334 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



greatly within families only the latter characters may be relevant 

 to the discussion of the relationships of the families. 



The assumption by Bertelsen (1951) that the Caulophrynidae 

 are isolated from all other ceratioids was based mainly on three 

 larval characters: ( 1 ) presence of pelvic fins; (2) absence of sexual 

 dimorphism in the illicial rudiments; and (3) lack of a distal 

 swelling of these rudiments representing the rudiment of an escal 

 bulb. The two latter character states indicate that the absence 

 of an escal photophore in caulophrynids is not due to a sec- 

 ondary reduction. As expressed by Pietsch (1979) who found a 

 number of additional resemblances between Caulophryne and 

 less derived Lophiiformes (lophiids and antennarioids): "That 

 these primitive character states suddenly reappeared in a lineage 

 that arose from an ancestorderivedinall, is highly improbable." 



The possibility, mentioned above, that the neoceratiids may 

 represent a similar isolated lineage derived from an ancestor 

 without escal photophore, is based on the same larval characters: 

 the absence of sexual dimorphism in their illicial rudiments 

 which lack distal swellings. However, the fact that neoceratiids 

 and caulophrynids share these primitive character states fur- 

 nishes no information on their relationship. In other larval char- 

 acters, especially body shape and size of pectoral fins, the two 

 families are extremely different. 



The assumption that the absence of escal photophore in the 

 gigantactinid genus Rhynchactis is due to a secondary reduction 

 is confirmed by the presence of a club-shaped illicial rudiment 

 in the female larvae. 



Little information on the relationships between the ceratioid 

 families can be obtained from other observed differences in 

 larval morphology. The greatly enlarged pectoral fins pres- 

 ent only in gigantactinids and caulophrynids may as assumed 

 by Pietsch (1979) represent a primitive character state which 

 has been retained separately in the two lineages. The most con- 

 spicuous derived character states of the larvae are the extreme 

 elongation of the body in neoceratiids, the hump-backed shape 

 of ceratiids, and the barbels of centrophrynids. Being each re- 

 stricted to a single family they only confirm the distinct sepa- 

 ration of these lineages. 



Within the families, inter-generic comparisons of larvae are 

 possible only in Gigantactinidae, Linophrynidae, Ceratiidae and 



Oneirodidae. Of the remaining seven families, four are mono- 

 typic and in each of three, which are divided into two genera, 

 larvae of only one is known. In each of these families very 

 distinct inter-generic differences in larval pigment patterns have 

 been found. 



This larval character, retained in juveniles of both sexes, has 

 been one of the main keys to the identification of the free-living 

 metamorphosed males and thus has contributed considerably 

 to the concept of the relationships within the ceratioid families. 



The fact that separation of larvae (and males) below generic 

 level has been possible only in those exceptional cases where 

 intra-generic differences above species rank (subgenera, species- 

 groups, etc.) have been observed, underlines that within this 

 suborder the term "genus" indicates a remarkably well-defined 

 and natural group. 



However, little information on phylogenetic relationships 

 within the families has been obtained from the study of the 

 larvae. The difficulties in interpreting their character states is 

 well illustrated in the Linophrynidae. Two apparently typical 

 derived larval character states occur in this family: (1) well 

 developed sphenotic spines (within larval Lophiiformes found 

 only in the linophrynid genus Borophryne and in one of the 

 three subgenera of Linophryne), and (2) a characteristic sub- 

 dermal pigment pattern (found only in the linophrynid genus 

 Haplophryne and in the two subgenera of Linophryne lacking 

 larval sphenotic spines). If it is assumed very unlikely that these 

 specializations have evolved independently in different genera 

 of the same family, the only alternative is that apparently prim- 

 itive character states are in fact due to three secondary reduc- 

 tions: (1) lack of sphenotic spines in two L/>!o/)/!n'«e subgenera; 

 (2) lack of barbels in female Borophryne, making this one more 

 subgenus oi Linophryne, and (3) lack of subdermal pigment in 

 one of the subgenera of Linophryne and in Borophryne. 



Ceratioids are still very incompletely known and future stud- 

 ies on additional characters and as yet unknown forms may 

 bring answers to at least some of the many questions about their 

 phylogenetic relationships. 



Zoological Museum, University of Copenhagen, 

 Copenhagen 2100 0, Denmark. 



Atherinomorpha: Introduction 



B. B. COLLETTE 



THE superorder Atherinomorpha (Greenwood et al., 1966) 

 includes the atherinoids (silversides and phallostethids), 

 cyprinodontoids (killifishes), and beloniforms (halfbeaks and 

 their relatives), first grouped together by Rosen (1964) as the 

 order Atheriniformes. The series Atherinomorpha was redefined 

 by Rosen and Parenti (1981) as including the Atherinoidei (of 

 uncertain rank), Cyprinodontiformes and Beloniformes. 



Utilizing 17 apomorph characters, Rosen and Parenti (1981) 

 found 1 synapomorphies uniting the atherinoids, Cyprinodon- 

 tiformes, and Beloniformes. Two of these involve early life his- 

 tory characters: complete separation of embryonic afferent and 

 efferent circulation by development of the heart in front of, 

 rather than under, the head and the presence of large demersal 

 eggs with long adhesive and short filaments and many lipid 



globules that coalesce at the vegetal pole. Four additional syn- 

 apomorphies between the Cyprinodontiformes and Beloni- 

 formes show the atherinoids to be the plesiomorphic sister group 

 of these two orders. 



Rosen and Parenti (1981) were unable to find derived char- 

 acters to unite the atherinoids as a monophyletic group but 

 White et al. (this volume) have discovered two early life history 

 characters which define the Atheriniformes as the plesiomor- 

 phous sister group of the Cyprinodontiformes plus Beloni- 

 formes. 



National Marine Fisheries Service Systematics 

 Laboratory , National Museum of Natural History, 

 Washington, District of Columbia 20560. 



