Desert ants learn vibration and magnetic landmarks

Abstract

The desert ants Cataglyphis navigate not only by path integration but also by using visual and olfactory landmarks to pinpoint the nest entrance. Here we show that Cataglyphis noda can additionally use magnetic and vibrational landmarks as nest-defining cues. The magnetic field may typically provide directional rather than positional information, and vibrational signals so far have been shown to be involved in social behavior. Thus it remains questionable if magnetic and vibration landmarks are usually provided by the ants' habitat as nest-defining cues. However, our results point to the flexibility of the ants' navigational system, which even makes use of cues that are probably most often sensed in a different context.

Figure 1. Experimental procedure.

(A) The ants' nest was connected with a tube to the training channel where the ants were trained to visit a feeder 1 m away from the nest entrance that was marked with either a magnetic, vibrational, visual, olfactory or no landmark. For size and shape of the solenoid, and for the application of the massaging rod next to the channel see Material and Methods. (B) Trained ants were displaced from the feeder of the training channel into the parallel test channel (displacement shown by dashed arrow) where the homing runs and nest searches of the tested ants were tracked and recorded. Blue filled circle, nest entrance; black filled circle, feeder; black empty circle, release point; blue empty circle, fictive nest position, red rectangle, landmark; blue dashed line, nest position as defined by path integration, red dashed line, nest position as defined by landmark. Nest-to-feeder distance, 1 m; landmark was 1 m behind fictive nest position in test channel. (C) Exemplar homing run and nest search. We analyzed the first six turning points (TP1–TP6) after the ants had crossed the nest-defining cue for the first time.

Figure 2. Individual test runs of homing ants.

Schematic nest searches of ants trained and tested with a nest-defining landmark that was either a magnetic, vibrational, visual or olfactory cue (red), control ants trained and tested without landmark (black) or naïve ants that experienced the landmark in the test for the first time (blue). Blue dashed line, nest position as defined by path integration; red dashed line, nest position as defined by landmark; point of release for each homing run at position -2 m from nest-defining cue. The first six turning points after the ants had passed the landmark for the first time were analyzed for their median position.

Figure 3. Ants learn magnetic, vibrational, visual and olfactory landmarks.

Box plot representation of the medians of the first six turning points of ants that were trained and tested with a landmark (red boxes), control ants trained and tested without a landmark (white box), and naïve ants that experienced the landmark during the test for the first time (blue boxes). Blue dashed line, nest position as defined by path integration; red dashed line, nest position as defined by landmark. Box plots show median, interquartile range and whiskers indicating the 90th and 10th percentiles. Kruskal-Wallis test and Dunn's multiple comparison tests were performed for selected pairs: control (training and test without landmark) versus naïve ants (landmark only during test), for each landmark type P>0.05; ants trained and tested with the landmark versus naïve ants, for each landmark type P<0.05; ants trained and tested with the landmark versus control ants, for each landmark type P<0.05. Numbers depict sample sizes.