MAY-FLIES. 41 



which I am acquainted approaching it in this respect. 

 The general colour of the larva is very protective, which 

 renders its detection a matter of considerable difficulty 

 when it is resting on stones in the water. These nymphs 

 are very common in some streams, and as they are always 

 to be found, they must provide the trout with a certain 

 amount of food, especially during the winter. On the 

 other hand, being carnivorous and very rapacious, they 

 destroy large numbers of the larvse of the smaller Ephe- 

 meridse, which are equally, if not more, serviceable as 

 food to the trout, than themselves, so that it is doubtful 

 whether the angler has much cause to thank this insect. 



When full grown this nymph crawls out of the stream 

 and clings tightly to the surface of a dry stone, the 

 winged insect afterwards emerging through a large rent 

 in the back of the thorax. In one specimen I observed 

 the process of exclusion lasted about eight minutes, but 

 the wings expanded to their full size almost as soon as 

 they were withdrawn from their cases. The May-fly 

 appears to give them one shake, and then places them in 

 the usual vertical position over its back. 



The sub-imago is a most beautiful insect. Its wings are semi-opaque 

 and of the most brilliant yellow, with a few faint blackish markings. The 

 legs and tails are considerably shorter than in the imago. The eyes are 

 bright green in colour. 



The insect remains in this condition for about three 

 days, during which interval of time it quietly rests 

 amongst the vegetation on the bank of the stream. The 

 colouring of the sub-imago is no doubt protective, the 

 beautiful yellow tint of the wings closely resembling the 

 hue of certain faded leaves. This resemblance is espe- 

 cially noticeable when the sub-imago is resting just after 

 emergence on the surface of the water. At this critical 

 moment the May-fly is in great danger of attack both 

 from fish beneath the water and from birds above it, 

 these creatures being only too ready to seize and devour 

 the helpless insect. It is therefore probable that a special 

 protective resemblance, such as the present species enjoys, 

 would be most efficient in preserving it from destruction 

 at this time ; and such a resemblance w T ould readily be 

 produced by natural selection continually preserving 

 those variations, which conformed in the closest possible 

 manner to the leaf -like pattern. 



The perfect insect appears in January, February, and 

 March. It is about for a much longer interval than most 



