Diversity of Ptychadena in Rwanda and taxonomic status of P. chrysogaster Laurent, 1954 (Amphibia, Anura, Ptychadenidae).

We assess the diversity of Ptychadena species in Rwanda based on re-examination of voucher specimens in museum collections and our own data from recent assessment of the species composition of amphibian communities in Rwanda. We recognize five species which we allocate to the following available names: P. anchietae, P. chrysogaster, P. nilotica, P. porosissima, and P. uzungwensis. We did not find evidence for the presence of P. grandisonae and P. oxyrhynchus which have been listed for the country. The five species can be distinguished by quantitative morphometrics (discriminant analysis, success rate: 100 %) and a number of qualitative characters of external morphology. We provide an identification key to the Rwandan species and describe the morphology of each species in detail. The taxonomic status and the phylogenetic position of Ptychadena chrysogaster are further assessed based on the partial sequence of the mitochondrial 16S rRNA. The species differs genetically from available homologous sequences from congeners by an uncorrected p distance of at least 4.2 % and appears to be most closely related to specimens assigned to P. porosissima, P. mahnerti, "P. aff. uzungwensis" and "P. aff. bibroni".

Introduction

Ridged Frogs of the genus Boulenger, 1917 are widespread in sub-Saharan Africa where approximately 50 species occur. Species of the genus share a similar general appearance and many are poorly delimited, having been described based on taxonomically doubtful characters. Several species names have been erroneously considered synonyms of others, thus confusing character diagnoses in subsequent accounts; and some taxa were described based on specimens later found to represent more than one species (e.g. Boulenger 1879; Loveridge 1932; Laurent 1954; Guibé and Lamotte 1957; Schmidt and Inger 1959; Lamotte 1967; Poynton 1970; Rödel 2000; Channing 2001; Channing and Howell 2006; Dehling and Sinsch in press). Therefore, even the local/regional diversity of these frogs is often difficult to assess. Herein, we address the diversity of Ridged Frogs in Rwanda. We have recently shown that three species ( [Bocage, 1868], [Seetzen, 1855], and [Steindachner, 1867]) inhabit the wetlands along the upper Nile (Dehling and Sinsch in press). Further taxa have been reported from Rwanda and it is currently unclear which species actually occur in the country. Nieden (1913) reported (referred to as Duméril & Bibron, 1841) from several localities in Rwanda. Based on his own collections, Laurent (1954) reported (Loveridge, 1932) and described Laurent, 1954 and Laurent, 1954 as new species, the latter now being considered a synonym of (Schmidt and Inger 1959). Poynton and Broadley (1985), Channing (2001), and Poynton and Channing (2004) stated that Laurent, 1954 occurs in Rwanda. Fischer and Hinkel (1992) listed only “” [= ]. Poynton et al. (2004) stated that was likely to occur in Rwanda but confirmed records were missing. According to Spawls et al. (2006), , , and occur in Rwanda but not and . Branch (2005) included Rwanda in the geographic range of (Smith, 1849). Recently, we collected in Rwanda and resurrected the name for the populations which occur along the Nile and in Central Kenya and Tanzania and which had been formerly referred to as or (Sinsch et al. 2012, Dehling and Sinsch in press).

In order to clarify how many and which species occur in Rwanda, we re-examined the specimens of in the herpetological collection of the Royal Museum for Central Africa in Tervuren, Belgium (RMCA), on which almost all previous Rwandan records are based. We herein report the results and compare the findings to our own data from recent assessment of the composition of amphibian communities at numerous locations in Rwanda. We further assess the taxonomic status and the phylogenetic position of based on examination of most of the available voucher material from Rwanda including the type series and on comparison of the partial sequence of the mitochondrial 16S rRNA gene with homologous sequences of its congeners.

Material and methods

Morphological examination

We examined voucher specimens deposited at RMCA. Additional specimens including our recently collected material are deposited in the collection of the . See Appendix 1 for a complete list of examined specimens.

Zoologisches Forschungsmuseum Alexander Koenig, Bonn, Germany

For the morphological analysis, we took the following 18 measurements to the nearest 0.1 mm using digital calipers, following Dehling and Sinsch (in press): (1) Snout-vent length (SVL); (2) tibiofibula length (TFL, measured with both knee and tibio-tarsal articulation flexed); (3) thigh length (THL, from vent to knee with thigh being held vertically to median body plane and knee flexed); (4) total hindlimb length (LEG, from vent to tip of fourth toe with leg fully extended and being held vertically to median body plane); (5) tarsus + foot length (TarL, from tibio-tarsal articulation to tip of fourth toe); (6) foot length (FOT, from proximal end of inner metatarsal tubercle to tip of fourth toe); (7) forearm + hand length (ARM, distance from elbow to tip of third finger); (8) hand length (HND, distance from proximal end of inner palmar tubercle to tip of third finger); (9) head width (HW, measured at the level of the jaw joint); (10) head length (HL, distance from posterior end of mandible to tip of snout); (11) interorbital distance (IO, shortest distance between upper eyelids); (12) upper eyelid width (EW); (13) horizontal eye diameter (ED); (14) horizontal tympanum diameter (TD); (15) eye to nostril distance (EN, distance between anterior margin of eye and centre of nostril); (16) nostril to snout distance (NS, distance between centre of nostril and tip of snout); (17) snout length (SL, distance between anterior margin of eye to tip of snout); (18) internarial distance (NN, distance between centres of nostrils). To avoid an inter-observer bias, all measurements were taken by JMD. Additionally, we recorded the following qualitative characters: (1) position of external vocal sac aperture in males; (2) number of longitudinal dorsal dermal ridges; (3) texture of ventral skin; (4) extent of nuptial pads in males; (5) number of supernumerary metacarpal tubercles; (6) size and shape of thenar and palmar tubercles; (7) extent of toe webbing; (8) presence of outer metatarsal tubercle; (9) relative size of inner metatarsal tubercle; (10) ventral colouration; (11) presence of light line on dorsal face of tibia; (12) presence of light band on dorsum; (13) presence of dark brown stripe on preaxial side of tarsus; (14) colour of external dorsal fold; (15) colour pattern on postaxial side of femur. Sex of males was determined by presence of secondary sexual characteristics (vocal slits, nuptial pads), that of females by either examination of gonads through dissection or size (female if larger than smallest 10 percent of adult males). The webbing formulae are given as proposed by Myers and Duellman (1982). Terminology for dermal dorsal ridges and orientation of external vocal sac aperture follows Perret (1979).

