120 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [May 18, 



might have been afforded to the general question by the accurate 

 knowledge of the characteristic fossils, should have been so much 

 weakened by the uncertainty which has gradually enveloped the sub- 

 ject, in consequence of the undue multipHcation of species, the con- 

 flicting nomenclature, and rival schemes of classification propounded 

 by those engaged in this investigation. These errors are, no doubt, 

 to a certain extent unavoidable. The labour and difficulty of con- 

 structing from fragmentary specimens extinct forms unlike any now 

 existing are very great. Again, when men of science in widely sepa- 

 rated localities are engaged upon the same subject, much time must 

 necessarily elapse before they can become mutually acquainted with 

 each other's progress ; and this knowledge, when at last obtained, is 

 frequently defective, having to be gathered, it may be, from foreign 

 languages, and from descriptions without figures, often meagre and 

 unintelligible. These are certainly extenuating circumstances. At the 

 same time a vicious practice is too prevalent — of coining genera and 

 species with undue haste from imperfect data, and so seeking to 

 establish what is termed the right of priority, however erroneous 

 the determination may have been. In some of the Ichthyolitic 

 beds of the Old Red Sandstone in Scotland, the specimens are so 

 well preserved, that the specific characters are easily seen ; in 

 others, on the contrary, such as those of Scats Craig and Elgin, 

 they are found in a most fragmentary and disjointed condition, con- 

 sisting of detached scales, broken plates, and variously formed teeth. 

 To determine with any degree of accuracy the characters and rela- 

 tions of these fragments is a task of great difficulty, and requires 

 the utmost patience. The proportions and ornament of a scale, or 

 the pattern of the sculptured surface of a dermal plate, are frequently 

 the sole features we have to rely upon ; and these are subject to much 

 variation. The relative proportions of the scales to each other may 

 vary in different parts of the same fish ; those on the back and belly 

 may be elongated or lozenge- shaped, while those on the flanks are 

 deep or curvilinear. The surface -ornament too is by no means 

 constant in character over the whole body. Age and sex may also 

 cause considerable modifications. The risk is therefore great of 

 forming new genera and species based upon data which may prove 

 of trivial moment as our knowledge of the true characters augments. 

 In our endeavours to steer clear of this Scylla, variety, we must be- 

 ware of falling into the Charybdis similarity. The surface-patterns 

 of the dermal plates of many of the fishes of the Devonian strata are 

 so much alike (as, for instance, in Asterolepis, Pterichthys, and Ce- 

 phalaspis), that fragments of each might be selected so similar in cha- 

 racter as to lead to the supposition that they were derived from one 

 and the same genus, and so the observer might be betrayed into the 

 suppression of generic and specific forms through an excessive exer- 

 cise of caution. Professor Agassiz has been accused of being guilty 

 of the former fault — that of multiplying genera and species without 

 due consideration ; but in all his writings and correspondence he 

 acknowledges this impeachment, and justifies it. In conducting a 

 great work on the classification and description of all Fossil Fishes 



