Food quantity and the intensity of the alarm signal combine to modulate the resource selection in a termite species

Maximizing food intake while minimizing risk is an important trade-off in the foraging behavior of most animals. In general, foragers are vulnerable and the ability to trade off benefits (food quantity) against costs (risk of being killed) may provide a considerable ecological advantage. Despite the increasing number of studies, the effects of food quantity and mortality risk signals on resource selection in eusocial insect is not well understood. Here, we investigated the combination of distinct levels of food quantity and the intensity of alarm signal on resource selection of a Neotropical termite, Nasutitermes corniger (Motschulsky) (Termitidae: Nasutitermitinae). Manipulative bioassays with binary and multiple choices were conducted over time to check the recruitment of termite groups among resources containing different levels of food quantity and alarm signals. Overall, our results showed that regardless of food quantity, termites avoid a food source if there is even a small amount of alarm signal. This work contributes to a better understanding of habitat use by termite species. Furthermore, it shows for the first time the combined effects of food quantity and alarm signals on the resource selection of an important ecological and economic termite species. 

Methods
Binary and multiple choice bioassays were conducted in experimental arenas under laboratory conditions following procedures described in Silva et al. (2021) with modifications. Experimental arenas consisted of a plastic pot (250 mL) with an internal bottom surface lined with filter paper.

In the binary choice bioassays, observations consisted of a central plastic pot connected in a straight line with a transparent tube (7 mm diameter x 5 cm of length) to two other plastic pots at the end containing the treatments (see Figure 1). In this bioassay, N. corniger groups were given access to different combinations of all treatments (LR+HAS, LR+LAS, HR+HAS, and HR+LAS) determined by a complete factorial as previously detailed above (see section “Manipulation of food resource quantity and levels of alarm signals”). For each combination of treatments (N= 6), three repetitions per nest (N= 8) were performed, totaling 144 experimental units. Control of binary choice bioassays consisted of the choice of termites between treatments versus the control solvent (sugarcane baits with only hexane; SC). For each treatment (N= 4), three repetitions per nest (N= 8) were performed, totaling 69 control bioassays. In addition, to test whether N. corniger groups may be able to distinguish arenas with and without resources, positive control was conducted with arenas containing sugarcane baits (with low and high resource quantity) versus empty arenas. The amount of food resources used in the controls was the same used in the treatments (i.e. low resource quantity [5 g] and high resource quantity [25 g]). For each treatment (N= 2), three repetitions per nest (N= 8) were performed, totaling 48 positive control bioassays. A total of 261 arenas were conducted in the binary choice bioassays. Figure 1 shows a scheme of all binary bioassays conducted.

Bioassays with multiple choices were conducted in arenas consisting of a central plastic pot connected with a transparent tube (7 mm diameter x 5 cm of length) to four other plastic pots at the end containing the treatments tested (see Figure 1). Therefore, bioassays aimed to check the choice of termite groups for all treatments tested. Three repetitions per nest (N= 9) were performed, totaling 27 experimental units. A second bioassay with multiple choices was conducted including all treatments tested and an option without food resources (empty arena) (see Figure 1). This second bioassay was conducted in arenas consisting of a central plastic pot connected with a transparent tube (7 mm diameter x 5 cm of length) to five other plastic pots at the end containing the treatments tested. Three repetitions per nest (N= 6) were performed, totaling 18 experimental units. A total of 45 arenas were conducted in the multiple-choice bioassays. Bioassays with multiple choices were designed to test the preference of termites for all treatments simultaneously.

Termite groups (four soldiers and 16 workers) were confined into the central plastic pot of experimental arenas (in both binary and multiple-choice bioassays). The choice of the caste ratio and termite numbers in each experimental arena was a compromise between natural caste proportions (soldier: worker=3:7; Thorne 1985) and optimal density to maximize inter-individual contacts and survival (Miramontes and DeSouza 1996). The number of termites, in the vicinity of the food sources, in the pots with different treatments was recorded over 48 h. Evaluations were conducted at one, two, four, eight, 24, and 48 hours after the beginning of bioassays. Individuals were used only once in each repetition, and they were randomly selected from each colony.

All bioassays were conducted in the laboratory under controlled conditions (25 ºC, 70% r.h.).