42 SHOULDER-GIRDLE AND BREAST-BONE. 



ginous interspace between this part and the scapula. It evidently began to ossify at the lower 

 selvedge, and here it has become fused with the aponenrotic plate of the interclavicle (i. cl.) ; 

 its upper angle, behind, is more than a right angle, and the unchanged hyaline cartilage continues 

 down behind it, forming the glenoid space for the lowest brachial (h. 4). Here, as in the genus 

 Clarias (fig. 2 A, cr. i. cl.) and as in the Ostracion (fig. 4 B, p. cr. i. cl.), there has been con- 

 siderable abortion of the body of the coracoid, caused, as it would seem, by the potent develop- 

 ment of the correlated splint-bone. ' ". 



There are four brachials, not differing much in size, but the uppermost is the smallest, and 

 the lowest but one the largest. They are bounded by a band of soft hyaline cartilage, both before 

 and behind, and the hinder band runs up behind the scapula ; the segmentation is arrested in this 

 selvedge. So also is the segmentation into a prse-brachial and a post-brachial series, and the 

 " fenestrse," which are arranged in a vertical line, only serve to notch these rays ; the hinder edge 

 of each ray is turned outwards, so that the pectoral dermal rays turn outwards as easily as they 

 lie backwards. What I have been describing in no degree answers to a Sclerogenous Fish, 

 like a Coitus or a Trigla, for their large brachials abort the coraco-scapular band ; and they show 

 no tendency to coalescence of exo- with endo-skeletal elements. But this genus does not 

 properly belong to the Thoracici ; it is an "Abdominal " type. The large ventral bones, to which 

 the ventral fins are attached, are not ischio-pubic elements, but sub-mesial abdominal ganoid 

 xcutes ; the serial hotnologues of the large inter-clavicular scutes. Each abdominal fin, in this and 

 several other species, is composed of one highly barbed dermal spine, and this articulates partly 

 with the abdominal scute ; close beneath its ascending portion ; and partly, like the pectoral spines 

 of the Siluroid, with the endo-skeletal bone ; this inner bone is the true " ischio-pubic " plate, 

 which is foreshortened in front, articulates by suture with its fellow, and becomes anchylosed to 

 the dermal plate outside (exactly as in the union of the coracoid and the inter-clavicle, and also of 

 the scapula and clavicle). Moreover, I find that the ventral fins are set on further backwards 

 than in Cattichthys littoralis ; in that Siluroid the "cotyloid" articulation i. c. of the ventral fin 

 with the " ischio-pubis" is between the fourth and fifth ganoid thoracic cincture, but in Gaste- 

 rosteus leiurus it is between the sixth and seventh, and the distance between the anterior end of 

 the pubis and the base of the clavicle is one third of an inch, or as far as the upper part of 

 the preopercular bone from the snout. I need not remind the reader that the ends of the 

 " pubes " lie between the lower part of the clavicles in Cottus, Agonus, Gobius, and in the 

 Acanthopteri generally. I shall make no apology for taking this Genus from amongst its false 

 relatives, and placing it side by side with the Siluroids, in close proximity to the old Ganoid 

 types ; meanwhile, however, it would well repay the morphologist who would work out the rest 

 of the osteology of this type, especially if it were done in a developmental manner. 



I must now turn to the higher types, the " Pisces Acanthopteri " of Miiller ; but this 

 Order might be divided into an atypical and a typical assemblage. The former should take 

 in the Trigloid, Cottoid, Gobioid, Blennioid and Lophioid Families ; all these are more or less 

 aberrant, and come into proximity to the Sub-ganoid types, and even to the true Ganoidei. 



