Herpetofaunal Diversity in a Dahomey Gap Savannah of Togo (West Africa): Effects of Seasons on the Populations of Amphibians and Reptiles

Abstract

The Dahomey Gap is a human-derived savannah zone, interspersed by patches of moist tropical forest, that separates the forest zone into two blocks, the Upper Guinean and the Lower Guinean forests. Community ecology aspects of amphibians and reptiles are still relatively unexplored in this ecological zone of West Africa. Here, the overall species richness and the variation of the diversity metrics (dominance, evenness) of a whole herpetofaunal community in Togo was studied, with emphasis on the effects of the seasons (wet and dry) on the population structure. Overall, we observed 998 amphibian individuals from 27 species: 148 individuals belonging to 11 species during the dry season and 849 individuals belonging to 25 species during the wet season. For reptiles, we observed 517 individuals belonging to 44 species: 323 individuals belonging to 41 species during the dry season and 194 individuals belonging to 28 species during the wet season. The analyses on the diversity metrics showed opposite patterns between amphibians and reptiles in each season. Indeed, the dry season rank–abundance curve was systematically higher in reptiles than in amphibians for each rank of abundance, while the opposite pattern occurred in the wet season rank–abundance curve. Singletons and doubletons were much more numerous in the reptiles. Concerning the diversity indices, the Dominance index was significantly higher in amphibians during the dry season than in all other pairwise comparisons, whereas the Shannon’s index was significantly lower in dry season amphibians and significantly higher in wet season reptiles. Evenness index was significantly lower in reptiles than in amphibians and the mean number of individuals was significantly higher in amphibians by wet season compared to dry season amphibians or reptiles during both seasons. The ecological implications of these data are discussed. Most species were of minor conservation concern.

1. Introduction

The so-called “Dahomey Gap” situated along the West African Gulf of Guinea coast in Benin, Togo, and eastern Ghana is a human-derived vegetation zone that originated in historical times and that is a savannah-like vegetation zone interspersed by patches of moist tropical forest. This human-derived vegetation zone is ecologically very important as it separates the forest zone that covers much of the south of the region into two separate forest regions, i.e., the Upper Guinean and the Lower Guinean forests [1].

Compared to the Upper and Lower Guinean forests, the organismal communities of the Dahomey Gap savannahs have been far less intensely studied, probably because the general perception is that these savannahs are poorly speciose and of lesser conservation interest, as they are human-derived and broadly altered by human activities. However, some studies focusing on the community composition and resource partitioning patterns of Dahomey Gap turtles [2,3] and lizards [4] are available, with other more comprehensive studies on the distribution records and taxonomy of the Togolese species [5,6,7,8,9]. Nonetheless, no previous study has reported any inventory of the species and the diversity metrics of a whole community of reptiles and amphibians at a given area within the Dahomey Gap savannah region. Thus, we still do not know much about (i) the overall species richness and (ii) the variation of the diversity metrics (dominance, evenness) of the whole herpetofauna communities of the Dahomey Gap by seasons (wet and dry). This study is the first contribution aiming at investigating these two aspects within a same article. Indeed, the objective of this study is (1) to carry out the most complete herpetological inventory possible in both the dry and the wet seasons, (2) to evaluate the quantitative differences between amphibians and reptiles in terms of diversity metrics, (3) to evaluate the effects of the season on the diversity metrics and the community composition of both amphibians and reptiles, and (4) to evaluate the presence and relative sighting frequency of priority species for conservation according to IUCN [10].

More specifically, because the African savannah habitats are strongly seasonal with prolonged droughts during the dry months [11,12,13], we hypothesize that:

(a)

The various diversity metrics should be very different between amphibians and reptiles, given their divergent tolerance to the climatic conditions (especially rainfall) between dry and wet seasons, and because of their different role within the savannah trophic chains. That is, the mean number of amphibian individuals per species should be higher than the mean number of reptile individuals per species, given that these latter are high-rank predators in the savannah trophic chains (for instance, some snakes, see [14]).

(b)

Amphibians should be more impacted than reptiles by the dry season months, as they are more linked to water bodies and humidity threshold than reptiles. That is, the number of amphibian species and individuals should be significantly lower by dry season, with consequent increase in the dominance and decrease in the evenness of the assemblage, whereas the same inter-seasonal difference should be less evident in reptiles.

