the anal have appeared and the ventrals have lengthened until they reach back 

 past the middle of the trunk and become three-rayed, whereas New England 

 larvae have shown a fourth dorsal ray before the third ventral appears (fig. 276). 

 A fifth dorsal ray next appears behind those preexisting, and a sixth in front 

 of them, all being connected with the membrane at their bases but free at 

 the tips. The pectorals assume a great breadth and fanlike outline, the second 

 dorsal, the anal, and the caudal fins take definite form, the ventral rays become 

 filamentous at their tips, streaming far behind the tail, and a complete row of teeth 

 appears in the lower jaw, with a few in the upper. The goosefish pictured in this 



" Baird. American Naturalist, Vol. V, 1871, pp. 785-786. 



" Agassiz and Whitman (1885) give the diameter as only 1.75 mm., but this may have been after preservation as Connolly's 

 eggs of 1.7 mm. were. 



'< Larval goosefish from New England have been described by Agassiz (1882, p. 280) and by Agassiz and Whitman (1885). Spec- 

 imens from the North Sea have been described by Bowman (Fishery Board for Scotland, Scientific Investigations, 1919 (1920), 

 No. II, pp. 1-42, Pis. I-VI, 2 charts), while a Nova Scotianeiample was described by Coimolly (1922), and others from the Adriatic 

 by Stiasny (.irbeit, Zoologische Institute Vienna, vol. 19, 1911, p. 71). 



