Introduction

Understanding animals’ movements is key to get a better comprehension of the species’ behavioural ecology (Horne et al. 2020). The home range of an individual encompasses the space where the individual conducts its daily activities, while the territory is the area within the home range that is defended (Burt 1943). By tracking animals’ movement, we can obtain information about their habitat preferences, as well as behaviour and mating system. For amphibians, the most endangered group of organisms with 40.7% species under threat (Luedtke et al. 2023), understanding their movement patterns also is key for developing conservation measures as the spatial requirements of a species is needed to determine suitable areas for the species survival.

In dendrobatids, home range behaviours over resources associated to reproductive success has been widely reported in genera like Ameerega, Allobates and Oophaga (Roithmair 1994; van Wijngaarden and van Gool 1994; Lima et al. 2002; Méndez-Narváez and Amézquita 2014; Neu et al. 2016). Understanding their home range behaviour has enhanced our comprehension on the habitat requirements and the resources that individuals defend, as well as their parental care strategies (Werner et al. 2012). In species with male parental care, males’ home ranges are highly associated to the distribution of reproductive resources, with higher densities of individuals in areas with higher number of resources (Pröhl and Berke 2001; Pröhl 2005; Poelman and Dicke 2008; Pašukonis et al. 2022).

The dendrobatid Phyllobates vittatus (Cope 1893), commonly known as the Golfo Dulce Poison Frog, is an endemic species of the southern Pacific rainforest of Costa Rica (Savage 2002; Garrido-Priego et al. 2023). Males of P. vittatus care for their clutches by hydrating them and, once they hatched, they transport the tadpoles on their back until they deposit them on a water pool in a small stream (Silverstone 1976). To date, there has only been one study focusing on the home range behaviour of the species, performed in a population located in Corcovado National Park (Summers 2000). This author suggested that P. vittatus males’ home range encompassed an average area of 6.0 (±7) m2 and observed male-male competitive interactions. The home range was estimated by plotting the capture-recapture sightings of different individuals on a graph paper with a grid and was analysed using the Minimum Convex Polygon method (herein, MCP). According to Summers (2000), males of P. vittatus might have higher chances of encountering neighbours, potentially resulting in escalated aggressive behaviours. Given this perspective, we would expect a high occurrence of overlapping home ranges among males and aggression encounters between conspecifics.

Our aim was to determine males’ home range size and the degree of overlap between individuals. For this, we used the harmonic direction finder (HDF) technique, which facilitates to track small amphibians thanks to the miniaturization of transmitters (Langkilde and Alford 2002; Rowley and Alford 2007; Pašukonis et al. 2014). Since HDF is highly directional (Langkilde and Alford 2002), it allows location of individuals equipped with these tags in close range. Hence, we expect to detect individuals more easily and have a higher recapture rate using HDF than with a more traditional capture-recapture method. This way we can obtain a more comprehensive baseline of the individuals’ locations to determine (1) males home range size and (2) the degree of overlap between individuals of P. vittatus. Furthermore, we also report courtship and aggressive behaviours observed in the field, which shed new light on the natural history of the species and brings up questions about P. vittatus’ males territorial behaviour and females social interactions with males carrying tadpoles.

Materials and methods

Study species

Phyllobates vittatus is small diurnal poison frog (average snout-to-vent length, SVL = 2.44 ± 0.09 cm) that lives alongside streams in primary lowland rainforest, from 20 to 550 m a.s.l. (Savage 2002). This species is catalogued as vulnerable by the IUCN (IUCN SSC Amphibian Specialist Group 2020), but recent studies suggest a revision on its status due to potential effects of habitat fragmentation and pesticides applied for agriculture (Garrido-Priego et al. 2023). Like other dendrobatid species, males of P. vittatus are territorial; they fight for their territory wrestling and pressing their body over the opponent (Leenders 2016; Summers 2000). Males also defend their territory by producing a high-pitched long call (Silverstone 1976; Summers 2000). Early morning duets of males can be heard in the forest (Duellman 1967). Males can be seen calling on top of rocks or roots to attract females (Savage 2002). Both males and females are active during courtship, which can last up to seven days (Savage 2002; Silverstone 1976). There is no amplexus in P. vittatus (Savage 2002). Clutches, typically comprising 7–21 eggs, are deposited in leaf litter basins, in tree holes or within small cavities on the ground (Silverstone 1976; Leenders 2016; Savage 2002). Males provide parental care attending the clutch, once the eggs hatch, males carry the tadpoles on their backs through the rainforest to small puddles near narrow streams (Raby Nuñez, pers. comm.; Silverstone 1976).

