Frogs and toads normally have four toes on the forelimbs and five on the hindlimbs. An exception is the Otton frog Babina subaspera, which has regained the pseudothumbs. The morphology and use of pseudothumbs in Otton frogs were compared between the sexes. The pseudothumbs, which are longer and thicker in males than in females, encase a prepollical spine that occasionally cuts through the tissue as it is projected from the sheath. The males had visible spines and demonstrated a jabbing response more often than did females. Males were observed to use their pseudothumbs in male–male combat over females or breeding nests, as well as during amplexus. There was no evidence that females use their pseudothumbs. It is suggested that pseudothumbs first evolved as an anchor for amplexus, but that they are now used as a weapon in combat as well, giving males higher fitness. The study of pseudothumbs in the Otton frog will facilitate further study of related topics of interest, including extra fingers in vertebrates, self-damaging structures and developmental constraints in the hands.
Hand morphology in vertebrates is widely recognized as being conservative (Sánchez-Villagra & Menke, 2005); however, the existence of exceptions became well-known with the discovery of the pseudothumb of the giant panda (Endo et al., 1999). Extra digit-like structures are also known in other taxa, including the ‘sixth toes’ of elephants, which have evolved to support weight (Hutchinson et al., 2011), and the mole's thumb, which lacks any apparent functional role (Sánchez-Villagra & Menke, 2005). Examination of the function of these unique characters has shed light on the developmental evolution of hands.
Anurans are one vertebrate taxon that possess extra digit-like structures (Fabrezi, 2001). Frogs and toads normally possess four toes on the forelimbs, but extra ‘finger’, sometimes called a ‘pseudothumb’ because of its physical appearance, occur in species of some families around the world (Wells, 2007). In the only detailed report of five-fingered frogs, Kluge (1981) showed that the pseudothumb of the male Hypsiboas rosenbergi encases a sharp prepollical spine, which is occasionally projected from the sheath. Shine (1979) suggested that the ‘sexually dimorphic development of spines … is an adaptation to male combat rather than to amplexus’ (Shine, 1979, p. 302). Male H. rosenbergi were witnessed using the pseudothumb and spine aggressively while fighting each other (Kluge, 1981). In contrast to Shine (1979), Kluge opined that ‘amplexus formed the adaptive basis for the origin of prepollical spines’ (Kluge, 1981, p. 22). These arguments have not yet been settled due to a lack of detailed data; moreover, these studies have generally focused exclusively on the pseudothumbs of males.
The Otton frog Babina subaspera (Barbour), which is endemic to the Amami Islands of south Japan, has pseudothumbs (Wells, 2007; Tokita & Iwai, 2010). In a study of hand morphogenesis in the Otton frog, Tokita & Iwai (2010) showed that the spine encased in a pseudothumb was a well-developed ossified prepollex, but the function of this unique character was no more than speculation. This was mainly because of the difficulty in obtaining detailed data because the Otton frog is an endangered rare species and highly sensitive to observers. Because the breeding habits of this species are similar to those of H. rosenbergi, it is possible that the Otton frog also uses its pseudothumb for male–male combat or amplexus (i.e. only males use it). However, unlike other five-fingered frogs, including H. rosenbergi in which female pseudothumbs are only slightly ossified, females of the Otton frog possess unambiguous pseudothumbs and associated ossified spines. This suggests that pseudothumbs in the Otton frog could be used in a way that provides a benefit to both sexes such as protection from predators and obtaining food. It is also possible that they are used by females in a different way than by males, or that the feature is present as a developmental or evolutionary relic but is not actually used by females.
The evolution of sexual dimorphism is generally thought to be driven by intrasexual selection (e.g. combat), intersexual selection (e.g. mate choice) and natural selection (Andersson, 1994). If the pseudothumb is used in intrasexual or intersexual selection, or if it is used differently between the sexes for utilizing resources, sexual dimorphism of pseudothumbs might be observed in the Otton frog. Although it may help in understanding the function of the pseudothumb, sexual dimorphism of the pseudothumbs in any frog species has not been studied.
The goals of this study were to reveal the function of the pseudothumbs and their associated spines, and to discuss the evolutionary significance of these features in Otton frogs, where they are present in both sexes. The morphology of the pseudothumb and pseudothumb-associated features were compared between the sexes to assess sexual dimorphism, and the practical use of pseudothumbs in Otton frogs was observed in the field.
