the "squamosal." Like Klein, they were using dried 

 skeletal material. 



Rosen (1916a) had some insight into this problem of 

 the "double-layered" condition, which is especially 

 prominent in the otic region of tetraodontids. In his brief 

 description of Sphoeroides testudineus he contradicted 

 Klein's statement that the epiotics are covered by the 

 parietals, for Rosen stated (1916a: 19) that "I have not 

 been able to find distinguishable such elements either in 

 adult specimens or in young ones (18 mm. in length)." 

 But, quite unaccountably, Rosen went on a few lines 

 later to say, without any elaboration, that there were 

 "Parietals present." Unfortunately, there were no figures 

 of plectognath skulls in that paper, and one can only 

 guess that Rosen was referring to a portion of the pos- 

 terior end of the frontal in Sphoeroides as the parietal. 

 Rosen's most interesting statements, however, were 

 about the sphenotic and pterotic in Sphoeroides, for he 

 described them as follows: "Sphenoticum. Sections of 

 specimens 18 mm in length show that this bone orig- 

 inated as two elements: a dermosphenoticum and an 

 autosphenoticum in the same mode as, e.g., squamosum. 

 Squamosum is formed by the fusion of a dermo- 

 squamosum and an autosquamosum." No other com- 

 ments were made about this condition of the otic region, 

 nor were any figures presented of it. Rosen was obviously 

 aware of the double-layered condition in the otic region, 

 but why he did not notice it in the epiotic is a mystery. 

 Rosen was probably incorrect in assuming that the upper 

 layers of the sphenotic and pterotic were of dermal 

 origin, for in the many young tetraodontids examined for 

 the present work, this upper layer is obviously ossified in 

 the upper region of the cartilage of the chondrocranium. 

 This is not to say, however, that there is not any dermal 

 ossification eventually incorporated into the external 

 surface of the upper endochondral layer of the otic bones, 

 but only that the great bulk of these bones is endochon- 

 dral in origin. In ParahoUardia and other triacan- 

 thodids, for example, most of the substance of the part of 

 the pterotic that overlies the sphenotic and epiotic ap- 

 pears to be of dermal origin, but this is only a very small, 

 and irregular, portion of the pterotic. 



A number of workers have erroneously referred to cer- 

 tain portions of the pterotic as the "opisthotic." Awati 

 and Bal (1937) did so in Fugu oblongus, and Gamaud 

 (1956) did likewise in Batistes. Kaschkaroff (1914a:351) 

 seems to be the one who began the mistake, for he said 

 that an opisthotic was present in scleroderms but not in 

 gymnodonts. Kaschkaroff s "opisthotic" is only a region 

 of the pterotic. 



A well-developed myodome has long been known to be 

 present in triacanthoids and balistoids, and more recent- 

 ly in triodontids (Tyler 1962a), but it was thought to be 

 absent in all other plectognaths. Klein (1884, 1885, 1886) 

 was the first person to point out the diagnostic value of 

 this structure (see part 1:150-152 and part 3:230-234) in 

 plectognaths, but Regan (1903a) rather unaccountably 

 stated that the Sclerodermi have the "basis cranii more 

 or less distinctly double" (p. 280) and that the Gym- 

 nodontes have the "basis cranii simple" (p. 291), which 



implies that the ostracioids have a myodome. Kasch- 

 karoff (1914a:351) described the longitudinal concavity 

 on the ventral surface of the basioccipital in Triacan 

 thus, Balistes, and Monacanthus, but he did not men 

 tion that this channel leads into the myodome. Holm 

 gren and Stensio (1936:488) contended that the 

 myodome of plectognaths and a few other fishes cor- 

 responds to the dorsal part of the myodome of Salmo 

 because it encloses only the M. recti externi. Be that as it 

 may, the important fact that Triodon possesses a well 

 developed myodome was overlooked by Dareste (1849) 

 and Hollard (1857b), and it is significant to know that at 

 least a rudiment of the dorsal roofing of the myodome is 

 present in some tetraodontids (Tyler 1963b). 



Frost (1930:621) gave cursory descriptions of the 

 otoliths of a few plectognaths and said that they "are 

 curiously aberrant in form, showing little affinity with 

 those of the other orders." 



In the orbital region there has been much controversy 

 in the literature concerning the bones with which the 

 frontal articulates posteriorly. The difficulty stems from 

 Hollard's (1857b) inaccurate figures of the top of various 

 gymnodont skulls. Gill's (1885, 1892a) reproduction and 

 use of these figures could only lead to erroneous diag- 

 noses, with the tetraodontids being said to have the fron- 

 tals in contact with the supraoccipital, except for 

 Canthigaster and Chonerhinus, which were supposed to 

 have these bones separated from one another by the in- 

 tervention of the sphenotics. This erroneous difference 

 between the tetraodontins and canthigasterins was used 

 by Jordan and his associates (Jordan and Edwards 1887; 

 Jordan and Evermann 1898; Jordan and Snyder 1901) 

 and thus it gained popularity. However, Regan (1903a) 

 showed that the differences were purely in Hollard's 

 figures and presented figures of his own which showed 

 the correct relationships of the bones of the top of the 

 skull in a tetraodontin and a canthigasterin. Fraser- 

 Brunner (1943) used the fact that the sphenotic separates 

 the frontal and epiotic in diodontids as one of the diag- 

 nostic characteristics of that group. Sexual dimorphism 

 in the size of the frontals has been pointed out by Ebina 

 (1932) in filefishes, and supposedly by Le Danois (1954) 

 in diodontids. In relation to this discussion of the frontal, 

 it is worthwhile to mention that none of the dermal bones 

 of the head develop in conjunction with the sensory canal 

 system, a fact first stated by Rosen (1916a:20). The der- 

 mal bones show no trace of canals, bridges, or troughs 

 that could in any way be associated with the cephalic 

 sensory canal system, which is confined entirely to the 

 skin in plectognaths. Le Danois' (1956) statements to the 

 contrary are unproven. 



A basisphenoid is present only in two groups of plectog- 

 naths — Triodon and molids. Neither Dareste (1849) nor 

 Hollard (1857b) observed the basisphenoid in Triodon, 

 but that of molids has been described often, although un- 

 der a variety of names. A few workers have described the 

 anterior region of the prootic in various scleroderms as a 

 "basisphenoid" (Klein 1884; Supino 1905; Rosen 1916a), 

 but, strangely, there has been agreement (as early as 

 Stannius 1854:61) that gymnodonts, except for molids. 



