QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bisch-Knaden, S. Articles by Wehner, R. Search for Related Content PubMed PubMed Citation Articles by Bisch-Knaden, S. Articles by Wehner, R.

Egocentric information helps desert ants to navigate around familiar obstacles

Sonja Bisch-Knaden and Rüdiger Wehner^*

Department of Zoology, University of Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland

View larger version (15K): [in a new window]   Fig. 1. (A) Training arrangement. The outbound and inbound paths of ants are indicated by black and green arrows, respectively. The heavy grey line marks a V-shaped barrier, each arm of which is 6 m long and 5 cm high, enclosing an angle of 120°. A sandy ramp (see B) attached to the inside of the V-shaped array enabled the ants to pass the barrier on their way from the nest N to the feeder F. In the opposite direction, F to N, smooth, light-brown tape glued to the feeder-facing side of the barrier forced the ants to detour around the left (El) or right (Er) end of the barrier. T, central tip of the V-shaped barrier. Lengths of path elements: NF, 13 m; FT, 5 m; TEl and TEr, 6 m; ElN and ErN, 7.2 m. (B) Transverse section of the barrier. (C) Data analysis. Example trajectory of a zero-vector ant tested with the barrier in the training orientation. The ant was released at T and its path was recorded for 5 min. The following variables were analysed: (i) the walking distance along the barrier immediately before the ant turned around Er (orange); (ii) the length of the subsequent path segment (green, ‘detour vector’) measured from the intersection of the trajectory with the imagined extension of the barrier until the ant made a sharp turn of more than 90° compared with its preceding running direction (arrowhead); (iii) the angular deviation of the detour vector from a line connecting Er and the fictive position of the nest relative to the barrier (). For this purpose, the intersections (small circles) of the trajectory of the ant with concentric circles (radii 1, 2, 3 and 4 m) centred on Er were determined.

View larger version (38K): [in a new window]   Fig. 2. Detour vectors associated with the barrier in the training orientation. (A) Vector ants (left side N=9, right side N=10) were released at the fictive position of the feeder relative to the barrier (5 m south of the tip of the barrier). (B) Zero-vector ants (left side N=13, right side N=12) were released at the tip of the V-shaped barrier. The green arrows point towards the fictive position of the nest relative to the barrier (). The fictive position of the nest relative to the home vector of the ant coincides with in A and is located at the tip of the V-shaped barrier (the release point) in B. The normalised path density histograms were obtained by determining the path lengths of the detour vectors in squares of 0.5 mx0.5 m. These values were then assigned to seven classes ranging from 0 % (white) to more than 1.5 % (red) of the total path length of the detour vectors at each of the ends of the barrier.

View larger version (21K): [in a new window]   Fig. 3. Angular distribution of detour vectors in vector ants compared with that of zero-vector ants. Circles represent the first intersections (open circles, vector ants; filled symbols, zero-vector ants) between the trajectories of the ants and concentric circles (radii 1, 2, 3 and 4 m) centred on the ends of the barrier. Vector ants (open arrowheads) and zero-vector ants (filled arrowheads) do not differ in their mean angles and angular variations (P>0.05, Mardia–Watson–Wheeler test) at each of the circles. In both groups of ants, the mean angles do not deviate from the line (0°, arrow) connecting the end of the barrier with the fictive position of the nest relative to the barrier (); angles are given clockwise from this direction. A comparison of the statistical details for vector ants and for zero-vector ants (values given in parentheses) at the 3 m circle for the left side of the barrier shows: mean angles 1.7° (359°); r-values 0.952 (0.941); number of ants 9 (13); 95 % confidence intervals 347.9–15.5° (348.9–13.1°); for the right side of the barrier, mean angles 5.4° (13.2°); r-values 0.969 (0.880); number of ants 10 (10); 95 % confidence intervals 355.1–15.7° (352.5–33.9°).