Materials and methods
The present study was conducted on Amami-Oshima, one of the two islands in southern Japan (Amami-Oshima and Kakeroma-jima) where the Otton frog is found. The island is covered with subtropical rain forests and provides habitat for many rare endemic species. Nine species of frogs inhabit Amami-Oshima, five of which are endemic to the Amami islands, including Otton frogs (Maeda & Matsui, 1999; Matsui, 2011). The Otton frog is a large frog with a mean snout-vent length (SVL) of 117 mm (Maeda & Matsui, 1999) that lives in mountainous areas. It builds a breeding nest that is a water-filled excavation with a rampart, and oviposition occurs in the nest.
Otton frogs were captured on Amami-Oshima on forest roads or at breeding sites from April to October 2010. Individuals were measured for weight, SVL, length and width of the pseudothumb and first finger, and forelimb width. If present on an individual, the extent and location of scarring was also recorded. Whether the prepollical spine could emerge from its encasing sheath was also tested by gently pulling down the sheath.
The behavior of adult frogs at the breeding sites was documented by direct observation and overnight videotaping with an infrared video camera (SONY, DCR-SR65/HDR-SR1, Tokyo, Japan). Videotaping was performed 50 times at the breeding sites; 16 oviposition events, 2 obvious male–male combat events and 5 predation scenes were captured.
Sexual dimorphism in SVL and weight was analyzed by a t-test. Forelimb width was compared by a general linear model using SVL as a covariant. The length and width of the pseudothumb were compared using the length or width of the first finger as a covariant. The rate of emergence of the spine from its sheath was compared between the sexes using Pearson's chi-squared test.
Sexual differences in morphology and behavior
In total, 79 males and 87 females were captured. The males were significantly larger than the females in terms of SVL and body mass (Table 1). Moreover, the males had larger forelimbs compared with females of the same SVL (Supporting Information Fig. S1; Table 2). The length and width of the pseudothumb compared with the first finger in the same individual also showed sexual dimorphism: the pseudothumbs of the males were longer and thicker than those of females with the same size of first finger (Figs 1 and 2; Table 2).
Table 1. Comparison of body size (mean ± sd) between male and female Otton frogs Babina subaspera
d.f., degrees of freedom; SVL, snout-vent length.
124.8 ± 8.6 (n = 79)
120.5 ± 8.5 (n = 87)
274.0 ± 51.7 (n = 78)
236.8 ± 38.9 (n = 87)
Table 2. Results of a general linear model analysis. Snout-vent length was used as the covariant for ‘forelimb width’ analysis, length of the first finger for ‘pseudothumb length’ analysis and width of the first finger for ‘pseudothumb width’ analysis
d.f., degrees of freedom.
When captured, individuals of both sexes tried to escape by turning their bodies and kicking the captor's hand. Once captured, they continued kicking with their legs and slapping with their forelimbs. However, when something irritated the chest, some individuals promptly pulled their arms toward their chest and jabbed their pseudothumbs with the spines projecting at whatever came within their embrace (Supporting Information Fig. S2). When the animals were touched on the chest with a finger, this response was observed more often in males (65 out of 76 males and 43 out of 78 females; χ2 = 16.98, P < 0.001). The strength of the response differed greatly among individuals. Some males showed an extreme response and did not stop pulling their arms inward even when they jabbed their spines into their own chests. If jabbed in the finger by a male's spines, the researcher responded by dropping the frog. In contrast, the response of female frogs was weak to absent and a jab to the bare finger did not elicit a pain response in the researcher.
Spines are generally encased in the pseudothumb. When the sheath was pulled down, the spine was released from the sheath and became visible in 44 out of 77 males and 4 out of 80 females (χ2 = 50.26, P < 0.001). The pseudothumbs of the males were sometimes wounded around the tip, seemingly injured when the spine emerged from the sheath (Fig. 3). No female had a wounded pseudothumb, and only the tip of the spine was ever visible out of the sheath, if it emerged at all.
Scars were found in 29 out of 78 males and 31 out of 78 females. This difference was not significant (χ2 = 0.108, P > 0.05), but the males had more scars on the dorsal area while the females had more on their sides (Table 3). The scars on the males were mostly scratches and were more widespread than those of females, which mostly had stab wounds on their sides under the arms (Fig. 4). Three females were captured soon after they had laid eggs, and all had stab wounds on their sides where the amplexing male had embraced them with his arms.
