We should undoubtedly also recognize Calceostomatidae, having 

 the 12 -hooked attaching disc, as a relatively young fannily. As we have 

 already indicated (see page 274), it is distributed very singularly both on marine and fresh 

 water fishes, and it is possible it is an artificial one, but at the same time 

 the Paleocene period should probably be considered as its earliest period 

 of appearance. Morphological peculiarities of the typical representatives 

 of the Calceostomatidae show that during the development of certain pro- 

 gressive traits characteristic for this family, the process of oligomerization 

 of the chitinous apparatus is characteristic, which is expressed not only by 

 th.e disappearance of one pair of edge hooks but also in the simplification of 

 the middle hooks and of their connecting apparatus which disappear even 

 among part of the genera (see page 361). If one is to take into consideration 

 and. reject this tendency in development we can nevertheless speak about the 

 considerable proximity of Calceostomatidae to Dactylogyridae. Both families 

 undoubtedly originate from very close if not from common ancestors. 



As was already indicated (see page 355 ) the family Diplectanidae 

 is somewhat more removed. Undoubtedly this progressive branch is charac- 

 terized by a number of neoplasms (new formation, nobis). Taking into 

 account the occurrence of Diplectanidae almost exclusively on the Perci- 

 formes and even more precisely on Percoidae^ one can think that separation 

 of this family took place not earlier than the Paleocene and inost probably p. 456 

 somewhat later. Apparently the ancestral forms of Diplectanidae separated 

 directly from Dactylogyridae just as among Protogyrodactylidae, or what 

 is more probable, from ancestors common with the latter. Certain traits, 

 sharply distinguishing Diplectanidae from Dactylogyridae and not permitting 

 us to derive the first from the second, speak for this supposition. In the 

 first place, the characteristic structure of the ovary which is not encountered 

 in Dactylogyridae is attributed to these traits. 



As regards the most numerous family, both from the generic 

 and specific point of view, Dactylogyridae, it undoubtedly occupies the 

 phylogenetically central position in the suborder. The interrelations be- 

 tween the subfamilies of Dactylogyridae are not fully clear. One can con- 

 sider it certain that Ancyrocephalinae derived from Dactylogyrus -shaped 

 ancestors; however, to link this subfamily with Dactylogyrinae would be 

 erroneous, not only because it is possible that certain forms of the latter 

 are derived by means of simplification from Ancyrocephalinae (see page 

 347 ), but because of the fact that the representatives of both of these 

 families have a number of morphological peculiarities which apparently 

 developed independently among these and others from common ancestors. 

 Consequently, one can think that both subfamilies develop more or less 

 simultaneously and appear historically almost at the same time. The time of 

 their appearance can be determined by the fact that the former are linked 

 most closely with Cypriniformes in their distribution whereas the latter, 

 most probably, first appeared on Perciformes. Both hosts of the family 

 under examination are known as fossils starting from the Paleocene, which 



547 



