Figure 1.
Apparatus for manipulation of substrate cues during colony emigration.
A) Side view shows a box containing the old nest below an arena mounted on risers. The new nest was fastened to the bottom of one corner of the arena, which ants reached by climbing a wooden dowel. A Fluon-coated collar guided the ants through an entrance hole to the upper surface of the arena. B) Top view shows the arena floor carpeted with a transparent acetate sheet. In the midst of each emigration, the sheet was rotated 60° clockwise. Emigrating ants were observed to determine if they continued to walk toward the new nest (dashed line) or shifted their headings to match the new substrate position (solid line).
Figure 2.
Substrate rotation does not affect orientation by emigrating ants.
A) The last trips before rotation (dashed blue line and open blue circles) were highly directed (Rayleigh test: r = 0.95, p = 6×10−8), with an average heading only 3° from the direction of the new nest (0°). After rotation (solid red line and closed red circles), headings remained highly directed (Rayleigh test: r = 0.93, p = 9×10−8), with an average value only 4° from the nest direction. The difference in mean headings before and after rotation was not significant (Watson's two-sample test: U2 = 0.1541, 0.10> p>0.05, n = 19). B) The mean change in heading by each ant could not be distinguished from zero (Hotelling test: F2, 17 = 0.32, p = 0.73). C) The duration of journeys from the old to the new nest was not affected by rotation (Wilcoxon paired ranks test: V = 62.5, n = 19, p = 0.30). In each box plot the closed circle shows the median, boxes delimit the 1st and 3rd quartiles, and whiskers show the range. The open circle is an outlier.
Figure 3.
Sample paths before and after substrate rotation.
Each panel shows a single ant's paths before rotation (dashed blue line) and afterwards (solid red line). Squares show the arena boundary. Small closed circles show locations of old nest entrance at arena center and new nest entrance at upper right corner.
Figure 4.
Substrate rotation plus blocking of visual cues causes disorientation.
A) Before rotation (dashed blue line and open blue circles), headings were significantly directed (r = 0.89, p = 8×10−5) and average heading deviated 11° from the direction of the new nest (0°). After rotation (solid red line and closed red circles), headings were highly dispersed, and the mean heading could not be distinguished from random (average heading = −65°, r = 0.27, p = 0.53). B) The change in heading was extremely variable across ants and the mean angular change could not be distinguished from random (average change = −72°, r = 0.13, p = 0.86). C) The duration of journeys from the old to the new nest increased sharply after blocking of visual cues (Wilcoxon paired ranks test: V = 0, n = 8, p = 0.0078). See Figure 2 for box plot details.
Figure 5.
Sample paths before and after substrate rotation plus blocking of visual cues.
Each panel shows a single ant's paths before manipulation (dashed blue line) and afterwards (solid red line). Squares show arena boundary. Small closed circles show locations of old nest entrance at arena center and new nest entrance at upper right corner.
Figure 6.
Apparatus for manipulation of visual cues during colony emigration.
A) Side view shows the arena, old nest, and new nest arranged as in Figure 1, except that the arena was circular and designed to minimize visual cues (see Materials and Methods). The arena was centered in a cylinder with a complex visual scene displayed on its inner surface (partially cut away to show the arena). The cylinder rested on a copy stand that also carried a video camera and a screen to mask visual cues from the ceiling. The copy stand was mounted on a turntable so that it, along with the cylinder, camera, and screen, could be rotated around the stationary arena. B) Top view shows the arena floor. In the midst of each emigration, the copy stand was rotated 90° clockwise. Emigrating ants were observed to determine if they continued to walk toward the new nest (dashed line) or shifted their headings to match the new position of the visual scene (solid line). An acetate sheet was placed on the arena floor for consistency with the substrate-cue experiments, but it was not rotated.
Figure 7.
Rotation of visual cues strongly affects orientation by emigrating ants.
A) The last trip before rotation (dashed blue line and open blue circles) was highly directed (Rayleigh test: r = 0.88, p = 6×10−7) and deviated only 3° from the direction of the new nest (0°). After rotation (solid red line and closed red circles), average heading remained highly directed (Rayleigh test: r = 0.88, p = 8×10−7) but shifted to 65° clockwise from the nest heading (Watson's two-sample test: U2 = 0.3883, p<0.001, n = 13). B) The average change in heading by each ant was 67° (r = 0.89, p = 5×10−7) and the mean change significantly differed from zero (Hotelling test: F2, 11 = 35.2, p = 2×10−5). C) The duration of journeys from the old to the new nest increased after rotation of visual cues (Wilcoxon paired ranks test: V = 2, n = 11, p = 0.0029). See Figure 2 for box plot details.
Figure 8.
Sample paths before and after rotation of visual cues.
Each panel shows a single ant's paths before rotation (dashed blue line) and afterwards (solid line). Paths after rotation are shown in red until they reach the edge of the acetate circle carpeting the arena, and in pink afterwards. Large circles show arena boundary. Small closed circles show locations of old nest entrance at arena center and new nest entrance at right side.