Table 3. Locations of the scars found in 29 males and 31 females
Behavioral observations in the field
The breeding season of Otton frogs lasts for half a year, from April until October. Males remained near the breeding site for several months making calls to attract females. Up to 15 nests could be observed in one breeding site, but a decreased number of males were present each night. The number of females that came to the breeding site in one night was small, and often zero. However, on some nights during the peak of the breeding season (June–July), multiple females came to the site on the same night.
Sixteen oviposition events were captured. While amplexing, males grabbed the base of the female's arms using their third and fourth fingers and placed their other fingers (including the pseudothumb) on the female's sides under the arms (Fig. 5). One pair was captured just after egg-laying; the two frogs were in an amplexing position when captured. When the pair was pulled apart, the spine on the male was found to have been jabbed into the side of the female (Fig. 6). Females did not show any use of the pseudothumbs during oviposition.
Predation behavior was captured in five scenes in which males jumped toward something moving and swallowed it. In one scene, the prey was an amplexing pair of the small frog Buergeria japonica, and in the other, it was a giant house centipede Thereuopoda clunifera. The Otton frogs did not use their pseudothumbs during any of the predation events.
Male–male combat was observed twice. The first observation was made on the night of 27 June 2010 at a breeding site. The area consisted of a 5 × 5-m artificial concrete barrage built next to a forest road. There were two separate pools inside the barrage (2 × 2 m and 1 × 1 m). The author first visited the site at 20:45 h, when there were five adult Otton frogs (three males, two unidentified). One of the three males had scars on his back. The infrared video camera (SONY, HDR-SR1) was set facing the corner of the smaller pool, which was being used as a substitute for the breeding nest and had a 3-day-old egg mass. Recording occurred from 22:42 to 08:50 h the next day. Over the course of the night, amplexus occurred eight times within the screen. Each time two frogs amplexed, another male jumped in and the amplexus was broken. Intense male–male combat occurred after the fifth bout of amplexus. At 03:28 h when the couple started laying their eggs, another male (male A) suddenly head-butted the amplexing male (male B), and the two grappled with a growl. Male A jabbed his arms into the head of male B while holding its head from two sides (Supporting Information Fig. S3). Male B struggled to escape from the grasp of male A, but male A continued jabbing. For more than 4 min, male B kept trying to escape from male A by kicking and flapping. While grappling, the two frogs floated deeper into the water away from the center of the screen. Unclear images of the two wrestling frogs and water movement continued until 03:43 h, when the two frogs separated. On this night, there seemed to be another fight after the sixth amplex broke up, but the scene occurred mostly outside of the camera's field of view, and only a growling sound and a portion of a head were observed. After the eighth amplex, oviposition occurred successfully at 04:32 h and ended at 04:44 h. Unfortunately, the identity of which male frog eventually fertilized the egg mass could not be determined. The fight scene is registered in the Movie Archives of Animal Behavior (http://www.momo-p.com; data # momo100928un01b).
The second observation was made on the night of 13 July 2010. It occurred in an Otton frog nest constructed at the edge of a 4 × 4-m pool in a concrete barrage. When the author first visited the area at 19:50 h, one male was inside the nest and another was sitting in front of the nest. The infrared video camera (SONY, DCR-SR65) was set facing the nest. Recording occurred from 19:52 to 09:50 h of the next day. At 20:48 h, the male sitting in front of the nest (male C) slowly walked to the nest edge at the side opposite the male in the nest (male D) and hid under the vegetation. Male D appeared motivated and called more frequently. At 21:15 h, male C came out of the vegetation and walked into the nest. Male D stopped calling and sat motionless. Male C sat just beside male D, facing his side. At 21:19 h, male C pounced on male D at the moment male D started to move to turn toward him. Male C embraced the waist of male D (Supporting Information Fig. S4), who then fought back by pulling both arms to his chest as if jabbing his pseudothumbs into the enemy (Supporting Information Fig. S4). His intention was not successful, as male C was holding male D lower than his chest, and the two frogs separated. After the first fight, the two males remained around the nest. Twice, one male jumped on the other, but the attacked male did not fight back and simply jumped away. At 21:58 h, however, one male jumped on the other's back and jabbed his arms into the other frog's side while in an amplex position. Two males stayed in this position for 30 s, separated and moved outside the view of the camera with a growling sound. No other fight was recorded, and the two males jumped away at 04:33 h the next morning. Five days later, a male Otton frog was found sitting in the same nest with a scratch on his side, which might have been due to the fight. The fight scene is registered in the Movie Archives of Animal Behavior (http://www.momo-p.com; data # momo100928un02b).
