balistid palatine, the foot now sutured to the ectoptery- 

 goid (and mesopterygoid) rather than simply ar- 

 ticulated closely to it through a short ligament, with the 

 top of the ostracioid palatine probably representing a 

 combination of the top part of the shaft and the pos- 

 terior part of the crossbar held by ligament to the eth- 

 moid-vomerine (especially the former) region in 

 balistids. 



The specialized nibbling incisors of monacanthids are 

 far less like the dentition of ostracioids than is that of 

 balistids. Balistids have four teeth in an outer series in 

 each half of the upper and lower jaws, while the derived 

 monacanthids have reduced the number to three in the 

 upper jaw and to three or, in some cases, only two in the 

 lower jaw. In ostracioids there are usually four or five 

 teeth in each half of both the upper and lower jaws. It is 

 unlikely that the reduced in number and relatively broad 

 incisors of monacanthids could be ancestral to the more 

 elongate and basically conical teeth in greater numbers 

 as found in ostracioids. 



While the number of teeth in the outer series in balis- 

 tids is similar to that in the single series in ostracioids, 

 their shapes, at least at present, are not. The teeth in 

 balistids are somewhat wider, thicker, and more notched 

 than in Recent ostracioids. However, the teeth of the two 

 species of Eocene ostracioids probably were wider, 

 thicker, and more notched than in the Recent species, 

 approximating the balistid condition. It is suggested here 

 simply that the Eocene balistids (for which there are 

 as yet no fossils from that period) and ostracioids could 

 easily have had rather similar numbers, sizes, and shapes 

 of teeth, those of balistids somewhat smaller and less 

 notched than at present and those of ostracioids some- 

 what larger and more notched. 



There are marked similarities between the balistid and 

 ostracioid parasphenoid, ethmoid, and prefrontal that do 

 not exist in the monacanthids. In balistids and os- 

 tracioids the prefrontal is relatively well developed and 

 extends ventrally to articulate with a thickened region of 

 the parasphenoid, while in monacanthids the prefrontal 

 is greatly reduced in size and is far removed from any 

 contact with the parasphenoid. In balistids and os- 

 tracioids the ethmoid has only a relatively shallow ven- 

 tral keel, if present at all, while in monacanthids the ven- 

 tral keel is always extremely well developed. In balistids 

 and ostracioids the parasphenoid is expanded dorsally 

 into a thick plate in front of the orbit, but no such ex- 

 pansion is present in monacanthids. The large, thick, 

 usually rhomboidal scale plates of balistids are far closer 

 to the larger, even thicker, usually hexagonal scale plates 

 of ostracioids than are the small, thin, usually more or 

 less rounded to rectilinear scale plates of monacanthids, 

 and in some balistids the scale plates are hexagonal. 



In all ostracioids there are 18 vertebrae, with the ex- 

 ception of one specialized genus with a secondary in- 

 crease to 19, and all balistids normally have 18 vertebrae 

 also, while monacanthids always have 19 or more verte- 

 brae. Balistids and ostracioids have mostly branched 

 rays in the soft dorsal, anal, and pectoral fins, while these 



rays are all unbranched in monacanthids. A supraneural 



is present in both balistids and ostracioids, although it is 

 possible that their origins are different, while a supra- 

 neural is absent in monacanthids. The lesser reduction in 

 number of pharyngobranchial elements in balistids ver- 

 sus monacanthids is more similar to the condition in 

 most ostracioids. In balistids and ostracioids the para- 

 sphenoid is only slightly, if at all, expanded laterally just 

 behind the orbit, while in monacanthids it is moderately 

 to greatly expanded there, except in one highly specializ- 

 ed genus. In balistids and ostracioids the postcleithrum 

 is sometimes composed of two pieces, while in mona- 

 canthids it is always a single piece. 



By contrast, there are only a few ways in which mona- 

 canthids are more similar to ostracioids than £U"e 

 balistids. In monacanthids and ostracioids the medial 

 edges of the pterotics on the ventral surface of the skull 

 are widely separated, while they are only narrowly 

 separated in balistids. However, balistids and mona- 

 canthids are similar in that the separation is by the para- 

 sphenoid and basioccipital, while in ostracioids it is by 

 the prootic and exoccipital, a rather different arrange- 

 ment. The posttemporal of monacanthids and os- 

 tracioids is more superficially held to the skull than is 

 that of balistids, which is always placed in a deep groove 

 on the skull. However, the posttemporal in ostracioids is 

 usually much larger and more extensively sutured to the 

 pterotic than is that of monacanthids, and the slight 

 resemblance between the posttemporal of monacanthids 

 and ostracioids is probably coincidental and of no phylo- 

 genetic significance. 



In monacanthids and ostracioids the supracleithrum is 

 not posteriorly expanded and the postcleithrum does not 

 have a dorsal prong, while in balistids a posterior expan- 

 sion and dorsal prong are present. However, these struc- 

 tures in balistids probably are involved with support of 

 the specialized tympanal region above the pectoral fin 

 base only found in balistids and not to be expected in 

 groups lacking a tympanum. In monacanthids and os- 

 tracioids the fifth ceratobranchial is toothless but has a 

 series of gill rakers along its anterior edge, while in 

 balistids the fifth ceratobranchial bears teeth but no gill 

 rakers. These differences are surely correlated with the 

 usually coarser diet of balistids versus monacanthids and 

 ostracioids, and the loss of teeth from and the gain of gill 

 rakers along the anterior edge of the fifth cerato- 

 branchial probably occurred independently in mona- 

 canthids and ostracioids. It is perhaps instructive that in 

 triacanthoids, from which the balistids are derived, the 

 fifth ceratobranchial also is toothed but lacks gill rakers. 



Thus, the evidence strongly indicates that it is the 

 balistids among the balistoids to which the ostracioids 

 are most closely related, and not to the monacanthids 

 that are derivatives of the balistids. Moreover, since the 

 ostraciids are clearly derived from the aracanids, and the 

 aracanids have remained far more generalized, the 

 closest relationship of the balistids to the ostracioids is to 

 the aracanids and not to the ostraciids. 



There are only a few ways in which balistids are more 



