Stumpffia lynnae, Mullin & Rakotomanga & Dawson & Glaw & Rakotoarison & Orozco-terWengel & Scherz, 2022

Mullin, Katherine E., Rakotomanga, Manoa G., Dawson, Jeff, Glaw, Frank, Rakotoarison, Andolalao, Orozco-terWengel, Pablo & Scherz, Mark D., 2022, An unexpected new red-bellied Stumpffia (Microhylidae) from forest fragments in central Madagascar highlights remaining cryptic diversity, ZooKeys 1104, pp. 1-28 : 1

Stumpffia lynnae sp. nov.

Holotype.

ZSM 1/2022 (field number KAMU2), an unsexed adult, collected by K. Mullin, M. G. Rakotomanga, M. L. C. Razafiarimanana and T. Raditra, on 21 January 2020, in one of the northern fragments in Ambohitantely Special Reserve (18.1517°S, 47.2886°E, 1540 m a.s.l.), Analamanga Region, central Madagascar (Fig. 5 ).

Diagnosis.

The new species is assigned to the genus Stumpffia based on its morphological and genetic affinities. Within the genus, it is distinguished by the unique combination of the following characters: (1) SVL 15.5-22.2 mm (adults in life), (2) limited digital reduction on the hands and feet such that first finger is reduced; other fingers not reduced and first toe is slightly reduced; other toes not reduced, (3) bright red to orange colouration confined to the ventral surfaces of the legs, posterior abdomen, and ventral arms, (4) absence of red markings on the lower jaw, and advertisement call with (5) inter-call intervals of 3498-5581 ms, (6) call/note duration 163-184 ms, (7) dominant frequency 2027-2044 Hz, and (8) distinct genetic divergence in the mitochondrial and nuclear genome to other known species.

Stumpffia lynnae sp. nov. can be distinguished from all other Stumpffia species except S. be , S. kibomena , S. meikeae , S. miovaova , S. nigrorubra , and S. roseifemoralis by the presence of bright red colouration on ventral surfaces of arms, legs, and abdomen. Among these species it can be distinguished from S. be by smaller body size (15.5-22.2 mm adults in life vs. 25.2 mm) and less expanded terminal discs on fingers and toes, as well as being in a different major clade of Stumpffia (Clade C rather than Clade A; Rakotoarison et al. 2017); from S. meikeae by the colour on the belly (bright red to orange vs. fainter more champagne to salmon colouration in S. meikeae ), by less expanded terminal discs on fingers and toes, as well as being in a different major clade of Stumpffia (Clade C rather than Clade B; Rakotoarison et al. 2017); from S. miovaova by a larger maximum body size (22.2 vs. 18.2 mm) and colour on the belly (bright red vs. fiery orange); from S. nigrorubra by the dorsal colouration and patterns (dark brown dorsolateral bands and pale brown colouration in adult S. lynnae vs. darker mottled black and dark iridescent dorsal colouration in S. miovaova , and its phylogenetic placement in Clade C2 rather than Clade C1 ( Rakotoarison et al. 2017); and finally from S. roseifemoralis by a larger maximum body size (22.2 vs. 18.4 mm), more vibrant ventral colouration, rougher dorsal skin in life, and less homogeneous dorsal colouration. Further, S. lynnae sp. nov. differs from all the above, other than S. kibomena , by the presence of dense blackish pigmentation on the throat, and different dorsal patterns.

Stumpffia lynnae sp. nov. is morphologically almost indistinguishable from S. kibomena (Table 3 ). However, an elongated red marking on each side of the lower jaw is absent from all observed specimens, but can be present in S. kibomena ( Glaw et al. 2015), and the upper arms of S. lynnae are usually brown versus usually red in S. kibomena . Direct comparison of these features can be seen in Fig. 6 , with additional 'in life’ images of the S. lynnae individuals sampled and S. kibomena specimens in Figs 7 , 8 , respectively, for further comparison. The new species also strongly differs from S. kibomena in bioacoustics (see Table 2 and the comparison of the two species in the previous section). It differs specifically from S. kibomena by the following base pair differences in the analysed 5' fragment of the 16S rRNA gene: position 385 T vs. C and 547 A vs. T. It further differs from S. kibomena and all other Stumpffia species in our alignment by the base pair differences as listed above: position 480 C vs. T; 545 T vs. C or A; 141 T (Ambohitantely specimens) vs. C; 196-197 CC/CT deletion (Anjozorobe specimens); and 456 A insertion (Ambohitantely specimens).