Statistical Analyses

Descriptive statistics depended on the outcome of the test for normality. Normally distributed data were described by the arithmetic mean and corresponding standard error and/or range, those deviating significantly by median and range. Principal component analyses were run on the morphometric data set including 18 variables and 89 observations each (: 15 males, 3 females; : 13 males, 10 females; : 13 males, 10 females; : 11 males, 7 females; : 6 males, 1 female). We compared the scores obtained for the principal components 2 and 3 describing shape to distinguish taxa without a priori assignment to taxa. The morphometric distances were adjusted for SVL by calculating a linear regression of each variable against SVL and storing the residuals as representatives of size-independent shape variables. This transformed data set was used for discriminant analyses with taxa as predefined groups to optimize distinction. To account for sexual dimorphism, discriminant analyses were run separately for males (n=58) and females (n=31). Significance level was set at alpha = 0.05. All calculations were based on the procedures of the program package STATGRAPHICS centurion for Windows, version XV.

DNA barcoding and phylogenetic analyses

We isolated DNA from a liver tissue sample from a specimen of (ZFMK 58797), collected in southern Rwanda by H. Hinkel in 1993. DNA was used to sequence a fragment of the 16S mitochondrial rRNA gene, a universal marker to barcode amphibian species (Vences et al. 2005). Protocols of DNA extraction, PCR, purification, and sequencing follow Dehling and Sinsch (in press). The obtained sequence was compared with those in GenBank using a standard nucleotide-nucleotide BLAST search and with our own sequences from Rwandan specimens and was incorporated into an existing alignment (see Dehling and Sinsch in press for a list of sequences and GenBank Accession numbers). Editing and alignment were completed in MEGA5 (Tamura et al. 2011). Sequences were trimmed to the same length. The final alignment consisted of 548 base pairs. Calculations of pairwise distances and phylogenetic analysis (Maximum Likelihood) were carried out in MEGA5. Maximum Likelihood analysis was run using the GTR + G + I model and the Nearest-Neighbor-Interchange with 1000 bootstrap replicates.

Results

Examination of specimens suggested that five morphologically distinct species were present in Rwanda to which we assign the following names: PageBreak, , , , and (Figures 1 and 2). For allocation of specimens to , , and and discussion thereof see Dehling and Sinsch (in press). The examined material included type specimens of both of and (Appendix 1). Allocation of other specimens to the latter two species is based on direct comparison with the type material. We re-assigned several specimens that had been deposited in the museum collections under wrong names. Noteworthy are two of the paratypes of (RMCA 41989, 41994) which belong in fact to .

Figure 1.

Males of from Rwanda in life. A B [Foto: E. Fischer] C D .

Figure 2.

A Preserved female holotype of (RMCA 109096) from Lac Karago, Rwanda; dorsal view (left) and ventral view (right) B Preserved male specimen of (RMCA 108993-108997) from Munini, Rwanda; dorsal view (left) and ventral view (right). Not to scale.

Morphological differentiation

The morphometric features of the five species are summarized in Table 2. Principal component analysis yielded three PCs accounting for 89.6% of total variation (Table 3A). PC1 represented variation in size, whereas the shape-related PC2 and PC3 were mainly loaded by features describing head morphology (Table 3B). In females, PC 2 unequivocally distinguished from the other taxa, and PC 3 unequivocally distinguished from , , and (Figure 3A). Also, and could be distinguished from each other. However, both species were represented by only few individuals (three and one, respectively) in the analysis. The two species did not differ significantly in shape from (Figure 3A). A similar pattern was observed in the analysis of the males (Figure 3B) but males of all five species were generally more similar to each other in shape. Males of could be distinguished unequivocally from males of and but not from some of the males of and (Figure 3B). Males of could be distinguished from males of all other species but some of the males of and were very similar in shape (Figure 3B). Males of did not differ significantly in shape from males of , , and . Gender-specific discriminant analyses based on the residuals of 17 SVL-adjusted morphometric variables had a classification success of 100 % among the five species in both males and females (Table 4A, B, C, Figure 4A, B).

Table 2.

Morphometric features of species from Rwanda. Data are given as arithmetic means and minimum and maximum values (in mm).

