﻿Xll 
  INTKODTJCTIOtf. 
  

  

  centra 
  are 
  never 
  fully 
  formed, 
  and 
  they 
  only 
  reach 
  even 
  the 
  annular 
  

   stage 
  when, 
  notwithstanding 
  specialization, 
  the 
  fish 
  retains 
  its 
  

   original 
  thick 
  and 
  continuous 
  squamation. 
  The 
  cheek-plates 
  are 
  

   apparently 
  never 
  lost, 
  hut 
  become 
  irregularly 
  subdivided 
  in 
  the 
  

   later 
  forms 
  ; 
  and 
  there 
  is 
  always 
  a 
  tendency 
  to 
  reduction 
  of 
  the 
  

   branchiostegal 
  apparatus. 
  The 
  scales 
  often 
  degenerate 
  or 
  disappear 
  

   in 
  certain 
  parts 
  — 
  on 
  the 
  back 
  in 
  Macrosemius, 
  on 
  the 
  tail 
  in 
  Aphne- 
  

   lepis, 
  AetJieoIejois, 
  and 
  Tetragonolepis. 
  Fulcra 
  are 
  least 
  conspicuous 
  

   in 
  the 
  genera 
  which 
  exhibit 
  most 
  tendency 
  to 
  such 
  degeneration 
  of 
  

   the 
  scales. 
  

  

  These 
  considerations 
  are 
  important 
  because, 
  if 
  truly 
  expressions 
  

   of 
  fact, 
  they 
  seem 
  to 
  determine 
  the 
  systematic 
  position 
  and 
  affinities 
  

   of 
  the 
  anomalous 
  family 
  of 
  Pycnodontidae. 
  As 
  amply 
  demonstrated 
  

   in 
  the 
  Catalogue, 
  these 
  fishes 
  exhibit 
  an 
  appendicular 
  skeleton 
  quite 
  

   distinct 
  from 
  that 
  of 
  the 
  Chondrostei, 
  but 
  identical 
  with 
  that 
  of 
  

   the 
  Protospondyli 
  and 
  Aetheospondyli, 
  and 
  only 
  distinguished 
  from 
  

   that 
  of 
  the 
  Isospondyli 
  by 
  the 
  large 
  number 
  of 
  basals 
  in 
  the 
  pectoral 
  

   fin. 
  The 
  additional 
  characters 
  of 
  the 
  axial 
  skeleton 
  of 
  the 
  trunk 
  

   and 
  the 
  mandible 
  place 
  them 
  unhesitatingly 
  among 
  the 
  Proto- 
  

   spondyli. 
  

  

  In 
  five 
  respects 
  they 
  are 
  merely 
  extreme 
  members 
  of 
  the 
  modified 
  

   series 
  of 
  deep-bodied 
  Protospondyli, 
  which 
  begins 
  with 
  Dapedius. 
  

   There 
  is 
  never 
  any 
  approach 
  to 
  the 
  development 
  of 
  vertebral 
  centra 
  ; 
  

   and 
  even 
  when 
  the 
  axis 
  acquires 
  special 
  rigidity 
  (e.g. 
  Palceoba- 
  

   listum 
  and 
  Pycnodus) 
  this 
  is 
  attained 
  by 
  the 
  interlocking 
  of 
  laminar 
  

   expansions 
  of 
  the 
  neural 
  and 
  haemal 
  arches. 
  The 
  cheek-plates, 
  so 
  

   far 
  as 
  known, 
  are 
  reduced 
  to 
  tesserae. 
  The 
  branchiostegal 
  rays 
  are 
  

   not 
  more 
  than 
  two 
  in 
  number. 
  The 
  fin-fulcra 
  are 
  still 
  more 
  

   insignificant 
  than 
  in 
  Tetragonolepis. 
  The 
  caudal 
  region 
  is 
  very 
  

   commonly 
  destitute, 
  or 
  nearly 
  destitute, 
  of 
  scales 
  behind 
  a 
  line 
  con- 
  

   necting 
  the 
  origin 
  of 
  the 
  dorsal 
  with 
  that 
  of 
  the 
  anal 
  fin. 
  

  

  "When 
  examined 
  more 
  closely, 
  it 
  is 
  not 
  difficult 
  to 
  perceive 
  that 
  

   even 
  the 
  apparently 
  unique 
  characters 
  of 
  the 
  Pycnodonts 
  are 
  also 
  

   the 
  ultimate 
  result 
  of 
  tendencies 
  to 
  specialization 
  already 
  exhibited 
  

   by 
  the 
  Semionotidge. 
  

  

  Firstly, 
  there 
  is 
  the 
  forwardly-displaced 
  mouth 
  with 
  crushing- 
  

   teeth, 
  which 
  is 
  much 
  like 
  that 
  of 
  the 
  latest 
  species 
  of 
  Lepidotus, 
  

   only 
  a 
  little 
  more 
  advanced 
  and 
  powerful, 
  and 
  thus 
  more 
  firmly 
  

   fixed 
  to 
  the 
  skull. 
  The 
  mode 
  of 
  fixation, 
  however, 
  resembles 
  that 
  

   observed 
  in 
  fishes 
  so 
  remotely 
  related 
  to 
  each 
  other 
  as 
  Gymnarchus, 
  

   Tetrodon, 
  and 
  Mormyrus 
  : 
  proving 
  that 
  it 
  is 
  merely 
  a 
  physiological 
  

  

  