Introduction

The Amami tip-nosed frog (Odorrana amamiensis) is a relatively large (56–101 mm) terrestrial species and endemic to the montane forests on Amami-Oshima Island and Tokunoshima Island, Japan (Matsui, 1994; Matsui et al., 2005; Matsui and Maeda, 2018). It is listed as an endangered species on the IUCN Red List of Threatened Species (Kaneko and Matsui, 2004), the national monument by Kagoshima prefecture, and the endangered species of wild fauna and flora by Amami city. It is negatively impacted by various human-induced factors such as the introduction of invasive species and habitat destruction through logging (Utsunomiya and Utsunomiya, 1983; Watari et al., 2008; Matsui and Maeda, 2018). Although action is needed to protect this species, the basic ecology of O. amamiensis is poorly understood.

Several studies on O. amamiensis have been conducted in recent decades. These studies focused on phylogenetic relationships (Matsui, 1994; Matsui et al., 2005), micro-satellite marker development (Igawa et al., 2015), population change caused by invasive mongoose (Watari et al., 2008; Watari et al., 2013), calling site preference (Iwai et al., 2018), and ecological character displacement (Komine et al., 2019). However, little is known on the breeding ecology and breeding sites of O. amamiensis. Observing its breeding activity is difficult because they breed around mountain streams for only a few days in autumn to spring (Iwai et al., 2015; Matsui and Maeda, 2018). A better understanding of breeding ecology and breeding sites would contribute to conservation efforts for this species. Although Iwai et al. (2015) reported body size and age of breeding individuals and Sakoda et al. (2007) reported habitat of wild tadpole, there is no study reporting detailed breeding process of O. amamiensis. This study describes temporal changes in the number of breeding individuals of O. amamiensis on Amami-Oshima Island, Japan.

Materials and Methods

Study area

Amami-Oshima Island, approximately 712 km² in size, is part of the Nansei Islands in southern Japan (Fig. 1). The average annual temperature is 21.5°C and average annual rainfall is 2,914 mm (Watari et al., 2008). The island is covered with broadleaved evergreen trees, such as chinquapin (Castanopsis sieboldii). The island has a subtropical climate and contains many endemic species. Because of its valuable biodiversity, part of the island is currently nominated to be inscribed on the World Heritage List of the United Nations Educational, Scientific and Cultural Organization.

Fig. 1.

Map showing the study location in Kinsakubaru area, Amami-Oshima Island in southern Japan.

Fig. 2.

Photos of the transect line at the breeding stream. The transect line at the stream consists of a flowing section (Fig. 2 [a]) and a cascade (Fig. 2 [b]). The photo (b), which was taken on 30 Nov. 2015, shows several individuals of Odorrana amamiensis.

Results

The number of O. amamiensis individuals at night and daytime was recorded for eight days (Table 1). Amplexus of individuals and egg mass were confirmed, hence the frogs apparently gathered at the stream to breed (Fig. 3a, b). Most individuals were males and only several females and amplexus were observed (Table 1). The number of individuals at night exceeded 200 on 29 Nov. 2015, and this extremely high density of breeding individuals at night remained until 2 Dec. 2015. The high density at night strongly decreased on the 3 Dec. and a few individuals were observed on the 4 Dec. The number of individuals recorded during the day exceeded 20 individuals on the 29 Nov, and the high density of breeding individuals continued until the 2 Dec. Only one female and no amplexus were observed during the day through this study period. No individuals were recorded on the 3 Dec. during the day.

Fig. 3.

The photo (a) shows Odorrana amamiensis in amplexus and the photo (b) shows frog eggs. Those photos were taken on 3 Dec. 2015.

Discussion

This study recorded over 200 O. amamiensis individuals breeding with lively mating calls around a small stream at night for three days (Table 1). Although the number of individuals around the stream before the 28 Nov. was not quantified, we confirmed that there were no lively mating calls on the 23 Sep.–2 Oct., 24 Nov. and 26 Nov, 2015 when we observed the forest road. Since the breeding site is just beside the forest road, we could notice the lively mating calls easily if there was one. Therefore explosive breeding should have started after the 27 Nov. at least. Because O. amamiensis did not seem to be usually around the stream, the number of breeding individuals greatly increased between the 27 and 28 Nov. The number of breeding individuals remained high until 1 Dec 2015, after which the density decreased rapidly. Most of the individuals left the stream by 4 Dec. 2015. Although there may be multiple explosive breeding at the same site in a year, most of the individuals left the stream by 4 Dec. 2015 which suggests that the single explosive breeding may have finished by this time. This finding agrees with a previous study describing that O. amamiensis mainly breed at mountain streams for a short time period (Iwai et al., 2015; Matsui and Maeda, 2018). The current study is the first to quantify the temporal changes in number of breeding individuals of O. amamiensis in a single explosive breeding activity.

Table 1.

Temporal changes in number of individuals of Odorrana amamiensis (a) at night and (b) at daytime. Number of males and females are only single individuals.

Although the number of breeding individuals during daytime was far less than at night, O. amamiensis did gather at the breeding site during daytime as well. It was unknown if some breeding individuals would be active during daytime because O. amamiensis is usually described as a nocturnal species (Watari et al., 2008). Our observation indicates that the activity pattern of some O. amamiensis individuals may change during the breeding period. This finding may be beneficial for the conservation of O. amamiensis. We observed no snakes both during the day and at night probably because this study site was seriously damaged by invasive mongoose (Watari et al., 2008; Komine et al., in press). However, it is generally risky to walk around the forests on Amami-Oshima Island at night because of the nocturnal poisonous snake Habu (Protobothrops flavoviridis) which inhabits mountain streams as well (Oumi, 2006). Our observation of O. amamiensis breeding activity suggests that searching for breeding sites during daytime may be a relatively safer and useful alternative. Finding breeding sites can lead to better protection of the area and a better understanding of the species' breeding habitat preference. Searching for breeding sites of O. amamiensis during daytime may contribute to conservation efforts for this species.

Fig. 4.

Annual changes of temperature (a), humidity (b) and rain fall (c) at Naze city close to Kinsakubaru area in Amami-Oshima Island in 2015. Grey bars represent observation period (28 Nov.–5 Dec.).

Although it is known that they bred from autumn to spring (Oct to May) (Matsui and Maeda, 2018), it is unknown what factors relate to their breeding activity. The current study does not sufficiently address the potential factors influencing the breeding activity of O. amamiensis, while breeding activity might be related to temperature. The observed breeding activity started on 27 or 28 Nov. when there was the lowest mean temperature (27 Nov: 16.2°C, 28 Nov: 16.4°C) since 26 Mar. (26 Mar: 15.9°C) (Fig. 4a). Decreasing temperature after summer might be related to the start of the breeding activity. On the other hand, it seemed to have little relationship between breeding activity and other climate factors (humidity and rain fall) (Fig. 4b, c). However, current study was conducted only a single year hence it is difficult to draw the general conclusion. Year-round monitoring over the years will be necessary to reveal the potential factors influencing breeding activity and the seasonal pattern.