﻿378 
  PROCEEDINGS 
  OF 
  THE 
  NATIONAL 
  MUSEUM 
  vol. 
  06 
  

  

  South 
  African 
  Miis., 
  vol. 
  30, 
  pp. 
  587-589, 
  1935) 
  has 
  reviewed 
  in 
  an 
  

   excellent 
  manner 
  the 
  problems 
  concerning 
  the 
  confusion 
  in 
  regard 
  to 
  

   African 
  mugilid 
  species 
  and 
  their 
  diagnoses, 
  but 
  he 
  did 
  not 
  attempt 
  to 
  

   solve 
  the 
  problem 
  as 
  to 
  what 
  genera 
  are 
  valid. 
  His 
  discussion 
  of 
  

   characters 
  used 
  in 
  describing 
  mullets 
  is 
  worthy 
  of 
  considerable 
  thought, 
  

   inasmuch 
  as 
  it 
  applies 
  in 
  general 
  to 
  generic 
  descriptions. 
  

  

  Despite 
  the 
  large 
  amount 
  of 
  material 
  examined, 
  I 
  have 
  experienced 
  

   great 
  difficulty 
  in 
  arranguig 
  the 
  species 
  of 
  Mugilidae 
  into 
  genera 
  of 
  

   concise 
  and 
  of 
  clear 
  definition, 
  owing 
  mostly 
  to 
  the 
  paucity 
  of 
  useful 
  

   taxonomic 
  characters 
  and 
  somewhat 
  to 
  the 
  inadequate 
  descriptions 
  

   that 
  abound 
  in 
  the 
  literature. 
  Nevertheless, 
  I 
  have 
  constructed 
  a 
  key 
  

   to 
  the 
  genera 
  that 
  is 
  practical, 
  though 
  artificial 
  in 
  itself, 
  and 
  that 
  I 
  

   believe 
  defines 
  the 
  natural 
  generic 
  units 
  as 
  observed. 
  

  

  The 
  type 
  of 
  Agonostomus 
  bryanti 
  Bean 
  and 
  Weed, 
  which 
  I 
  have 
  

   examined, 
  belongs 
  to 
  the 
  genus 
  Hypseleotris 
  Gill 
  in 
  the 
  family 
  Eleo- 
  

   tridae. 
  It 
  is 
  not 
  a 
  mullet. 
  

  

  The 
  family 
  Mugilidae 
  as 
  recognized 
  by 
  authors 
  is 
  remarkably 
  con- 
  

   stant 
  in 
  anatomical 
  structures 
  as 
  far 
  as 
  investigated 
  by 
  me 
  ; 
  the 
  num- 
  

   ber 
  of 
  vertebrae 
  is 
  usually 
  24 
  or 
  25, 
  with 
  11 
  to 
  13 
  abdominal 
  and 
  11 
  to 
  

   13 
  caudal. 
  The 
  relative 
  positions 
  of 
  the 
  fjns 
  are 
  uniform, 
  and 
  there 
  

   is 
  no 
  outstanding 
  example 
  of 
  a 
  great 
  increase 
  in 
  number 
  of 
  fin 
  rays 
  in 
  

   any 
  genus. 
  Among 
  other 
  characters 
  the 
  shape 
  of 
  body, 
  with 
  depressed 
  

   head 
  anteriorly, 
  spiny 
  preorbital, 
  bi"oad 
  scaly 
  interorbital, 
  large 
  scales 
  

   on 
  head 
  and 
  body, 
  two 
  dorsal 
  fins 
  well 
  separated, 
  the 
  first 
  with 
  pecu- 
  

   liarly 
  arranged 
  spines, 
  are 
  characteristic 
  of 
  all 
  mullets. 
  

  

  Among 
  the 
  characters 
  studied 
  it 
  appears 
  that 
  the 
  mouth 
  parts 
  of 
  

   mugilid 
  fishes 
  have 
  evolved 
  and 
  specialized, 
  whereas 
  the 
  other 
  an- 
  

   atomical 
  structures 
  have 
  remained 
  more 
  or 
  less 
  constant 
  for 
  most 
  of 
  

   the 
  genera. 
  Therefore, 
  I 
  have 
  studied 
  the 
  preorbital, 
  nostrils, 
  teeth, 
  

   jaws, 
  and 
  other 
  mouth 
  parts 
  rather 
  carefully 
  in 
  search 
  of 
  characters 
  

   suitable 
  for 
  defining 
  and 
  recognizmg 
  mugilid 
  genera. 
  In 
  most 
  genera 
  

   the 
  teeth 
  remain 
  fairly 
  constant 
  in 
  structure 
  at 
  all 
  sizes 
  beyond 
  the 
  

   very 
  immature, 
  but 
  in 
  Mugil 
  there 
  is 
  evidence 
  that 
  some 
  of 
  the 
  tips 
  

   are 
  simple 
  in 
  the 
  young, 
  becoming 
  bifid 
  or 
  even 
  trifid 
  in 
  very 
  large 
  

   adults. 
  It 
  is 
  a 
  well-established 
  fact, 
  also, 
  that 
  in 
  certain 
  species 
  

   young 
  mullets 
  have 
  two 
  anal 
  spines, 
  but 
  half 
  grown 
  and 
  adults 
  have 
  

   three 
  anal 
  spines. 
  Recent 
  American 
  autliors 
  refer 
  to 
  the 
  immature 
  

   stage 
  of 
  the 
  mullet 
  with 
  two 
  anal 
  spmes 
  as 
  the 
  "querimana 
  stage." 
  

   The 
  adipose 
  eyelid 
  is 
  another 
  character 
  that 
  develops 
  with 
  age, 
  usually 
  

   appearing 
  at 
  standard 
  lengths 
  of 
  30 
  to 
  50 
  mm. 
  and 
  reaching 
  the 
  highest 
  

   development 
  in 
  the 
  adults, 
  or 
  else 
  being 
  absent 
  at 
  all 
  ages. 
  

  

  The 
  key 
  that 
  follows 
  is 
  based 
  on 
  specim.ens 
  of 
  40 
  mm. 
  and 
  over, 
  but 
  

   one 
  should 
  be 
  able 
  by 
  it 
  to 
  identify 
  mullets 
  to 
  genera 
  at 
  lengths 
  of 
  

   half 
  that 
  or 
  shorter. 
  1 
  1 
  is 
  a 
  preliminary 
  step 
  in 
  the 
  definition 
  of 
  mugilid 
  

   genera, 
  based 
  on 
  the 
  genotypes 
  and 
  other 
  species 
  referred 
  to 
  the 
  genera 
  

   recognized. 
  • 
  

  

  