Study area

The study areas was at the Osa Verde Biological Station (known locally as Piro; 8.40388 N, 83.33661 W; 25 m elevation), located in the south of the Osa Peninsula, Costa Rica. This is a private and protected land where, within its 1330 ha, there is old-growth forest, secondary forest, and grassland (Whitworth et al. 2019). Fieldwork was carried out between 2018 and 2019 during Costa Rica’s rainy season, from June to September. We studied a population of P. vittatus from the same location that had previously been investigated for skin alkaloid toxicity (see Protti-Sánchez et al. 2019).

Home range data collection

We began surveys searching for calling males early each morning (05:00–10:00). Once a calling male was found, it was collected, measured its SVL to the nearest 0.5 mm with callipers, and weighed to the nearest 0.01 g on an electronic portable scale. If individuals weighed enough (> 2 g) to be able to carry the transmitter, we equipped them with the tracking tags. We used the handheld detector R9 by RECCO (RECCO R9 Handheld Detector, www.recco.com), and the Schottky diodes from the same company to build the tags (Langkilde & Alford 2002; Pašukonis et al. 2014). A 13 cm antenna was welded to the cathode of the diode and another 3 cm antenna to the anode (Gourret et al. 2011). To avoid damaging individuals, we wrapped the diode in silicone. The reflector tags built weighted an average of 0.11–0.01 g, less than 10% of the average body mass of the individuals tracked (2.49–0.45 g), as recommended by Richards et al. (1994). We used clear bead cord (0.8 mm), thread seal tape and cotton thread to build the belts. The cotton thread was used so it secured the belt to the transmitter and, in case the individual was not found, the cotton deteriorates and frees the frog from the transmitter (Langkilde and Alford 2002). To prevent the transmitter from turning and disturbing and/or hurting individuals, we used the thread seal tape to secure the belt by making a loop around the back legs. The thread seal tape did not cause any skin abrasion to any individual and we also left it a little bit loose so it would not feel too tight. Males carrying tadpoles were not tracked on this study since did not want to disturb parental care and maybe influence tadpole mortality.

Once the individual was released, we marked the releasing location on a GPSmap 64 s (Garmin International, Inc.) and used flagging tape as reference. We followed these individuals for two entire days, allowing them to adapt to the transmitters (Rowley and Alford 2007). If we observed that the tracked individuals showed abnormal behaviour (ex. not moving from the same site where we released them after attaching the transmitter), severely reduced mobility or any kind of skin damage, we removed the transmitter immediately. We evaluated whether there was any skin damage by recapturing the individuals after the first day and checking their skin in contact with the belt. We found three individuals with skin damage while we were testing for different materials to attach the transmitter belt to the legs, until we decided to try the thread seal tape around their legs, so the belt was fixed. After the acclimating period, we started recording the location of individuals every hour for five days. Locations were recorded for nine hours (05:00–11:00/14:00–18:00), using a laser measuring device (Bosch Blaze One) and a precision compass (Suunto Tandem) (Pašukonis et al. 2018). To avoid disturbing individuals while recording their locations, we first measured distance and bearing from our position to the releasing point marked with a flagging tape, and then, distance and bearing from our position to the individual’s location.

Space use analysis

To obtain the coordinates of the individuals tracked from the data collected in the field, we used the R package geosphere (Hijmans 2019). Using this package, we calculated each individual’s geodesic coordinates when at the initial position, and estimated its bearing and distance travelled (Karney 2013). Home range of each individual was calculated using the method of the minimum convex polygon (MCP; Mohr 1947) using the function mcp of the package adehabitatHR (Calenge 2006). We used this technique since MCP is a common method used for amphibians and hence allows for comparisons among species. Kernel density estimates, on the other hand, are measured differently across different studies and there is no standardisation. Nevertheless, aiming to provide a better insight of the space use of the individual (Silverman 1986), we also calculated 50% (k50) kernel density estimates (KDE) using the least-squared cross-validation method with the function kernelUD of the package adehabitatHR. We used this method to complement the MCP method. Overlapped of home ranges was calculated using the ‘intersect’ tool in QGIS.

In order to expand the knowledge of the natural history of the species, we recorded behaviours that occurred while we tracked P. vittatus’ males. Using the tracking data, we took note of the activity patterns and the substrate from which males would call to attract females. Moreover, we also recorded social interactions in untracked individuals.. The description these interactions are described below.

