Kerivoula hardwickii’s roost choice

In each study site we monitored the occurring K. hardwickii individuals for 30.0 ± 18.3 days (mean ± s.d.) by daily checking all potential roosts (furled leaves and Nepenthes pitchers below a height of 2.5 m) and additionally by catching individuals with harp traps. We radio-tracked on average 5.5 ± 3.8 (range: 0–12) individuals per site. Parts of the radio-tracking data have already been published^11,12. Additionally, individuals could easily be identified from outside the roost with a handheld PIT-tag reader (LID-575 Midrange Reader, Trovan, UK). Of special interest to us was study site “Airport” where bats not only use pitchers of the species N. hemsleyana and N. bicalcarata but also furled leaves (M. muluensis, Zingiber kelabitianum, Plagiostachys albiflora, Plagiostachys strobilifera) as roost. In this study site we radio-tracked three K. hardwickii individuals (two males, one female) from furled leaves and one male individual from a pitcher for an overall mean of 8.50 ± 2.87 days.

In a flight arena (length and width 3.5 m, height 2.5 m, Fig. 1) we conducted three types of behavioural experiments where we investigated if bats generally prefer a certain plant species, if they prefer the species in which we had found them or if they randomly choose between different plant species.

In all experiments, we randomly arranged pitchers and furled leaves within the flight arena (distance to each other = 0.5 m; height = 1.5 m). To prevent the plants from excessive damage by cutting pitchers and leaves, we offered the same experimental leaf/pitcher to up to three bats. We tested each bat only once per experiment and released them within 24 hours of capture into their original habitat. Before and after the experiment we fed the bats with mealworms and offered them water ad libitum. We excluded pregnant and lactating females as well as juveniles.

We filmed (Sony HDR-CX560VE) all experiments to determine how the bats react to different roost types (pitchers or furled leaves) or plant species, how often they approach to the different potential roosts and which of them they finally choose. We defined an approach as hovering flight in front of an object within a distance of 10 cm. Three bats in the first, one bat in the second and twelve bats in the last experiment did not choose a roost within the maximum time span of 30 min per trial and thus were excluded from the analyses of the bats’ final roost selection.

First, we simultaneously offered one N. hemsleyana, one N. bicalcarata, and one N. ampullaria pitcher, and additionally a pitcher of N. rafflesiana, which is not used by the bats, as well as a plastic tube (Fig. 2a). We tested 41 bats (12 males, 29 females) from areas where K. hardwickii only used pitchers as roosts, although furled leaves were available. Sixteen individuals derived from an area where the bats exclusively use N. hemsleyana pitchers, the other bats were captured at study sites where the bats exclusively roost in N. bicalcarata pitchers or where they use both pitcher plant species (see Table ^S1 for roost availabilities per plant species and site). In the latter case, we only tested individuals that exclusively roosted in N. bicalcarata pitchers during a radio-tracking period of 5–13 days (9.82 ± 2.64 days; for details see¹²).

Similarly, in a second experiment we tested 14 bats (10 males, 4 females) that had roosted in furled leaves (Fig. 2b). For the experiment we simultaneously offered a total of three furled leaves, one of each species: A. ligulata, B. grandis, and M. muluensis.

Finally, we tested how 68 bats (56 males, 12 females) reacted to different roost types (pitchers versus furled leaves). We offered one N. hemsleyana, one N. bicalcarata, and one N. ampullaria pitcher, one furled leaf (A. ligulata, B. grandis or M. muluensis; for the 47 bats that we had found in furled leaves we used the respective roost plant species; in the case of the 21 bats found in pitchers we used the plant species that occurred within a distance of 20 m from the roost) as well as the plastic tube as roost (Fig. 2c).