ESPECIALLY MYCTOPHOIDS 235 
covering the scales and dermal bones in pholidophorids (Nybelin, 1966). The loss 
of enamel has made possible the sinking inwards of dermal bones of the skull and the 
production of uroneurals in the caudal fin skeleton in teleosts, This latter feature, 
concerning the internal uroneurals attached to the ural vertebrae, has been used by 
Patterson (1967c) provisionally as a diagnostic feature of teleosts. 
The leptolepids have convincingly been shown to have their origin within the 
Pholidophoridae (Rayner, 1948 : 338), and several authors (Rayner, 1948 : 340 ; 
Griffith and Patterson, 1963 : 40) have proposed that the Leptolepidae gave rise to 
the remainder of the teleosts. Other authors have supposed only the clupeoids to 
be direct derivatives of the leptolepids. Gardiner (1960 : 351) has indicated that 
both the Clupeidae and Chirocentridae can be derived from the leptolepids, all three 
groups having no bone cells in their scales, whereas elopoids and salmonoids, which 
still retain bone cells in the scales, must have evolved from pholidophoroids. 
Greenwood, ef al. (1966) have stressed that the Clupeomorpha have several unique 
characters which have so far not been elucidated in the leptolepids. Patterson 
(1967c) has considered the origin of the Clupeomorpha in detail and has refuted 
Bardack’s (1965) hypothesis concerning the derivation of the Chirocentridae from 
forms like the ichthyodectids and the Upper Jurassic Thrissops and Allothrissops. 
The origin of the Clupeidae according to Saint-Seine (1949) from a form like Lepto- 
lepis coryphaenoides has also been refuted by Patterson (1967c). Woodward 
(1942b : 908) postulated the derivation of Diplomystus from a pholidophoroid and 
not from a leptolepid because of the presence of ridge scutes behind the pelvics in the 
diplomystids and in some pholidophorids. This character of Woodward’s (1942b) is 
insignificant, but nevertheless his idea of deriving the clupeoids from a pholidophorid 
and not a leptolepid is sound. Arambourg (1954), in describing the new genus 
Clupavus, related it to the clupeoids, in particular the Dussumieriidae. Patterson 
(1967c) has convincingly shown this to be incorrect and has demonstrated the 
closeness of Clupavus to the leptolepids, and their dissociation from the clupeo- 
morphs. Greenwood, ef al. (1966 : 360) are in agreement with Patterson (1967Cc), 
and both propose a connection between Clupavus and Leptolepis in a convergent 
evolutionary trend to the Clupeomorpha. This provides an instance of what 
Schaeffer (1965 : 322) maintained when he stated, “‘ mosaics which in retrospect 
are involved in a transition to a higher level, may be modified in much the same way 
in a number of related lineages ”’. 
The Elopiformes represent the most primitive of living teleosts and this group 
retains the greatest number of halecostome characters. Gardiner (1960 : 351) has 
indicated that on scale structure alone the Elopiformes could be derived most 
satisfactorily from a pholidophoroid but not from a leptolepid. Saint-Seine (1949) 
has described a late Jurassic ‘ protelopid’ Eoprotelops, but Gardiner (1967b) con- 
siders this to be a halecostome, and Patterson (1967b) also indicates that it is not an 
elopoid. Bertin and Arambourg (1958 : 2200) had previously appended Eoprotelops 
to the family Leptolepidae. 
An elopid caudal skeleton has been described from the late Jurassic by Nybelin 
(1963) and it is probable that the elopiform lineage was in existence throughout 