(c)

In reptiles, dominance and evenness should be relatively independent on season, although excessive heat and drought may also depress the reptilian assemblages.

2. Materials and Methods

2.1. Study Area

The present study was conducted in Togo (West Africa) as part of a herpetological fauna assessment relating to the construction of a hydroelectric dam on the Mono River in an area between the locality of Tetetou in the south and the Nangbeto dam in the north (Figure 1). Ecologically, the study area is located in the Guinean lowland climate zone, within the Benin-Togolese plain east of the Atakora range.

The main habitats are shrub savannahs with sparse woodlands, with main vegetation being dominated by Combretaceae and Andropogoneaceae such as Daniellia oliveri, Terminalia macroptera, Combretum spp., Pterocarpus erinaceus, Parkia biglobosa, Vitellaria paradoxa, etc. There are also scattered islands of semi-deciduous dry forests as well as forest galleries along the main rivers and streams, whose main species are Cynometra megalophylla, Parinari congensis, Pterocarpus santalinoides, etc. This common vegetation in the region includes some endangered tree species such as Pterocarpus erinaceus, for example.

Savannah and dry forest habitats are under pressure at the study area from rampant agricultural expansion and population growth leading to deforestation and degradation of environments for planting and crops and for harvesting firewood.

2.2. Protocol

The study was carried out from 1 to 14 October 2021 (end of the wet season) and 28 February to 14 March 2022 (end of the dry season). During each season, a total of 105 man-hours were spent in the field.

Visual Encounter Surveys (VES) were carried out to assess the specific herpetofauna richness of the study area as this method assures a good determination of the communities of species at the local level, and to estimate the relative abundances of species within a community [15]. We explored all the available habitat types, including temporary pools, gallery forests, rivers, but also savannas and some fallow land. Reptiles were inventoried in all the habitats present in the study area.

The search techniques consisted of a visual sweep of the terrain and the inspection of potential shelters: the leaves, trunks and branches of trees for arboreal species, bodies of water for aquatic species, all potential shelters for terrestrial species (plant debris, trees, burrows), ground shelters for burrowing species. Signs of presence were noted (droppings, burrows, footprints, manifestations, moults, shells, skeletons). The search for both amphibians and reptiles took place day (for at least 4 h per day) and night (for at least 3 h per day), with a significant portion of reptile species having nocturnal activity. At night, flashlights were used to observe specimens in the field. We conducted surveys also following hard rains as these are usually very productive in tropical forest habitats. Local guides also helped in collecting herpetofauna individuals. For amphibians, the fieldwork was mainly concentrated in habitats with the presence of water. The different individuals were mostly noted opportunistically during the visual surveys. The songs of male amphibians observed at night were used for the identification and counting of species.

The identifications of the different species were carried out following the identification keys/criteria of [16,17] for snakes, and [18] for other reptiles. For amphibians, species identifications were carried out following the identification keys/criteria of [19]. The taxonomy follows [20].

On the different sites, for each sighting, we recorded, in an Excel spreadsheet: (i) species name, (ii) date of observation, (iii) number of individuals, (iv) age and sex (if determinable), (v) habitat and habitat quality, (vi) sign of presence (seen/heard), (vii) GPS coordinates, (viii) photograph number if individual is photographed, etc.

2.3. Statistical Analyses

Several distinct measures of community diversity were calculated for both amphibians and reptiles [21,22]:

(a)

Species richness, the total number of taxa recorded into each plantation type at each study area and during the wet or the dry season;

(b)

Dominance,

D = 1 − impson index

(1)

and ranges from 0 (all taxa are equally present) to 1 (one taxon dominates the community completely);

(c)

Shannon H’ index, varying from 0 for communities with only a single taxon to high values for communities with many taxa, each with few individuals [22]. Shannon’s index (H’) is

H’ = − S(fr) × [ln(fr)],

(2)

where S is the total number of amphibian or reptile species recorded, fr = n/N, n is the number of individuals of each species in each season and N is the total number of individuals of each taxon (amphibians or reptiles) in the study area.

