LYGODACTYLUS DEPRESSUS Schmidt, 1919

(Figs 13–16, Table 3; Supporting Information, Table S6)

Lygodactylus picturatus depressus: Loveridge (1947), Wermuth (1965).

Lygodactylus depressus: Pasteur, 1965(1964); Kluge (1991, 1993, 2001); Rösler (2000); Chirio and LeBreton (2007); Röll et al. (2010); de Lisle et al. (2013, 2016)

Lygodactylus depressus was described from Medje, Haut-Uele Province, DRC, by Schmidt, who described the species as an intermediate form between L. picturatus and L. gutturalis: ‘ … The relation of this species with L. picturatus picturatus (Peters) appears to be close; its coloration is in some respects similar but does not seem to fall within the wide variation … chevrons of the throat equally distinct in the female (only two V’s in three of the paratypes) narrower than in gutturalis … Venter and enlarged subcaudals immaculate yellow in life—an apparently constant distinction from Lygodactylus picturatus gutturalis ’ (Schmidt 1919: 466). However, Loveridge (1947: 227), relegated this taxon to a subspecies of L. picturatus, stating: ‘ … a somewhat doubtful race, differing from gutturalis only in gular and subcaudal markings … ’. This taxonomic action was ignored by Pasteur, 1965(1964), but followed by Wermuth (1965), both authors without justification. Kluge (1993) included L. depressus as a full species, in his checklist again without any justification of his taxonomic decision.

Since then this species has been collected in different regions of the Congo Basin. However, this material was overlooked in many publications that revisited the L. picturatus group (Röll et al. 2010, Malonza et al. 2016, 2019). We recovered a genetic clade that is sister to L. gutturalis, from the rainforest of the Congo Basin (Figs 1, 6) that differs by c. 9.04% (16S uncorrected p-distance; Table 2) from its sister taxon. Among this material, we included specimens from Lake Tumba, where historical material of L. depressus was previously collected. Consequently, we revised historical material collected from the Congo Basin, tentatively attributed to L. gutturalis and L. depressus (Supporting Information, Table S2). After morphological examination and comparison of preserved specimens (faded), we observed a complete overlap between specimens ascribed to these two taxa, failing to identify any morphological differences based on pholidosis or gular ornamentation. In his description of L. depressus, Schmidt (1919: 466) remarked: ‘… coloration dark blueish grey, irregularly mottled with black, more heavily anteriorly …’. However, he also made reference to colour variation recorded by Herbert O. Lang (collector): ‘… one specimen was entirely black when caught, turning blueish grey when injected …’. We observed that two specimens allocated to L. depressus and collected at Lotende (a river at Mabali, the site of a colonial-era research station at Lake Tumba; Marlier 1958), Lake Tumba (RMCA 1981.065.R.005) by Laurent, exhibit both colour morphs described for L. depressus (mottled and uniform). In contrast, two specimens (RMCA R.8575/A and RMCA R.8575/B) collected from Flandria, Équateur Province, DRC (~ 110 km north-east from the previous locality) and identified by de Witte in 1923 (RMCA unpublished data) as L. gutturalis, show the same two dorsal patterns as the previous specimens mentioned above (Fig. 13). Two additional specimens (RMCA 1038/B and 1038/C) were collected from the Ituri Forest (without a precise locality within this region, but in the same region as the type locality of L. depressus) and tentatively assigned to L. gutturalis by de Witte in 1923 (RMCA unpublished data). However, these specimens have morphological characters that are intermediate between these two taxa. Thus, material from DRC and Angola (previously attributed to L. gutturalis), shows substantial morphological overlap with topotypic material of L. depressus, making it impossible to differentiate them from the type series (Figs 15, 16). The specimen collected from Angola (MNCN 50772) has a dark blueish coloration in life (Fig. 15G), as described in L. depressus. Consequently, due to the high morphological and geographic overlap, we consider that this clade, sister to L. gutturalis s.s. (Fig. 1), represents L. depressus. Nevertheless, we recommend caution until topotypic material can be included in a phylogenetic context, because the possibility of further cryptic diversification within this group cannot be excluded.

A photographic record from Molondou, East Region, Cameroon, reported by Chirio and LeBreton (2007), and attributed to L. depressus, does not agree with the diagnostic dark blueish coloration mentioned by Schmidt (1919). Therefore, because of the highly diverse character of the picturatus subgroup, records from Cameroon and other localities west and north of the Ubangi and Congo rivers might belong to an undescribed taxon.

Holotype (Fig. 14): AMNH 10345, adult male with original tail, collected at Medje, Ituri Forest, Haut-Uele Province, DRC, on 5 July 1914 by Herbert O. Lang.

