Figure 128.— Hypothesized phylogenetic relation- 

 ships of the genera of Monacanthidae. 



dorsal spines are placed posteriorly on the head. Thus, 

 Laputa is as generalized as Stephanolepis, Mona- 

 canthus, and Paramonacanthus except for the shorter 

 and more flattened condition of the supraoccipital. 



In Chaetoderma there are only two encasing scale seg- 

 ments, a slight specialization, but minute remnants of 

 the fin-ray element remain and the encasing scale 

 segments are flexible, the minuteness of the fin-ray ele- 

 ment also being a slightly specialized condition. 

 Moreover, Chaetoderma has a relatively well developed 

 forward extension of the supraoccipital, almost to the 

 front of the eye and only slightly shorter than in 

 Stephanolepis, Monacanthus, and Paramonacanthus. 

 The forward extension of the supraoccipital in Chaeto- 

 derma bears a thin low crest over the middle and pos- 

 terior regions of the eye which is probably a remnant of 

 that found in the three genera in which the crest is well 

 developed. 



In two other genera besides Chaetoderma the supraoc- 

 cipital extends forward only slightly less than the front of 

 the eye, but both are obviously specialized by the total 

 loss of the pelvic fin-ray element and the dorsal lobe of 

 the pelvis, as well as by either the total loss of the en- 

 casing scale series or its reduction to a single scale that is 

 lost in large adults. These two genera are Alutera, in 

 which all four species have a low supraoccipital crest on 

 the forward extension, and the single species of 

 Brachaluteres examined (trossulus), in which there is no 

 crest in the smaller of the two specimens but a very low 

 crest present from the middle of the eye posteriorly in the 

 larger specimen. 



One other genus must be mentioned in connection 

 with the shape of the supraoccipital, that being Psilo- 

 cephalus, in which there is no true long forward and 

 mainly medial extension of the supraoccipital as found in 

 the genera just discussed. In Psilocephalus the supraoc- 

 cipital is represented only by the relatively larger more 



laterally expanded posterior portion. But probably 

 because of the relative shortening of the posterior part of 

 the skull and the great elongation of the snout, the an- 

 terior end of the supraoccipital in Psilocephalus does ex- 

 tend forward to the level of the dorsal end of the pre- 

 frontals, although still short of the front of the eye. 



In my opinion, the combination of the most generaliz- 

 ed condition of the pelvic apparatus and that of the 

 supraoccipital supporting the basal pterygiophore and its 

 dorsal spines placed over the rear of the eye points to 

 Stephanolepis, Monacanthus, and Paramonacanthus as 

 representing the most generalized group of mona- 

 canthids. Of the other features listed in the opening 

 paragraph of this section as being considered generalized 

 a priori on the basis of the structure of the ancestral 

 balistids, these three genera have terminal mouths with 

 the most typical monacanthid number of teeth (the 

 minimum reduction from balistids), 19 vertebrae (the 

 minimum increase in number from balistids), and an up- 

 per free hypural, except that only a minority of 

 specimens of one of the species {ciliatus) of Mona- 

 canthus have a free upper hypural. 



However, Stephanolepis, Monacanthus, and Para- 

 monacanthus are among those majority of mona- 

 canthids with 1-1-4 branchiostegals, rather than the 2 + 

 4 arrangement found in balistids and in Navodon, 

 Acanthaluteres, Alutera, Cantherhines, and Amanses 

 among the monacanthids, and in having no relatively 

 distinct and deep groove medially behind the basal 

 region of the dorsal spines into which the latter are 

 received and held more or less flush with the surface, as 

 found in balistids and in the monacanthids Acantha- 

 luteres, Cantherhines, Amanses, Meuschenia, Oxy- 

 monacanthus, Eubalichthys, and most but not all 

 species of Pervagor. 



The 2 + 4 versus 1 + 4 number of branchiostegals and 

 the degree of development of a groove for the reception of 

 the dorsal spines seem in this case to be less important 

 indicators of the overall generalized condition of mona- 

 canthids than do the other outstanding features dis- 

 cussed above, this especially in light of the fact that, as 

 described here, Monacanthus ciliatus at least sometimes 

 has the 2 + 4 branchiostegal number and that among the 

 apparently closely related species of Pervagor, most have 

 a groove behind the dorsal fin, while at least one species 

 (tomentosus, according to Fraser-Brunner 1941b: 183) 

 does not. 



The three most generalized genera, Stephanolepis, 

 Monacanthus, and Paramonacanthus, seem very closely 

 related, and only a single genus has often been recog- 

 nized for the species now placed in Stephanolepis and 

 Monacanthus. The latter two genera have been best dif- 

 ferentiated by Berry and Vogele (1961:63, 68), mostly on 

 the basis of the spinulation of the scales. Monacanthus 

 has unbranched and relatively fewer spinules per scale 

 than does Stephanolepis, which has branched spinules, 

 although the number of spinules increases greatly with 

 increasing specimen size in both genera, just as in most 

 other monacanthids. Berry and Vogele also pointed out 

 that the ventral dewlap between the end of the pelvis and 



178 