(d)

Pielou’s Evenness index, calculated as

e = H’/Hmax,

(3)

where

Hmax = lnS,

(4)

with H’ representing Shannon’s index, and S the total number of amphibian or reptile species recorded. Hmax corresponds to the maximum value of diversity (i.e., when all species are equally represented in our samples [21]).

(e)

Chao-1 index that calculated the predicted number of species at a given site given the observed sample size and the observed number of species. Chao-1 index is

Chao-1 = S + F1(F1 − 1)/(2 (F2 + 1),

(5)

where F1 is the number of singleton species and F2 the number of doubleton species. A singleton is a species occurring only once in the total sample, and doubleton is a species occurring just twice in the total sample.

The various indices were calculated using the raw data for the two seasons, because the survey field effort was identical at each season (hence unbiased). Bootstrap analysis was applied to generate upper and lower confidence intervals of all indices, with 1000 random samples being generated, each with the same total number of individuals as in each original sample [22].

We performed individual rarefaction curves to evaluate the completeness of the samples done by season and for both amphibians and reptiles. This module estimates how many taxa you would expect to find in samples with a smaller total number of individuals. In other words, by this method, it is possible to read out the number of expected taxa for any smaller sample size [23]. We also graphically compared diversities in the several samples (i.e., in both amphibians and reptiles and in both seasons) by the diversity profile models [24]. We also used diversity profiles because they are not subjected to any issues that may arise from the arbitrary choice of a given diversity index to describe a given study community. We bootstrapped the data (giving a 95% confidence interval) with 2000 replicates.

We evaluated the statistical differences in the diversity indices between seasons and between taxa (amphibians versus reptiles) using One-Way Analysis of Similarities (ANOSIM). ANOSIM is roughly analogous to an ANOVA in which the univariate response variable is replaced by a dissimilarity matrix, i.e., with distances that were converted to ranks. Significance was computed by permutation of group membership, with 9999 replicates, and Bray–Curtis was used as distance measure. We used a sequential Bonferroni correction for post hoc pairwise comparisons between all pairs of groups. ANOSIM was performed in R-software, using Vegan package [25].

Differences in the mean number of individuals per species by season and by taxon (amphibians versus reptiles) were analyzed by one-way ANOVAs. All other analyses were performed using PAST software, with alpha set at 5%.

3. Results

3.1. Amphibians

The checklist of the amphibians and their conservation status is given in

Table 1. Number of individuals of reptiles and amphibians (including both alive and dead specimens) observed at the study area, during the wet and the dry season, including their IUCN (2022) Red List status, and the type of record.

