just in front of the prootics found in Aracana. Capro- 

 pygia and Caprichthys differ almost exclusively on the 

 basis of the caudal peduncular scales, although Caprich- 

 thys sometimes develops a supraorbital spine whereas 

 Capropygia never does. In both Caprichthys and Capro- 

 pygia the small scale plates that make up what are the 

 anterior peduncular saddles in other genera are less con- 

 solidated and do not form distinctive saddles that are 

 functionally single pieces. This can be considered a 

 reduction from the condition seen in Strophiurichthys 

 and Anoplocapros. The posterior series of peduncular 

 plates forms a complete ring around the peduncle in 

 Capropygia, just as in Anoplocapros, while the posterior 

 plates are completely lost in Caprichthys. 



It seems evident that Capropygia is a derivative of an 

 Anoplocapros -like ancestral stock specialized by the 

 reduction of the anterior peduncular plates, a posterior 

 elongation of the carapace, a reduction in the size of the 

 caudal fin, the development of prominent spiny 

 processes on the carapace, the reduction in the pharyn- 

 geal dentition and the development of a partial hard 

 palate over the front of the oral cavity. Caprichthys is 

 simply a closely related derivitive of Capropygia in which 

 the anterior peduncular plates are further reduced in size 

 and the posterior plates completely lost, and with a ten- 

 dency for one more carapace spine to develop. 



Capropygia and Caprichthys have a slightly higher 

 number of dorsal and anal fin rays than do the other 

 genera, and a relatively large number of fin rays might be 

 expected to be a primitive character in a group derived 

 from a posttriacanthid and prebalistid ancestral stock, 

 both of which families have long-based and many-rayed 

 dorsal and anal fins. However, Proaracana has only 10 or 

 11 dorsal and anal fin rays, like all of the Recent genera 

 except Capropygia and Caprichthys. I suspect that the 

 number of dorsal and anal fin rays in Capropygia and 

 Caprichthys is a de novo increase associated with the 

 greater dependence on these two fins for locomotion 

 demanded by the longer posterior extension of the 

 carapace and concomitant reduction in lateral flexibility 

 of the caudal peduncle, at the end of which is a caudal fin 

 less well developed than ii; all of the other genera. 



Relationship to the Ostraciidae. — The deepwater 

 aracanids are so obviously closely related and ancestral 

 to the shallow water, more speciose (about 20 species), 

 ostraciids, that it is needless to belabor the point. The 

 changes from the aracanid to ostraciid levels of or- 

 ganization mostly are associated with the increased 

 length of the carapace and its more complete enclosure 

 posterior to the dorsal and anal fins of the more elongate 

 body concomitant with a reduction in the size and 

 strength of the orbital and postorbital regions of the skull 

 and an extensively sutured, inflexible, generally weaker 

 vertebral column within the carapace, with four or five 

 rather than only two vertebrae involved in the variously 

 fused and sutured complex at the rear of the skull, and a 

 general reduction in the size of most of the neural and 

 haemal arches. The caudal fin supporting apparatus in 



the more generalized ostraciids is similar to that of 

 aracanids, except that there are only 10 fin rays rather 

 than 11, while the apparatus becomes highly specialized 

 in most other ostraciids. 



Differences in the type of food available in deep versus 

 shallow water that the aracanids and ostraciids are 

 specialized in feeding on may account for the slight dif- 

 ferences in dentition between the two families, ostraciids 

 having a much reduced pharyngobranchial dentition but 

 usually a few more teeth in the jaws, especially in the up- 

 per jaw, relative to aracanids. The differences in the 

 structure of the ethmoid supporting the upper jaw and of 

 the completeness of the parasphenoid forming a hard 

 palate over the roof of the oral cavity between the two 

 families are also undoubtedly related to differences in 

 diet, the shallow-water ostraciids perhaps making 

 greater use of plant material in their omnivorous diet 

 than do the deepwater aracanids. 



Other evolutionary trends from the aracanids to the os- 

 traciids are the reduction in the size of the postcleithral 

 apparatus, perhaps associated with a lesser need for but- 

 tressing the side of the carapace in the more completely 

 and solidly enclosed ostraciid body, the reduction of the 

 size of the uppermost pectoral fin ray, the development 

 variously by the basal pterygiophores and neural and 

 haemal spines of a firmer system of support for the 

 carapace around the dorsal and anal fins, the diversion of 

 the haemal canal away from the midline and its close ap- 

 position to the ventral surface of the centra, perhaps a 

 space saving device in a crowded abdominal region from 

 which even muscles are largely excluded laterally, the 

 loss of the ventral keel of the carapace and the reduction 

 in the number of isolated scale plates on the caudal 

 peduncle posterior to the carapace. 



With the aracanids obviously ancestral to the os- 

 traciids, a critical question is whether the latter can be 

 shown to have evolved from one or the other of the two 

 main evolutionary lines postulated here as diverging 

 from a Strophiurichthys-\ike ancestral stock, or whether 

 it was from such a generalized form as Strophiurichthys 

 or Proaracana. This is discussed under the section on 

 generic relationships of the Ostraciidae, with the con- 

 clusion being that the ostraciids, and more particularly 

 the lactophrysins, the more generalized of the two sub- 

 families, probably evolved from one of the two lines 

 possessing a laterally expanded parasphenoid shelf over 

 the roof of the oral cavity, but whether from the line with 

 the shelf placed posteriorly (Aracana) or anteriorly 

 (Capropygia and Caprichthys) cannot be determined. 



It is of interest that in both the triacanthoids, in which 

 the deepwater Triacanthodidae are clearly ancestral to 

 the derived shallow-water Triacanthidae, and in the os- 

 tracioids, in which the deepwater Aracanidae are clearly 

 ancestral to the derived shallow-water Ostraciidae, all of 

 the families occur together in the same strata of the up- 

 per portion of the lower Eocene of Monte Bolca, Italy. 

 The Triacanthodidae are represented there by 

 Spinacanthus cuneiformis, Protobalistum imperiale, and 

 Eoplectus bloti, all so different from the Recent species 



