BERTELSEN: CERATIOIDEI 



333 



male Ceratiidae; and very different high specialization of the 

 illicial and jaw mechanisms of female Neoceratiidae, Thau- 

 matichthyidae and Gigantactinidae. 



The interrelationships of the ceratioid families have been dis- 

 cussed by Regan (1912a, 1926), Regan and Trewavas (1932), 

 Bertelsen (1951), and Pietsch (1972,1 979), the latter illustrating 

 with branching diagrams, alternative proposals for phylogenetic 

 relationships of the families. However, no detailed analysis or 

 full discussion of the basis for these proposals has been pre- 

 sented. For this reason the dendrogram shown in Fig. 1 7 1 should 

 be regarded only as a very schematic compilation of the ex- 

 pressed views, following most closely Pietsch ( 1 979: fig. 26) with 

 some modifications discussed below. 



In accordance with Bertelsen (1951) and Pietsch (1979) it is 

 assumed that sexual parasitism has been established indepen- 

 dently in different phylogenetic lineages. The observation of a 

 parasitic male (character 26) in a representative of one of the 

 1 5 oneirodid genera (Pietsch, 1 976) makes it extremely unlikely 

 that the five families in which such males have been observed 

 represent a monophyletic line. (Furthermore, this observation 

 underlines the possibility that sexual parasitism might be found 

 in other families as well.) It seems that the evolutionary step 

 from the temporary attachment of the male to the female, by 

 means of the denticular teeth present in all ceratioid males (and 

 resulting in a close and protracted contact between the dermis 

 of the pair), to a fusion of their tissues is a less drastic event 

 than it might be supposed and has been established indepen- 

 dently in different taxa and possibly even facultative in some, 

 as proposed by Pietsch (1976). 



Presence of an escal photophore (no. 1 1) is presumed to be a 

 synapomorphy separating the other families from Caulophryn- 

 idae (and ?Neoceratiidae), a primitive sister-group. This implies 

 that the similarity of some derived character states (nos. 4. 6, 

 9, 15) to those of one or more of the families Linophrynidae. 

 Gigantactinidae, and Ceratiidae is due to convergence in these 

 bone reductions. The alternative, proposed by Pietsch (1979), 

 that these families were derived from a caulophrynid-like ances- 

 tor, would imply that the escal photophore has been evolved 

 independently in two separate lineages. Morphologic and his- 

 tologic studies of these organs in different families show simi- 

 larities in such details that this seems extremely unlikely [cf for 

 instance Brauer. 1908 (Gigantactis); O'Day, 1974 (Oneirodes); 

 Hansen and Herring, 1977 (Linophryne); and Munk and Ber- 

 telsen, 1980 (Chaenophryne)]. 



Based on a number of shared derived character states (nos. 

 6, 7, 15, 20, and presence of teeth externally on the jaws) it has 

 been assumed that Neoceratiidae are closely related to Gigan- 

 tactinidae. However, they differ considerably in other characters 

 (nos. 5, 9, 13, 14, 21, and 24) and especially in the illicial and 

 jaw mechanisms of the females. While the complete loss of 

 illicium in neoceratiids undoubtedly is a derived character state 

 it remains uncertain whether this family is derived from ances- 

 tors with or without escal photophores. As discussed in the 

 following section, some larval characters might indicate the lat- 

 ter possibility. In reference to this the numerous characters 

 shared by the two genera of gigantactinids leaves no doubt that 

 the lack of photophore in Rhynchactis is due to secondary re- 

 duction (Bertelsen et al., 1981). While none of the highly spe- 

 cialized families Linophrynidae, Gigantactinidae, and Cerati- 

 idae appear closely related, their shared derived character states 

 may indicate a common descendence as shown in Figure 171. 

 As pointed out by Pietsch (1972) Centrophrynidae has retained 



a number of primitive character states but may be more closely 

 related to Ceratiidae than to any other family, and Thaumati- 

 chthyidae are most probably derived from an oneirodid-like 

 ancestor (Bertelsen and Struhsaker, 1977). The remaining four 

 families Melanocetidae, Himantolophidae, Diceratiidae, and 

 Oneirodidae appear more similar to each other than the more 

 specialized families mentioned above, but as their shared char- 

 acter states are nearly all primitive their interrelationships are 

 uncertain. The position of Melanocetidae in the dendrogram 

 (Fig. 171) is based on the presumption that a reduction of the 

 number of dorsal fin rays to less than 10 is synapomorphic 

 within the following series of families. 



Except for the significance of observed sexual parasitism 

 the characters of the males have not been considered in previous 

 discussions of the interrelationships of the ceratioid families. 

 The presence of denticular teeth shared by all families is a de- 

 rived character state in relation to all other Lophiiformes. The 

 absence of such denticles on the snout observed in caulophryn- 

 ids and neoceratiids may represent a primitive state within the 

 suborder. In accordance with the classification based on the 

 characters of the females or shared by the sexes, the males are 

 highly but very differently specialized in the families Ceratiidae, 

 Gigantactinidae, and Linophrynidae while the least number of 

 presumed derived character states are found in Melanocetidae, 

 Himantolophidae, and Diceratiidae. 



Within the families the inter-generic relationships appear close 

 and relatively simple in the four families divided into two gen- 

 era. In each of these one of the genera shows more derived 

 character states in reductions and specializations than the other. 

 (Rohia in Caulophrynidae; Phrynichthys in Diceratiidae; Tfiau- 

 matichthys in Thaumatichthyidae; Cryptopsaras in Ceratiidae 

 and Rhynchactis in Gigantactinidae). Among the five genera of 

 the well-defined family Linophrynidae, L/«op/;n'««' appears the 

 most derived (females with photophore carrying barbels). Bor- 

 ophryne and Acenlrophryne seem closely related to this genus 

 (very similar osteology and dentition) while each of the genera 

 Edriolychmts and Photocorynus appear more isolated; the latter 

 has retained a number of primitive or less derived character 

 states (nos. 10, 13, 15, 17). 



In contrast to the other ceratioid families no conspicuous 

 distinctive characters have been found which are common to 

 the large assemblage of genera united in the family Oneirodidae. 

 However, the presence of quadrate and articular spines in most 

 of the genera and shared only with the closely related thau- 

 matichthyids might be significant and their absence in some 

 genera could be due to secondary reduction. On the basis of 

 osteological characters the evolutionary relationships of 9 of 

 the 1 5 genera were studied by Pietsch ( 1 974) and notes on some 

 of the others have been added by Bertelsen and Pietsch (1975) 

 and Pietsch (1975). According to these studies Spiniphryne ap- 

 pears the most primitive of these genera, having retained well- 

 developed dermal spines, among a number of other primitive 

 character states. Among the most specialized genera are Lo- 

 phodolos (reduction or loss of some elements of the skeleton 

 and enlargements of others) and Chaenophryne (lack of sphen- 

 otic, quadrate and articular spines, shape of opercular bones and 

 a unique structure of ossifications; Pietsch, 1975). 



Contribution of early life history stages.— Apart from meristic 

 and osteological characters shared with adults, the larvae of 

 ceratioid taxa differ from each other only in pigmentation and 

 to some extent in morphology. As the pigment patterns vary 



