( 132 ) 



On the Classification of Fishes. By L. Agassiz. 



Far from participating in the opinion of those who regard 

 Our classifications merely as an artificial scaffolding calculated 

 to facilitate our researches, by assisting the memory, I ani 

 firmly persuaded that the progress of the natural sciences will, 

 sooner or later, lead to the establishment of a system which 

 shall be the true and complete expression of the various re- 

 lations by which the entire series of created beings are con- 

 liected with each other. But such a system cannot be estab- 

 lished until we acquire a more complete acquaintance with the 

 innumerable variety of objects in this vast field of inquiry. The 

 attempts hitherto made to attain to it, appear to me to be only 

 the first foundation of the edifice, a provisional means for re- 

 cognising when we are among the varied forms which must be 

 registered according to their diverse affinities. Even the 

 very principles which must guide us in this operation have 

 not been definitively settled. We may compare the efforts of 

 naturalists desirous of grouping natural bodies in the most con- 

 venient manner, to the labours of engineers wishing to repre- 

 sent the aspect of a country in a map. They first fix a few 

 salient points from which they can command the whole. 

 From thence they enter upon the details by subdividing the con- 

 siderable extent comprised within their first triangulation ; they 

 then study each new section by itself, by traversing it in every 

 direction. It is then only that they can begin to note the pe- 

 culiarities which form the special character of these restricted 

 compartments. With such materials alone a good map may 

 be constructed. But one observer would wish further to learn 

 the heights of the mountains which are indicated in relief ; an- 

 other would desire to know what parts of the surface are wooded, 

 or what is cultivated, and what not ; while another might de- 

 sire information on the diff*erent climatological phenomena, and 

 find nothing to guide him. In these respects, therefore, a 

 map constructed on the basis mentioned, would be insufficient 

 for the wants of science. New researches would become ne- 

 cessary ; the results of geological and meteorological studies 

 would have to be combined with geodesical details ; draw- 

 ings to represent the accidents of the formations would have 