Family	Species	Type of Record	Observed Number in Dry Season	Observed Number in Wet Season	IUCN (2022)
Amphibia
Arthroleptidae	Arthroleptis poecilonotus	O	0	21	LC
Arthroleptidae	Leptopelis spiritusnoctis	O	8	35	LC
Arthroleptidae	Leptopelis viridis	O	0	3	LC
Bufonidae	Sclerophrys maculatus	O	0	51	LC
Bufonidae	Sclerophrys regularis	O	25	110	LC
Hyperoliidae	Afrixalus dorsalis	O	0	30	LC
Hyperoliidae	Afrixalus vittiger	O	0	23	LC
Hyperoliidae	Afrixalus weidholzi	O	0	3	LC
Hemissotidae	Hemisus marmoratus	O	3	1	LC
Hyperoliidae	Hyperolius concolor	O	0	50	LC
Hyperoliidae	Hyperolius igbettensis	O	0	40	LC
Hyperoliidae	Hyperolius nitidulus	O	0	18	LC
Hyperoliidae	Kassina fusca	B			LC
Hyperoliidae	Kassina cassinoides	O	0	2	LC
Hyperoliidae	Kassina senegalensis	O	0	17	LC
Phrynobatrachidae	Phrynobatrachus latifrons	O	20	17	LC
Phrynobatrachidae	Phrynobatrachus natalensis	O	5	0	LC
Pipidae	Xenopus fischbergi	O	0	36	LC
Pipidae	Xenopus tropicalis	B			LC
Ranidae	Amniana galamensis	O	3	16	LC
Dicroglossidae	Hoplobratrachus occipitalis	O	0	32	LC
Ptychadenidae	Ptychadena bibroni	O	16	1	LC
Ptychadenidae	Ptychadena oxyrhynchus	O	30	56	LC
Ptychadenidae	Ptychadena pumilio	O	6	43	LC
Ptychadenidae	Ptychadena sp.	O	24	46	LC
Ptychadenidae	Ptychadena tellinii	O	5	7	LC
Microhylidae	Phrynomantis microps	B			LC
TOTAL			145	658
Reptilia
Pelomedusidae	Pelusios castaneus	O	32	18	LC
Pelomedusidae	Pelomedusa subrufa olivacea	O	23	16	LC
Testidinidae	Kinixys nogueyi	O	9	11	VU
Tryonichidae	Cyclanorbis senegalensis	O	2	48	VU
Tryonichidae	Trionyx triunguis	O	21	0	VU
Crocodylidae	Crocodylus suchus	O	6	2	VU
Crocodylidae	Osteolaemus tetraspis	E			VU
Leptotyphlopidae	Leptotyphlops bicolor	B			LC
Typhlopidae	Typhlops punctatus	B			LC
Pythonidae	Python regius	O	3	0	LC
Pythonidae	Python sebae	O	2	0	LC
Atractaspididae	Atractaspis aterima	B			LC
Atractaspididae	Atractaspis dahomeyensis	B			LC
Atractaspididae	Atractaspis irregularis	O	2	0	LC
Colubridae	Afronatrix anoscopus	B			LC
Colubridae	Amblyodipsas unicolor	B			LC
Colubridae	Chamaelycus fasciatus	B			LC
Colubridae	Crotaphopeltis hotamboeia	O	3	2	LC
Colubridae	Dasypeltis fasciata	O	0	1	LC
Colubridae	Dasypeltis gansi	O	2	0	LC
Colubridae	Grayia smithii	O	1	0	LC
Colubridae	Lamprophis fuliginosus	O	3	0	LC
Colubridae	Lamprophis lineatus	O	6	0	LC
Colubridae	Lycophidion irroratum	O	0	1	LC
Colubridae	Lycophidion semicinctum	B			LC
Colubridae	Mehelya crossii	B			LC
Colubridae	Mehelya poensis	B			LC
Colubridae	Meizodon coronatus	B			LC
Colubridae	Meizodon regularis	B			LC
Colubridae	Philothamnus irregularis	O	2	0	LC
Colubridae	Philothamnus semivariegatus	O	2	0	LC
Colubridae	Prosymna meleagris	B			LC
Colubridae	Psammophis elegans	O	2	1	LC
Colubridae	Psammophis phillipsi	B			LC
Colubridae	Psammophis sibilans	B			LC
Colubridae	Rhamnophis aethiopissa	B			LC
Colubridae	Rhamphiophis oxyrhynchus	O	3	0	LC
Colubridae	Scaphiophis albopunctatus	B			LC
Colubridae	Telescopus variegatus	B			LC
Colubridae	Toxicodryas blandingii	B			LC
Colubridae	Toxicodryas pulverulenta	B			LC
Elapidae	Dendroaspis viridis	O	2	1	LC
Elapidae	Elapsoidea semiannulata	B			LC
Elapidae	Naja melanoleuca	O	2	3	LC
Elapidae	Naja nigricollis	O	0	1	LC
Viperidae	Bitis arietans	O	2	3	LC
Viperidae	Causus maculatus	O	5	2	LC
Viperidae	Echis ocellatus	O	6	2	LC
Agamidae	Agama agama	O	44	36	LC
Agamidae	Agama parafricana	B			LC
Agamidae	Agama sankaranica	O	3	0	LC
Chamaeleonidae	Chamaeleo gracilis	B			LC
Chamaeleonidae	Chamaeleo senegalensis	O	2	2	LC
Gekkonidae	Cnemaspis spinicollis	B			LC
Gekkonidae	Hemidactylus albituberculatus	O	1	0	LC
Gekkonidae	Hemidactylus angulatus	O	10	7	LC
Gekkonidae	Hemidactylus mabouia	O	9	10	LC
Gekkonidae	Hemidactylus matschiei	O	4	1	LC
Gekkonidae	Hemitheconyx caudicinctus	O	3	1	LC
Gekkonidae	Lygodactylus conraui	O	2	0	LC
Lacertidae	Heliobolus nitidus	O	3	0	LC
Scincidae	Mochlus guineensis	O	12	0	LC
Scincidae	Panaspis togoensis	O	5	3	LC
Scincidae	Trachylepis affinis	O	18	8	LC
Scincidae	Trachylepis maculilabris	O	7	4	LC
Scincidae	Trachylepis perrotetii	O	8	15	LC
Scincidae	Trachylepis quinquetaeniata	O	20	14	LC
Varanidae	Varanus exanthematicus	O	2	1	LC
Varanidae	Varanus niloticus	O	6	3	LC
Varanidae	Varanus ornatus	B			LC
TOTAL			300	217

