jor outer series (present in both the upper and lower jaws 

 of a few triacanthodids and in all triacanthids, diodon- 

 tids, and molids; present only in the upper jaw of 

 balistids and monacanthids; present in the upper jaw of 

 many tetraodontids and, to a lesser extent, in the lower 

 jaw; absent in all aracanids and ostraciids); gill rakers 

 along anterior edge of first gill slit (present only in 

 molids); gill rakers along anterior edge of fourth arch (ab- 

 sent only in a few diodontids); gill rakers along posterior 

 edge of fourth arch (absent in tetraodontids and diodon- 

 tids); gill rakers along anterior edge of fifth arch (present 

 in monacanthids, aracanids, ostraciids, and molids); 

 scales (present on at least a part of the body in all species 

 except a few tetraodontids). 



The few major features in the soft anatomy of plec- 

 tognaths treated here are about as variable between or 

 within various families as the bony parts just discussed. 

 Of most importance here, neglecting for the moment the 

 musculature that has been so superbly studied by Win- 

 terbottom (1974), is the presence of four gills and a gill 

 slit between the fourth and fifth arches in all groups ex- 

 cept tetraodontids and diodontids, which have lost the 

 gill of the fourth arch and the slit between it and the 

 fifth. A pseudobranch is present in all species, but the 

 degree of development and number of lamellae are highly 

 variable, the number of lamellae ranging from only 4 or 5 

 to about 70, with balistids and a few tetraodontids tend- 

 ing to have the pseudobranch the least well developed. 

 The olfactory epithelium is in the form of a rosette in 

 triacanthodids, triacanthids, and triodontids, but it is 

 variously modified in all other families as folds (often ex- 

 tensive but never as a rosette), ridges, pits and 

 plications, or relatively smooth. There are two nostrils in 

 all families, except in tetraodontids and diodontids in 

 which there are either one or two nostrils. There is 

 basically a single lateral line on the body in all families, 

 except in tetraodontids in which there are one, two, or 

 three lateral lines on the body. A well-developed air 

 bladder is present in all families, except that it is absent 

 in molids, at least as adults. An inflatable diverticulum 

 of the gut is well developed only in tetraodontids and 

 diodontids, while an expansible dewlap of skin between 

 the end of the rotatable pelvis and the anus is moderately 

 developed in most balistids, well developed in many 

 monacanthids, and greatly developed in triodontids. 

 Whether a slighly inflatable diverticulum of the gut is 

 present in a few monacanthids (especially 

 Brachaluteres) is still debateable (see Thilo 1899b, 1914; 

 Rosen 1912; Breder and Clark 1947; Clark and Gohar 

 1953; Arbocco 1957), but, if so, it is far less well de- 

 veloped than in tetraodontids and diodontids and has 

 undoubtedly been independently evolved. Molids may 

 be unique among plectognaths by having a single ovary 

 (according to Cleland 1862 and Wahlgren 1867 for Mola, 

 and Pellegrin 1912 for Ranzania). With the possible ex- 

 ception of molids, all plectognaths have hard otoliths of 

 normal size, but these are usually so corroded or other- 

 wise dissolved in the formaldehyde-preserved, cleared, 

 and stained study material that they cannot be ade- 

 quately described or figured, and are not further dis- 



cussed here. Cleland (1862) said that otoliths were ab- 

 sent in Mola, while Cuvier (1805) and Thompson (1888; 

 repeated by Kaschkaroff 1914a) found that the otoliths in 

 Mola were represented by small calcareous granules 

 grouped together. It is not known whether this is also the 

 case in Masturus and Ranzania, but there is no evidence 

 of otoliths in the present study specimens. 



The large number of bones not mentioned above in 

 either of the two lists (constant versus variable occur- 

 rence) that are often found in perciform fishes are ab- 

 sent in plectognaths and will help form a large segment 

 of the defining characteristics of the Order Plectognathi 

 when the other orders and suborders of acanthopteryg- 

 ian fishes are better known anatomically. Some of these 

 structures typical of perciforms that have been lost by 

 plectognaths are: anal fin spines; suborbital bones; 

 parietals; at least one and usually several of the other- 

 wise five pelvic fin rays, and in all but the Eocene 

 triacanthodid Eoplectus of the branched and cross- 

 striated structure of the fin rays that are retained; 

 nasals; tabulars; scale bones; intermuscular bones other 

 than 5 to 11 epipleurals; intercalars or "opisthotics"; 

 supramaxillae; any sheath or surface concavity for the 

 upper jaw; canals in the skull bones for the lateral line 

 system; vomerine and palatine teeth, and of any other 

 teeth except those of the dentary, premaxillary, pharyn- 

 gobranchials, and fifth ceratobranchials; reduced or- 

 namentation of the skull bones and lack of spiny 

 processes; simplification of the posttemporal to a 

 relatively flat rod sutured along all of its medial surface 

 to the skull (mostly to the pterotic); development of a 

 small and only slightly, if at all, protractile mouth, al- 

 though often with massive jaws, and the relatively great 

 restriction of the gill opening (never ending ventrally 

 very far below the pectoral fin base); the complete cover- 

 ing by scaly to scaleless skin of the branchiostegal region, 

 and the usual development of relatively spinulose or 

 heavy scales probably derived from ctenoid scales. 



The two suborders of plectognaths (Balistoidei and 

 Tetraodontoidei) are difficult to comparatively diag- 

 nose, primarily for two reasons. One reason is that both 

 groups contain an anatomically highly diverse assem- 

 blage of families. For example, the diagnosis of the Bal- 

 istoidei must encompass such widely anatomically dif- 

 ferentiated families as the Triacanthodidae and Ostra- 

 ciidae, and that of the Tetraodontoidei has to include 

 such similarly diverse families as the Triodontidae and 

 Molidae or Diodontidae. Moreover, the diagnoses must 

 also take into account the great diversity found within 

 certain families, such as the differences between a gener- 

 alized monacanthid such as Stephanolepis and a highly 

 specialized one like Psilocephalus, or the differences be- 

 tween a generalized tetraodontid such as Sphoeroides 

 and a highly specialized one like Xenopterus. However, 

 the most important reason that the two suborders of plec- 

 tognaths are difficult to comparatively diagnose is that 

 we are fortunate in having alive today a very generalized 

 or primitive Recent member of the Tetraodontoidei, 

 Triodon, and that we also have available the speciaHzed 

 Eocene fossil subfamily of triacanthodid Balistoidei, the 