View larger version (32K): [in a new window]   Fig. 4. Detour vectors associated with rotation of the barrier through 45°. (A) Top: barrier oriented as during training; the green arrows point to the position of the nest relative to the barrier (). Bottom: barrier rotated through 45° to the west (left) and east (right); the arrows indicate the hypothetical directions of detour vectors according to hypothesis 1 (allocentric system of reference, blue) and hypothesis 3 (egocentric system of reference, green). (B) Zero-vector ants were released at the tip of a barrier rotated either to the west (left side N=9, right side N=9) or to the east (left side N=13, right side N=12). The path density histogram shows the trajectories from the east-rotated barrier together with the mirror-reversed trajectories from the west-rotated barrier. For further explanation, see Fig. 2.

View larger version (22K): [in a new window]   Fig. 5. Angular distribution of detour vectors associated with the rotated barrier. Detour vectors of zero-vector ants are superimposed and rotated in such a way that 0° marks the direction of the detour vector defined within an egocentric system of reference (hypothesis 3, green arrow). The arrow at 45° indicates the direction of the detour vector defined within an allocentric system of reference (hypothesis 1, blue arrow). The filled circles represent the first intersections of the trajectories of the ants with concentric circles (radii 1, 2, 3 and 4 m) centred on the end of the barrier (El/Er). The mean directions (arrowheads) do not deviate from 0°, but they are significantly different from 45°. Statistical details for the 3 m circle: mean angle 7.0°; r-value 0.926; number of ants 42; 95 % confidence limits 359.9–14.1°.

View larger version (22K): [in a new window]   Fig. 6. Angular distribution of zero-vector ants released within the test area without barrier. The ants (N=22) were released within a test area devoid of landmarks, and their trajectories were recorded for 5 min each. The first intersections (filled circles) of their trajectories with concentric circles (radii 1, 2, 3 and 4 m) centred on the point of release () do not show any directional preference (P>0.2 for all circles, Rayleigh test).

View larger version (10K): [in a new window]   Fig. 7. Detour experiments with a channel. Recently published data (see Fig. 3d–g in Collett et al., 1998) were re-evaluated according to the method used in the present study. (A) Training array. Desert ants were trained to travel the first half of their homeward journey from a feeder F to the nest N inside an east-pointing channel (marked by the heavy black line) hidden in a trench. The ants were therefore forced to deviate from their straight homebound path (indicated by the dashed line). The thin arrow marks the paths taken by ants when travelling from the exit of the channel southwards to N over open ground. (B–D) Test situations (top) and results (bottom). Ants were captured close to the nest entrance and released at a feeder F' inside test channels of different lengths (B, 8 m; C, 4 m; D, 2 m). The channels were rotated through 45° to the south (B, C and D) or north (C). The arrows indicate the hypothetical directions of detour vectors defined in egocentric (90° relative to the channel, green) and allocentric (south, blue) systems of reference. Circular diagrams show the angular distributions of the trajectories of the ants after they had left the rotated channels. Dots represent the first intersections of these trajectories with a circle (radius 1 m) centred on the exit of the channel (cross). Runs that did not reach this 1-m circle were excluded from the evaluation (B, 3 runs out of 17; C, 11 runs out of 43; D, 11 runs out of 25). The remaining trajectories are superimposed in such a way that 0° marks the direction of detour vectors defined in egocentric systems of reference and 45° marks detour vector direction using allocentric systems of reference. In B and C, the angular distributions seem to be bimodal, so mean directions could not be calculated. In D, the mean direction (arrowhead) coincides with the allocentrically defined direction of the detour vector (hypothesis 1) but deviates significantly from the egocentrically defined one (hypothesis 3; mean angle 45°; r-value 0.970; number of ants 14; 95 % confidence limits 23–67°).

View larger version (34K): [in a new window]   Fig. 8. Trajectories of vector ants released within the test area without a barrier. The ants (N=18) were captured at the feeder and released within a test area devoid of landmarks. During training (see Fig. 1), the ants chose to detour around the left (N=10, red) or the right (N=8, green) end of the barrier. The dashed lines mark the fictive position of the barrier relative to the point of release (). The trajectories are shown until the ants started to search. , fictive position of the nest relative to the point of release.