1 899.] BLOOD-TESSELS OF TELEOSTEAN FISHES. 945 



■with very similar vascular arrangement, but with widely different 

 epipharyngeal dentition, may be mentioned the Salmon and the 

 Mackerel. In the Salmon there are a pair of small patches ot 

 teeth borne by the fourth epibrancliial bones, lying immediately 

 ventral to the aortic extremities of the third efferent branchial 

 vessels; but in the Mackerel there are a pair of great den- 

 tigerous pads underlying the mesial ends of the second, third, 

 and fourth efferent branchial \essels, as well as a considerable 

 part of the circulus cephalicus. In the Carp the dorsal aorta 

 obtains the necessary protection by actually traversing the gi-eat 

 horn-covered bony projection of the basioccipital, against which 

 the lower pharyngeal teeth bite. In the Wrasse the vascular 

 arracgement conforms to a very average type, being appai-eutly 

 quite unaffected by the large and elaborate pharyngeal mill 

 developed in the vicinity of the posterior half of the circulus 

 cephalicus. In Exoccetiis, again, the epipharyngeal pad is of 

 relatively enormous size, underlying the upper extremities of all 

 four efferent branchial vessels, the anterior ends of the aorta and 

 coeliaco-mesenteric artery, and the w hole of the circulus cephalicus 

 except the extreme anterior part ; yet there is nothing very 

 remarkable in the disposition of these vessels which might bs 

 accounted for by then- relation to the epipharyngeal apparatus. 

 Only in Gohius, Cottus, and ilolva, of the forms examined, are the 

 epipharyngeal dental pads situated entirely within the circulus 

 cephalicus. The evidence afforded by forms devoid of epi- 

 pharyngeal teeth is probably inconclusive, on account of the great 

 possibility of the edentulous condition having been arrived at 

 independently in different groups of fishes. In the Sprat and 

 Pipe-fish the types of vascular arrangement are widely divergent. 

 The Sprat closely I'esembles the Herring (fig. 2) ; the vessels of 

 Syngnathus are shown in figure 3^. 



In order to discuss intelligibly the different forms of vascular 

 arrangement met with, some form of classification, however 

 artificial, is essential ; and after careful consideration I have 

 found it convenient to adopt the following scheme, based on the 

 relations of the efferent branchial \ essels to the circulus cephalicus 

 and the dorsal aorta. In consideration of the unsatisfactory 

 nature of the present classification of Teleostean fishes, a purely 

 artificial scheme, founded upon the single character which forms 

 the basis of the present communication, is likely to prove of more 

 permanent utility for purposes of subsequent reference, than one 

 which relies upon a classification which sooner or later may prove 

 to be an unnatural grouping. 



The great majority of the forms studied will be seen to come 

 under the headings B and C, while the rarer and more extreme 

 modifications occupy terminal positions in the classificatory 

 scheme. The division of the groups B and C into the subgroups 

 b and c, differing only in the transverse or oblique position of the 

 confluent third and fourth efferent branch'a vessels, appears 

 arbitrary, but in practice there are very few torms which fai to 



