The dense patch of setae found on the caudal peduncle 

 of male Rudarius ercodes is similar to that found in 

 males of some species of Cantherhines (others have pairs 

 of retrorse spines; see Randall 1964 and Tyler 1970e), 

 while the much longer posteriorly directed spines on the 

 caudal peduncle of male R. minutus are undoubtedly a 

 specialized development from the ancestral R. ercodes- 

 like form and are not in any way related to the even 

 longer quills further forward on the body of Amanses. 

 Similarly, the large dermal flaps found in R. minutus 

 have a much smaller counterpart in some species of 

 Cantherhines (see Randall 1964). 



Navodon shares many characteristics with Rudarius, 

 as pointed out by Fraser-Brunner (1941b:178-179), and 

 the two are probably closely related, both specialized in 

 their own way, Rudarius with a relatively deep small 

 body and only two teeth in each dentary and Navodon 

 with a larger more elongate body and a full complement 

 of three dentary teeth but a more forwardly placed gill 

 slit and reduced size of the ventral dewlap. 



Four genera not examined for this work and not dis- 

 cussed above should be briefly mentioned, these being 

 Arotrolepis, Pseudomonacanthus, Blandowskius, and 

 Eubalichthys. According to the data in Fraser-Brunner 

 (1941b), these four genera can be characterized as fol- 

 lows: 



Arotrolepis has the dorsal spines originating behind 

 the middle of the eye, a flexible encasing scale series 

 (presumedly of three segments and with a fin-ray ele- 

 ment) of small size, the ventral dewlap of no more than 

 moderate size, and the scales usually with a single strong 

 central spinule and arranged in distinct longitudinal 

 tracts. Fraser-Brunner thought Arotrolepis most closely 

 related to Monacanthus, although in his key to the 

 genera (which he cautions does not show the true rela- 

 tionships of some of the genera) it is coupled with 

 Chaetoderma. I suspect that he is right and that 

 Arotrolepis is on the Monacanthus-Chaetoderma line. It 

 obviously would be of great interest to know the form of 

 the supraoccipital in Arotrolepis. 



Pseudomonacanthus (most data from Fraser-Brunner 

 1940c) has the dorsal spines originating over the middle 

 or behind the middle of the eye and an inflexible short 

 series of encasing scales (presumedly of less than three 

 segments and with the fin-ray element absent) and no 

 dorsal pelvic lobe. 



Blandowskius, which Fraser-Brunner thought was 

 most closely related to Pseudomonacanthus, has no 

 encasing scales at all and the dorsal spines over the 

 rear of the eye, with both genera being intermediate 

 between the Rudarius and the Alutera levels of organiza- 

 tion. 



Eubalichthys, with the dorsal spines over or in front of 

 the middle of the eye and followed by a groove for their 

 reception, and with the inflexible encasing scale series 

 much reduced in size until it is minute and incon- 

 spicuous, was thought by Fraser-Brunner to be a 

 derivative of the Cantherhines-Amanses line. 



In short, many of the generic relationships within the 

 monacanthids remain unclear, and will probably con- 

 tinue to be unclear until many more species, represent- 

 ing all genera, have been examined osteologically than it 

 has been possible to do here. However, it does seem clear 

 that the most generalized forms are those with an 

 anteriorly elongate supraoccipital bearing a vertical crest 

 and with a well-developed series of flexible encasing 

 scales of three segments covering over a rudimentary fin- 

 ray element, and with the dorsal spine placed behind the 

 middle of the eye. The genera sharing these conditions 

 are Stephanolepis, Monacanthus, and Paramona- 

 canthus. 



A Stephanolepis-like stock is probably ancestral to 

 both Monacanthus and Paramonacanthus. Chaeto- 

 derma (and probably Arotrolepis) seems a derivative of a 

 Monacanthus-like stock, while Laputa is probably deriv- 

 ed from Paramonacanthus along with Pervagor. It is sus- 

 pected that a Pervagor-\\ke stock is ancestral to the 

 various moderately specialized genera with inflexible en- 

 casing scale series and no fin-ray element. An Alutera- 

 like stock is suspected to have evolved from some 

 derivitive of a Paramonacanthus-like stock, perhaps 

 from a form like Rudarius or Navodon, and to have 

 diverged in two directions, one leading to Alutera and 

 Psilocephalus and the other to two lines leading to 

 Brachaluteres and Paraluteres and to Oxymonacanthus 

 and Pseudaluteres. 



SUPERFAMILY OSTRACIOIDEA 



Comparative diagnosis (contrast with that of the 

 Balistoidea). — Head and most of body encased in a 

 relatively inflexible (except on cheek in front of gill slit) 

 cuirass of especially thick and mostly hexagonal scale 

 plates whose apposed edges are articulated by minute in- 

 terdigitations, the portion of the body posterior to the 

 carapace with or without isolated scale plates of various 

 shape but never continuously covered by scales; body 

 outline in cross section less laterally compressed and 

 always with two or more angles or ridges; no spiny dorsal 

 fin; soft dorsal and anal fins short-based, with 9 to 13 

 rays and a slightly lesser number of basal pterygio- 

 phores; caudal fin with 10 or 11 rays; pelvis and pelvic fin 

 absent; teeth relatively small and more or less conical or 

 only slightly compressed basally but often constricted 

 distally, between 6 and 17 in a single series in both jaws; 

 lateral line not associated with grooves or spiny processes 

 on the scale plates; branchiostegals usually 2-1-4 but 

 sometimes 1 -I- 4 or 2 -I- 3, at least some of the rays in the 

 posterior division as broad and laterally compressed as 

 those in the anterior division; distal end of last bran- 

 chiostegal ray always articulated to the inner surface of 

 the suboperculum; elements of the hyoid arch more com- 

 pacted together anteroposteriorly and most elements less 

 elongate and more deep bodied; urohyal much reduced in 

 size, a more or less flattened plate or a slightly curved 

 rod, a ventral flange either absent or very poorly 

 developed; fifth ceratobranchial (lower pharyngeal) al- 