Clear sexual differences were observed in the morphology and behavior of Otton frogs. For example, males had larger and thicker pseudothumbs than females. In addition, the proportion of individuals whose prepollical spine could project from the pseudothumb was higher in males than in females. Only the spine tip was visible in females, whereas in males, the spine was clearly visible, sometimes with a wound near the tip of the pseudothumb. A higher proportion of males showed a jabbing response than females, and the response by females, if it occurred, was relatively weak. These observations suggest that the pseudothumb is used mainly by males. Field observations supported these findings, showing that pseudothumbs were used in male–male combat and during amplexus.
Male–male combat occurred during competition for access to females or oviposition nests. The Otton frog has a long breeding season; thus, the chance of having a female at a breeding site on each night is small. Unlike explosive breeders, where multiple males aggregate to a female and fertilize eggs relatively by chance, this species lays and fertilizes eggs in a nest as a single pair. Therefore, obtaining females at each female-visit and having a good nest position to increase the chance of accessing females is highly important and likely leads to higher fitness in Otton frogs. The breeding habits, giving benefit only to limited males that successfully obtained females, might have led to the evolution of intense male–male combat in this species. Body, forelimb and pseudothumb sizes then became large in males as a consequence of physical combat: larger sizes would have advantages in combat allowing stronger jab and giving more damage to the opponent.
Kluge (1981) noted that some males of H. rosenbergi were found dead after violent aggression. He observed that the unsheathed pseudothumb spines were jabbed at the eyes and ear drums of the opponents, and the injuries were considered to be critical. In Otton frogs, however, although many males were observed to have scars possibly resulting from combat, none was found to have died from these wounds. They jab toward something within their embrace, not necessarily to eyes or ear drums; thus, the injuries may be less critical. Moreover, male Otton frogs have a raised patch on their sides where they sometimes have scratches or stub wounds (Maeda & Matsui, 1999). This structure is observed only in males and may serve to guard the spine, which might make their combat less lethal. It seems that the intensity of combat in Otton frogs is finely balanced so as not to result in critical or mortal injuries, yet it remains aggressive enough to establish a clear victor.
Use as a weapon in male–male combat was not the only role of the pseudothumbs in Otton frogs; they were used in amplexus as well. Male Otton frogs cling to the sides of the female by jabbing their pseudothumbs into her. Amplexus in the Otton frog occurred with one male and one female in an oviposition nest, and dense mating aggregation never occurred. In 2 out of 16 oviposition events, the disturbed male was observed to instantly release the female when an intruder male appeared, rather than hanging on. Pseudothumb use by males seems to play a supportive role in fastening to the females during amplexus and oviposition, but it is not used for clinging to the female while attacking an intruder.
In derived frog families, males usually clasp the female behind the front legs (Wells, 2007), and nuptial pads are clasped against the female's belly (Peters & Aulner, 2000) for stronger coupling. Otton frogs do have nuptial pads, but they use their pseudothumb and spines in amplexus as well. The observed finger use of Otton frogs in amplexus caused injury to females, and thus does not seem very beneficial to females. Despite the disadvantage, however, such finger use in Otton frogs may have evolved because of the larger body size of males relative to females. If males are larger than females, a male has to hang forward over a female during oviposition in order to place his cloaca at the upper position to that of the female so that the sperm can reach the ova when they are released from the female. Jabbing pseudothumbs into the side of the female might serve as an anchor point from which to hang forward.
Another use of pseudothumbs may be for obtaining food or protection from predators. If the pseudothumbs of Otton frogs serve these functions, the observed sexual dimorphism suggests that males use their pseudothumbs more often or more intensely than females while hunting for food or during anti-predator behaviors. However, the habitat range, food items and active period, all of which can lead to such differences, appeared to be the same between the sexes. This was confirmed by field observations. The male Otton frogs did not use their pseudothumbs for predation, and a reported observation of predation behavior in a female also did not mention the use of pseudothumbs (Iwai, 2010). Whether Otton frogs use their pseudothumbs against predators could not be confirmed because no observation of an Otton frog under predation was made during more than 70 nights of surveying. The only reported predator is the large snake Protobothrops flavoviridis, which preys on the Otton frog at a rate as low as 0.2% (Mishima, 1966). It appears that predation on Otton frogs does not happen often. Even when predation does occur, pseudothumbs may not be effective against predators because they face inward and the projected spines can only attack something within their arms. Also, most of the Otton frogs did not aggressively attack humans with their pseudothumbs when captured; aggression occurred only when their chest was irritated, which can be considered a reflex related to male–male combat or amplexus. Thus, the possibility of pseudothumb use for obtaining food or protection is slight.
