562 ME. E. w. L. HOLT ON THE [Jane 21 , 



Sp, 7 is oue of a group of ova and larvae which Kaffaele 

 considered to exhibit Murasnoid affinities ; and Grassi has practically 

 conhrmed the correctness of this view in the case of at least one 

 species, No. 10, which he has connected with Anguilla vulgaris. 

 Moreover it appears probable that all Muricnid larvoe pass through 

 a LejAocephahts-stage, losing the buccal armature of \\hat Grassi 

 terms the pre-larval condition. I imagine that it is impossible to 

 connect the Faeroe larvae with either end of a Leptocephaline 

 metamorphosis ; while the condition of the intestine and the 

 caudal fin suggest for them affinities which are not Muraenoid. 

 The presence of pelvic tins can hardly be held to prove that they 

 are not Muraenoids ; at least until Grassi shall have found that 

 such structures never occur as vestigial phenomena in the 

 development of Eels '. 



In 1893 my friend Captain F. Klotz, s.s. ' Dominican,' brought 

 me a number of young fish which he had taken at the surface off 

 the West Horn of Iceland on the 27th July. They range in size 

 from 36 to 57 mm., and in general shape have much the 

 appearance of Sand-eels (Ammodytes). The collection is suffi- 

 ciently serial to show that only oue species is present, while the 

 largest appear to associate themselves with the Capelin, Mallotus 

 villosus. 1 have figured the head of the smallest (fig. 9), a 

 specimen o£ 42-5 mm. (fig. 10), and the largest (fig. 11). The 

 radial formula of the largest appears to be D. 12 (or a few more), 

 A. 21. This specimen has 64 myomeres (perhaps more, as the 

 pectoral region is lacerated) exclusive of the peduncular part of 

 the tail, where a few others are probably present, though not 

 sufficiently defined to be counted. About 49 are abdominal. 

 From the ocular region backward the head is distinctly trigonal 

 in section, the upper surface being flat while the sides approach 

 each other ventrally. Though this is rather less marked in the 

 buccal region, thei'e is a distinct approach to the conformation (a 

 three-sided pyramid) described by tSmitt (Hist. Scand. Fish. ed. 2, 

 p. 877) as characteristic of the head of the adult Capelin. The 

 sides of the body are compressed and flattened, while the dorsum 

 is also rather flat. Malhtus has the radial formula D. 12—16, 

 A. 18-25 ; the vertebrae are from 65 to 70. In general proportions 

 and in the relative position of the fins the oldest Iceland specimen 

 is in agreement with Mallotus (compai'e iSmitt's ligures of the 

 latter, op. cit. pi. xli. with my figure 11 ). The Iceland specimens 

 are a good deal damaged and none have any scales on the body, 

 but there are traces of them on the gill-cover of the largest. The 

 teeth are small, and there is no distinct notch in the premaxillary 

 region for the reception of the mandibular extremity. 



Beyond a few remarks of Collett's, quoted by Smitt, I have not 

 found any description of the young stages of Mallotus. Our 



' Liitken (" Spol. Atlant., Cbangenients de forme eliez les Poiesons," Yid. SeUk. 

 Skr. 5. E«kke, 1880, p. .'i94) considers that pelvic fins probably exist in the 

 young of all species of Trichiunis, though their presence is only indicated in 

 the adult of one species. 



