1236 MR. C. S. ELTON ON THE 



among mites. It seems remarkable that eight species of scarlet 

 mites belonging to six genera occurred in Railway Ditch, Two 

 of these species are members of genera in which the other species 

 are mostly non-red {Limnesia and Fiona). The only species which 

 is not bright red, occurring in the ditch, has a special defensive 

 habit, not found among other mites. The absence of any other 

 non-red species is remarkable, since there is no dispersal difficulty 

 to prevent them from immigrating from the Thames. The shape, 

 size, and mode of swimming of all these species are much the 

 same, except that Eylais trails its legs behind. 



If we exclude the possibility of coincidence, the explanation 

 might be that the red mites possess some advantage which is not 

 shared by the non-red ones. Now, there may be two important 

 factors affecting the evolution of a group of animals in the 

 direction of having a common colour : — 



(1) In the words of Prof. Poulton, "the feasibility of certain 



colours and patterns depending upon their efi'ect on the 

 vertebrate eye and thus giving the enemies as easy an 

 education as possible " (4). That is to say, scarlet may 

 have a more striking effect on the retina of the Stickle- 

 back's eye than any other colour, and therefore be more 

 effective as a warning. 



(2) Reinfoi-cing this would be the co-operative advantage of 



Miillerian mimicry as usually understood. 



The early stages in the evolution of a set of Miillerian mimics 

 would be easier to conceiA^e of if there was any evidence that the 

 first factor is an important one, and its effect would be greater in 

 the case of animals which, like these scarlet mites, are of one 

 uniform colour and do not possess elaborate patterns. (The 

 latter, Avhere they occur, cannot be explained without the aid of 

 the Miillerian theory.) Therefore, if it can be shown that 

 Sticklebacks remember scarlet mites better than those with other 

 colours, the case in favour of Miillerian mimicry will be corre- 

 spondingly strengthened. The following experiments show that 

 this may be so : — 



Expt. 10. .The same fish was used as in Expt. 9, after being 

 starved for four days. As was seen in that experiment, a scarlet 

 Hydraraclina paludosa was refused after one trial. A Simocephalus 

 was then put in and eaten at once. A specimen of Hygrobates 

 longipcdpis Harm, was introduced. This mite is coloured black 

 and yellow, and is quite conspicuous. The fish snapped it up and 

 immediately spat it out. This it repeated fourteen times at 

 intervals. Another Simocephahhs was put in and eaten. The 

 mite was again taken in and rejected seven times. Another 

 SimocejyJialus was eaten. The fish now approached the mite 

 four times without touching it. After that the mite was more 

 or less ignored. A Limnesia imididata Miill. was now put in. 

 This species is yellow with black or brown spots. The fish took 



