Research Article |
Corresponding author: Günter Gollmann ( guenter.gollmann@univie.ac.at ) Academic editor: Peter Mikulíček
© 2020 Markus Pail, Lukas Landler, Günter Gollmann.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Pail M, Landler L, Gollmann G (2020) Orientation and navigation in Bufo bufo: a quest for repeatability of arena experiments. Herpetozoa 33: 139-147. https://doi.org/10.3897/herpetozoa.33.e52854
|
Research on navigation in animals is hampered by conflicting results and failed replications. In order to assess the generality of previous results, male Bufo bufo were collected during their breeding migration and translocated to two testing sites, 2.4 and 2.9 km away, respectively, from their breeding pond in the north of Vienna (Austria). There each toad was tested twice for orientation responses in a circular arena, on the night of collection and four days later. On the first test day, the toads showed significant axial orientation along their individual former migration direction. On the second test day, no significant homeward orientation was detected. Both results accord with findings of previous experiments with toads from another population. We analysed the potential influence of environmental factors (temperature, cloud cover and lunar cycle) on toad orientations using a MANOVA approach. Although cloud cover and lunar cycle had small effects on the second test day, they could not explain the absence of homeward orientation. The absence of homing responses in these tests may be either caused by the absence of navigational capabilities of toads beyond their home ranges, or by inadequacies of the applied method. To resolve this question, tracking of freely moving toads should have greater potential than the use of arena experiments.
Amphibia, behavioural ecology, Bufonidae, direction following, MANOVA, migration
The question of how animals navigate has been investigated in numerous species over many decades (
While the function of the olfactory system is well understood, the mechanisms underlying magneto-sensation are still debated. Currently, there are three main hypotheses of magnetoreception: 1) the light-dependent mechanism, based on the formation of spin-correlated radical pairs (
One might wonder why amphibians should possess such elaborate spatial capabilities at all, as they are usually regarded as small and slow-moving animals. Nevertheless, they can accomplish quite surprising spatial tasks and home from large distances, compared to their size. The common view might need some rethinking. Typical amphibian home ranges might only cover a few hundred meters, but some (e.g. red-bellied newts and water frogs) have been shown to home from up to 4 km and even 15 km (
From all anuran species the European common toad (Bufo bufo) is arguably the best investigated one in terms of its homing abilities. To quickly summarize the cornerstones of previous common toad migration studies: They have a tightly controlled and highly active (explosive) breeding migration (
In earlier experiments, we investigated whether we could elicit navigational responses in the common toad (
Also, in follow-up experiments where we translocated toads from the same migration route 2.1 km to an indoor testing set-up, toads showed direction following behaviour when tested at the same night. However, when we left the toads at the testing site for 3 days, presumably enough time for the toads to update their internal map, they oriented randomly (
For the present study, we collected toads migrating to another pond in the north of Vienna and tested them in the same arena at two different sites, located approximately 120° apart with respect to the pond. Our aims were twofold: first, to examine whether we could replicate the direction following behaviour immediately after collection; second, whether we could elicit a ‘true’ navigation response in an outdoor situation, after the toads had been kept at the testing site for 4 days, presumably enough time to update their positional information.
Male toads (B. bufo) were collected during their spawning migration close to their breeding pond (on Bisamberg, Vienna, Austria, 48.31294N, 16.38474E) and translocated to one of two testing sites; ‘site 1’, 2.9 km away (backyard in Floridsdorf, Vienna, Austria, 48.30458N, 16.42174E, homeward direction: 292°), and ‘site 2’ (Seeschlacht in Langenzersdorf, Lower Austria, Austria, 48.29834N, 16.36147E, homeward direction: 45°), 2.4 km away (Fig.
Map of the breeding pond and the two test locations. Toads were collected nearby the breeding pond and translocated to one of the two testing locations. The blue arrows indicate the mean migration direction (all toads were collected in the north-west of the pond), the black arrows the direction of possible homeward orientation.
Experiments took place from 10 to 23 April 2013 from dusk to approximately midnight. Mean testing temperature was 10 °C (SD: 4 °C) and mean cloud cover was 30% (SD: 40%). Following collection, toads were placed in uncovered plastic buckets and transported to one of the two testing sites by car. Testing began immediately after arriving at the testing site; the testing order was identical to the order of collection. Before, and after testing, plastic buckets with toads were placed 10 to 20 m away from the testing rig. Toads were tested in a visually symmetrical circular arena (diameter: 121 cm, height: 60 cm), which had been used in previous studies (
After this first day of experiments, animals were held for 4 days at the same location in the plastic buckets; the toads were kept wet the whole time, in order to prevent desiccation. On day 4 toads were tested again (second day of experiments), in order to test for a homing response (‘true navigation’ sensu
Infrared lights and an infrared camera (Panasonic NV-DS28EG) were used to record the trials. From the recorded videos screenshots were taken using the VLC media player 2.0.3 and then the image manipulating software GIMP 2.8. An inner radius was used (85% of the whole arena diameter) to determine the directional preference of each toad. Earlier experiments had shown that toads tend to follow the wall when being close to it, without immediately touching it, leading to a less clustered orientation. The direction for each toad was defined as the direction where the toad crossed the 85% criterion circle and measured to the nearest 5°.
