200 THE COCKROACH : 



in the waters of the sea. This is shown by the slow rate of 

 change in freshwater types. Many of our genera of land and 

 freshwater shells date back at least as far as to Purbeck and 

 Wealden times, while our common pond-mussel is represented 

 in the Coal Measures. The comparative security of fresh 

 waters is probably the reason why so many marine fishes enter 

 rivers to spawn. 



More important, and less open to question, is the direct action 

 of the sphere of life. The cheap method of turning the 

 embryo out to shift for itself can seldom be practised with 

 success on land. But in water floating is easy, and swimming 

 not difficult. A very slightly-built larva can move about by 

 means of cilia, and a whole brood can disperse far and wide in 

 search of food, while still in a mere planula condition hollow 

 sacs, without mouth, nerves, or sense-organs. Afterwards the 

 little locomotive larva settles down, opens a mouth, and begins 

 to feed. Nearly the whole of its development is carried on at 

 its own charge. 



The extra risks to which marine animals are exposed also 

 tell in favour of transformation, for they are met by an increase 

 in the number of ova. Marine species commonly lay more eggs 

 than freshwater animals of like habits. The Cod is said to 

 produce nine million eggs ; the Salmon from twenty to thirty 

 thousand ; the Stickleback only about one hundred, which are 

 guarded during hatching by the male. The Siluroid fish, Ariits, 

 lays a very few eggs, as big as small cherries, which the male 

 carries about in his mouth. 



Without laying stress upon such figures as these, which 

 cannot be impartially selected, we can safely affirm that marine 

 forms are commonly far more prolific than their freshwater 

 allies. But high numbers increase the difficulty of providing 

 yolk for each, and thus tend to early exclusion, and subsequent 

 transformation. We may rationally connect marine habitat 

 with small eggs, poorly supplied with yolk, segmenting regu- 

 larly, and producing larva which develop with metamorphosis. 



In fresh waters dispersal can seldom be very effective. The 

 area is usually small, and communicates with other freshwater 

 basins only through the sea. Migration to a considerable distance 

 is usually impossible, and migration to a trifling distance use- 