The pseudothumbs of the male Otton frogs were sometimes wounded. This seemed to be because the spine pierced its sheath during use. Otton frogs jab their pseudothumbs into their opponents so strongly that the spines emerge by cutting through the sheath. When the author pulled down the sheath, the spine emerged and became visible in more than half of the male Otton frogs. Presumably, those with a visible spine might have used them recently in combat, whereas those that were not visible had not been used recently (at least for more than the period during which the wound healed). Piercing of the skin while using spines or claws has been observed in other frog species (Blackburn, Hanken & Jenkins, 2008) and in salamanders (Brodie, Nussbaum & DiGiovanni, 1984). Blackburn et al. (2008) showed that the claws of Astylosternus and Trichobatrachus pierce their way to functionality, and Brodie et al. (1984) showed that Echinotriton andersoni has sharp ribs that protrude through the body wall against predators. Blackburn et al. (2008, p. 356) stated that the bony ribs of E. andersoni are the only comparative structures to the claws of Astylosternus and Trichobatrachus, and that the claws were not analogous to the prepollical spines of five-fingered frog species as ‘the spines … appear to grow through the skin rather than traumatically pierce it’. However, the spines of Otton frogs do not grow through the skin, but rather pierce the sheath traumatically. Thus, the claws of Astylosternus and Trichobatrachus, the ribs of E. andersoni and the prepollical spines of Otton frogs might have some common developmental features. Although amphibians are known to have remarkable regenerative capacity (Brockes & Kumar, 2005), a structure that damages the animal itself in its use does not seem adaptive. This topic needs to be examined further and will be an interesting case study for the development of self-damaging structures.
Although females do not appear to use their pseudothumbs and spines, they are still present, and a few individuals had spines that projected slightly from the sheath or showed a weak jabbing response. This could be because formation of the pseudothumb is linked developmentally with other important traits. The fact that the development of pseudothumbs in Otton frogs occurs at a fairly early stage, even in larvae (Tokita & Iwai, 2010), supports this idea. Corresponding formation of spines in females was also observed in some adult females of H. rosenbergi: ‘slight ossifications in the comparable thenar regions were observed’ (Kluge, 1981, p. 21), although the structure does not seem as prominent as that of female Otton frogs. Thus, developmental linkages might be common in frogs. Further study of this topic may reveal why hand morphology in vertebrates is recognized as so conservative (Sánchez-Villagra & Menke, 2005).
The earliest anuran had five toes, whereas modern frogs have four (Roček & Rage, 2000). This study was the first to examine the detailed morphology and function of the pseudothumb in modern frogs. It was revealed that the Otton frog uses its pseudothumb in a dual manner: as a weapon in male–male combat and as an anchor in amplexus. A scenario for the evolution of pseudothumbs in Otton frogs is proposed. First, their breeding habits led to the evolution of intense male–male combat, in which larger males had advantages. Subsequently, males became larger than females and the need of an anchor for amplexus arose. Those males with a better structure such as a slightly ossified, sharp, inwardly facing spine had higher fitness as they were able to fertilize more eggs. As the structure became larger, it was co-opted as a weapon in male–male combat, resulting in more damage from combative jabbing. The male Otton frog may have regained its pseudothumb in this manner. To confirm this hypothesis, a comparison of body size and amplexus position in the Otton frog with those of other frog species is needed.
This study succeeded in more than revealing the function of pseudothumbs in Otton frogs: it also showed the academic potential of the study of pseudothumbs in frogs, which will facilitate further research of related topics of interest such as extra fingers in vertebrates, self-damaging structures and developmental constraints in hands.
The author would like to thank Shohei Oumi and Kazuto Kawakami for their help. This study was carried out under permit no. 566 from the Kagoshima Education Commission and was financially supported by the JSPS Research Fellowship and Research Fund.