Morphometric character	Ptychadena anchietae		Ptychadena chrysogaster		Ptychadena nilotica		Ptychadena porosissima		Ptychadena uzungwensis
	Males<br/> N = 15	Females<br/> N = 3	Males<br/> N = 75/14*	Females<br/> N = 23/11*	Males<br/> N = 13	Females<br/> N = 10	Males<br/> N = 11/20*	Females<br/> N = 9	Males<br/> N = 6	Females<br/> N = 1
Snout-vent length	40.4<br/> (38.0–42.4)	49.0<br/> (46.7–51.3)	43.3<br/> (36.3–49.5)	53.7<br/> (48.0–57.7)	42.0<br/> (37.2–45.2)	49.1<br/> (45.6–53.1)	41.2*<br/> (37.3–44.5)	46.4<br/> (39.0–52.1)	34.7<br/> (33.3–35.7)	43.3<br/> -
Hindlimb length	79.6<br/> (74.2–84.9)	98.9<br/> (96.0–101.2)	88.1<br/> (83.5–93.0)	108.2<br/> (102.3–114.1)	77.2<br/> (70.1–85.2)	89.6<br/> (78.0–103.9)	78.2*<br/> (72.8–85.5)	88.3<br/> (74.3–94.1)	66.7<br/> (61.9–72.7)	81.2<br/> -
Femur length	23.0<br/> (21.9–24.5)	29.0<br/> (28.3–29.6)	23.9<br/> (22.5–25.0)	29.9<br/> (28.1–32.0)	21.5<br/> (19.4–24.0)	25.5<br/> (23.0–28.5)	21.7*<br/> (20.2–24.1)	24.4<br/> (19.8–27.6)	18.7<br/> (17.2–19.6)	23.1<br/> -
Tibiofibula length	26.3<br/> (24.4–28.0)	33.1<br/> (31.9–33.8)	28.4<br/> (24.6–32.0)	35.2<br/> (32.4–38.5)	23.4<br/> (21.1–26.1)	27.5<br/> (23.6–32.1)	24.7*<br/> (23.3–26.5)	28.7<br/> (25.2–31.0)	21.7<br/> (20.3–23.4)	26.9<br/> -
Tarsus length	34.5<br/> (31.6–36.5)	42.7<br/> (40.6–44.1)	40.9*<br/> (38.9–44.1)	49.3*<br/> (46.5–51.3)	36.2<br/> (32.4–40.1)	42.8<br/> (35.9–49.6)	35.5<br/> (33.3–38.8)	40.0<br/> (32.6–43.5)	30.5<br/> (27.9–32.2)	36.0<br/> -
Foot length	24.7<br/> (22.4–26.1)	30.5<br/> (29.1–31.3)	28.4<br/> (24.2–30.4)	34.5<br/> (32.9–36.1)	25.7<br/> (22.8–28.2)	29.3<br/> (25.6–33.9)	24.4<br/> (23.0–26.7)	27.4<br/> (23.1–29.3)	21.0<br/> (19.5–21.9)	28.6<br/> -
Forelimb length	16.9<br/> (15.9–18.1)	21.0<br/> (20.0–21.7)	18.3*<br/> (17.4–19.1)	22.0*<br/> (20.9–23.9)	17.7<br/> (16.0–19.3)	20.5<br/> (17.9–23.7)	17.2<br/> (16.0–18.6)	18.7<br/> (16.1–20.9)	13.5<br/> (12.8–14.4)	16.3<br/> -
Hand length	10.0<br/> (9.5–10.7)	12.3<br/> (12.1–12.6)	10.4*<br/> (9.6–11.1)	12.4*<br/> (11.8–13.3)	10.4<br/> (9.5–11.6)	11.9<br/> (10.6–13.8)	9.7<br/> (8.9–10.8)	10.7<br/> (9.1–11.7)	7.9<br/> (7.4–8.5)	9.0<br/> -
Head width	13.8<br/> (12.5–15.4)	16.7<br/> (16.4–17.0)	13.9<br/> (12.8–14.7)	16.9<br/> (15.4–17.5)	13.8<br/> (12.4–15.9)	16.4<br/> (14.2–18.8)	14.1<br/> (13.1–15.0)	15.1<br/> (12.5–17.1)	11.4<br/> (11.0–11.7)	13.8<br/> -
Head length	15.5<br/> (14.1–17.6)	18.7<br/> (18.2–19.1)	15.6*<br/> (14.7–16.4)	18.6*<br/> (17.9–19.8)	16.2<br/> (14.7–18.4)	18.3<br/> (16.6–20.4)	15.5<br/> (14.2–17.8)	17.1<br/> (13.6–19.1)	13.3<br/> (12.3–14.0)	16.1<br/> -
Interorbital distance	2.8<br/> (2.4–3.1)	3.1<br/> (3.0–3.2)	3.5*<br/> (3.1–3.9)	4.1*<br/> (3.8–4.5)	2.2<br/> (1.9–2.5)	2.5<br/> (2.2 - 2.7)	2.6<br/> (2.3–2.7)	3.1<br/> (2.8–3.6)	2.7<br/> (2.4–3.1)	2.9<br/> -
Eyelid width	2.8<br/> (2.5–3.1)	3.2<br/> (3.0–3.3)	2.7*<br/> (2.4–3.0)	3.3*<br/> (2.9–3.7)	2.7<br/> (2.3–3.2)	3.0<br/> (2.4–3.5)	2.8<br/> (2.3–3.1)	3.1<br/> (2.6–3.5)	2.6<br/> (2.3–2.8)	2.4<br/> -
Eye diameter	4.3<br/> (3.6–4.9)	5.2<br/> (5.0–5.5)	4.3<br/> (3.7–4.9)	5.0<br/> (4.6–5.6)	4.5<br/> (4.1–5.1)	5.1<br/> (4.8–5.5)	4.2<br/> (3.9–4.5)	4.6<br/> (3.9–5.1)	3.9<br/> (3.6–4.1)	4.5<br/> -
Tympanum diameter	3.3<br/> (2.9–3.6)	4.1<br/> (3.7–4.6)	3.7<br/> (3.2–4.2)	4.5<br/> (4.3–5.1)	3.7<br/> (3.3–4.1)	4.1<br/> (3.7–4.7)	3.1<br/> (2.9–3.49	3.6<br/> (3.3–3.9)	2.9<br/> (2.5–3.3)	3.5<br/> -
Eye–nostril distance	4.1<br/> (3.7–4.3)	5.2<br/> (4.9–5.7)	3.9*<br/> (3.5–4.4)	4.6*<br/> (4.0–5.0)	3.6<br/> (3.4–4.0)	4.2<br/> (3.7–4.9)	3.5<br/> (3.1–3.9)	4.2<br/> (3.6–4.7)	3.5<br/> (3.2–3.6)	4.2<br/> -
Snout–nostril distance	3.5<br/> (2.9–4.0)	4.4<br/> (4.3–4.5)	3.9*<br/> (3.5–4.3)	4.3*<br/> (3.9–4.9)	3.4<br/> (2.9–3.8)	3.7<br/> (3.3–4.1)	3.4<br/> (2.9–3.7)	4.1<br/> (2.8–4.7)	3.6<br/> (3.2–3.9)	4.3<br/> -
Internarial distance	3.9<br/> (3.5–4.3)	4.8<br/> (4.7–4.9)	4.3*<br/> (4.0–4.5)	4.9*<br/> (4.5–5.2)	3.4<br/> (3.0–3.6)	3.7<br/> (2.0–4.5)	3.6<br/> (3.4–4.0)	4.3<br/> (3.7–4.7)	3.0<br/> (2.9–3.2)	3.7<br/> -
Snout length	7.5<br/> (6.6–8.1)	9.5<br/> (9.0–10.1)	7.4*<br/> (7.0–7.9)	8.7*<br/> (8.4–9.4)	7.0<br/> (6.3–7.7)	7.9<br/> (7.1–9.3)	7.0<br/> (6.3–7.8)	8.1<br/> (6.2–10.1)	6.7<br/> (6.4–7.0)	8.0<br/> -

