glass a hair 16 centimetres long, with a volume of liquid not exceeding its diameter. Replacing the oil of olives with water, I obtained the same result. The hair was 0·06 of a millimetre in diameter, and the weight suspended may be calculated to have been 0·00045 of a gramme. Repeating the experiment with horse-hairs, I found that a hair 7·5 centimetres in length remained suspended under the same conditions. The hair was 0·12 of a millimetre thick, consequently the weight adhering to the glass was 0·00085 of a gramme. A hog's bristle 0·18 of a millimetre in diameter was suspended, although, being 55 millimetres long, it represented a weight of 0·00132 of a gramme.
I also experimented with a hair ending in a bulb, which I formed by holding the hair to a flame. I fixed to the glass a hair 0·06 of a millimetre in diameter, terminating in a bulb 0·12 of a millimetre in diameter, and weighing 0·00085 of a gramme, or the same as the horse-hair previously used.
The results of these experiments added weight to my supposition that the liquid does not have to be viscous to enable the flies to stick. To gain an absolute conviction, I weighed a number of flies, and found their mean weight to be 0·045 of a gramme. I then ascertained the number of hairs on the lower part of the pulvilli, and the size of the extremities which they brought to bear upon the glass. It is not an exaggeration to put the number on each pelote at 800 or 1,000; this would give the fly a total of 10,000 or 12,000 hairs, by means of which, with the assistance of a minute drop of liquid, it could support itself on a solid body. It is proper to add, however, that a fly running on a window has only three or four of its feet on the glass at a time, and that therefore only half of its hairs, or 5,000 or 6,000 of them, are serving it at once. I repeated my experiments, to determine the weight hairs are capable of supporting when suspended in the manner I have described, and found again that a hair 0·06 of a millimetre in diameter will bear a weight of 0*00045 of a gramme; of 0·12 millimetre, 0*00085 of a gramme; and of 0·18 millimetre, 0·00132 of a gramme, when the air is in motion. Then, according to my calculations, a fly would be able to walk upon glass, even if it weighed 0·020 of a gramme more than it actually does. I tested this by pasting little papers on the wings of flies to increase their weight. They still kept themselves on the glass; but they walked upward with some difiiculty when their weight was doubled.
I perceived in the course of my experiments that the flies, especially the weighted ones, ceased to adhere to the glass when it was moistened with the breath. Blackwall had essayed to explain this fact by assuming that the sticky substance by means of which he supposed they adhered mingled with the water, and was so much diluted by it as to cease to be effective. I found, by examination with the microscope, that this was not the case; no mixture or dilution took place, but rather a repulsion of the oily liquid by the water, and that that, or