Orientation data for each of the sites and test days were analysed using standard circular statistics (
In order to test potential influences of weather or lunar cycle on orientation we performed a MANOVA (using the function lm together with Manova from the package car (
Circular plots were generated using an adapted plot.circular function derived from the package circular (see R Code in Suppl. material
Out of 116 toads 96 were successfully tested and reached the arena wall in time. The individual d-axis directions of the toads tested at both sites were tightly clustered around 110° with respect to north (Fig.
On both test days and sites toads oriented randomly, when analysed towards geographic north (Fig.
In contrast, when analysed relative to the individual former migration direction, toads showed significant axial orientation along the expected direction at the evening of collection. Four days later toads showed weakly significant unimodal orientation towards the d-axis direction at site 1 but no significant orientation at site 2 (Fig.
Weather and the lunar cycle had only minor effects on the orientation. On the first test day the selected model included cloud cover and location, however, none of the two reached significance and the effect size (eta2) was well below 0.1 for both factors (Table
Also, on the second test day the effect size for all factors was below 0.1, however, cloud cover and the cosine of the lunar cycle reached significance (Table
Toad orientation on the first test day at site 1 (A) and site 2 (B) and the second test day at site 1 (C) and site 2 (D), relative to geographic north. The arrows represent the mean vectors of the distributions (radius of the circle corresponds to a vector length of 1), none of distributions reached significance (p-values (p) shown in the plots). Each dot represents the orientation of a single toad.
Toad orientation on the first test day analysed for site 1 (A) and site 2 (B) and the second test day for site 1 (C) and site 2 (D), relative to the d-axis. The arrows represent the mean vectors (the circle’s diameter equals r = 1, doubled headed arrows in case of axial orientation). P-values (p) are given in each plot. Dotted lines indicate bootstrap 95% confidence intervals for significant orientations. Each dot represents the orientation of a single toad.
MANOVA table showing results from the first test day after AIC-based model selection. Degrees of freedom (df), Pillai test statistics (test statistics), approximated F statistics (approx. F), degrees of freedom for the numerator (num df), degrees of freedom for the denominator (den df), p-values (p) and effect sizes (eta2) are shown for selected model.
Factor | df | test statistics | approx F | num df | den df | p | eta2 |
---|---|---|---|---|---|---|---|
Cloud cover | 1 | 0.056 | 2.719 | 2 | 92 | 0.071 | 0.056 |
Location | 1 | 0.060 | 2.975 | 2 | 92 | 0.056 | 0.061 |
MANOVA table showing results from the second test day after AIC-based model selection. Degrees of freedom (df), Pillai test statistics (test statistics), approximated F statistics (approx. F), degrees of freedom for the numerator (num df), degrees of freedom for the denominator (den df), p-values (p) and effect sizes (eta2) are shown for selected model.
Factor | df | test statistics | approx F | num df | den df | p | eta2 |
---|---|---|---|---|---|---|---|
Cloud cover | 1 | 0.071 | 3.481 | 2 | 91 | 0.035 | 0.071 |
cos_lunar | 1 | 0.090 | 4.524 | 2 | 91 | 0.013 | 0.090 |
sin_lunar | 1 | 0.047 | 2.241 | 2 | 91 | 0.112 | 0.047 |
In the present experiment we confirmed direction following (d-axis) orientation behaviour in toads tested at the evening of collection (Fig.
Axial responses, however, are common in behavioural orientation studies (
In contrast, the toads did not orient towards their home pond. There are two possibilities to explain such a result: First, B. bufo might be unable to ‘truly’ navigate. Second, the method we used to explore ‘true navigation’ is unsuitable for this species.
One argument supporting the first possibility is that the resolution of a magnetic map might be 10 km at its best (
If the navigational abilities of the toads were limited, however, why are there many reports indicating surprising homing performances consistent with the use of map-like navigation systems (
We collected toads from their way to the breeding pond, not directly out of the pond. Our rationale for collecting animals during their migration was that we surmised migrating toads to be highly motivated to reach the pond, whereas toads already present there might eventually lose the motivation to return later in the breeding season (
Problems with replicability of research results are not restricted to navigation studies, but have triggered intense discussions of conceptual and statistical questions in the behavioural, biomedical and social sciences (
We thank Judith Pail, Doris Reitschmidt and Danijel Vukancic for assistance in the field, and Hannes Pail for technical support. The market town Langenzersdorf generously provided a location for the experiments as well as storage space. Collection of toads for the experiments was authorized by permit MA22-3726/2009 from the municipality of Vienna. LL was partially funded by the Austrian Science Fund (FWF, Grant Number: P32586). Open access funding provided by University of Vienna.
R-script 1. R code to reproduce analysis and plots
Data type: R file
Explanation note: Can be opened using R, open source and free statistical computing software. In order to run it supplementary materials 2, 3 and 4 have to be placed in the working directory.
Table S1. Raw data for the first test day
Data type: csv-file
Explanation note: This file is called when running supplementary material 1 in R.
Table S2. Raw data for the second test day
Data type: csv-file
Explanation note: This file is called when running supplementary material 1 in R.
R-script 2. Adapted circular functions not included in R packages
Data type: R file
Explanation note: This file is called when running supplementary material 1 in R.