Table 3.

Principal component Analysis based on 18 standardized morphometric features of 89 specimens belonging to five species from Rwanda. Morphometric parameters accounting strongly for discrimination among species are highlighted in bold.

A: Statistical significance
Principal component	Eigen-value	Relative percentage	Cumulative percentage
1	13.84	76.9	76.9
2	1.46	8.1	85.0
3	0.82	4.6	89.6
B: Standardized coefficients of the principal components
Parameter	Principal component 1	Principal component 2	Principal component 3
Snout-vent length	0.256	-0.107	-0.084
tibiofibula length	0.253	0.212	-0.011
foot length	0.255	0.046	-0.278
tarsus + foot length	0.257	0.056	-0.246
total hindlimb length	0.261	0.109	-0.137
thigh length	0.260	0.048	-0.033
forearm + hand length	0.253	-0.164	-0.210
hand length	0.244	-0.231	-0.139
head width	0.243	-0.224	-0.016
head length	0.232	-0.319	0.105
interorbital distance	0.171	0.578	-0.237
upper eyelid width	0.201	-0.142	0.272
horizontal eye diameter	0.211	-0.325	0.162
horizontal tympanum diameter	0.231	-0.112	-0.284
eye to nostril distance	0.228	0.016	0.353
nostril to snout distance	0.195	0.326	0.412
snout length	0.234	0.070	0.469
internarial distance	0.227	0.324	0.072

Figure 3.

Morphological shape differentiation among 89 specimens representing five species, as assessed by principal component analysis (Table 3). A Individual scores obtained for 31 females B Individual scores obtained for 58 males.

Table 4A.

Gender-specific discriminant functions based on 17 SVL-adjusted morphometric features (residuals) to distinguish among five species from Rwanda. Statistical significance:

Discriminant function	Eigen-value	Relative percentage	Canonical correlation	Wilks Lambda	Chi-squared	Degrees of freedom	Statistical significance
Male 1	19.79	58.76	0.975	0.0003	362.2	68	P < 0.0001
Male 2	8.49	25.23	0.945	0.0079	222.6	48	P < 0.0001
Male 3	3.29	9.77	0.875	0.0750	119.0	30	P < 0.0001
Male 4	2.10	6.24	0.823	0.3223	52.0	14	P < 0.0001
Female 1	45.49	65.38	0.989	0.00005	187.7	68	P < 0.0001
Female 2	14.84	21.33	0.967	0.0023	114.7	48	P < 0.0001
Female 3	6.87	9.89	0.934	0.0377	62.2	30	P < 0.0001
Female 4	2.36	3.40	0.838	0.2973	23.0	14	P = 0.0596

Table 4B.

