ABOUSSOUAN AND LEIS: BALISTOIDEI 



459 



though in the latter it is illustrated as a serrate preopercular 

 border. However, this structure is present in all balistids and 

 Morph AB monacanthids examined for the present study, and 

 because it is an inconspicuous structure it is most likely that it 

 is present in previously described taxa and has been overlooked 

 (see Fig. 250). 



Aracanidae 

 Nothing is known of aracanid eggs or larvae. 



Ostraciidae 



Ostraciid eggs are pelagic, large, slightly ovoid, have one or 

 more oil droplets (Table 1 16) and hatch in two to four days. 

 There is some chorion ornamentation surrounding the micro- 

 pyle. In Indo-Pacific species (Ostraciinae) this consists of a par- 

 tially raised field of small bumps surrounding a small pore-like 

 depression containing the micropyle (Fig. 244). In the single 

 Atlantic species examined (Lactophrysinae), only the pore-like 

 depression is present. Development of larvae at hatching is rel- 

 atively advanced, but there is some interspecific variation in 

 how advanced: jaws are totally unformed to formed and ap- 

 parently functional, the eye is unpigmented to partially pig- 

 mented, dorsal and anal fin anlagen may be present. Moderate 

 pigment is present, much yolk remains, the gill opening is re- 

 stricted to a pore, and an inflated vesicular dermal sac encloses 

 head and trunk (Fig. 244). The dermal sac disappears before 

 flexion. The larvae are deep-bodied and the tail is compressed 

 (Fig. 247). Depending on species, the body may be moderately 

 (Rhinesomus) to very wide (Ostracion) (Fig. 251): the lacto- 

 phrysine species examined were more narrow-bodied than the 

 ostraciine species. Larvae tend to become wider with growth, 

 but never become as wide as deep. At hatching, ostraciine larvae 



are rotund with head and trunk a single ball-like unit, and they 

 have a small tail. Lactophrysine larvae attain this condition 

 within a few days of hatching. The tail progressively becomes 

 relatively smaller with age until after flexion, and the ball-like 

 shape of the body is retained. The notochord tip is small. The 

 lips have an unusual flared structure. The fins form P.-D = A- 

 C. The dermal ossifications do not pass through a spinule stage, 

 but form directly starting as thickenings in the dermal sac which 

 ossify and grow out from their centers. These eventually coalese 

 into the mosaic-like armoured carapace characteristic of adults. 

 The individual carapace units that eventually produce spines 

 and other ornamentation tend to be larger and with more relief 

 than other carapace units. The ossifications become visible well 

 before flexion, and larvae are fully armoured by the end of 

 flexion. Pigment is moderate to heavy and generally uniform 

 on head and trunk, with the tail often unpigmented. 



Le Danois (1961) describes the juvenile development of sev- 

 eral ostraciid species. 



Chorion ornamentation of ostraciid eggs previously has been 

 reported only for Hawaiian taxa (Watson and Leis, 1974; Leis, 

 1977, 1978), however it is present in all ostraciid eggs examined 

 in the present study (Fig. 244), albeit reduced to a pore in Acan- 

 thostracion quadncornis (Table 1 1 6). The ornamentation is sub- 

 tle and confined to a small portion of the chorion, and we feel 

 it is probably present in all taxa, but has been overlooked in 

 previous descriptions. 



(A. A.) Station Marine D'Endoume et Centre D'Oceanog- 

 RAPHiE, Rue Batterie des Lions, 13007, Marseille, 

 France; (J.M.L.) Department of Ichthyology, The 

 Australian Museum, P.O. Box A285, Sydney, N.S.W., 

 2000, Australia. 



Tetraodontiformes: Relationships 

 J. M. Leis 



IN this contribution I construct a phylogeny of tetraodontiform 

 fishes based on early life history (ELH) characters and con- 

 trast this with phylogenies based on adult characteristics. The 

 ELH characters of tetraodontiform fishes are summarized in the 

 preceding two contributions (Aboussouan and Leis, and Leis, 

 this volume). Although in many cases there is little information 

 available, I have assumed that which is available is represen- 

 tative, and that new information will not change the conclusions 

 herein. This is unlikely, and for this reason, the present treat- 

 ment must be viewed with caution. 



Inter-ordinal Relationships 



The tetraodontiform fishes are usually presumed to have been 

 derived from perciform ancestors, with the Acanthuroidei being 

 the popular choice for closest relative (Tyler, 1980; Winterbot- 

 tom, 1974a; Lauder and Liem, 1983). However, D. E. Rosen in 

 an unpublished study (pers. comm.) presents evidence sup- 

 porting a relationship between zeiform and tetraodontiform fishes 

 (see also Winterbottom, 1974a). 



There is little in the early life history of tetraodontiform fishes 

 to indicate they are the sister group of the acanthuroid fishes. 

 The few ELH characters acanthuroids and tetraodontiforms share 

 (small mouth, gas bladder present, relatively few myomeres, 

 large head, oviparity, spherical eggs with unsegmented yolk) are 

 very widespread in the perciform fishes, and the larvae are not 

 even generally similar (see Leis and Rennis, 1983). Certain char- 

 acter states (e.g., scale development) are shared by the acanthu- 

 roid fishes and some groups of tetraodontiform fishes. This sit- 

 uation could be interpreted as indicating a relationship between 

 acanthuroids and tetraodontiforms, whereupon the character 

 state involved would be viewed as primitive for the Tetra- 

 odontiformes as a whole. Therefore, the presence of an alternate 

 character state in some tetraodontiform families would be viewed 

 as a derived condition. This type of interpretation, while prob- 

 ably realistic, is avoided here as it is fraught with opportunities 

 for circular reasoning. 



Too little is known of ELH characters in zeiform fishes (Tighe 

 and Keene, this volume) to enable a proper evaluation of the 



