The circadian calling activity of a lebinthine cricket with high-frequency calls is unaffected by cicada choruses in the day

Introduction

Acoustic signalling for mate attraction and territory defence is common across animals, and signals often follow predictable circadian rhythms that reflect the ecology of each species (Schmidt & Balakrishnan, 2015; Rivas, da Bauzer & Meireles-Filho, 2016). Many factors can influence the circadian rhythm of acoustic signalling, such as changes in light and temperature (Clink et al., 2021; Feng & Bass, 2016), endogenous time-keeping mechanisms (Fergus & Shaw, 2013), social cues from conspecifics (Favreau et al., 2009; Fuchikawa et al., 2016), the presence of other acoustically signalling species (Hart et al., 2021), or a combination of these factors. When co-occurring species have similar circadian rhythms, the Acoustic Niche Hypothesis (ANH) predicts that selection will favour differences in signalling behaviour to help reduce errors in recognition and avoid acoustic interference (Krause, 1993; Farina, 2014). These differences may include producing sounds at different frequencies (Krishnan, 2019; Allen-Ankins & Schwarzkopf, 2022) or avoiding temporal overlap on short time scales (Hart et al., 2021). Community-wide partitioning of acoustic space in time and frequency suggests long-term adaptation to local acoustic environment (Farina, 2014). Little is known, however, about whether similar acoustic signalling circadian rhythms between species are due to a common response to abiotic factors, such as light and temperature, or could be synchronised by direct acoustic interactions between co-occurring species.

Tropical forests, due to their high biodiversity, tend to have more severe competition for acoustic niche-space than temperate habitats (Hart et al., 2021). In many Southeast Asian forests, the choruses of crickets and cicadas are among the most ubiquitous and conspicuous sounds. Ground crickets and cicadas in these forests tend to call at similar low frequencies (peak frequency <10 kHz) (Riede, 1996, 1997). These choruses are usually separated in time, with crickets being mainly nocturnal and cicadas mainly diurnal. There are, however, exceptions to these patterns in song carrier frequency and circadian rhythm. Cricket species in the tribe Lebinthini (Gryllidae: Eneopterinae) differ from other crickets by producing high-frequency (some even ultrasonic) calls (e.g., Robillard & Desutter-Grandcolas, 2004; Tan et al., 2021). Based on 24-h recordings in the laboratory, many of these lebinthine crickets appear to be diurnal or have complex acoustic circadian rhythms that peak during the day and night (Tan & Robillard, 2021).

Unlike in their natural environment, crickets recorded in isolation in the laboratory are not affected by acoustic competitors (both conspecific and heterospecific) or eavesdropping predators. The patterns observed in artificial conditions are thus more likely to reflect the intrinsic circadian clock (Tan & Robillard, 2021) or abiotic factors. Therefore, by comparing patterns observed in the field and laboratory, we can screen for potential biotic factors, such as acoustic competitors, influencing the circadian rhythm in the natural environment. Although Riede (1996, 1997) investigated the acoustics interactions of crickets and cicadas in Southeast Asia, studies combining laboratory and field data have not been undertaken for eneopterines and other Southeast Asian crickets.

In this study, we focus on the lebinthine cricket species Lebinthus luae Robillard & Tan, 2013 from Singapore, which produces broad-spectrum calls with a peak frequency of 18 kHz (Robillard & Tan, 2013) and exhibits two diurnal and one nocturnal peaks in calling activity in the laboratory (Tan & Robillard, 2021). In forests where L. luae are found, other singing insects exist that could potentially interfere with crickets’ ability to communicate. At least two cicada species commence singing around sunrise, generating a chorus that coincides with the increase in calling activity of L. luae in the laboratory. There are also low-frequency calling cricket and katydid species that have been described as largely nocturnal. We had two goals for this study. First, we tested whether the circadian rhythm in L. luae calling activity differs significantly between data collected in the field and in the laboratory. Second, we investigated potential overlap in the timing and frequency of L. luae calls with cicada choruses (i.e., ensembles of cicadas calling at the same time) and calls of other species of crickets and katydids in the same localities, particularly focusing on the sunrise period.

