he yellow, sometimes over the red. I also put four thicknesses of violet glass, so that it looked almost black.

Under these circumstances, the pupa were placed under the red 7 times, dark yellow 5, once they were half under each, but never under the violet, purple, light yellow, dark or light green.

The following day I placed over the same nest, in the sun, dark green glass, dark red, and dark yellow. In nine observations the pupa were carried three times under the red and nine times under the yellow.

I then tried a similar series of experiments with Lasius niger, using a nest in which were about 40 pupa, which were generally collected in a single heap all together. As before, the glasses were moved in regular order after each experiment; and I arranged them so that the violet followed the red. As far, therefore, as position was concerned, this gave violet rather the best place. The glasses used were dark violet, dark red, dark green, and yellow, the yellow being distinctly the most transparent to our eyes. Experiment

Experiment
8. Pupa under green.

10.

11.

green.

9

39

I now put two extra thicknesses of glass over the

The result is very striking, and in accordance with the observations on Formica fusca. In 40 experiments the pupæ were carried under the yellow 19 times, under the red 16 times, and under the green 5 times only, while the violet was quite neglected. After the first twenty observations, however, I removed it.

I then tried a nest of Cremastogaster scutellaris with violet glass, purple glass, and red, yellow, and green solutions, formed respectively with fuchsine, bichromate of potash, and chloride of copper. The purple looked almost black, the violet very dark; the

red and green, on the contrary, very transparent, and the yellow even more so. The yellow was not darker than a tincture of saffron. The latter indeed, to mv eye, scarcely seemed to render the insects under them at all less apparent; while under the violet and purple I could not trace them at all. I altered the relative positions as before. The nest contained about 50 larvæ and pupæ.

I made thirteen trials, and in every case the larvæ and pups were brought under the yellow or the green -never once under any of the other colours.

Again, over a nest of Formica fusca containing about 20 pupa I placed violet glass, purple glass, a weak solution of fuchsine (carmine), the same of chloride of copper (green), and of bichromate of potash (yellow, not darker than saffron).

I made eleven trials, and again, in every case the pupa were brought under the yellow or the green.

I then tried a nest of Lasius flavus with the purple glass, violet glass, very weak bichromate of potash, and chloride of copper as before.

With this species, again, the results were the same as in the previous cases.

In all these experiments, therefore, the violet and purple light affected the ants much more strongly than the yellow and green.

It is curious that the coloured glasses appear to act on the ants (speaking roughly) as they would, or,

I should rather say, inversely as they would, on a photographic plate. It might even be alleged that the avoidance of the violet glass by the ants was due to their preferring rays transmitted by the other glasses. From the habits of these insects such an explanation would be very improbable. If, however, the preference for the other coloured glasses to the violet was due to the transmission and not to the absorption of rays-that is to say, if the ants went under the green rather than the violet because the green transmitted rays which were agreeable to the ants, and which the violet glass, on the contrary, stopped-then, if the violet was placed over the other colours, they would become as distasteful to the ants as the violet itself. On the contrary, however, whether the violet glass was placed over the others or not, the ants equally readily took shelter under them. Obviously, therefore, the ants avoid the violet glass because they dislike the rays which it transmits.

But though the ants so markedly avoided the violet glass, still, as might be expected, the violet glass certainly had some effect, because if it were put over the nest alone, the ants preferred being under it to being under the plain glass only.

I then compared the violet glass with a solution of ammorio-sulphate of copper, which is very similar in colour, though perhaps a little more violet, and arranged the depth of the fluid so as to make it as nearly as posible of the same depth of colour as the glass.

In another experiment with Lasius niger I used the dark yellow glass, dark violet glass, and a violet solution of 5 per cent. ammonio-sulphate of copper, diluted so as to be, to my eye, of exactly the same tint as the violet glass; in 8 observations the pupa were three times under the violet solution, and 5 times under the yellow glass. I then removed the yellow glass, and in 10 more observations the pupa were always brought under the solution.

It is interesting that the glass and the solution should affect the ants so differently, because to my eye the two were almost identical in colour. The glass, however, was more transparent than the solution.

To see whether there would be the same difference between red glass and red solution as between violet glass and violet solution, I then (Aug. 21) put over a nest of Formica fusca a red glass and a solution of carmine, as nearly as I could make it of the same tint. In 10 experiments, however, the ants were, generally speaking, some under the solution and some under the glass, in, moreover, as nearly as possible equal numbers.

August 20.-Over a nest of Formica fusca con