Results

Home range estimation

Sixteen individuals were tracked from August to September 2019. On average, individuals were observed 55 (± 5.9) times for 4–5 days. Tracked males were hidden during periods of elevated temperatures (approx. 11:00 until 13:00–14:00) and after sunset. The calling peaks were usually between 6:00 and 09:00 and from 16:00 until 18:00. Some males were found in shelters like holes under tree roots, crevices under fallen trunks or rocks. Males would call on top of roots and/or fallen trunks. On average, individuals showed a MCP area of 55.7 (±10.4) m2 ranging from 36.95 to 67.64m2 (Fig. 1a). Five individuals overlapped in their MCP areas, one of these individuals having more than 50% of its MCP area overlapping with that of other three individuals (Fig. 1b). On average, k50 area was 26.75(± 7.31) m2 ranging from 12.7 to 39.25 m2 (Fig. 1a). The k50 area indicated that two males had overlapping areas, which covered an area of less than 0.5 m2 (Fig. 1b).

Fig. 1
figure 1

a Map of the distribution of the home ranges of the different individuals of Phyllobates vittatus in the study area in the South Pacific of Costa Rica (1: 500 scale and Coordinate Reference System (CRS) WGS84). b Zoom in of the individuals with their MCP and kernel areas overlapped within the area of heliconias (1: 210 scale and Coordinate Reference System (CRS) WGS84)

Full size image

Aggressive behaviours

On 30th of August of 2019, at 10:04, we observed two males with tadpoles on their backs, close to a big burrow on the ground. Both males jumped several times over each other (Video 1). If one male was getting into the burrow, the other male jumped in, forcing the other male out of the burrow. After more than one hour fighting, both males went into the burrow, and shared it.

Defensive behaviours

On 11th of May of 2018, at 07:54 h, we observed a male and a female in courtship (Video 2). The female jumped on top of the male, with the male’s snout touching the female’s cloaca. The male kept emitting calls while the female was on top. A few minutes later, both separated and the male went away. We recorded another male–female interaction at 10:34 h. During a playback reproduction, a male with tadpoles on the back appeared and jumped inside a burrow, followed by the female. The male started calling, while the female placed herself behind the male’s back, covering the tadpoles (Video 3, Fig. 2). Few seconds later, the male with tadpoles went deeper in a burrow and another male without tadpoles went inside the same burrow (Fig. 2b). At 10:57 h, after touching each other’s snout (Fig. 2c Video 4), the male without tadpoles left the burrow and started calling. The female jumped nearby and went away, followed closely by the male without tadpoles.

Fig. 2
figure 2

a Phyllobates vittatus female behind male with tadpoles, both hidden in the shelter. b Male with tadpoles and female inside the shelter and male without tadpoles in front of it. c Male without tadpoles interacting with a male with tadpoles, this one still hidden in the shelter

Full size image

Discussion

In this study, we investigated the home range of P. vittatus males and its overlap. Results of the MCP indicates that the home range of males of P. vittatus had an average area of 55.7 m2 (range 36.95–67.64 m2) and a k50 area of 26.75 m2 (range 12.7–39.25 m2). We found five individuals with an overlapping MCP area, and two of them with overlapping k50 area. While no aggressive behaviours were observed between the males tracked, we observed an aggression event between two males that were carrying tadpoles. We also recorded a possible defensive behaviour from a female in response to playback.

Home range estimation

The discrepancy between our results (X̄ = 55 m2) and Summers (2000: X̄ = 6 m2) on the area estimates for P. vittatus, is probably due to the different methods of data collection. While in Summers (2000) searched for individuals by scouting the study area and recording the individuals sighted, we used HDF technique, which allowed us to locate individuals tracked even when they were not exposed. When comparing our estimates with those obtained in other studies with similar tracking techniques, our data seems more similar to those of the poison frogs Allobates femoralis (MCP: X̄ = 595.3 ± 520.41 m2, range 54.57–1842.75 m2; k50: X̄ = 19.75 ± 11.9 m2, range 5.27–45.94 m2), Ameerega trivittata (MCP: X̄ = 96.33 ± 93.69 m2, range 32.53–286.76 m2; k50: X̄ = 70.47 ± 22.61 m2, range 51.32–135.03 m2), Dendrobates tinctorius (MCP: X̄ = 680.58 ± 581.58 m2, range 197.28–1787.45 m2; k50: X̄ = 41.51 ± 44.53 m2, range 7.62–39.45 m2) (Neu et al. 2016; Pašukonis et al. 2022). We must note that the difference in magnitude of the MCP estimates between our results and those from these three species may be because in our study we did not tracked individuals carrying tadpoles. Males with tadpoles on the back expand their home ranges in order to find suitable water pools for tadpole deposition (Pašukonis et al. 2019; Fouilloux et al. 2021). These three species have similar behaviours to that of P. vittatus, as they all present male parental care by transporting tadpoles on their backs, males are highly territorial, and present aggressive behaviour against conspecifics. Similar behaviours are observed in other dendrobatids with paternal care (Pröhl 2005; Poelman and Dicke 2008).