Gender-specific discriminant functions based on 17 SVL-adjusted morphometric features (residuals) to distinguish among five species from Rwanda. Morphometric parameters accounting strongly to discrimination among species are highlighted in bold. Standardized coefficients of the discriminant functions:

parameter (residuals)	discriminant function 1<br/> (males)	discriminant function 2<br/> (males)	discriminant function 3<br/> (males)	discriminant function 4<br/> (males)	discriminant function 1<br/> (females)	discriminant function 2<br/> (females)	discriminant function 3<br/> (females)	discriminant function 4<br/> (females)
tibiofibula length	0.909	-0.221	0.302	0.871	0.824	-0.143	0.196	0.384
foot length	0.268	-0.253	0.292	-1.332	2.079	-3.834	-2.308	-0.942
tarsus + foot length	-0.798	2.031	-0.026	1.308	-3.311	4.158	1.119	-0.410
total hindlimb length	0.310	-0.029	-0.147	-0.929	-0.617	0.188	0.361	0.999
thigh length	0.161	-1.001	0.367	-0.108	0.750	0.009	-0.268	0.196
forearm + hand length	-0.172	0.142	-1.406	-0.005	0.468	-0.492	0.483	0.105
hand length	-0.669	-0.506	0.383	-0.189	-0.879	-0.007	0.088	0.328
head width	0.009	-0.137	-0.370	0.326	-0.778	0.997	0.472	0.534
head length	-0.282	-0.341	0.252	-0.389	-0.293	-1.250	-0.563	-0.734
interorbital distance	0.366	0.144	0.285	0.136	1.030	0.561	-0.269	-0.019
upper eyelid width	-0.022	0.315	-0.221	0.328	1.579	0.473	0.810	-0.234
horizontal eye diameter	-0.296	-0.048	0.420	-0.193	1.013	0.232	0.808	0.089
eye to nostril distance	0.417	-0.532	0.473	-0.287	1.016	-0.555	0.242	-0.166
nostril to snout distance	0.072	0.529	0.225	0.124	1.162	-0.599	0.727	-0.818
snout length	-0.291	-0.063	-0.123	0.422	-0.176	0.035	-0.337	0.767
internarial distance	0.666	0.135	-0.861	-0.567	0.594	0.087	-0.026	-0.213
horizontal tympanum diameter	-0.052	-0.157	0.707	-0.187	-1.365	0.088	-0.621	0.488
Constant	0.909	-0.222	0.302	0.871	0.824	-0.143	0.196	0.384

Table 4C.

Gender-specific discriminant functions based on 17 SVL-adjusted morphometric features (residuals) to distinguish among five species from Rwanda. Classification success.

predicted species<br/> actual species	Ptychadena anchietae	Ptychadena chrysogaster	Ptychadena nilotica	Ptychadena porosissima	Ptychadena uzungwensis
Ptychadena anchietae. male<br/> female	15 (100%)<br/> 3 (100%)	0	0	0	0
Ptychadena chrysogaster. male<br/> female	0	13 (100%)<br/> 10 (100%)	0	0	0
Ptychadena nilotica. male<br/> female	0	0	13 (100%)<br/> 10 (100%)	0	0
Ptychadena porosissima. male<br/> female	0	0	0	11 (100%)<br/> 7 (100%)	0
Ptychadena uzungwensis. male<br/> female	0	0	0	0	6 (100%)<br/> 1 (100%)

Figure 4.

Morphological shape differentiation among 89 specimens representing five species, as assessed by discriminant analyses (Statistical details are given in Table 4). A Individual scores obtained for 31 females B Individual scores obtained for 58 males.

The five Rwandan species can be distinguished unequivocally from each other using a combination of qualitative morphological characters (Table 1, Figure 5; see also Dehling and Sinsch in press). An identification key based on these characters is given below. Detailed morphological descriptions of the species are in Appendix 2.

Table 1.

Distinguishing qualitative characters of species from Rwanda.

