AmphibiaWeb - Hemisus marmoratus


(Translations may not be accurate.)

Hemisus marmoratus (Peters, 1854)
Shovel-nosed Frog, Marbled Snout-burrower, Isinana esipendiwe (Zulu)
family: Hemisotidae
genus: Hemisus
Species Description: Peters, W. C. H. (1854). Diagnosen neuer Batrachier, welche zusammen mit der früher (24. Juli und 18. August) gegebenen Übersicht der Schlangen und Eidechsen mitgetheilt werden. Bericht über die zur Bekanntmachung geeigneten Verhandlungen der Königlich Preußischen Akademie der Wissenschaften zu Berlin 1854: 614–628.
Hemisus marmoratus
© 2015 Alberto Sanchez-Vialas (1 of 10)

sound file   hear call (764.0K MP3 file)
sound file   hear call (8413.1K WAV file)

[call details here]

Conservation Status (definitions)
IUCN Red List Status Account Least Concern (LC)
National Status None
Regional Status None


Berkeley mapper logo

View distribution map in BerkeleyMapper.
amphibiandisease logo View Bd and Bsal data (26 records).

A plump, almost cylindrical frog with a very pointed snout and short but sturdy limbs. It has short limbs and a large, flanged inner metatarsal tubercle. Males measure 22–34 mm (SVL), weighing 1.3 to 5.0 g. Females bearing no eggs weigh 5.8–6.3 g, otherwise 7.7–12 g (SVL: 37–49 mm). A bulging dermal fold separates the neck and the snout regions. Males have a single subgular vocal sacs which are usually covered by a small gular flap. They have small eyes with vertically slit pupils. These frogs have no visible tympanum. Both the hardened snout and the large inner metatarsal tubercles have been transformed into digging tools. The inner metatarsal tubercle reaches 0.9–1.9 of the shortest toe length. The thighs reach 0.2–0.4, the lower legs 0.3-0.4 and the foot, including the longest toe, 0.5–0.6 of the SVL. Neither toe-tips nor finger-tips are enlarged. No traces of webbing. The skin is usually smooth, occasionally finely granular. Loveridge (1929) cites a female from Kenya with 52 mm SVL. Voucher specimens: SMNS 8963 1–14 + tadpoles; SMF 78649; ZFMK 68418 + 5 specimens.

Coloration: The snout is beige to dark olive. Yellow patches form a complex and highly variable meander pattern covering all of the dorsal part of the animal. Most often, brown and yellow pigments are equally distributed throughout. Rarely, one of these colors appears less conspicuous, so that yellow or black dots are scattered on a contrasting background. The vocal sac of the male is dark violet, whereas in the female, the throat shares the grayish-white color of the venter. Coloration in alcohol is almost equal to the live coloration, except that the frogs appear somewhat paler, and the colors usually turn to brown. On some animals, the mottled pattern of the back disappears completely so that the dorsum turns uniform brown or gray. The male throat either turns black, or it looses all its pigments. Finally, the neck fold loses definition in alcohol.

Voice: A loud buzzing call similar to the sound produced by crickets. Its frequency is around 4kHz. The call lasts 1.5–2 sec. The same call duration was reported by Schiøtz (1964c). He gives a dominant frequency that initially is 2.8 kHz, but quickly reaches a constant level of 3.1 kHz. The call comprises numerous pulses. In Nigeria animals uttered a warning vibration (Schiøtz 1963). Regarding the body dimensions given by Schiøtz, the latter observation most likely refers to H. guineensis. According to Amiet (1973a), the advertisement call is a shrill whistling sound.

Distribution and Habitat

Country distribution from AmphibiaWeb's database: Benin, Botswana, Burkina Faso, Cameroon, Central African Republic, Chad, Cote d'Ivoire, Eritrea, Ethiopia, Gambia, Ghana, Guinea-Bissau, Kenya, Malawi, Mozambique, Namibia, Nigeria, Senegal, Somalia, South Africa, South Sudan, Sudan, Swaziland, Tanzania, United Republic of, Togo, Zambia, Zimbabwe