Additional material examined (12 specimens): • DRC (10 specimens): RMCA R.3216 (formerly AMNH 10346) and MCZ R45987 (formerly AMNH 10344), males (paratypes), with same collection data as the holotype; UTEP 22579 / UTEP 22582 (ELI 2128 / ELI 2152), males, and UTEP 22580–81 / UTEP 22583 (ELI 2129–30 / ELI 2153), females, collected at Npenda Village, north-east of Lake Tumba, Équateur Province, S00.7465, E18.2243, 311 m a.s.l. on 6 July 2013 by locals and brought to Eli Greenbaum; UTEP 22578 (ELI 1547), female, collected at Lake Tumba, Équateur Province, c. S00.80, E18.15, 300 m a.s.l. on 13 February 2010 by Chifundera Kusamba; UTEP 22595 (ELI 3624), female, collected at Katopa, Maniema Province, S02.75128, E25.10403, 455 m a.s.l. on 4 July 2015 by locals and brought to Eli Greenbaum; UTEP 22597 (ELI 3585), male collected from Katopa, ICCN Camp, Maniema Province, S02.74769, E25.10323, 450 m a.s.l. on 2 July 2015 by Eli Greenbaum. • Angola (two specimens): MNCN 50771–72 (P1-307 and P1-306), males, collected at Barra do Cuanza, Luanda Province, S09.19518, E13.20327, 7 m a.s.l. on 10 September 2021 by Pedro Vaz Pinto and Timoteo Julio.

Diagnosis: Lygodactylus depressus is a large-sized Lygodactylus with a maximum SVL 37.8 mm (mean 35.6 ± 1.4 mm) that has the typical gular pattern of the gutturalis group. Seven to nine supralabials and 6–7 infralabials. Dorsal pholidosis with granular scales that become flattened, larger, and imbricate in original tails. Large triangular mental followed by 2–3 symmetric postmental scales (Fig. 16). Males with 7–8 precloacal pores. Ventral pholidosis with large, flattened, and imbricate scales. Ventral scales usually with small denticulation posteriorly. Digits elongated with five terminal scansors on the fourth toe (Supporting Information, Table S6).

Like other members of the L. gutturalis subgroup, this species can be easily differentiated from L. angularis group members and from other members of the picturatus subgroup by the gular pattern and dorsal colour pattern (Fig. 5). It should be noted that L. depressus has a dark blueish grey dorsum in life, unlike the light blueish dorsal coloration in the picturatus subgroup.

Lygodactylus depressus can be differentiated from other species within the gutturalis subgroup by having a dark blueish grey coloration of the dorsum, with some specimens having a mottled pattern on the dorsum, vs. brownish or light grey without a mottled pattern in L. gutturalis. However, blackish specimens can only be differentiated by subtle morphometric and meristic data, being almost impossible to differentiate in the field. Lygodactylus depressus can be partially differentiated from L. gutturalis as follows: eye proportionally smaller in L. depressus (OD/HL ≤ 0.25 vs. 0.26–0.31 in L. gutturalis); and narrower snout (IN/HW ≤ 0.29 vs. 0.30–0.34 in L. gutturalis). It also differs by a minimum of c. 9.04% uncorrected p-distance for 16S (Table 2), lacks any nuclear haplotype sharing with L. gutturalis in RAG1 (Fig. 2), and habitat (rainforest for L. depressus and sub-Saharan savannah for L. gutturalis). It can also be differentiated from L. dysmicus by having fewer precloacal pores (7–8 vs. 9 in L. dysmicus) and nostril never in contact with rostral scale vs. nostril contacting the rostral in L. dysmicus. For a distinction from other species not included above, described herein, see the respective diagnoses below.

Coloration: In life (Fig.15), dorsal colour is highly variable; predominantly dark grey, with lighter grey blotches surrounded by black dots on the flanks, which can form a diffuse dorsolateral band; dark V -shaped line between the eyes (absent in darker specimens), and black line from snout to anterior insertion of the forelimb; tail with diffuse bars of lighter grey; throat usually with white background and black chevrons; venter vivid yellow, orange, or cream from tail tip to posterior part of the gular region. In preservative (holotype; Fig. 14), dorsum grey with dark brown dots from head to midbody; venter uniform light cream to yellow with whitish colouration on all digits.

Variation: Meristic and morphometric data are summarized and depicted in the Supporting Information (Table S6) and Figures 15–16. Lygodactylus depressus is the most variable taxon within the L. gutturalis group with respect to coloration, particularly the gular pattern (Fig. 16). Some specimens lose the dorsal pattern, resulting in a uniform dorsal coloration. However, other specimens have a mottled dorsal pattern (Fig. 13).

Habitat and distribution (Figs 6, 15): This species is widely distributed within the Congo Basin, from the Ituri Forest in the north-east, through the heart of the Congo River at Npenda and Lake Tumba, south to Angola at the southernmost extreme of its range. The material from Angola was collected from mangroves growing at the mouth of the Cuanza River. However, all material from DRC was collected in dense Congolian rainforest.

Natural history: Individuals in Angola were observed actively moving and hunting between the branches and trunk of trees during the day, but due to the difficult access to the mangroves, specimens were collected at night while sleeping on thin branches. Specimens from Lotende, DRC, were collected in the canopy, as recorded by Laurent in his field notes.