Legend: IUCN (2022): LC: Least Concern, VU: Vulnerable. Type of record: O = observed; B = reported in the literature but not directly observed in the wild; E = species reported by village survey.

. Overall, we observed 803 individuals belonging to 24 species (Figure 2): 145 individuals belonging to 10 species during the dry season and 658 individuals belonging to 23 species during the wet season (Table 1). An additional three species were not directly observed but were cited in the literature (Table 1). Saturation curves revealed that the plateau phase was reached in both seasons (Figure 3a), with 12 species predicted to occur in the area by dry season and 26 by wet season according to Chao-1 index. Diversity indices revealed a much higher diversity and evenness by wet season than by dry season (

Table 2. Diversity indices for amphibian and reptile communities at the study area during the wet and the dry season.

	Dry	Wet	Dry	Wet
	Amphibia		Reptilia
Species richness	10	23	41	28
Number of individuals	145	658	323	194
Dominance	0.1339	0.08656	0.05867	0.08864
Shannon	2.184	2.75	3.199	2.749
Evenness	0.7398	0.6256	0.5978	0.5578
Chao-1	12	26	41	33

), and diversity profile curves showed that the two seasonal samples were very different from each another (Figure 3b). There were no singletons and only one doubleton (Kassina cassinoides). Several species were very abundant, with Phrynobatrachus latifrons, Ptychadena sp., Ptychadena oxyrhynchus, and Sclerophrys regularis being the most common.

All amphibian species were non-threatened by IUCN (2022) and not protected in Togo (Table 1).

3.2. Reptiles

The checklist of the reptiles and their conservation status is given in Table 1. Overall, we observed 517 individuals belonging to 44 species (Figure 4): 300 individuals belonging to 41 species during the dry season and 217 individuals belonging to 28 species by wet season (Table 1). In addition, 25 species were not directly observed during our field surveys but are cited in the available literature (Table 1). The sightings of a gekkonid species (Hemidactylus matschiei) represented a considerable extension to its known distribution within Togo (Figure 5). Saturation curves revealed that the plateau phase was reached in both seasons (Figure 3c), with 41 species predicted to occur in the area during the dry season and 33 during the wet season according to Chao-1 index. In other words, Chao-1 estimates revealed that our sampling was more efficient during the dry season. Diversity indices revealed a much higher diversity and evenness during the dry season (Table 2), and diversity profile curves showed that the two seasonal samples were similar (Figure 3d). There were five singletons (four snakes: Dasypeltis fasciata, Grayia smithii, Lycophidion irroratum, Naja nigricollis; one lizard: Hemidactylus albituberculatus) and seven doubletons (the snakes Dasypeltis gansi, Atractaspis irregularis, Hapsidophrys smaragdinus, Philothamnus irregularis, Philothamnus semivariegatus, Python sebae, and the lizard Lygodactylus conraui), thus showing that the distribution of local rarity was much different from that of sympatric amphibians. The most common species were two lizards (Agama agama and Trachylepis quinquetaeniata) and three turtles (Cyclanorbis senegalensis, Pelusios castaneus and Pelomedusa subrufa).

In conservation terms, most reptile species were non-threatened by IUCN (2022) but five are listed as VU in the IUCN Red List (Table 1).

