FISHES FROM THE SEA-BOTTOM 425 



only found one (^M. parallelus) which is common to the Indian, 

 Atlantic, and Pacific Oceans. All the 19 species taken at the 

 Sandwich Islands are known only from that locality. Some 

 species, like M. ar77iatus and M. filicmtda, have a wide distribu- 

 tion, but these are exceptions from the rule. Thus, in his 

 opinion, there are no species common to both sides of the Atlantic. 

 The only exceptions then known i^M. siimtlus, M. goodei, 

 M. berglax, and M. rupestris) are explained by him as being due 

 to these species following the cold Labrador current from their 

 normal habitat, the eastern side of the ocean. 



Brauer attempts to explain the peculiar distribution of the 

 Macruridae. He considers that the Macruridse have originated 

 from coast-fishes, and only commenced to migrate towards the 

 abyssal region after a great variety of coast-forms had been 

 developed. ** The fact," he observes, "that only a few species 

 have penetrated into the abyssal plain, while the main body of 

 the species still remains on the slope, tends to show that in 

 most cases the migration towards the abyssal plain is still going 

 on, that it is very slow, and that it has not yet reached the 

 borders of the abyss ; or else it indicates that the abyssal plain 

 tends to limit further distribution, acting as an almost in- 

 surmountable obstacle." 



We have seen that all the deep-sea expeditions, prior to the 

 "Michael Sars," captured only 35 individual "bottom-fishes," 

 and that these belonged to twenty-one species. Our present 

 knowledge must therefore be very imperfect. We have not yet 

 learnt to fish to perfection at 2000 or 3000 fathoms, and we 

 have as yet made too few fishing experiments at such depths. 

 The short cruise of the "Michael Sars" in the Atlantic has 

 essentially altered Brauer's ideas of the distribution of deep-sea 

 fishes, and it appears desirable to give the interesting question 

 raised by him a fresh trial, in view of the large amount of 

 information which we now possess regarding the migrations 

 of many fishes. When, for instance, we find the cod of the 

 Norwegian Sea at one season spawning near the coasts of 

 Norway, at another season migrating to Spitzbergen, or to 

 the slopes of the coast - plateau, we must conclude that 

 fishes may undertake horizontal as well as vertical migrations 

 of enormous extent in a short space of time. Seeing that 

 Macrurus sclero^'-JiyncJms has the enormous bathymetrical range 

 of from 540 to 3655 metres, we can hardly suppose that the dis- 

 tribution of deep-sea fishes down the slope and on the abyssal 

 plain could have been prevented by "lack of time." We have 



