378 Mr. F. Merrifield and Mr. E. B. Poulton on 



of a second-floor window about 3 feet wide and 6 feet high, 

 they were facing the W.N.W. and looking into an open 

 country, and were often screened by a muslin window- 

 blinil during hot sunshine. No direct light from the sky 

 reached any part of the interior except a small part of the 

 bottom and of the sides. 



Much orange light came through the roof of the first 

 compartment, ver}^ little greenish light through the top of 

 the green compartment, and much white light througli 

 the top of the white compartment ; in addition to white 

 direct light, much was reflected from the coloured interior 

 of these three compartments, but the black compartment 

 was very dark inside except the part close to the glass 

 front. 



The effect was that the black compartment was for the 

 most part very dark ; the white, green and orange-yellow 

 compartments much lighter, as the tops transmitted much 

 light, and the colours were such as to reflect a great deal 

 of light. In the green compartment, however, while 

 covered Avith the fourfold art muslin, very little light 

 came through the top. In all cases a little light came in 

 through the chinks of the doors, aud a very little through 

 the draped perforated zinc at the back. In all four com- 

 partments there were shady regions in the angles of the 

 framework, and I found so many of the larvae had a dis- 

 position to select these shadier regions for pupating, where 

 of course the coloured light would operate less strongly, 

 that I found it expedient to transfer them, when I could 

 do so in time, to receptacles where they were exposed to 

 stronger light. For these purposes I prepared glass 

 cylinders of the dimensions before mentioned, covered at 

 the bottom and for about two-thirds of the outside circum- 

 ference with paper of the appropriate colour, leaving clear 

 the one-third next the window, and the tops being usually 

 covered with paper of the same colour, but sometimes with 

 clear glass. In this way I fitted up cylinders for the follow- 

 ing colours : black, white, green, yellow, orange, Dutch 

 " gold," all provided with sticks covered with paper of the 

 corresponding colour. These cylinders were placed on thin 

 pieces of wood or cork carpet with tintacks driven through 

 the bottom, forming spikes on which the coloured sticks 

 were fixed. The bottoms, as well as about two-thirds of 

 the circumference opposite the light, and the tops, were 

 covered with single or double paper of the proper colour 