Materials and Methods

Acoustic recordings in the field

Lebinthus luae occurs only in a few isolated patches of forests in Singapore, usually close to the coast (e.g., Tan, 2010, 2012, 2013; Tan, Ngiam & Ismail, 2012). To monitor the circadian rhythm of the calling activity of L. luae in Singapore, four isolated localities were studied: secondary hillside forest in Hindhede Nature Park (N1.3485, E103.7755), coastal forests in Labrador Nature Reserve (N1.2659, E103.8032) and Sentosa Island (i.e., Imbiah Nature Trail) (N1.2570, E103.8138) and secondary forest in the southern parts of Ubin Island (N1.4089, E103.9723) (Figs. 1A and 1B). Sampling of the four localities was limited to the inter-monsoon period between mid-September to early November 2021 to minimise the effect of varying rainfall and temperatures between the different recording periods and the effect of heavy rain during the monsoon period on the recordings. Permission for collecting material and setting up in-situ recorders was granted by the National Parks Board [NP/RP18-064-2a], Singapore.

Wildlife Acoustics Song Meter (SM) Mini acoustic recorders were attached to tree trunks at knee height and programmed to record continuously for 48 h with a sampling frequency of 96 kilo-samples per second and channel gain of 18 dB (16 bit) (Fig. 1C). The microphone was placed at around 45° downward to better sample the forest floor where the crickets dwell. Two recorders were programmed to record two different parts of the site simultaneously at each time. The two recorders were placed at least 20 m apart to ensure that they were not recording the same individual cricket(s) at the same time.

After 48 h (hence two consecutive circadian cycles), sound data were retrieved from both recorders and recordings were repeated in two other parts of the locality for the subsequent 48 h. Recorders were never placed in the same position within each locality. In total, six repeats of 48-h recordings were done for each recorder. Hence, a total of 24 circadian cycles of sound data were obtained (six repeats × two circadian cycles per repeat × two recorders) for each locality. After accounting for failed recordings (e.g., incorrect settings, problems with SD cards), 88 circadian cycles were analysed and used to feed the models. Sunrise and sunset times for each day were obtained from Meteorological Service Singapore (2022). The ambient temperature was recorded using a HOBO 8K Pendant® Temperature logger (model: UA-001-08; Onset, Bourne, MA, USA): 28.9 ± 1.6 °C.

Results

Over the circadian cycle, the model (adjusted R² = 0.36, n = 81 circadian cycles, Table 1) did not have a significant interaction between time of day (HR) and data source (F-value = 0.0, P-value = 0.640), indicating that calling activity circadian rhythm does not differ significantly between data collected in the field and in the laboratory. In both datasets, three peaks were observed (Fig. 1D). The highest peak was in the morning (around 08:00), a smaller peak occurred in the afternoon (around 15:00), and another small peak was seen at night. For the field data (n = 44 circadian cycles), an average of 20 ± 1 echemes were detected in the morning peak (08:00), 7 ± 0 echemes in the afternoon peak (15:00) and 1 ± 0 echeme in the night peak (23:00 and 0:00). For laboratory data (n = 37 circadian cycles), an average of 7 ± 0 echemes per individual were detected in the morning peak (07:00), 5 ± 0 echemes per individual in the afternoon peak (14:00) and 5 ± 0 echemes per individual in the night peak.

Around the sunrise period, the model (adjusted R² = 0.53, n = 73 circadian cycles, Table 1) indicated a moderate but significant interaction between time of day and data source (F-value = 0.054, P-value = 0.016), indicating that calling activity around sunrise slightly differs between data collected in the field and in the laboratory (Fig. 2E). This can be attributed to the earlier morning peak observed in laboratory than in field data, as indicated by the line in Fig. 1D.

Our field recordings contained cicada choruses starting in the morning (Fig. 2A) with calls of Purana sp. (Fig. 2D) commencing on average at 06:52 (SD = 13 min, earliest = 06:34, latest = 07:39) and calling throughout the day. Additionally an unidentified species that we call “cicada ^∧” sang for an average of 12 ± 5 min at dawn, on average between 06:45 (SD = 6 min, earliest = 06:37, latest = 07:03) and 07:01 (SD = 7 min, earliest = 06:49, latest = 07.19) (Fig. 2C). The cicada choruses produced almost continuous sound due to their long calls and multiple individuals singing at the same time (Fig. 2A). Therefore, L. luae calls were almost always overlapping with cicada calls even on short time scales (Fig. 2E). Unlike L. luae, other sympatric cricket species (e.g., Gymnogryllus spp., Velarifictorus aspersus, and Ornebius spp.) call at lower frequencies (3–8 kHz) and overlap with the first and second harmonics of the cicada choruses (Fig. 2B). These crickets generally stop calling around the same time as the cicada choruses start in the morning (on average 07:08, earliest = 05:21, latest = 08:27) (Table S1). In contrast, the calling activity of L. luae increased during this period (adjusted R² = 0.87, n = 72 circadian cycles, Fig. 2B, Table 1).