We found that five individuals had overlapping MCP areas, with one of them having more than 50% of its MCP area overlapping with that of other three individuals (Fig. 1). This could be due to the spatial patterns of resource availability, as the males that had overlapping home ranges were located on an area where there were many heliconias and debris that offered numerous shelters and calling sites (pers. Obs. by first author). These crevices are key, as P. vittatus deposits its eggs on the leaf litter within small cavities on the ground (Summers 2000; Leenders 2016). On the other hand, males that had their MCP and k50 area far away from other males were located on open areas, and they used holes under the roots of big trees as shelter. In other species like Dendrobates ventrimaculatus and Oophaga pumilio similar patterns have been found: where there was high density of resources, there was higher density of individuals (Pröhl and Berke 2001; Poelman and Dicke 2008; Meuche et al. 2012). Further studies on whether space use patterns of males of P. vitattus are also shaped by females’ distribution are recommended.

Aggressive behaviours

Aggressive behaviours in P. vittatus were believed to be more likely to occur since there were higher chances of encountering neighbours due to home range proximity or even high degree of overlap (Summers 2000). However, we did not record any aggression among the males tracked. Aggressive behaviours recorded in Summers (2000) happened in other populations (Raby Nuñez, pers. comm.). Aggressive behaviours in P. vittatus could context-dependent, like in D. tinctorius (Rojas & Pašukonis 2019), where males were only observed being aggressive against conspecifics during mating events. We suggest that aggression in males of P. vittatus might be context dependent and might exhibit the ‘dear enemy’ effect (Fisher 1954). By this hypothesis, males of P. vittatus will be less aggressive to neighbours than to strangers, which will reduce the cost of territory defence. This is expected to occur in species with limited reproductive resources (Tumulty and Bee 2021). Recently, it has been suggested that males of A. beebei are territorial over reproductive resources such as pools with high-quality characteristics for tadpole deposition (Fouilloux et al. 2023). Considering the hypothesized spatial association between higher density of males with the higher density of heliconias in P. vittatus, males may also be territorial towards this limited resource and, therefore, be less aggressive to neighbours which already have access to these resources. Being dependent on the ecological context (Hyman 2005; Briefer et al. 2008), aggressive behaviour might change when males are carrying tadpoles and the priority is the protection of the offspring. This will also imply that males can recognize other individuals (Bee 2016). So far, in dendrobatids it has been shown how in some species males have distinctive calls but, only one species, A. beebei, has been shown to exhibit the dear enemy effect (Bee 2016; Tumulty 2018; Tumulty & Bee 2021). Intrusion experiments to assess the territorial behaviour of P. vittatus are needed to understand the territorial behaviour of this species.

We did not observe aggressive behaviours between the males tracked. Consequently, we did not observe any aggressive behaviour between the courting male and the male carrying tadpoles. Since the male with tadpoles was not calling, probably searching for a suitable water pool to release its tadpoles, it was not considered as rival (Pröhl 2005). Similarly, in O. pumilio, a male was observed calling towards a frog carrying tadpoles, and no physical aggression was observed (González-Santoro and Mateus-Cruz 2023). Whether males can detect tadpoles transport and whether it works as a cue for males’ reproductive status remains unknown. The only aggressive behaviour registered was between two males that were carrying tadpoles. This encounter was at 10:00, when temperatures are usually high, and the weather is drier (pers. Obs.). Male-male competition for limited resources is quite common in the sexual selection hypothesis (Andersson 1994). After competing to gain exclusive access to the shelter, the low alternative shelter options and the environmental circumstances probably forced them to share the available shelter to ensure their own survival and the survival of the tadpoles.

Female defensive behaviour

We observed a female placing herself behind the male’s back, covering the tadpoles, during a playback reproduction. While female parental care is not rare in dendrobatids (Ringler et al. 2015; Furness and Capellini 2019), P. vittatus is a species with only male parental care in which the males transport the tadpoles to water bodies for further development. The only explanation we could suggest about the observed behaviour is that the female may have considered the presence of a male intruder as a potential threat to the tadpoles. By covering the tadpoles, the female might be attempting to deter the intruding male, protecting the tadpoles from potential harm. Although this is only an anecdotal observation, it is, nonetheless, a rare behaviour that provide valuable insight into the complexity of the species behaviour. Further research focus on the mating system and the females’ behaviour in P. vittatus is highly recommended.

Conclusions

This study provides a platform for future research on the space use and social behaviour of the poison frog P. vittatus. Here, we compiled information about home range and space use patterns, but further experiments are needed to understand in detail its territorial behaviour. Space use patterns of a species provides information related to habitat preferences and resources for which individuals compete. Moreover, it can help us to prioritize areas to protect and considered their necessities when planning conservation strategies. This is key for threatened and endemic species like P. vittatus, for which information is scarce.