ioo MARVELS OF FISH LIFE 



separate and distinct. At three months old he is very 

 much the same shape that he will be in after-life, except 

 that he is not so thick in the body. At this age the 

 goldfish is a beautiful little object, and can be trans- 

 ferred from the cage into a glass vessel to be watched. 

 The best way to pick out larval fishes is with a glass 

 pipette. 



All the early stages of the members of the carp 

 family are very similar, so far as I have at present 

 examined them, and these include the carp, the bream, 

 the roach, the rudd and the minnow, but there is suffi- 

 cient variation to enable one to be distinguished from 

 the other with the aid of the microscope. 



I used to wonder how it was that the early life 

 histories of our pond fishes had never been described, 

 until I started to work out that of the roach for 

 myself. 



The first year I tried in an aquarium which had 

 been prepared months beforehand and contained abund- 

 ance of microscopic food. The eggs hatched, but all 

 the larvae were dead within a day or two. Next time 

 I arranged six large tubs in a row, connecting them 

 together with three-inch lead pipe, and again bred in 

 these all the minute forms of pond life. This time the 

 roach eggs were placed in an open wooden tank and 

 the water from the tubs was made to circulate through 

 it. Again the eggs were hatched, and again in a day or 

 two all the larvae were dead. The third attempt resulted 

 in complete success, for, in addition to using floating 