Species	Ptychadena anchietae	Ptychadena chrysogaster	Ptychadena nilotica	Ptychadena porosissima	Ptychadena uzungwensis
relative length of Toes III and V	tips reaching to knee or slightly beyond, distal subarticular tubercle never reaching knee	tips reaching beyond knee, distal subarticular tubercle reaching knee	tips reaching beyond knee, distal subarticular tubercle reaching knee or beyond	tips reaching to knee or slightly beyond, distal subarticular tubercle never reaching knee	tips reaching to knee or slightly beyond, distal subarticular tubercle never reaching knee
position of vocal sac aperture	inferior, at ventral edge of arm insertion	inferior, at ventral edge of arm insertion	superior, above dorsal edge of arm insertion	inferior, at ventral edge of arm insertion	semi-inferior, at level of centre of arm insertion
spiny tubercles on venter	absent	present in males, very small	absent	present in males, comparatively large	present in males, very small
median dorsal ridge on snout	absent	absent	absent	absent	present
outer metatarsal tubercle	very faintly visible	very faintly visible, rarely distinct	distinctly present, rarely faintly visible	faintly visible, rarely distinct	faintly visible
inner metatarsal tubercle size (Fig. 5)	about half the length of metatarsus of Toe I	less than half the length of metatarsus of Toe I	less than half the length of metatarsus of Toe I	more than half the length of metatarsus of Toe I	about half the length of metatarsus of Toe I
supernumerary metacarpal tubercles (Fig. 5)	one below each finger	one below each finger	only one below Finger IV, often indistinct	one below Fingers I, II, and IV; two, rarely one below Finger III	one below Fingers I and IV, two below Finger II, two to four below Finger III
palmar and thenar tubercles (Fig. 5)	inner and outer palmar tubercle more or less equal in length; thenar tubercle oval, slightly longer than palmar tubercles	outer palmar tubercle longer than inner; thenar tubercle elongate, about as long as outer palmar tubercle	outer palmar tubercle longer than inner; thenar tubercle elongate, about as long as outer palmar tubercle	outer palmar tubercle longer than inner; thenar tubercle elongate, longer than outer palmar tubercle	inner and outer palmar tubercle more or less equal in length; thenar tubercle elongate, longer than palmar tubercles
toe webbing (Fig. 5)	I0.5-2II0.5-2III(0.5-1)-2IV2-0.5V	I2-2.5II(1.5-1.75)-3III(2-2-)(3.25-3+)IV3-(1.5-2)V	I(1.5-1.75)-(2-2.25)II1.5-(2.75-3)III(1.75-2)-3IV2.75-(1-1.5)V	I(1.75-2)-2.25II1.5-3III1.75-(3-3.25)IV3-(1-1.5)V	I2-(2.25-2.5)II1.5-3-III(1.75-2-)-3IV3-(1+-1.25)V
ventral colouration	head white, trunk yellow	head and trunk yellow	head white, mottled with grey; trunk yellow	head and trunk yellow	colours in life unreported
dark brown stripe on preaxial side of tibia	absent	present, continuous or almost continuous	absent in most specimens; few specimens with dark mottling, not forming continuous stripe	absent	absent
light tibial line (Figs 1 & 2)	absent	usually present, rarely absent	present or absent	present	absent
light dorsal band	absent	usually present, rarely absent	present or absent	present or absent	present
dark spots on dorsum<br/> (Figs 1 & 2)	usually absent; if present, small and narrow	usually present, small and narrow, sometimes forming longitudinal lines; rarely absent	present, large and wide, sometimes fused with neighboring ones	present, large and wide, sometimes fused with neighboring ones	present, large and wide, often fused with neighboring ones
light, prominent dorsolateral fold (Figs 1 & 2)	usually absent	present	present	present	present
Colour pattern on postaxial side of femur	irregularly delimited, reticulated, longitudinal bands, alternately yellow and dark brown coloured	irregularly delimited, reticulated, longitudinal dark bands on light background; colours in life unreported	relatively sharply delimited longitudinal bands, alternately yellow and black coloured	yellow spots diffusely arranged in longitudinal rows on dark brown background	irregularly delimited, reticulated, longitudinal light bands on dark background; colours in life unreported

Figure 5.

Volar view of hands (top) and plantar view of feet (bottom) of males of (a), (b), (c), (d), and from Rwanda. See also Table 1.

Phylogenetic analyses

Comparison of the mitochondrial 16S rRNA gene sequences corroborated the status of PageBreakPageBreakPageBreakPageBreakPageBreakPageBreakPageBreakPageBreak as a distinct species. The partial sequence of this species differed from all available comparative sequences by an uncorrected p distance of at least 4.2 %. The p distance to sequences from Rwandan specimens of and was 13.1–13.3 % and 13.6 %, respecitively. The lowest values were observed in comparison with specimens assigned to “ aff. uzungwensis” (4.2 %), “” from South Africa and Rwanda (4.7–4.9 %), “ aff. porosissima” from Tanzania (6.0–6.9 %), “” (6.2 %), and “P. aff. bibroni” from Gabon (6.9 %). The consensus tree yielded by Maximum-Likelihood analysis indicated that is most closely related to the aforementioned species (Figure 6). The clade consisting of these species is well supported by bootstraping (value 0.90; Hillis and Bull 1993), whereas the relationships within the clade are not resolved (bootstrap values <50 %).

Figure 6.

Maximum likelihood phylogram of species in the genus and as outgroup, based on comparison of 548 base pairs of the mitochondrial 16S rRNA gene. Included are specimens from Rwanda and samples taken from GenBank (see Dehling and Sinsch in press for a complete list of sequences and accession numbers). Numbers above nodes are bootstrap support values (only values >0.50 are shown).

Discussion

Among the examined material we identified five distinct species of : , , , , and . The five species are distinguishable from each other unambiguously using quantitative morphometric as well as qualitative morphological characters. The comparison of the partial 16S rRNA sequence of a specimen of with sequences from congeners corroborated its distinct specific status. Sequences from specimens of , , and from Rwanda differ from each other considerably by an uncorrected p distance of more than 10 % (Sinsch et al. 2012, Dehling and Sinsch in press). Unfortunately, no homologous sequence of from Rwanda and only a sequence of a specimen with doubtful identity from Tanzania ( aff. uzungwensis, GenBank# DQ525945) were available for comparison.

We did not find any PageBreakPageBreak individual collected in Rwanda which was assignable to in the collections of the RMCA and the ZFMK. Three specimens from Kisenyi (= Gisenyi, nowadays Rubavu; RMCA 51565–67), Rwanda, had been deposited under the name but were re-identified as males of . Nieden (1913) reported on several specimens of (as ) which Schubotz had collected in what today is northwestern Tanzania, at Kifumbiro and in the Mpororo area, close to the present border with Rwanda. We have not examined Schubotz’ material but if his specimens are indeed it is possible that the species can be found in Rwanda as well, given its vast distribution in eastern Africa and the fact that the herpetofauna of the northeastern part of Rwanda has been poorly sampled so far.