Berkeley mapper logo

View distribution map in BerkeleyMapper.
amphibiandisease logo View Bd and Bsal data (26 records).
Habitats: This species mainly inhabits savannas, but it is also found in gallery and island forests (Wake & Kluge 1961, Walker 1968, Rödel et al. 1995). Walker (1968) found this species most often where the substrate was well-cultivated and irrigated. However, his report gives no clear indication as to whether this frog is actually more common in those habitats, or whether it is simply found more regularly because of the regular human activities. However, a preference for humid soils was equally observed by Poynton & Broadley (1985a) and Van Dijk (1982). On the contrary, Loveridge (1933) writes that he found the nests of this species at rather dry sites.
Range: According to Laurent (1972a) and Frost (1985), subspecies of H. marmoratus live all over the savanna regions south of the Sahara. The Comoé National Park forms a part of the range of H. marmoratus sudanensis (Steindachner, 1863). According to Laurent (1972a), this subspecies inhabits an area south of the Sahara stretching from Senegal to Eritrea. For West Africa, he cites Sierra Leone, Ivory Coast, Nigeria and Chad. I found this subspecies in Burkina Faso, too (Böhme et al. 1996). According to Laurent (1972a) the range of the nominate subspecies considerably overlaps that of H. m. sudanensis. If the two subspecies should not cross-breed. H. sudanensis ought to be considered as a valid species. Otherwise the actual subspecies names should not be used any longer. I believe that reports from the following countries actually refer to H. marmoratus: Senegal, Gambia, Guinea Bissau, Guinea, Burkina Faso, Ivory Coast, Ghana, Benin, Nigeria, Cameroon, Chad, Central African Republic, Somalia, Eritrea, Ethiopia, Kenya, Tanzania, Zanzibar, Malawi, Zambia, Zimbabwe, Mozambique, South Africa, Namibia (Pfeffer 1893, Boulenger ?1906, 1910, Lönnberg 1910, Nieden 1915, Chabanaud 1919b, Loveridge 1925, 1929, 1930, 1936, 1942, 1955c, 1957, Parker 1930, Andersson 1937, Mertens 1955b, Wake & Kluge 1961, Schiøtz 1963 in part, 1964a in part, c, 1967 in part, Poynton 1964a, c, 1966, 1991, Lamotte 1966 in part, 1969, 1971, 1998, Barbault 1967, 1974d, Vuattoux 1968, Walker 1968, ?Euzet et al. 1969, ?Maeder 1969, Broadley 1971, 1991, Amiet 1973a, Stevens 1974, Lanza 1978, Poynton & Broadley 1985a, Frost 1985, Wager 1986, Branch 1988, Lambiris 1988, 1989, Channing 1989, Joger 1990, Channing & Griffin 1993, Simbotwe & Mubemba 1993, Pickersgill 1994, Passmore & Carruthers 1995, Rödel et al. 1995, Böhme et al. 1996, Rödel 1996, 1998b, Largen 1997a, c, Spieler 1997a, Kaminsky et al. 1999).

Life History, Abundance, Activity, and Special Behaviors
Biology: The diet nearly exclusively consists of ants and termites (e.g. Loveridge 1933, 1936, Andersson 1937, Lamotte 1967b, Barbault 1974d). The anatomy of the tongue and the mechanics of food intake have been analyzed by Ritter & Nishikawa (1995). H. marmoratus seems to be a predominantly subterranean species. Unlike most other burrowing species, which burrow by means of the metatarsal tubercles of their hind limbs (transformed into shovels), digging themselves backward rather slowly, Hemisus species invariably burrow head first, even when the male clasps his mate in inguinal amplexus. However, it is not sure whether these frogs are active or forage for prey below the surface. Although they are very likely to do so, we have not yet found evidence. H. marmoratus is encountered at the surface mainly at night, or immediately before or after rainfall when the air is saturated with humidity. This observation is confirmed by Poynton & Broadley (1985a). Otherwise, the frogs are hardly ever active at the surface.

During the first rainy nights, large numbers of shovel-nosed frogs will migrate to the vicinity of the ponds. Reproductive behavior starts long before the ponds are filled with water. Calling however, seems to be triggered by rain (Rödel et al. 1995, Kaminsky et al. 1999). Quite regularly, amplectant pairs are already observed on these nights. Males are known to call both below ground and on the surface. In Namibia males call from the edge of pans during wet weather (Channing & Griffin 1993). The large number of solitary females arriving at the ponds most probably indicates that pairs are formed mainly at these sites. The hands and forearms of males apparently secrete sticky fluids so that pairs can be separated only by "force". Each pair will dig a burrow near the water, and the eggs are deposed in this subterranean refuge. Most probably, the males leave the burrow soon after. We found just a single cavity where a male was occupying a small chamber below the nest site (Rödel et al. 1995). On the other hand, the females keep sitting on the clutch which would dry up otherwise. Eggs stuck on the back of the female (Passmore & Carruthers 1995) are almost certainly accidental. Rainfall is supposedly not the single factor triggering reproduction. For example, we found a nest in rather sandy soil, about 30 cm below the surface, long before the onset of the rainy season. The adjacent pond, separated from the nest by distance of about 5 m, was completely dry at that time. Tadpoles found in ponds that had been filled only hours ago unmistakably show that the reproductive activities of H. marmoratus can begin long before rainfall occurs (Rödel et al. 1995).

