OBSERVATION AS AGAINST UNDUE GENERALIZATION. I065 



of ca. 100 atmospheres)." In other words, eels cannot mature their gonads 

 nor breed in fresh water, yet there are many persons to the present day who 

 maintain that they must do so, because they can not perceive how eels could be 

 found in ponds and waters isolated from rivers communicating with the ocean. 

 There a.re many ways in which they might be diffused, but that need not concern 

 the biologists; that they must have originated in the ocean is certain. 



No eels which have once spawned have been found in fresh waters, but 

 because large eels have been seen at some place pursuing an upward course, 

 strenuous claims have been made that some do ascend rivers after spawning; 

 here again we have a difficult)- (but an extremely slight one) balanced against a 

 counterfact. 



We have still much to learn about our most common and longest-known 

 species. The Apogon imberhis or rex-mullorum is a Mediterranean fish which 

 had never been regarded as of much interest. Several years ago (1903), how- 

 ever, a French naturalist (L. Vaillant) found its own eggs in the mouth of the 

 male of a related Caribbean fish {Cheilodipterus affinis) and quite recently the 

 United States Deputy Commissioner of Fisheries (Hugh M. Smith) found also 

 in the waters of the Philippine Archipelago a number of species exercising oral 

 incubation. This present month (September i, 1908) Dr. L. Plate records the 

 discovery of a small species of the same group {Apogonichthys strombi) as a 

 commensal of the large whelk known as Strombus gigas. 



With these facts discovered respecting congeneric species, renewed obser- 

 vations should be made. It would be another example of the undue generali- 

 zation which has been deprecated to assume that the Mediterranean fish agreed 

 with its relatives in oral incubation — very much more that it was a commensal. 

 It should be reexamined till something definite can be learned of its habits 

 during the breeding season. Fishermen may often have found individuals 

 with eggs in the mouth and assumed that they had been taken in as food, so 

 that the fact of none of these fish having been recorded with such eggs is a 

 matter of minor consequence. We would have reason for surprise if it should 

 be found that the Mediterranean A /jogon does not exercise oral incubation, and 

 also if other species have commensal habits like the Apogonichthys strombi, but 

 positive assumption is illegitimate in both cases. 



The relationship of fishes to other animals is a subject which will repay 

 future investigation, and search may be rewarded by many cases scarcely less 

 expected than the parasitic habit of the Apogon. Certain tropical poma- 

 centrids of the genera Aynphiprion and Premnas use actinizoans for shelter; 

 the butterfish of the American coast (Poronotus triacanthus) harbors during its 

 early vouth under the disk of a medusa, and so does also the scad (Trachurus 

 trachurus) of Europe. Still more remarkable are the fierasfers which seek 



B. B. F. 190S— Pt 2—25 



