112 COLORATION OF THE SKINS OF FLAT-FISHES. 
in this Journal, vol. i, No. 4, 1890. <A circulation was kept up in 
the bottle over the mirror in the experiment by connecting it with 
another bottle by means of a siphon outflow tube, the aperture of the 
tube being protected by silk bolting cloth, so that the little fish could 
not escape. 
I was absent from Plymouth in July and the early part of August 
in 1890. When I returned I noticed that the little flounders 
continually clung with their lower sides to the darkened sides of the 
bottle, so that the object for which the apparatus was arranged was 
to a great extent defeated. I tried to prevent this by confining the 
fish beneath a horizontal partition of coarse cloth fitted into a 
cylindrical glass vessel substituted for the bottle, but the cloth did 
not allow of sufficient renewal of the water beneath it, and the fish 
were found all dead one morning, having been killed by suffocation. 
There were thirteen of these fish, and all except one had some 
pigment on the lower side. ‘The greatest extent of the pigmenta- 
tion was over the region along the edges of the lower side, from the 
base of the dorsal and ventral fins inwards, the region correspond- 
ing to the muscles of the fins. As far as could be observed in the 
course of the experiment (it was not possible to make a minute 
examination without risking the hfe of the fish), the pigmentation 
present at the end of the experiment was not due to a retention of 
the pigment present on the lower side before the transformation of 
the larval fish was complete, but the original chromatophores had 
disappeared from the lower side as usual, and had been redeveloped 
under the action of hght. 
In my next experiment I took four flounders belonging to the same 
brood as those of the first experiment. These were some of a number 
which had been reared under ordinary conditions, which had long 
passed their transformation and had no pigment on their lower sides. 
They were about five or six months old, and between 2 and 3 inches 
long. JI removed the covering from the sides of the vessel, and left 
off using any partition inside it, keeping only an opaque cover on the 
top. In consequence of this the fish could not protect their lower 
sides from the light by clinging to the sides of the vessel, and their 
upper sides were illuminated by light passing through the sides, as 
well as their lower sides by the light from the mirror. At the 
beginning of 1891 I had made a wooden tank with a plate-glass | 
bottom, which is still in use, and is shown in the figure illustrating 
this article. It is 33 feet long, 2 feet 3 inches broad, and 11 inches 
deep, and I procured large mirrors to place beneath it. In this 
tank the four flounders lived and grew. A recurrence of the old 
difficulty of the fish clinging to the opaque sides took place, and I 
met this as far as possible by keeping the water in the tank very 