There is no specimen of among the material Laurent collected in Rwanda. Laurent (1954) described the species based on type specimens from Muita in Angola, from Kanzenze and Kansenia in Katanga (DRC), and from Bitare in “Urundi” [= Burundi]. The latter is a town in central Burundi, about 20 km north of Gitega at 3°15'S, 29°54'E. Several authors, however, have stated that occurs in Rwanda (Poynton and Broadley 1985, Channing 2001, Poynton and Channing 2004), and Poynton and Channing (2004) even stated that there is no record from Burundi. This misinformation was very likely caused by Laurent himself in a paper which was cited by all above mentioned authors, an account on the “Reptiles et Amphibiens de l’Angola” (Laurent 1964). Therein, Laurent (1964: 139) cited one of the type localities of wrongly as “Bitare (Ruanda)”. On the same page, the locality is correctly given as “Bitare […] (Urundi)” in the account on . Thus, the often cited record of from Rwanda in fact refers to specimens from Burundi (RMCA 109036–37; Appendix 1). So far, there is no evidence for the occurrence of in Rwanda.

The available evidence indicates that only five species of PageBreak occur in Rwanda. At present, three of these species (, , and ) are widespread and can be found abundantly in both wetlands of the eastern lowland between 1300 and 2000 m elevation which drains into the Nile River and the western lowland on the shore of Lake Kivu which drains into the Congo River. The species inhabit higher elevations of up to 2300 m in deforested, cultivated areas, but are absent from dense forest habitats at similar elevations which at present only remain in the Volcano and Nyungwe National Parks and in the Gishwati Forest. is known from Rwanda from only few specimens, five males from “Kumunini” [= Munini, South Province, 2°42'S, 29°32'E] and a female from “Astrida” [= Butare/Huye, South Province, 2°36'S, 29°44'E], collected in 1952 and 1951, respectively, and has not been found since. Assuming the species is still extant in Rwanda, its distribution is apparently restricted to the south of the country.

There are large series of from various localities in Rwanda in the collection of the RMCA (Appendix 1), collected by Laurent in 1951–1952, indicating that the species was abundant at that time. We repeatedly conducted surveys at several of these localities including the type locality at Lac Karago (1°37'S, 29°30'E) but did not encounter individuals of . Our survey periods (February to April, September to October) were at similar times of the year to those of Laurent (Janurary, February, and October). Species of are among the most conspicuous frogs in areas they inhabit, usually occurring in high numbers and easy to detect. Although the absence of a species from a certain area cannot be proven ultimately, our observations indicate that has disappeared from these areas or at least is much less common than it used to be. The human population in Rwanda has grown from little more than 2 million people in 1950 to approximately 11 million in 2011 (United Nations, Population Division 2011). Nowadays almost every cultivatable area except the three national parks and few small forest patches has been altered to farmland (pers. observation). Gishwati Forest has been reduced to a small patch of a few square kilometres, but until the mid-1990s it had covered a large area in northwestern Rwanda and its extensions reached the shores of Lac Karago. The former presence of forest habitat at the lake is still indicated by the occurrence of two forest-dwelling frog species, and ,which call from bushes and groups of small trees at the shore of the river (own unpublished data; see also Sinsch et al. 2011). Judging from its collection sites, appears to occur primarily in wetlands within or at the edge of forest. Instead of , we found at Lac Karago and and in Huye (formerly Astrida and Butare) and in the vicinity of Muzanze (formerly Ruhengeri) during our recent surveys, two species that Laurent had not collected in Rwanda. Both species are known to be able to cope with habitat alteration and are often found in disturbed habitats and in human settlements (Spawls et al. 2006, pers. observation). It is possible that habitat alteration promoted population decline in and its replacement by otherspecies. The distribution of in Rwanda is currently under study. If our preliminary observations are affirmed, the Red List classification of would have to be changed to a “threatened” category and it would call for conservation measures.

Our recent efforts to untangle the diversity of PageBreak in Rwanda are a first step to clarify the complicated taxonomy of the genus in sub-Saharan Africa. The results of our studies show that species of can be easily distinguished, if standardized diagnostic schemes are applied, which has also been demonstrated by previous studies (e.g. Perret 1979, Bwong et al. 2009). Integrative approaches combining data from morphology, bioacoustics, and molecular genetics will be the best way to address the existing taxonomic problems. Doubtful delimitations of species were often caused by assigning specimens to the wrong species based on non-diagnostic characters. Thereby, states of possible diagnostic characters were mixed up in subsequent accounts on these species, rendering them difficult to distinguish from each other. In the case of , two of its paratypes were in fact , a severe confusion by Laurent (1954). The latter species, however, was described by Laurent in the same paper as yet another new species, . When even the describer cannot reliably distinguish the species, subsequent workers must fail. When reviewing the taxonomic status of certain species, it is mandatory to critically question decisions made by earlier authors by carefully re-evaluating proposed diagnostic characters and re-examining not only the holotypes, but also the material on which accounts discussing the variation within species and keys to species were based.

