776 MR. 6. A. BOtTLETfGER OK THE T-ISHES [NoV. 15, 



which latter genus has even been raised, without sufficient 

 justification, to family rank. Although naturally adverse to the 

 multiplication of genera, I cannot hesitate, in this case, to restore 

 most of those previously proposed by the above-mentioned authors, 

 and even to add to their number, as I am doing in a publication 

 dealing with the Congo species, which will appear simultaneously 

 with the reading of this paper. I believe that 1 have succeeded 

 in giving definitions of the genera clear and precise enough 

 to greatly facilitate the study of this highly remarkable family. 

 I can confidently affirm that the genera here admitted are fully 

 equivalent to, and at least as clearly defined as, those universally 

 admitted in the families Chipeidce, Cyprinidce, and CJiaracinidce. 

 The difference between Momiyrus and Hyperopisus, for instance, 

 is strictly comparable to that between Alhida and Baihythrissa 

 among the Clupeines ; yet, in the same classification (Giinther's 

 ' Study of Fishes,' 1880), the two former types are not allowed even 

 the rank of genera, while the two latter are unnecessarily referred 

 to distinct families. 



The union of the genus Mormyrops with Mormynis can have 

 been suggested only in ignorance of the marked differences in the 

 vertebral column to which attention was first drawn by Hyrtl. 

 The vertebral column shows a greater degree of specialization 

 in the latter than in the former, the posterior prsecaudal vertebrae 

 being devoid of those remarkable hsemal bridges to the extremity 

 of which the corresponding ribs are attached, a morphological 

 difference the importance of which would alone justify generic 

 separation. Moreover, as also pointed out by Hyrtl, and since con- 

 firmed by Peters and by Fritsch, the il/orm?/ropes are true carnivores, 

 like Gymnarchus, while the other Mormyrs feed exclusively or 

 mainly on vegetable matter and minute animals. 



In the more generalized character of the vertebral column, 

 Mormyrops agrees witli Gymnarchus, and \^'e may regard the latter 

 as an ultra-specialized, nuguillifoi'm modification of the former, 

 the other Mormyroids being also modified, but in other directions, 

 from such a type. This conception is supported by a con- 

 sideration of other characters. Thus I represent to myself the 

 hypothetical primitive type of Mormyr as elongate like an 

 Alhula (1), with the dorsal (2) and anal (3) fins elongate (basally), 

 large ventral and caudal fins (4), a fairly large mouth (5), and with 

 the prsemaxillary (6), parasphenoid (7), and glossohyal (8) bones 

 armed with several rows of small conical teeth. As many as 7 

 out of these 8 points (1, 2, 3, 4, 5, 7, 8) have been retained by 

 some of the Mormyropes, no. 2 in common with Gymnarchits ; 

 no. 6 has been preserved in one genus only, Myomyrvs ; while it 

 is interesting to observe the interchange in nos. 2 and 3 exhibited 

 by Momiyrus and Hyperopisus. It is also of importance, as 

 bearing on this question, to note the reduction in the size of the 

 caudal fin that takes place within the genus Mormyrops, a feature 

 which may even lead us to speculate on the possible discovery of 

 forms that would fill the gap now existing between the Mormyrs 

 with well-developed homocercal tail and the Gymnarchus, in which, 



