HENSLEY AND AHLSTROM: PLEURONECTIFORMES 



685 



in Norman's Pleuronectidae. Members of this subfamily are 

 dextral or apparently secondarily indiscriminate (Hubbs and 

 Hubbs, 1945). They apparently have a monomorphic optic 

 chiasma. However, most character states which species of this 

 subfamily share appear to be plesiomorphic for the order or 

 bothoids. e.g.. symmetrical or nearly symmetrical ventral-fin 

 placement and fin-base lengths, anus on or close to the mid- 

 ventral line. We have examined the caudal osteology of about 

 half of the pleuronectine genera. All have the bothoid hypural 

 pattern (6) and one or possibly two free epurals. We have found 

 no synapomorphies in the caudal fin for this group. 



Larval characters 



In the previous discussion, many doubts were raised con- 

 cerning pleuronectiform interrelationships as expressed in the 

 Regan-Norman model. Unfortunately, larvae for many of these 

 groups are unknown. A second problem is that surveys for many 

 characters where larvae are known have been incomplete and 

 inconsistent. Most descriptive larval research has dealt with 

 characters useful for identification and has not involved com- 

 parative work of sufficient detail to determine homologous states. 

 Such work is sorely needed before distnbutions of homologous 

 states can be determined for many characters. 



Below is a list and discussion of certain characters and com- 

 plexes. Selection of these was based mainly on the amount of 

 available information. 



Preopercidar spines. — The presence of preopercular spines ap- 

 pears to be plesiomorphic for the order and some pleuronecti- 

 form groups. This is based on the observation that the slate is 

 widespread among flatfish and percomorph larvae. 



Neurocranial spines. —Spines occur in some regions of the neu- 

 rocranium in some pleuronectiform larvae. Most of these are 

 said to occur in the otic or frontal regions. However, determining 

 homologies here is difficult due to a general lack of detailed 

 osteological study of the bones carrying these spines. Spines in 

 the otic and frontal regions appear to be of two types. One of 

 these is where spines are associated with neurocranial ridge 

 systems. These are known for larvae of achirines (Houde et al., 

 1970; Futch et al., 1972), some scophthalmids (Jones, 1972), 

 and some pleuronectines (Pertseva-Ostroumova, 1961). In the 

 second type, spines occur singly or in small groups but are not 

 part of a pronounced ridge. These have been said to occur on 

 various bones of the otic region (epiotics, autosphenotics, au- 

 topterotics) or on the frontals. Tucker (1982) was not able to 

 determine the origin of such spines in the larvae of Citharichthys 

 and Etropus and referred to them as frontal-sphenotic spines. 

 Although thorough studies are needed before neurocranial spines 

 can be used to infer or test pleuronectiform interrelationships, 

 certain patterns are noteworthy: (1) Spines that are not part of 

 some pronounced ridge system appear to be limited to some 

 bothoids (some species of the Paralichthys group, Cyclopsetta 

 group, Pseudorhombus group. Scophthalmidae, Pleuronectinae, 

 and Bothidae). (2) Within the Bothidae, only the larvae of En- 

 gyophrys. Taeniopsetla, and Trichopselta (Taeniopsettinae; lar- 

 vae of Perissias are unknown) are known to have otic spines 

 (Amaoka. 1979). In these genera, the spines are on the same 

 bones (epiotics and autosphenotics) and are probably homol- 

 ogous. (3) Within the Cyclopsetta group, a relatively well-de- 



veloped otic or frontal spine occurs in Cyclopsetta and Syacium 

 (Aboussouan, 1968b; Gutherz, 1970; Ahlstrom, 1971; Futch 

 and HoflT, 1971; Evseenko. 1979), while series of small spines 

 occur in Citharichlfiys and Etropus (Tucker. 1982). 



Urohyal, basipterygial. and cleithral spines. Spines on these 

 bones are limited to certain genera of the Bothidae. Thus, they 

 are considered apomorphic at the pleuronectiform and bothoid 

 levels of universality. 



Early-forming elongated dorsal-fin rays. — The presence of elon- 

 gated dorsal-fin rays in pleuronectiform larvae has been exten- 

 sively and justifiably used for identification purposes. However, 

 use of these structures for phylogenetic interpretations is pres- 

 ently difficult and generally premature. There are several reasons 

 for this. Surveys for these characters are inadequate, since larvae 

 for many groups are unknown. Characters and character states 

 have never been adequately defined to allow proper compari- 

 sons to be made. The only pattern here that is clear and phy- 

 logenetically interpretable is the state in bothids. All species of 

 this family for which larvae are known show elongation of only 

 the second dorsal-fin ray. This state is known only in this family 

 and thus appears to be apomorphic within the order and both- 

 oids. 



Early-forming elongated ventral- fin rari.— Ocular ventral-fin 

 rays which are elongated relative to those of the blind side are 

 limited to certain species of the Cyclopsetta group. Due to the 

 restricted occurrence of these, they are probably apomorphic 

 for the order and bothoids. However, within the Cyclopsetta 

 group, the distribution of elongated ocular ventral-fin rays does 

 not conform to generic groups based on adult morphology. At 

 least one species of cynoglossid is known to have elongated rays 

 in the ventral fin of the blind side (Kyle, 1913; Padoa, 1956k). 



Size at metamorphosis. — MosX flatfishes metamorphose in the 

 size range of ca. 10-25 mm. When size at metamorphosis has 

 been discussed in regard to evolution in pleuronectiforms, the 

 usual hypothesis has been that certain species and groups have 

 evolved mechanisms for prolonging larval life for greater dis- 

 persal, and others have actually shortened larval life for re- 

 cruitment to limited habitats (Amaoka, 1979; Moser, 1981). 

 There are several implications in this hypothesis that are rele- 

 vant here; (1) There is some size range for transformation that 

 is plesiomorphic for the order. This is usually implied to be ca. 

 10-25 mm because most pleuronectiforms metamorphose in 

 this range. (2) Metamorphosis at markedly smaller (e.g., Achir- 

 inae) or larger (e.g., Bothidae, some pleuronectines) sizes are 

 derived states. (3) According to the Regan-Norman model, pro- 

 longed larval development must have developed independently 

 in several lines. Although metamorphosis at large sizes is most 

 common in bothids, it is also known for some Pleuronectinae, 

 the Poecilopsettinae, some species of the Cyclopsetta group, and 

 some cynoglossids. 



Size at metamorphosis is an important character for larval 

 identification, but its use for inferring phylogenetic relationships 

 in most instances is premature. Exceptions may exist in the 

 Bothidae, where the extremely long premetamorphic lengths 

 exhibited by some genera are probably apomorphic within the 

 family and can be used for phylogenetic information. 