Holotype description.

A specimen in a good state of preservation, except for skin loss where left thigh tissue muscle was removed for tissue sample. The body is elongated with the head wider than long, but not wider than the body. The snout is roughly rounded in both dorsal and lateral views. The nostrils are not protuberant and are closer to the tip of the snout than the eye. The tympanum is distinct and large (75% of eye diameter), and the supratympanic fold is indistinct. First finger short, others not reduced (Fig. 9 ). Inner metacarpal tubercle low, without a distinct prepollex, outer metacarpal tubercle indistinct and pale. Forelimbs are slender, the hand is without webbing, presents a relative finger length 1<2<4<3 with the fourth finger slightly longer than the second; and fingertips are not expanded into discs. First finger reduced; other fingers not reduced. The hind limbs are slender and the tibiotarsal articulation reaches the eye when the hind limb is adpressed along the body. There is no webbing between toes, and the relative toe length is 1<2<5<3<4 with the fifth toe slightly shorter than the third. First toe slightly reduced; other toes not reduced. Inner metatarsal tubercle oblong and indistinct, and outer metatarsal tubercle absent, lateral metatarsalia connected. The dorsal skin was slightly bumpy in life, without distinct dorsolateral folds, and the ventral skin was granular on the abdomen but smooth on the throat. The tongue is long, broadening posteriorly, attached anteriorly, not notched. Maxillary teeth and vomerine odontophores are absent, and the choanae are large and oval shaped.

Measurements (in mm). Snout-vent length 20.5, maximum head width 6.5, head length from tip of snout to posterior edge of snout opening 5.0, horizontal tympanum diameter 1.5, horizontal eye diameter 2.0, distance between anterior edge of eye and nostril 1.5, distance between nostril and tip of snout 0.5, internarial distance 2.0, forelimb length (from limb insertion to tip of longest finger) 12.0, hand length, to the tip of the longest finger 4.0, hind limb length (from the cloaca to the tip of the longest toe) 32.0, tibia length 9.0, foot length including tarsus 14.5, foot length 8.5 (Table 3 ).

In life the dorsum was pale brown, with dark brown markings (Fig. 5 ). These markings include a dark brown triangular marking between the eyes which extends posteriorly. A beige coloured band runs between the eyes at the base of the triangular marking. As in the S. kibomena holotype, two continuous narrow black bands extend posteriorly from level of the arms’ insertion, converging towards the mid-dorsum and then diverging posteriorly without touching towards the inguinal region (referred herein as dorsolateral bands). The throat is black with few scattered white spots and these white spots continue as the ventral side becomes bright red. The pupil is black with a copper iris and golden colouration at the top. All four limbs are coloured red on the ventral sides. After two years in preservative, the red ventral colouration has faded.

Variation.

Seven additional frogs were sampled for DNA and basic measurements were taken. Details of these frogs and their collection sites are shown in Table 1 and images in Fig. 7 . There is some variation in colour/patterns with some individuals being greyer on the dorsal side (rather than brown) (KAMUS60 and KAMUS74), also with dark brown markings. The markings of the juvenile are black. The dorsolateral bands are not always present, for example KAMUS167 has speckled markings. This individual also did not have the beige band across the head between the eyes. The female from Anjozorobe (KAMUS370), KAMUS256 and KAMUS200 from Ambohitantely had pale brown markings outlining the dark brown dorsolateral bands. All specimens have the same ventral colouration and markings (black throat, red belly, and white spots), though the tone of the belly varies from orange, dark orange to red (KAMUS167; see Fig. 7 ).