1	external vocal sac apertures and nuptial pads on dorsal side of metacarpals and phalanges of Fingers I–III present	adult males...2
–	external vocal sac apertures and nuptial pads absent	adult females and subadults...6
2	vocal sac aperture superior; only one supernumerary metacarpal tubercle proximal to Finger IV, often indistinct; longitudinal, alternately black and yellow coloured bands on postaxial side of femur (spiny tubercles on venter absent; inner metatarsal tubercle less than half the length of metatarsus of Toe I; distal subarticular tubercles of Toes III and V reaching to knee; toe webbing I(1.5–1.75)-(2–2.25)II1.5-(2.75–3)III(1.75–2)-3IV2.75-(1–1.5)V; ventral side of head white, mottled with grey)	Ptychadena nilotica
–	vocal sac aperture inferior or semi-inferior; at least one supernumerary metacarpal tubercle proximal to each finger; colouration on postaxial side of femur different	3
3	spiny tubercles on venter absent; toe webbing reaching distal phalanx on postaxial sides of Toes I, II, and III and on preaxial side of Toe V; external dorsal ridge usually not light and prominent (vocal sac aperture inferior; distal subarticular tubercles of Toes III and V never reaching knee; inner metatarsal tubercle about half the length of metatarsus of Toe I; ventral side of head white, trunk yellow; light tibial line and light dorsal band absent; dark spots on dorsum usually absent, if present, small and narrow; irregularly delimited, reticulated, longitudinal, alternately yellow and dark brown coloured bands on postaxial side of femur)	Ptychadena anchietae
–	spiny tubercles on venter present; toe webbing not reaching distal phalanges on toes; external dorsal ridge light and prominent	4
4	median dorsal ridge extending to level between nostrils on dorsal side of snout; vocal sac aperture semi-inferior; two supernumerary metacarpal tubercles proximal to Finger II; inner and outer palmar tubercle more or less equal in length; inner metatarsal tubercle about half the length of metatarsus of Toe I	Ptychadena uzungwensis
–	median dorsal ridge extending to level between eyelids only; vocal sac aperture inferior; one supernumerary metacarpal tubercle proximal to Finger II; outer palmar tubercle longer than inner; inner metatarsal tubercle either more than or less than half the length of metatarsus of Toe I	5
5	foot large, tips of Toes III and V reaching distinctly beyond knee, their distal subarticular tubercles reaching knee; ventral tubercles tiny, hardly visible with naked eye; inner metatarsal tubercle less than half the length of metatarsus of Toe I; dark brown stripe present on preaxial side of tibia; thenar tubercle approximately as long as outer palmar tubercle; webbing not reaching beyond distal subarticular tubercle on postaxial side of Toe III; dorsal spots small and narrow; irregularly delimited, reticulated, longitudinal dark bands on light background on postaxial side of femur	Ptychadena chrysogaster
–	foot smaller, tips of Toes III and V at most reaching slightly beyond knee, their distal subarticular tubercles not reaching knee; ventral tubercles large, visible with naked eye, palpable with finger; inner metatarsal tubercle more than half the length of metatarsus of Toe I; dark brown stripe absent on preaxial side of tibia; thenar tubercle longer than outer palmar tubercle; webbing reaching beyond distal subarticular tubercle on postaxial side of Toe III; dorsal spots large and wide; yellow spots, diffusely arranged in longitudinal rows on dark brown background on postaxial side of femur	Ptychadena porosissima
6	median dorsal ridge extending to level between nostrils on dorsal side of snout; two supernumerary metacarpal tubercles proximal to Finger II (inner metatarsal tubercle about half the length of metatarsus of Toe I; distal subarticular tubercles of Toes III and V not reaching to knee; toe webbing I2-(2.25–2.5)II1.5-3-III(1.75–2)-3IV3-(1+–1.25)V; light tibial line absent; light dorsal band present; dark spots on dorsum large and wide, often fused with neighboring ones; light, prominent dorsolateral fold present)	Ptychadena uzungwensis
–	median dorsal ridge extending to level between eyelids only; one or no supernumerary metacarpal tubercle proximal to Finger II	7
7	toe webbing reaching to distal phalanx on postaxial sides of Toes I, II, and III and on preaxial side of Toe V; light prominent external dorsal ridge usually absent; inner metatarsal tubercle about half the length of metatarsus of Toe I (tips of Toes III and V at most reaching slightly beyond knee, their distal subarticular tubercles not reaching knee; ventral side of head white, trunk yellow; light tibial line and light dorsal band absent; dark spots on dorsum usually absent, if present, small and narrow; irregularly delimited, reticulated, longitudinal, alternately yellow and dark brown coloured bands on postaxial side of femur)	Ptychadena anchietae
–	toe webbing not reaching to distal phalanx on toes; light prominent external dorsal ridge present; inner metatarsal tubercle either less than or more than half the length of metatarsus of Toe I	8
8	inner metatarsal tubercle more than half the length of metatarsus of Toe I; tips of Toes III and V at most reaching slightly beyond knee, their distal subarticular tubercles not reaching knee; thenar tubercle longer than outer palmar tubercle; yellow spots, diffusely arranged in longitudinal rows on dark brown background on postaxial side of femur	Ptychadena porosissima
–	inner metatarsal tubercle less than half the length of metatarsus of Toe I; tips of Toes III and V reaching distinctly beyond knee, their distal subarticular tubercles reaching knee; thenar tubercle about as long as outer palmar tubercle; colouration on postaxial side of femur not consisting of spots	9
9	dorsal spots small and narrow; one supernumerary metacarpal tubercle proximal to each finger; ventral side of head and chest yellow; dark brown stripe present on preaxial side of tibia; irregularly delimited, reticulated, longitudinal dark bands on light background on postaxial side of femur; webbing not reaching beyond subarticular tubercle on Toe I	Ptychadena chrysogaster
–	dorsal spots large and wide; only one supernumerary metacarpal tubercle proximal to Finger IV, often indistinct; ventral side of head and chest white; dark brown stripe on preaxial side of tibia absent, few specimens with dark mottling, not forming continuous stripe; longitudinal, alternately black and yellow coloured bands on postaxial side of femur; webbing reaching beyond subarticular tubercle on Toe I	Ptychadena nilotica