﻿4 
  PROCEEDINGS 
  OF 
  THE 
  NATIONAL 
  MUSEUM 
  vol.98 
  

  

  Japan, 
  India, 
  Australia, 
  Tasmania, 
  the 
  south 
  and 
  east 
  coasts 
  of 
  South 
  

   Africa, 
  and 
  Chile. 
  Its 
  range 
  should 
  be 
  materially 
  extended 
  as 
  more 
  

   collecting 
  is 
  done 
  in 
  the 
  ocean 
  surf. 
  Perhaps 
  from 
  an 
  ancestral 
  stock, 
  

   represented 
  by 
  the 
  Atherioninae, 
  with 
  the 
  monotypic 
  genus 
  Atherion, 
  

   the 
  Tropidostethinae 
  could 
  have 
  evolved. 
  Both 
  appear 
  to 
  prefer 
  surf 
  

   or 
  near-surf 
  conditions. 
  

  

  In 
  the 
  Menidiiuae 
  I 
  have 
  recognized 
  18 
  genera, 
  fresh-water 
  and 
  ma- 
  

   rine, 
  all 
  of 
  which 
  are 
  American. 
  It 
  too 
  is 
  a 
  New 
  World 
  subfamily, 
  

   ranging 
  on 
  the 
  Atlantic 
  side 
  from 
  Nova 
  Scotia 
  to 
  Brazil 
  and 
  on 
  the 
  

   Pacific 
  side 
  from 
  Mexico 
  to 
  Peru 
  and 
  the 
  Galapagos 
  Islands. 
  The 
  

   Menidiinae 
  contain 
  18 
  genera 
  with 
  the 
  most 
  variable 
  characters 
  in 
  any 
  

   subfamily 
  of 
  the 
  silversides, 
  and 
  these 
  characters 
  overlap 
  somewhat 
  

   among 
  a 
  few 
  closely 
  related 
  genera. 
  There 
  are 
  eight 
  genera 
  found 
  only 
  

   on 
  the 
  Atlantic 
  side 
  and 
  seven 
  on 
  the 
  Pacific. 
  Three 
  genera, 
  Melano- 
  

   rhinus, 
  Thyrino'ps^ 
  and 
  Coleotrofis^ 
  have 
  species 
  on 
  both 
  sides 
  of 
  Cen- 
  

   tral 
  America. 
  Many 
  of 
  the 
  genera 
  in 
  this 
  family 
  are 
  peculiar 
  to 
  cer- 
  

   tain 
  stream 
  systems 
  of 
  the 
  New 
  World. 
  

  

  The 
  Atherinopsinae 
  are 
  a 
  New 
  World 
  subfamily 
  containing 
  seven 
  

   known 
  genera, 
  six 
  of 
  which 
  are 
  confined 
  to 
  the 
  Pacific 
  coastal 
  region 
  

   ranging 
  from 
  Oregon 
  to 
  Mexico 
  and 
  from 
  Peru 
  to 
  Patagonia. 
  The 
  

   other 
  genus, 
  Odontesthes^ 
  occurs 
  from 
  Argentina 
  to 
  Brazil. 
  Both 
  

   fresh- 
  water 
  and 
  marine 
  species 
  are 
  found 
  in 
  this 
  American 
  subfamily. 
  

   So 
  far 
  none 
  of 
  the 
  genera 
  have 
  been 
  collected 
  on 
  the 
  Atlantic 
  or 
  Pacific 
  

   sides 
  of 
  Central 
  America. 
  Four 
  genera 
  occur 
  from 
  Oregon 
  to 
  Baja 
  

   California 
  and 
  the 
  Gulf 
  of 
  California, 
  and 
  the 
  other 
  three 
  genera 
  occur 
  

   from 
  Peru 
  to 
  Patagonia, 
  the 
  Falkland 
  Islands 
  northward 
  to 
  Brazil. 
  

   As 
  far 
  as 
  known, 
  the 
  subfamily 
  is 
  not 
  tropical, 
  occurring 
  in 
  the 
  temper- 
  

   ate 
  and 
  subtropical 
  regions. 
  

  

  The 
  New 
  World 
  forms 
  the 
  sole 
  habitat 
  of 
  28 
  genera 
  of 
  silversides 
  and 
  

   shares 
  1 
  additional 
  genus, 
  AUaneUa, 
  with 
  the 
  tropical 
  Western 
  Pacific 
  

   and 
  Indian 
  Oceans 
  and 
  the 
  tropical 
  Atlantic 
  Ocean. 
  There 
  remain 
  11 
  

   genera 
  occurring 
  in 
  the 
  Old 
  World. 
  Nine 
  of 
  these 
  are 
  found 
  in 
  the 
  

   Western 
  Pacific 
  and 
  Indian 
  Oceans, 
  leaving 
  two 
  for 
  Europe 
  and 
  Asia 
  

   Minor. 
  

  

  My 
  kej^ 
  to 
  the 
  genera, 
  though 
  somewhat 
  artificial 
  in 
  itself, 
  is 
  prac- 
  

   tical 
  and 
  defines 
  the 
  natural 
  generic 
  units 
  as 
  observed 
  by 
  me. 
  Unfor- 
  

   tunately, 
  those 
  who 
  classify 
  and 
  base 
  genera 
  mostly 
  on 
  external 
  char- 
  

   acters 
  will 
  find 
  this 
  family 
  a 
  little 
  too 
  complex. 
  Ichthyological 
  studies 
  

   must 
  be 
  increasingly 
  more 
  carefully 
  done 
  if 
  the 
  various 
  genera 
  are 
  to 
  

   be 
  thoroughly 
  understood. 
  Though 
  it 
  is 
  necessary 
  to 
  use 
  certain 
  in- 
  

   ternal 
  structures, 
  this 
  method 
  need 
  not 
  be 
  too 
  complex 
  for 
  practical 
  

   purposes. 
  

  

  The 
  premaxillary 
  may 
  be 
  pulled 
  forward 
  to 
  expose 
  the 
  nature 
  of 
  the 
  

   ascending 
  processes, 
  and 
  the 
  tissues 
  should 
  be 
  dissected 
  off 
  to 
  expose 
  

  

  