Spawn: The eggs are deposited in subterranean burrows, with each clutch having a diameter of up to 4 cm and comprising 88–242 eggs. The unpigmented eggs have a diameter of 4.96–7.30 mm and are rich in yolk. The clutches are usually covered by a layer of infertile eggs. They are further covered by a layer that in appearance is similar to foam or parchment. The tadpoles hatch three to six days later, showing no traces of external gills. According to Kaminsky et al. (1999) the tadpoles hatch about one week after egg deposition. Wager (1986) reports approximately 200 eggs in a nest. In Tanzania, Loveridge (1933) found 110 tadpoles in a single nest.

Experiments have shown that the tadpoles may remain up to two months in those chambers. However, they can survive in open water at the age of some days (Rödel et al. 1995, Kaminsky et al. 1999). As the time to be spent in the burrow is not fixed, they may react very rapidly and flexibly as soon as rainfall occurs. Therefore, they can reach the ponds before any other tadpoles hatch, and they begin their aquatic life larger than their competitors and most of their potential predators. This is especially important as very young H. marmoratus tadpoles seems to be extremely vulnerable to predators, whereas larger larvae have a much higher survival probability (Rödel 1998b). However, the breeding season of H. marmoratus is not strictly restricted to the early phase of the rainy season. We found young larvae in the ponds throughout the rainy season, sometimes completely independent of actual rainfall (Rödel et al. 1995).

Contrary to Amiet (1991c), the mother both ensures an adequate level of humidity and protection for the larvae. These are actively defended against intruders (Rödel et al. 1995). When the mother was separated from the clutch, the tadpoles either desiccated, or they were eaten by ants. The tadpoles usually adhere to the mother as long as they remain in the burrow. Females carrying tadpoles have also been observed by Budgett (in Bliss 1907), Bourquin (1985) and VanDijk (1985). In the older literature, only one reproductive mode, based on a paper of Wager (1929), has been published for this species. The female was said to dig a subterranean tunnel to the nearest pond, thereby enabling the tadpoles to reach water (e.g. Wager 1929, 1986, Duellman & Trueb 1986, Lambiris 1989). At Comoé National Park, we have succeeded in demonstrating that this phenomenon is virtually impossible in most breeding places. The banks of most ponds are so flat that long tunnels would be necessary to prevent the cavities from being inundated from the very beginning (Rödel et al. 1995, Kaminsky et al. 1999). Some nests were situated from the nearest pond by distances of up to 100 m. Comparable situations found at ponds in South and East Africa have been described by Loveridge (1936), Van Dijk (1985), and Channing & Griffin (1993). For example Loveridge (1936) found a pair beneath the stem of a fallen coconut palm, and a female on an egg mass in similar conditions. He mentioned that the rainy season had just begun and the whole area seemed to become flooded in the course of a week or two. According to our data (Rödel et al. 1995), the tadpoles seemed to be, at least in some cases, transported to the pond by their mother. The tadpoles found in the nests usually adhered to their mothers, and we encountered larvae in artificial concrete ponds where no tunnel could have been dug. Since 1995, I found additional tadpoles inhabiting rock-pools. In most cases, however, the female formed a superficial slide enabling the young to wriggle into the water (Rödel et al. 1995, Kaminsky et al. 1999). Breeding chambers are even dug at the bottom of future ponds (Kaminsky et al. 1999). These nests are flooded after heavy rains. Most probably, shovel-nosed frogs follow highly variable strategies in response to the respective environment.

At Comoé National Park, the tadpoles are found in any type of water, with the exception of the river Comoé, e.g. in residuary puddles in the dry beds of brooks, savanna ponds of all sizes, and forest ponds. I even captured them in a brook on a rocky plateau. They may be found in any sector of these waters, but almost invariably near the bottom (Rödel 1998b). In clear waters with dark bottom substrates, they are usually almost black. On the contrary, I have found white larvae in rather muddy water. These tadpoles are capable of changing their color within a few minutes. Most probably, their respective colors provide a certain degree of protection as the tadpoles have got to inhale atmospheric oxygen quite often. They are supposed to feed mainly on tiny algae and detritus. The highly variable sizes of metamorphosing tadpoles possibly indicate that the time when metamorphosis takes place is variable, so that they can minimize the risk of desiccation. (Rödel,

In Lamto, Hemisus marmoratus was captured during the dry season in the soil below 2.6 % of all Rônier palms, Borassus aethiopum. During June and July they reproduce in savanna ponds. During the rest of the year they live 15 cm beneath the soil at the feet of the palms. In January (dry season) one specimen was discovered 50 cm beneath the soil (Vuattoux 1968; Rödel,

Tadpoles: Average length of hatching tadpoles is 11.1 mm (s.d. + 0.4 mm). This stage does not possess an oral disc. The rather elongate young larvae resemble tiny eels, with a large yellow yolk sac shimmering through the ventral skin. Several days pass before the differentiation of the oral disc is completed. This process proceeds more rapidly on tadpoles transferred to open water than on those which are left in the mother’s burrow (Rödel et al. 1995).
The complete keratodont formula reads 1 / 4+4 // 4 or 1 / 4+4 // 1+1 / 3. The oral disc is lined both laterally and caudally by one or two rows of short papillae touching the craniad border of the anterior teeth row. Some long cylindrical papillae are arranged in a caudad position, too. The horny teeth are elongate, and their distal ends bear tiny flat shovels with numerous tips.