Call description.

The calls were recorded from a male (KAMUS371) at Anjozorobe during heavy rainfall on the 15th of February 2020, at 19:30, at an air temperature of 20.5 °C. The male was calling while sitting on the leaf litter on a slope in primary forest close to a female. The male was not collected as a specimen due to lack of permits to allow this, but swabs for DNA analysis were collected. The advertisement call of S. lynnae sp. nov. is structurally similar to that of other Stumpffia species in that it is a single melodious note that is repeated at regular intervals in call series. The call is simple, with the call composed of a slightly amplitude-modulated single tonal note, with multiple calls repeated in series at regular intervals. The calls are evenly spaced across the call series with silent intervals between calls. A definitive number of calls per call series cannot be determined given the sample size of one recording. Frequency is distributed across one band for each note, and frequency modulation is relatively equal across the note. No harmonics were seen in this call recording. The individual emitted eight calls, with a call-repetition rate of 0.25 calls per second. Call/note duration was 163-184 ms, and the duration of intervals between calls was 3498-5581 ms (n = 7). The dominant frequency range was 2027-2044 Hz (Table 2 ).

Etymology.

This species name is a matronym honouring Lynne Mullin, to whom we are pleased to dedicate this attractively colourful species in recognition of the unconditional support she has provided to the first author. The origin of Lynn/e is from Celtic language, with the meaning waterfall, pond, and lake. Given the popular waterfall in the centre of Ambohitantely Special Reserve where this species was first found, this name seems appropriate. The name has further relevance to this beautiful red-bellied frog with the Spanish meaning of the feminine name ‘pretty’. The species epithet is defined as a genitive noun with the ‘e’ removed for easier pronunciation.

Distribution.

While just eight individuals were recorded, the six at Ambohitantely were distributed across four fragments (three in addition to the core forest block) including a very small (3.5 ha) fragment at the southern end of the reserve (Fig. 1 ). This suggests that they are widely distributed across the protected area. Surveys in the two forest fragments at Ankafobe did not detect the species, but this is not surprising given the size of the fragments and the reduced diversity at Ankafobe compared to Ambohitantely ( Mullin et al. 2021). However, it cannot be ruled out that this species may exist in relict forest fragments in the area surrounding Ambohitantely Special Reserve, and between Ambohitantely and Anjozorobe. This species’ elevational range (1432-1586 m) is greater than S. kibomena 's range at Andasibe (900-950 m).

Natural history.

The encounters of the nine frogs (including the one individual found by Razafindraibe et al. 2021) were during both morning and evening surveys from December to May, suggesting they are active throughout the day and the wet season. No surveys have been undertaken outside of this wet season window, so their activity during the dry season is unknown. The holotype was resting on top of the leaf litter under a Pandanus sp. screwpalm in riparian habitat at an elevation of 1540 m a.s.l. All eight frogs from this study were found in slope or riparian forest, between 1 and 10 metres from a water source. All individuals were found on the forest floor; seven were on leaf litter and one was found on bare soil under a large rock. The individual recorded by Razafindraibe et al. (2021) was found guarding eggs in a water filled bamboo node. This was the first record of a Stumpffia laying eggs in a bamboo hole but supports evidence that closely related Stumpffia may reproduce opportunistically in water-filled cavities close to the forest floor ( Razafindraibe et al. 2021). The advertisement call was recorded during heavy rain in the evening during February. Combined with the eggs observed in December, this suggests that this species may be reproductively active throughout the wet season. Given the low number of individuals found across the high number of survey hours, we conclude that this species either has cryptic habitats, or is indeed rare.

Conservation.

