derivation, being related to the isospondylous fishes; and 

 Canthigaster is related to the ostracioids rather than the 

 tetraodontids. Le Danois' classification of the "Orbicu- 

 lati" is as follows: 



Sous-ordre Orbiculati 

 Division Orbispiniformes ou Tetraodontoidea 



Families: Xenopteridae, Colomesidae, Diodontidae, 

 Tetraodontidae, Lagocephalidae. 

 Division Ostracioniformes ou Ostraciontoides 

 Families: Canthigasteridae, Aracanidae, Ostra- 

 cionidae. 

 Division Moliformes ou Moloidea 

 Families: Ranzanidae, Molidae. 

 Division Triodontiformes ou Triodontoidea 

 Famille: Triodontidae. 



To what order the "Orbiculati" belong is not stated. 

 Le Danois' work can be dismissed, except nomen- 

 claturally. 



Although not directly concerned with the general phy- 

 logeny and higher classification of the plectognath fishes, 

 several outstanding recent works on the systematics of 

 limited groups of plectognaths have made certain species 

 complexes far better known than in the past, and repre- 

 sent the basic building blocks upon which such revisions 

 as this monograph are based, de Beaufort (1962) sys- 

 tematically treated all of the more common species of the 

 Indo-Pacific plectognaths (104 species) for the first time 

 since Bleeker, immensely aiding the often difficult task 

 of specifically identifying the plectognaths of that region. 

 Berry and Vogele (1961) carefully analyzed the sys- 

 tematics of the Atlantic species of monacanthids, and 

 Berry and Baldwin (1966) did the same for the Pacific 

 balistids. Shipp (1970, 1972a, b, 1974) and Shipp 

 and Yerger (1969a, b) have brought order to the nu- 

 merous Atlantic species of Sphoeroides, discussing ex- 

 ternal morphological features, delimiting distributional 

 patterns, and describing new species. With such exem- 

 plary work as that of Berry and Shipp and colleagues, the 

 American balistoids and tetraodontoids have become the 

 best known of their respective families. In Asia the ex- 

 ceptionally speciose and mainly freshwater pufferfishes 

 of the genus Tetraodon received their first systematic 

 review by Dekkers (1975), who placed them in five 

 species groups and unraveled most of the systematic 

 snarls that previously surrounded them. Similarly, the 

 marine species (primarily of Arot/iroM and Canthigaster) 

 of tetraodontids from Taiwan were redescribed by Shen 

 and Lim (1974a) and Shen et al. (1975) in that general 

 area for the first time since the series of articles by Abe 

 (1942, 1944, 1949a, b, c, 1950-51, 1952, 1954, 1960), while 

 Shen and Lim also reviewed the ostracioids (1973) and 

 balistids (1974b) of Taiwan. Allen and Randall (1977) pro- 

 vided an excellent and superbly photographed review of 

 the Indo-Pacific pufferfishes of the subfamily Canthigas- 

 terinae, recognizing 22 species, of which 7 were described 

 as new, in this group of notoriously difficult tetraodontids, 

 while Leis (1978) has performed a similarly fine service for 



the five species of Diodon he recognizes worldwide. The fos- 

 sil and Recent triacanthoids were redefined and revised by 

 Tyler (1968), which monograph was reviewed by Myers (1970). 



The provisional reclassification of fishes by Green- 

 wood et al. (1966), which is wisely accepted as the con- 

 temporary standard, does not attempt to include new in- 

 formation on the plectognaths. While I agree with their 

 basic subgroupings of the Recent Plectognathi, I consider 

 their families as superfamilies, a matter of opinion that I 

 hope is supported by the numerous differences given here 

 between the 6 superfamilies and their included 10 

 families as recognized in this monograph. 



In Gosline's (1971) combination of functional mor- 

 phology and fish classification, the plectognaths are 

 recognized in 7 superfamilies with 12 families (the 

 Aluteridae recognized as distinct from the Mona- 

 canthidae, and the Canthigasteridae from the 

 Tetraodontidae) in the usual two suborders, while Gos- 

 line (1968:16) had previously related the acanthuroids 

 and plectognaths. 



The tentative classification of the plectognaths 

 proposed by Tyler (1974), with 7 superfamilies and 11 

 families, is modified here by the combination of the 

 Canthigasteridae with the Tetraodontidae and of the 

 Diodontoidea with the Tetraodontoidea. 



It is pleasant to be able to conclude this historical 

 review of the higher classification of the plectognaths 

 with praise for the latest significant publication, that by 

 Winterbottom (1974) on "The Familial Phylogeny of the 

 Tetraodontiformes (Acanthopterygii: Pisces) as Evidenc- 

 ed by Their Comparative Myology." This is a beautifully 

 illustrated and thoroughly analyzed survey of the 

 muscles of numerous representative plectognaths and of 

 its implications to the phylogeny and classification of the 

 order. It is unquestionably the finest comparative 

 myological study of any major group of fishes. A phy- 

 logenetic interpretation is given for the condition of every 

 main muscle (75 in toto) in representatives of every plec- 

 tognath family and of various reinterpretations of these 

 data. It is a forcefully reasoned analysis based on an ex- 

 ceptional amount of myological data. 



A hallmark of Winterbottom 's work is a rigorous 

 application of the principal that only synapomorphies 

 (shared specialized characters) can be used to link 

 together related groups, while symplesiomorphies 

 (shared generalized or primitive characters) are not valid 

 as phylogenetic indicators. The excellence of Winterbot- 

 tom 's contribution can best be seen by perusal of his 

 brilliant monograph, and only the highlights of his phy- 

 logenetic conclusions and an outline of his proposed 

 major reclassification can be given here. 



Winterbottom (1974:ii) states in his abstract that 

 "... using only those characters of a shared specialized 

 (synapomorph) nature . . . suggests that the Triacantho- 

 didae and Triacanthidae are sister groups, and separated 

 off first from the ancestral stock. The remaining families 

 are divided into two lines. On the one hand, the 

 Balistidae form the sister group of the Monacanthidae, 

 and together form the sister group of the Aracanidae and 

 Ostraciidae. In the other linage, the Triodontidae 