The dominant frequencies of the cicada species did not overlap with the dominant frequency of L. luae. For “cicada ^” and Purana sp., the fourth harmonics (15.1 ± 0.3 kHz and 12.0 ± 0.2 kHz, respectively) have the highest energy that are also closest to the spectrum of L. luae calls. The first to third harmonics of these cicadas are below 10 kHz. Following a Kruskal-Wallis rank sum test (χ² value = 85.2, d.f. = 2, P-value < 0.001, n = 29 “cicada ^∧”, 33 Purana sp., 35 L. luae), pairwise Mann–Whitney U-tests showed that (all pairwise P-values < 0.001 after Bonferroni correction) the dominant frequency of L. luae does not overlap significantly with those of the fourth harmonics of the two cicada choruses (Fig. 2G).

Discussion

We found that the circadian rhythm of calling activity does not generally differ significantly between crickets recorded in the field or in the lab. This result validates the idea that recording eneopterine crickets in isolation under artificial laboratory conditions can still offer meaningful information about their natural circadian rhythms. This also suggests that the observed pattern is predominantly influenced by the individual’s intrinsic biological clock (Fergus & Shaw, 2013) and abiotic environmental cues such as light (Clink et al., 2021; Feng & Bass, 2016), rather than social cues from conspecifics (Favreau et al., 2009; Fuchikawa et al., 2016) or the presence of other acoustically signalling species (Hart et al., 2021).

The nocturnal peak, however, appears to be marginally greater in the laboratory than in the field data. We postulate two plausible explanations for this difference. First, in the wild, L. luae have to contend with predation risks, such as nocturnal palm civets (Fung, Tan & Sivasothi, 2018) and reptiles (M. K. Tan, 2021, personal observations), and they might reduce calling activity in response to predator cues (e.g., ter Hofstede & Fullard, 2008). Second, couples might form during daytime calling with mating occurring more at night (M. K. Tan, 2021, personal observations), but under isolated laboratory conditions in the absence of females, males continue to sing. These hypotheses could be tested using playbacks of potential predator sounds in the presence or absence of females.

The second main finding of this study is that L. luae calls alongside the daytime cicada chorus, unlike low-frequency calling crickets that partition their calling activity with cicadas by calling predominately after dusk when the cicada choruses have stopped (Riede, 1996, 1997). By comparing circadian rhythms of calling in the laboratory and field, our results suggest that daytime calling activity by L. luae is not drastically affected by cicada choruses. Frequency partitioning, as predicted by ANH, may have allowed L. luae, with high-frequency calls (peak frequency = 18 kHz), to call alongside cicada choruses of slightly lower frequencies (maximum frequency of cicada choruses are 16 kHz) (Fig. 2). By avoiding frequency interference with cicadas, L. luae may still communicate acoustically with conspecifics during the day without having to segregate their calling time.

The divergence and diversification of the crickets from the tribe Lebinthini have been attributed to the clade’s novel communication system, including shifting to calls with high or ultrasonic frequencies (ter Hofstede et al., 2015; Benavides-Lopez, Ter Hofstede & Robillard, 2020; Tan et al., 2021). This divergence also coincided with the origin of echolocation in bat ancestors, implying that the novel communication system in the lebinthines can be a form of adaptation to avoid detection and predation by bats at night (Hill et al., 2021; Tan et al., 2021). Based on our study, we may offer another plausible hypothesis, that is, lebinthines can avoid frequency jamming with cicada choruses in the day owing to the shift to high-frequency calls in accordance with ANH.

Conclusions

Here, we show that the circadian rhythm of L. luae calling activity observed in the laboratory is representative of that in the natural environment. This suggests that the individual’s intrinsic biological clock and environmental cues play an important role in the circadian patterns, more so than influences from other acoustically signalling animals. This can be further supported by our observations that the calling activity of L. luae does not appear to be affected by the commencement of cicada choruses after sunrise. These crickets and cicadas can call at the same time because the dominant frequency of the L. luae call does not overlap with the maximum frequency of the cicada calls, and hence the crickets can still communicate acoustically without being masked by the cicada choruses. Further investigations into the circadian rhythm of lebinthine crickets and cicadas, including playback experiments and comparative studies that account for phylogeny, can offer more direct evidence for ANH in this system and further insights into the interactions.

Supplemental Information