Corresponding author: Cédric Devigne (
Academic editor: J. Štrus
Terrestrial isopods are known to be sensitive to humidity, brightness or temperature. Until now, aggregation was assumed to depend on these sensitivities as a result of individual preferences. In this paper, we show that the social component is also important in the isopod aggregation phenomenon. In experimental arenas with two identical shelters up to nearly 90% of woodlice aggregated under shelters. This aggregation was quick as in 10 minutes most of the animals aggregated, irrespective of their density. Nonetheless, 10–15% of the animals walked around the arena, rarely forming very small and short-lasting aggregates outside shelters. Woodlice aggregated preferably under one of the shelters in 77% of experiments. Indeed, almost 80% of the animals out of 40, 60 or 80 animals in the arena aggregated under one shelter. In arenas with 100 individuals the aggregations were proportionally smaller (70%). Our results revealed that 70 animals was a maximum number of woodlice in an aggregate. We concluded that the location of aggregates is strongly governed by individual preferences but the dynamics of aggregation and collective choice are controlled by social interaction between congeners. The tested densities of the animals in the arena did not impact the aggregation patterns.
Woodlice are mainly detritivorous organisms feeding on leaf litter, decayed wood, fungi, and bacteria. They are one of the most important groups of organisms driving the dynamics of soil (
The rough woodlouse,
The experimental set-up consisted of a circular arena (diameter193mm) with two dark shelters (
Experimental set-up.
Shelters consisted of a small glass plate (diameter 35mm placed at 5mm of soil). Darkness in shelters was achieved by adding to glass plates, two layers of red ROSCO® filters (ref. Roscolux #19 Fire – this filter changed the spectrum of light by transmitted to nearly only red energy). The set-up was lit with 156 lux and the brightness under both shelters was only 41 lux. Both shelters in the arena were strictly identical in size, darkness and contact surface with the edge of the arena. No bias between the number of woodlice observed under the left and the right shelter could be found by analyzing the whole data (Wilcoxon’s test, p=0.263, N=87).
Before the experiments, woodlice were placed in groups of 40 (N=29), 60 (N=20), 80 (N=20) or 100 (N=19) individuals in the centre of the experimental arena in a small removable central arena (diameter 65mm –
In order to determine whether woodlice selected one shelter preferentially, binomial tests were carried out with H0 assuming an equal distribution of woodlice between both shelters. After this binomial test, it is possible to define the “
X² test was used to compare the proportion of experiments with choice of one shelter according to density.
Since our data did not meet conditions for parametric tests, comparisons of results obtained with different densities were carried out with a Kruskal-Wallis test followed, if necessary, by a Dunn’s test.
GraphPad software InStat 3 was used to carry out the statistical tests.
First of all, only one of all 88 replicates did not show any aggregation during the 45 minutes of observation. Hence, this replicate was not considered. In all the other replicates, regardless of the density, nearly 90% of woodlice were observed to aggregate under shelters after 45 minutes (
Proportion of aggregated woodlice and proportion of woodlice under shelters or outside shelters at the end of experiments.
|
|
|
|
|
40 woodlice | 88.2 (± 7.1) | 87.1 | 12.9 | 29 |
60 woodlice | 87.4 (± 7.5) | 87.4 | 12.6 | 20 |
80 woodlice | 88.4 (± 7.0) | 88.4 | 11.6 | 20 |
100 woodlice | 89.1 (± 5.6) | 89.1 | 10.9 | 18 |
Experiments showed that groups of woodlice generally selected one of both shelters (
In order to understand the aggregation dynamics, separate analyses of replicates with a clear choice of one single shelter (77%, N=87) and replicates where isopods reparted almost equally among the two shelters, i.e. no selection of one shelter (23%) were necessary. However, the number of replicates without choice was low and were evenly distributed among the four densities tested (
Woodlice showed a strong thigmotactic behaviour; just after their release, woodlice walked in the arena, generally near the edge and they quickly entered under both shelters (
The densities used in this study do not impact the aggregation process. Indeed, no differences were observed between density conditions in the dynamics of aggregation, the collective choices and the rates of selection of only one shelter. Aggregation in woodlice is very frequent (
In more than 77% of experiments, a choice of one shelter was made by groups of woodlice. In such experimental conditions, these selections can only be explained by the social interactions between congeners (
Concurrently to the absence of an enhanced selection rate, our results did not show the expected acceleration of aggregation dynamics, in conjunction with higher density. Indeed, this phenomenon being driven, in part, by social interactions between congeners, aggregation in a preferred shelter should happen faster at higher densities. In our experiences, whatever the number of woodlice, the aggregation was very quick (in less than 10 minutes most of woodlice are aggregated) and did not differ between densities. Possibly, the aggregation process was already very quick even at our lowest density used (1325 individuals/m² corresponding to 40 individuals) so that the phenomenon could not happen any faster. The density used in these experiments corresponded to the high values observed in nature (
The number of woodlice aggregated under the “winning” shelter increased with the number of woodlice within the setup. However, from a number of 80 woodlice in the set-up, the number of woodlice under the “winning” shelter reaches a plateau around 70 woodlice (no difference was found between 80 and 100 woodlice set-ups). This result firstly suggests a saturation of shelters at 70 woodlice. This may result from the shelter carrying capacity. Nevertheless, a stable aggregation under a shelter whatever the density, often extended beyond the edge of that shelter. As a consequence, some woodlice belonging to the aggregation were not in the darker area. Keeping in mind that at the 100 woodlice condition, a second stable aggregate grows under the losing shelter, two “functional” hypotheses, deserving new investigations, can explain this maximal number of woodlice in an aggregate. Firstly, it is possible that competition in the aggregate increases with the number of woodlice and beyond 70 woodlice, it could be better for a woodlouse to join a smaller aggregate (
These results were in accordance with the existence of aggregation pheromone coming from faeces suggested by the past (Kuenen and Noteboom 1963,
In the field, in woodlice and most of the organisms, the local population densities depend on characteristics of their environment (litter,
Moreover, a complete understanding of the woodlice aggregation and its characteristics needs a theoretical approach of the costs and benefits of the aggregation in order to evaluate the differences for woodlice between optimal and stable sizes of clusters (
Social interactions in woodlice and different environmental parameters (such as maximum carrying capacity of shelters or maximum size of aggregates) are important to understand the distribution of woodlice in the environment. In natural conditions, a local peak of population (in case of binary choice, the population is higher on one side) may result from the coupling between the response to the environmental heterogeneities and the social interaction. Moreover, even if more investigations are necessary to decipher the mechanisms explaining, the velocity of gathering in aggregates, the maximum size of clusters and the social signals used we suggest that similar observations could be made now in field.
Since woodlice are often used as bioindicators for pollution, the explanation of the collective decision making and patterns of aggregation of woodlice population could inform us about quality of environment (
P. Broly is supported by a FRIA grant (Fonds pour la Recherche dans l’Industrie et dans l’Agriculture). J.-L. Deneubourg is Senior Research Associate from the F.R.S.-FNRS. The authors warmly thank Mark Pritchard and Fiona Houghton for English corrections as well as Martin Thiel, Ivan Tuf and two anonymous referees for their helpful comments and suggestions. This study has been carried out with support from the Catholic University of Lille in the framework of the fund-raising campaign «Ensemble Innovons». This financial help generally allows the authors to buy small materials and pay students. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was partly supported by the grant “Action de Recherche Concertée”: Individual and collective issues in dispersal and aggregation: from proximal causes to ultimate consequences at contrasting scales.