known to exist. In the attempt to give an estimate I can onlj 

 approach the truth, as in numerous cases it is impossible to deciae 

 from imperfect accounts whether an author has described a distinct 

 species or one previously known. Again, the views of ichthyologists 

 on species diverge so much that one will give a number several 

 times as great as another*. I consider a species to be well 

 established only when it is founded on characters which, from 

 an examination of numerous examples, are found to be perma- 

 nent, not subject to gradiud variation, and not dependent on 

 season, sex, or age — or which are known to be so from the ex- 

 amination of allied forms. A character of this kind is in general 

 constantly accompanied by another, which would appear to be 

 insignificant by itself. By this principle I have been guided 

 throughout the work, and in naming the species of the Collection 

 of the Museum ; and this should be borne in mind in comparing 

 the numbers of species given by me with the estimates prepared 

 by other natiu'alists. 



" In the present work 6843 species are regarded as well esta- 

 blished and described, whilst 1682 others are doubtful and re- 

 feiTed to by name only. Assuming, then, that about one half of the 

 latter will be ultimately admitted into the system, and that, since 

 the publication of the volumes of this work, about 1000 species have 

 been described elsewhere, we may put the total number of fishes knoivn 

 at present as about 9000. 



•' At the respective dates of the preparation of the eight volumes 

 of the present work 4219 species were available for autoptical exa- 

 mination. To these were afterwards added 958 species which were 

 received too late for insertion ; so that the Collection of the British 



* " Prof. Agassiz states, in iSilliui. Anier. Jom-n. 1854, xvii. p. 360, that he 

 knows at least 20 North-American specips of Lcpidosfetis ; I can distinguish 

 three or four only. In Prof. Kaup's ' Catalogue of Apodal Fishes ' the two 

 European Eels admitted by me are split into 20 species. Prof. Dumeril, finally, 

 has published the names of some 80 Sturgeons distinguished by him ^Jvouv. 

 Arch. Mus. d'Hist. Nat. iii. 1867) ; I cannot recognize more than twenty. Such 

 nominal species rarely survive their author ; but, before they are merged again 

 in the synonymy, they are the cause of much unnecessary trouble, and, being 

 founded on slight individual peculiarities, tliey are frequently mistaken, rarely 

 recognized." 



