greater diversity in the pelvic apparatus of monacan- 

 thids. The number of segments of scales in the encasing 

 series ranges from three to two to one to none, and a rudi- 

 mentary fin ray represented by a pair of bony splints 

 above and below a plug of cartilage at the end of the 

 pelvis is variously present or absent. The encasing scales 

 are either flexible or fixed, the dorsal lobe to the pelvis 

 varies from greatly to only slightly developed or absent 

 altogether, and the pelvis itself ranges from relatively 

 massive and extending to the region of the anus to 

 slender and reduced in length. 



In what can be considered the most generalized condi- 

 tion of the pelvic apparatus in monacanthids (i.e., the 

 least reduced from that of balistids), the pelvis is well- 

 developed and has at least a moderate dorsal lobe, the 

 encasing scales are flexible and in three series, and a fin- 

 ray element is present as two bony splints. This general- 

 ized condition is found in Monacanthus, Paramonacan- 

 thus, Stephanolepis, Pervagor, and Laputa. Tendons 

 attach to the splints and run forward to muscles on the 

 dorsal and ventral surface of the pelvis, the dorsal tendon 

 passing through a horizontal tube in the basal region of 

 the dorsal lobe, as described by Tyler (1962b) for Mona- 

 canthus ciliatus. The splints representing the fin-ray ele- 

 ment are of moderate size, about like that illustrated by 

 Tyler (1962b) for M. ciliatus, in all of the species examin- 

 ed of the above genera, except that they are longer in M. 

 mylii and M. chinensis and shorter in Paramonacanthus 

 barnardi. In one other genus a fin-ray element and flexi- 

 ble encasing scales are present, this being the monotypic 

 Chaetoderma, in which there are only two series of en- 

 casing scales and the fin-ray splints are minute. The only 

 genus not examined which has a flexible encasing scale 

 series, Arotrolepis, can be expected to have fin-ray 

 splints. 



In all of the genera examined in which the encasing 

 scales are fixed or absent, the fin-ray splints are absent, 

 although in a few species tendons still attach to the small 

 cartilaginous plug at the end of the pelvis, as described 

 by Tyler (1962b) for Cantherhines sandwichiensis. Tyler 

 (1962b) attributed three series of encasing scales to that 

 species. Examination of additional specimens of C. sand- 

 wichiensis and of other species of Cantherhines makes it 

 unclear to me now whether the scales are in two or three 

 series; perhaps it should be counted as two and a half 

 series, the half series being a pair of scales scarcely if at 

 all meeting in the midline ventrally between the larger 

 anterior and posterior pair. Whether these represent two 

 or three series of scales is problematical, but, in any case, 

 when the encasing scale series is inflexible, the length of 

 the series is shorter than in those genera in which there 

 are clearly three series and in which flexibility is pres- 

 ent. 



Of the genera examined with an inflexible series, there 

 are two and a half or clearly only two series of encasing 

 scales in Amanses, Cantherhines, Nauodon, and Acan- 

 thaluteres, and there is no fin-ray element. This is 

 probably also the case in the other genera with fixed en- 

 casing scales: Scobinichthys, Pseudomonacanthus, 

 and Meuschenia; and, questionably, in Eubalichtys, as 



discussed below. In Rudarius there are two series of en- 

 casing scales and no fin-ray element in R. ercodes, but 

 only a single series and no ray in R. minutus, the reduc- 

 tion of the series into a single, although highly spinous, 

 series in the latter perhaps being associated with the ex- 

 tremely small adult size of that species. In Oxymona- 

 canthus there is no fin-ray element and no precisely 

 delineated series of encasing scales; rather, the scales 

 bordering the ventral surface of the posterior end of the 

 pelvis simply gradually become larger and more spinous 

 distally so that the encasing scales are of indeterminate 

 number. 



The genus Alutera provides an example of the process 

 of loss of the encasing scales. No fin-ray element is pres- 

 ent, but Berry and Vogele (1961) have shown that in 

 small specimens of three of the four species (A. 

 heudelotii, A. scripta, and A. monoceros) there is a single 

 enlarged spinous scale present on the midline below the 

 ventral edge of the pelvis, this being the last remnant of 

 the encasing scales. This scale becomes unornamented 

 and then lost in large adults. In A. schoepfi there is no 

 rudimentary encasing scale at any size (Berry and Vogele 

 1961). 



In Acanthaluteres, the specimens of the single species 

 here examined {spilomelanurus, if I have it properly 

 identified, and I am not sure) lack encasing scales, 

 whereas Fraser-Brunner (1941b: 178) gives as a key char- 

 acter for the genus "Pelvic shield present at all stages 

 (sometimes minute or inconspicuous)." It is possible that 

 there are allometric changes in the encasing scales, just 

 as in Alutera. This may also be true of Eubalichthys, of 

 which Fraser-Brunner (1941b: 180) says: "Pelvic shield 

 minute, inconspicuous." 



There are no encasing scales or fin-ray elements in 

 Psilocephalus, Brachaluteres, Paraluteres, and Pseuda- 

 luteres, and, according to Fraser-Brunner (1941b:181), in 

 Blandowskius. 



The structure of the pelvis itself is highly diversified in 

 monacanthids. In those species with flexible encasing 

 scales, the pelvis is always, relatively well developed and 

 has a dorsal lobe of at least moderate height. The dorsal 

 lobe is largest in the genus Monacanthus, the species of 

 which are noted for the especially large size of the fan of 

 distensible skin with enlarged scales that can be flared 

 out between the anus and pelvis when the latter is 

 rotated downward and forward. However, the presence of 

 a dorsal lobe is not absolutely correlated with the expan- 

 sion of the abdominal dewlap, for one genus (see below) 

 without a lobe has the ability to flare a remarkably large 

 fan. 



In general, however, most of the genera (Pseuda- 

 luteres, Paraluteres, Brachaluteres, Alutera) without a 

 dorsal lobe also do not flare fans, the exception being 

 Psilocephalus, which not only has no dorsal lobe but also 

 the most slender and weakly developed pelvis in the 

 family, and yet can flare as large a fan as in Monacan- 

 thus. The presence of a dorsal lobe is probably more 

 strictly correlated with the presence of enlarged scales 

 along the posterior edge of the fan, these scales being 

 supported by the lobe. The scales along the posterior 



154 



