560 MK. E. w. I., iioi/r (jx TiTE [June 21, 



Secondly, like many obher cold-water forms, it appears to be 

 enrvbathic in liigb latitudes ; the difference in temperature between 

 the superficial and deeper \vat«r being comparatively small, and 

 offering no marked thermal barrier to its descent. 



Lastly, as regards the Faeroe Channel, it is noticeable that no 

 specimens, larval or adult, were taken at the actual surface in 

 twenty-five hauls ; that the smallest specimen of all was captured 

 nearest to the surface, between 100-0 fathoms (Sta. 13 L); that 

 other larvaj were taken in six out of the thirteen deep hauls, and 

 may thus fairly be ranked among Mesoplankton. One (?) adult 

 specimen was taken in a haul which began at 530 fathoms and 

 finished at the surface ; this unfortunately gives us no help. 

 Although none of the ' Eesearch ' specimens were captured at the 

 surface, still if, as Mr. Holt suggests, Dr. Giinther's Scopdus 

 scoticus is identical with these larvae, some larvie come to the 

 surface at night in the Faeroe Channel. 



Though more observations are required for confirmation, still it 

 seems probable that Scopelus glacialis, at any rate as regards the 

 Faeroe Channel, falls into the category of animals which have an 

 early epiplanktonic stage, but frequent grenter depths when adult 

 (cf. 1^. 578, infra). Even in higher latitudes the adult has been 

 most frequently recorded either from considerable depths, or as 

 dead and floating if at the surface. — G. H. F.] 



[Note added Aug. 1898.— The stages shown in figs. 5 & 6 are 

 connected by an intermediate specimen of 7'5 mm., received too 

 late for description in the text. The proportions of the head, eye, 

 and snout are as in the specimen of 8 mm., but the general form 

 is more slender. — B. W. L. H.] 



Impeefectlt chabacterized I/ARTa with very elongate abdo- 

 men. ? JIallotus tillosus Miiller. Capelin. (Plate XL VII. 

 figs. 8-11.) 



These very elongate larvae have at first sight much the 

 appearance of young Eels, but closer inspection soon dispels this 

 illusion. They measure respectively 17, 19 (ca.), and 24-5 mm., 

 from the snout to the extremity of the notochord. I have figured 

 the most advanced, which on the whole is the most perfect 

 specimen of the series. The others differ little in general 

 conformation, but the smallest has the caudal extremity still 

 practically diphycercal, and the marginal fin terminates, without 

 spatidate expansion, in a sharp lanciform process. The propor- 

 tionate lengths of the abdominal and caudal regions are shown in 

 Plate XL VII. fig. 8 ; it will be seen that the abdomen is about 

 twice as long as the tail, the rectum being thus given off at a point 

 far posterior to median. The fore-brain extends but little in 

 front of the eye, which is only of moderate proportions. The 

 considerable bluntly-rounded rostral region is occupied anteriorly 

 by a large olfactory pouch. The angle of the jaws is opposite the 

 front of the eye. The pectorals are small. The pelvics are 

 indicated by a pair of membranous lobes supported anteriorly by 