Depending on the respective weather, the tadpoles either are transferred to the ponds within a few days, or they remain in the burrow for up to two months. Older larvae usually appear sturdier than younger stages. Rarely elongate individuals with almost parallel flanks have been sampled, too. The keratodont formula of the latter corresponds to that of the typical morph. The thickened base of the tail axis bears a black strip which stretches to the tip. This feature is just outlined in burrow stages. As for the rest, the coloration of the tadpoles is highly variable, depending most likely on the respective environment (compare below). Once in the pond, their development proceeds rapidly, especially during the first 8–10 days spent in water. Their hind limbs begin to emerge at a BL of 13.5–18.1 mm (TL: 41–53.3 mm). Metamorphosis sets in at a BL of 16 mm (SVL: 46mm). One of the largest tadpoles ever collected measured 20 mm (BL) and 56 mm (TL). In the humid residuary puddles of a brook, we found eight tadpoles whose TL amounted to more than 60 mm All these animals already possessed well developed hind limbs. The following exemplary data are meant to give an impression of the sizes-weight ratio (TL in mm / weight in g): 11.4/ 0.01; 16.6/ 0.02; 24/ 0.15; 30/ 0.28; 40/ 0.87; 53/ 0.85). Having spent three to four weeks in the water, the metamorphosed frogs measured 11.5 to 25 mm SVL. They leave the ponds to burrow themselves near the banks. Their coloration corresponds to that of their parents, being at most somewhat less contrasting. Some individuals even show uniform black backs, white bellies and grayish throats.

Wager (1986) and Lambiris (1988) give the following keratodont formula for tadpoles collected in South Africa: 2 / 3+3 // 1+1 / 3. The descriptions published by Guibé & Lamotte (1958a), Schiøtz (1963) and Perret (1966) most probably relate to tadpoles of H. guineensis that are quite similar to the above-mentioned ones. Besides, Perret (1966) gives the keratodont formula of an obviously mutilated individual: 4+4 // 4+4. In Malawi, the young frogs leave the water after a period of six weeks (Stewart 1967).

This account was taken from Rödel, M.-O. (2000), Herpetofauna of West Africa vol. I. Amphibians of the West African Savanna, with kind permission from Edition Chimaira publishers, Frankfurt am Main. For references in the text, see here.

This species was featured in news of the week 24 October 2022:

Changes to food webs can have unforeseen effects on other members of an ecosystem. Demare et al. (2022) documented how increased vegetation, presumably from the overhunting of large herbivorous mammals, changed the larval amphibian assemblage in tropical savannas in Comoé National Park, Ivory Coast (Cote d’Ivoire). Comparing surveys conducted before and after two civil wars, during which poaching caused vast reductions in large herbivorous mammals, the authors found significant increases in vegetation cover around ephemeral water bodies and a shift in dominant amphibian species in their larval form. The switch in amphibians favored species that use vegetation during breeding (Afrixalus spp., Hyperolius spp., and Hemisus marmoratus) and was detrimental to ground breeding species (Ptychadena spp.), which had previously been dominant. While increased vegetation likely provides more structure and food to aquatic systems, this shift was most closely correlated with the composition and structure of aquatic predators. It is unclear if this correlation is causal. However, this system provides a unique opportunity to better understand food web and ecosystem dynamics. (Written by Ann Chang)


Phaka, F.M., Netherlands, E.C., Kruger, D.J.D., Du Preez, L.H. (2019). Folk taxonomy and indigenous names for frogs in Zululand, South Africa. J Ethnobiology Ethnomedicine 15, 17. [link]

Rödel, M. O. (2000). Herpetofauna of West Africa, Vol. I. Amphibians of the West African Savanna. Edition Chimaira, Frankfurt, Germany.

Originally submitted by: Marc-Oliver Rödel (first posted 2001-05-02)
Edited by: Arie van der Meijden, Vance T. Vredenburg, Michelle S. Koo (2023-05-31)

Species Account Citation: AmphibiaWeb 2023 Hemisus marmoratus: Shovel-nosed Frog <> University of California, Berkeley, CA, USA. Accessed Jul 17, 2024.

Feedback or comments about this page.


Citation: AmphibiaWeb. 2024. <> University of California, Berkeley, CA, USA. Accessed 17 Jul 2024.

AmphibiaWeb's policy on data use.