1899.] JJLOOD- VESSELS OF TELEOSTEAK FISHES. 953 



except that the subclavian arteries arise farther from the median 

 line, just as they do in Molva. In Lota, as described and figured 

 by Hyrtl (7. pi. i. fig. ii), there is a tendency for the last two 

 efferent branchial vessels to unite before opening into the circulus. 

 The fact that the coeliaco-mesenteric artery arises from the circulus 

 is shown in Hyrtl's hgure, and is quoted by Stannius (22. p. 103, 

 footnote 3) and by Owen (19. p. 490). 



In Blennius (fig. 35) the circulus cephalicus is narrow in front, 

 the coeliaco-mesenteric artery arises from the aorta immediately 

 behind the circukis, and the subclavian arteries are just behind this. 

 In Orthcujoriscus (fig. 3'i) the aorta is unsyra metrical, running to 

 the right side of the vertebrae ; the coeliaco-mesenteric arterv is 

 formed by the union of a pair of vessels arising from the common 

 trunks formed on each side by the confluence of the third and 

 fourth eiferent branchial vessels. The mesial ends of all the 

 eft'erent vessels are closely approximated. The gills in Orthwjoriscus 

 are remarkably prolonged in a backward direction. The efferent 

 branchial vessel of each emerges from near the middle of the full 

 extent of the gill, and not, as is more usual, from the upper end. 

 It is formed by the union of one vessel coming from the lower part 

 or gill proper, with another from the dorso-posterior prolongation, 

 in a manner already made clear by Alessandrini (2) and Milne- 

 Edwards (15. p. 335, footnote 2). Judging from Alessandrini's 

 description, the peculiar mode of formation of the coehaco-mesen- 

 teric artery in 0. truncatus does not occur in the species examined 

 by him ( 0. mola). 



In Tetrodon (fig. 37) there are only three gills on each side, and 

 three efferent branchial vessels, the fourth of the normal series 

 being absent. The reason for putting the genus in the subgroup 

 D c, characterized by the union of the third and fourth efferent 

 vessels into a common trunk, is to be found in the relation of the 

 coeliaco-mesenteric artery to the third efferent vessel of the ri^ht 

 side. The association is such that, were the fourth vessel present, 

 it could not reach the circulus between the third vessel and the 

 coeliaco-mesenteric. Both the right and left subclavian arteries 

 arise from the right side of the circulus, and the latter is nearly 

 circular iu shape. 



Conclusions. 



It will be seen from the foregoing obser\'ations that very con- 

 siderable diversity in the arrangement of the efferent branchial 

 blood-vessels is to be met with in the Teleostean fishes. The type 

 of vascular arrangement is constant for different species of the same 

 genus, and does not vary to any considerable extent iu different 

 genera of the same family. If, as in the Siluroid fishes, some 

 widely divergent types are included in the same family, there are 

 to be found intermediate types which act as connecting-links 

 between these extremes. 



With regard, however, to families which, in the at present ac- 

 cepted taxonomy of the group, are brought into close relationship, 