3.3. Comparisons between Amphibians and Reptiles

Comparing the rank–abundance diagrams, we can notice remarkable trajectory differences between amphibians and reptiles and between seasons (Figure 6). In amphibians, the curves were very different between seasons with a high dominance of the two most abundant taxa in the wet season, otherwise with a relatively smooth evenness among all other species (Figure 6A). On the other hand, in reptiles, the curves were similar between seasons, although with constantly higher abundances for all species sharing a same rank during the dry season (Figure 6B). Interestingly, the seasonal patterns were opposite between taxa: the dry season curve was systematically higher than the wet season curve for a same given rank in the reptiles but was systematically lower for a same rank in the amphibians (Figure 3).

Concerning the diversity indices:

(i)

Dominance index was significantly higher in amphibians during the dry season than in all other pairwise comparisons, i.e., wet season amphibians and both seasons’ reptiles (ANOSIM: mean rank within seasons types = 102.4; mean rank between taxa = 1361.4; R = 0.311, P = 0.0012);

(ii)

Shannon’s index was significantly lower in dry season amphibians and significantly higher in wet season reptiles (ANOSIM: mean rank within seasons types = 116.3; mean rank between taxa = 118.3; R = 0.246, P = 0.0023);

(iii)

Evenness index was significantly lower in reptiles than in amphibians (ANOSIM: mean rank between taxa = 99.4; R = 0.221, P = 0.023).

The mean number of individuals was significantly different between amphibians and reptiles and between seasons (one-way ANOVA: F_3,138 = 15.24, P = 0.0001), and Tukey HSD post hoc tests showed that the number of individuals was significantly higher in amphibians during the wet season compared to dry season amphibians or reptiles during both seasons.

4. Discussion

4.1. General Considerations

If we consider the taxa whose presence is likely, the Togolese herpetological fauna consists of 60 species of amphibians [9] and 160 species of reptiles [6,7,8]. This study reveals that about 50% of Togolese amphibians do occur at the study area, whereas the fraction is much lower for the reptiles. Considering that (i) Guinea savannah is the dominant vegetation type in both the study area and Togo in general, and that (ii) this vegetation type is quite homogeneous all throughout the Dahomey Gap in terms of ecological characteristics, the percentages of locally found species appear to be relatively low. We suggest that this result was due to the strong anthropization of the study area, given that our analytical approach (saturation curves and Chao-1 estimates) confirmed that we detected the great majority of the species potentially occurring at the study area from a statistical point of view.

Overall, the encountered amphibian species formed communities that are typical of West African savannahs [26]. In Togo, they have a wide distribution over all the different ecological zones of the country [9]. These are often arboreal (Hyperolidae) or terrestrial-aquatic (Ptychadenidae) species. Most of the species belonging to the Hyperolidae family were active only in the wet season, whereas the species of the Ptychadenidae family showed a higher capacity to adapt to less humid conditions and so they were also frequently observed during the dry season.

In general, the species of reptiles occurring at the study area were typically West African savannah species [7,8,16,17,18]. For instance, the community composition of our turtle assemblage was very similar to that observed in the Dahomey Gap savannahs in both Ghana [3] and Benin [2]. However, in our study area, we found at a single site (Nagbeto dam) Trionyx triunguis that is usually very rare in savannah water bodies [3]. We suppose that the presence of gallery forests along the river banks may explain the presence of this otherwise rare turtle at the study area [7]. Concerning lizards, the great majority of the species recorded in the present study were also recorded in previous studies on the lizard communities of Togo savannahs [4]. As the study area was at least partially anthropized by agricultural activities and human settlements, the destruction of habitats has certainly contributed to the proliferation of generalist savannah species, although in the best-preserved site places (with remnants of a forest ecosystem, notably at Kpetuhoe) forest taxa were also observed.

Ecologically, among the species of reptiles listed, we found a dominance of terrestrial species, with aquatic species that were locally abundant (Pelusios castaneus, Pelomedusa subrufa olivacea, Cyclanorbis senegalensis) and arboreal species that occurred especially in the gallery forests growing along the banks of the Mono and Ra rivers (Dendroaspis viridis, Philothamnus irregularis and Philothamnus semivariegatus, for example). Although these latter were seen more rarely than terrestrial and aquatic species, it should be noted that this is likely an outcome of their highly elusive habits while hiding on trees.