This species is known to occur in two locations with different conservation situations. Ambohitantely Special Reserve, currently managed by Madagascar National Parks is highly fragmented, and is threatened by cattle grazing, illegal logging and forest activities, forest burning for charcoal, and forest fires and fire suppression activities (Goodman, Raherilalao and Wohlauser 2018 pp 1338-1340). The number of forest fragments surrounding the Reserve boundary has substantially declined since 1996 ( Vallan 2000b), as has the forest cover inside the reserve, having decline 6.3% from 1996-2016 ( Goodman et al. 2018). Reforestation efforts within the Reserve have been blighted by fires (KM pers. obs.). Further, an invasive caterpillar is currently causing canopy leaf loss across the Reserve (S. Goodman pers. comms) which has been noted to dry out the leaf litter which could be devastating for amphibians (KM pers. obs). Meanwhile, Anjozorobe is still connected to large expanses of continuous forest, and is managed by the local-led Association Fanamby; however, it is subject to similar threats and lost 33.2% of its forest between 1996 and 2016 ( Goodman et al. 2018: 1362-1365). Stricter conservation actions and management are required at both sites. We can assume that the populations of S. lynnae are declining due to the ongoing habitat loss and severe fragmentation of the populations. Given the very low numbers of individuals found during surveys we assume this species is naturally rare.

Under the IUCN Red List criterion B (Geographic range), we believe this species should be listed as Endangered for both B1, extent of occurrence (EOO), and B2, area of occupancy (AOO) ( IUCN 2012). Its EOO is <5,000 km2 at 530 km2; however, much of this area is inhospitable savannah grasslands and farmland between Ambohitantely and Anjozorobe. More relevant is its AOO, which is <500 km2, estimated to be ~ 100 km2. The species fulfils two further criteria to be considered Endangered. B(a) 'severely fragmented OR number of locations’ as it exists in severely fragmented locations/populations with no connectivity between the fragments at Ambohitantely, or between Ambohitantely and Anjozorobe. It was found in two locations across five forest fragments, one of which was just 3.5 hectares in size. Given that all the forest fragments it was found in are frequently burned and logged, they are predicted to reduce in size, with some disappearing in the near future. This degradation will influence all of the subcriteria within B(b) 'continuing decline observed, estimated, inferred or projected in any of; (i) extent of occurrence; (ii) area of occupancy; (iii) area, extent and/or quality of habitat; (iv) number of locations or subpopulations; and (v) number of mature individuals.' This suggested Red List status is in line with two microendemic amphibian species that are found at Ambohitantely, Anilany helenae and Anodonthyla vallani , which are both listed as Critically Endangered given they are found in just Ambohitantely Special Reserve ( A. vallani ) (IUCN 2020a) and Ambohitantely and surrounding isolated forest fragments ( A. helenae ) (IUCN 2016, 2020b; Mullin et al. 2021).

Figures

Figure 1. Map of Ambohitantely Special Reserve showing known Stumpffia lynnae sp. nov. distribution. A Ambohitantely, Anjozorobe, and Andasibe relative to one another B map of Madagascar showing the wider distribution of Stumpffia lynnae sp. nov. and its sister species Stumpffia kibomena, Stumpffia sp. Ca 11 and Stumpffia sp. Ca 34. Satellite imagery Google Earth (2015).

Figure 5. Images of the holotype ZSM 1 / 2022 (KAMU 2) from Ambohitantely Special Reserve in life A, B dorsal lateral view C ventral view D dorsal view.

Figure 6. Direct comparison between the holotype of S. lynnae ZSM 1 / 2022 (KAMU 2) and a female S. kibomena specimen from An'ala (photo Frank Glaw). Boxes highlight the morphological differences highlighted; elongated red marking on each side of the lower jaw on S. kibomena, absent on the observed S. lynnae, and the red arms on S. kibomena compared to the brown arms on S. lynnae.

Figure 7. Images of the additional observed specimens of Stumpffia lynnae sp. nov. including the calling male (KAMUS 371).

Figure 8. Images of Stumpffia kibomena (top row from the species description of Glaw et al. (2015) and the bottom row from inaturalist. org taken by Leonard Bolte and Jordan Broadhead, used with permission). Note the red marking on the lower lips and the red arms which differ from S. lynnae sp. nov.

Figure 9. Holotype hands and feet of S. lynnae (row A) and S. kibomena (row B). Not to scale.