an inch in depth. Over this moat I then placed a paper bridge, one end of which rested on some fine mould. I then put an ant to the honey, and soon a little crowd was collected round it. I then removed the paper bridge; the ants could not cross the glycerine; they came to the edge and walked round and round, but were unable to get across, nor did it occur to them to make a bridge or bank across the glycerine with the mould which I had placed so conveniently for them. I was the more surprised at this on account of the ingenuity with which they avail themselves of earth for constructing their nests. For instance, wishing, if possible, to avoid the trouble of frequently moistening the earth in my nests, I supplied one of my communities of Lasius flavus with a frame containing, instead of earth, a piece of linen, one portion of which projected beyond the frame and was immersed in water. The linen then sucked up the water by capillary attraction, and thus the air in the frame was kept moist. The ants approved of this arrangement, and took up their quarters in the frame. To minimize evaporation I usually closed the frames all round, leaving only one or two small openings for the ants, but in this case I left the outer side of the frame open. The ants, however, did not like being thus exposed; they therefore brought earth from some little distance, and built up a regular wall along the open side, blocking up the space between the upper and lower plates of glass, and leaving only one or two small openings for themselves. This struck

me as very ingenious. The same expedient was, moreover, repeated under similar circumstances by the slaves belonging to my nest of Polyergus.

The facility or difficulty with which ants find their way, while it partly falls within the section of the subject dealing with their organs of sense, is also closely connected with the question of their general intelligence.

Partly, then, in order to test how far they are guided by sight, partly to test their intelligence, I made various observations and experiments, the accompanying woodcuts being reduced copies of tracings of some of the routes followed by the ants during the course of the observations.

I may here note that the diagrams Figs. 12-17 are careful reductions of large tracings made during the experiments. Though not absolutely correct in every minute detail of contour, they are exact for all practical purposes. As the ants pursued their way, pencil-markings in certain instances, and coloured lines in others, were made so as to follow consecutively the paths pursued.

Experiment 1.-February. On a table communicating with one of my nests (see Fig. 12) I placed upright a common cylindrical lead pencil inch in diameter and 7 inches long, fastened with sealing-wax to a penny piece. Close to the base of the pencil (a) I brought the end of a paper bridge (B) leading to the nest, and then placed a shallow glass with larvæ at C,

4 inches from the base of the pencil. I then put an ant to the larvæ; when she had become acquainted with the road, she went very straight, as is shown in the woodcut (Fig. 12). In one case, at the point E, she dropped her larva and returned for another. When

Routes followed in experiment No. 1, as detailed above.

A, position of pencil. B, paper bridge. C and D, glass with larvæ. E, point where larva dropped, the opposite arrow and loop marking return route. 1, 2, 3, 4, comparatively straight paths to the glass. 5, 5, circuitous route on shifting of glass. different access to nest.

*

she returned on the next journey and was on the glass, I moved it 3 inches, to D, so that the end of the glass was 6 inches from the base of the pencil. If she were much guided by sight, then she would have had little

or no difficulty in finding her way back. Her pathway, however (No. 5), which is traced on the paper, shows that she was completely abroad; and, after all, she got back to the nest by a different route,

Routes followed in experiment No. 2, as mentioned in text. B, paper bridge leading to nest. C, glass tray with larvæ, in its first position; and D in its position when shifted. 1, 2, 3, 4, thin white lines indicating the comparatively straight routes. 5, thick white line, and 6, dotted line showing tortuous paths when glass had been altered in position. The arrows indicate directions travelled.

I then varied the experiment as subjoined, and as shown in the woodcut (Fig. 13).

Experiment 2.-I connected the table with the nest by a paper bridge, the end of which is shown at в (Fig. 13), and which came down about an inch. from the pole supporting the nest (see Fig. 1). This pole rose 18 inches above the table. I then put the glass tray (c) with larvæ as before, 12 inches from the base of the pole, and put an ant to the larvæ. When she had learnt her way I traced four of her routes, as shown in the thin lines 1, 2, 3, 4. I then on her next journey (5, thick white line), when she was on the tray (C), moved it three inches to D, as shown in the figure, and again traced her route. The contrast is very striking between the relatively straight thin white lines 1, 2, 3, 4 of the four journeys when familiar with the road; whereas in the broad white line No. 5 the zigzag twistings show how much difficulty the ant experienced in finding her way. When she returned I again moved the tray as before, and the dotted sinuous white line (6) shows the course she followed.

Experiment 3.-I then again varied the experiment as follows:-I placed the larvæ in a small china cup on the top of the pencil, which thus formed a column 7 inches high. The cross line close to the arrows (Fig. 14) is as before, the base of the paper bridge leading to the nest. c shows the position of the penny on which the pencil was supported. The dotted white lines 1, 2, 3, 4 show the routes of a marked ant on four