Although the study area did not contain any endemic species, during this survey, we observed a very rare species, Hemidactylus matschiei (at Gbanvikpli and at Amoukpli). This gekko species is known to occur only in Togo and Nigeria, exclusively in the few sites where the savannah zone is naturally associated with the forest zone. It was described in 1901 [27], and a specimen of the species was recently collected by Trape from the type locality (Yégué) [18]. This new collection makes it possible to enhance the known distribution area for this species.

4.2. Diversity Metrics

Consistently with our a priori hypotheses, we observed that the diversity metrics were remarkably different between amphibians and reptiles, and that there were clear seasonal directions of these differences. First, the number of sympatric species was much higher in reptiles. As said above, this could not be surprising given that the species richness of reptiles is much higher than that of amphibians in Togo in general (see references given above).

Second, the mean number of individuals per species was lower in reptiles than in amphibians, thus mirroring our a priori hypothesis (a) in the Introduction. More specifically, (i) the number of amphibian individuals peaked strongly during the wet season, thus by far outnumbering those of reptiles, and (ii) the number of individuals of the larger predators (Python sebae, Bitis arietans, other large snakes) was always low. Pattern (i) had a clear seasonal effect of the abundance of water and available food resources for amphibians during the wet months, and pattern (ii) shows that the number of large predators is relatively lower in the Dahomey Gap savannah habitats compared to smaller predators.

It should be remarked that pattern (i) confirms our a priori hypothesis (b) that amphibian populations should be more negatively affected than reptiles by the dry season climatic conditions. In fact, we recorded a total of 513 more amphibian individuals in the wet season than in the dry season. As predicted, the inter-seasonal differences in amphibian population size also affected the diversity metrics, with the consequent increase in the dominance and decrease in the evenness of the seasonal assemblages.

As predicted in our hypotheses (b) and (c), the same inter-seasonal differences were less evident in reptiles: the difference between the two seasons was only 83 individuals (in favor of the dry season), and the values of dominance and evenness were much less different between the seasons than in amphibians.

Interestingly, the distribution of local rarity was intriguingly different: there were several singletons and doubletons in reptiles but very few in amphibians. In this case, the relative species richness of the two groups cannot explain this pattern: indeed, 0% of amphibian species were singletons and just 3.7% were doubletons. Conversely, 11.4% of reptiles were singletons, and 13.6% were doubletons. These differences were so strong that could not be due to chance. We suppose that (i) many reptile species are indeed rarer at the study area (also possibly an effect of a resource partitioning patterns for food or space; see [28,29]), and (ii) they are more elusive than amphibians, especially during the wet season. Overall, we obtained indirect evidence at least for pattern (ii): indeed, some singletons/doubletons are usually abundant species with a niche that is not shared by any potential competitor at the study area (for instance, the strictly aquatic fish- and frog-eater Grayia smithii, [30], or the frequently snake-eater Naja nigricollis [31,32,33] or the gigantic Python sebae [14,34]), thus making us theorize that in their apparent rarity there are no competitive mechanisms with other sympatric species. However, if we observe more carefully within seasons (in savannah habitat, the available resource is certainly more limited in the dry than in the wet season), we can notice some aspects that suggest the occurrence of resource partitioning patterns in our assemblage of reptiles: (i) within the genera Agama, Hemidactylus and Trachylepis there were high abundance differences among species by both dry and wet season; (ii) the tryonichid turtles Cyclanorbis senegalensis and Trionyx triunguis, occurring syntopically in the Nagbeto dam, were found in alternate seasons in a clearly statistically significant way. Thus, although more research is needed, we suppose that both elusiveness and synecology may explain the observed reptile patterns at the study area.

4.3. Conservation Considerations

With the exception of a few turtles and crocodiles, all the other species of both classes encountered on the study area are not considered priority taxa for conservation. Indeed, only Kinixys nogueyi (VU), Cyclanorbis senegalensis (VU), Trionyx triunguis (VU), Osteolaemus tetraspis (VU) and Crocodylus suchus (VU) are listed on the IUCN (2022) Red List. However, since (i) live specimens of some of the reptile species are exploited in international trade (especially pythons, monitor lizards and freshwater turtles) and (ii) monitor lizards, turtles and crocodiles are also hunted for local consumption, their populations should be monitored regularly in order to prevent any unnoticed decline.