Tarentola protogigas, JOGER 1984

Vasconcelos, Raquel, Perera, Ana, Geniez, Philippe, Harris, D. James & Carranza, Salvador, 2012, An integrative taxonomic revision of the Tarentola geckos (Squamata, Phyllodactylidae) of the Cape Verde Islands, Zoological Journal of the Linnean Society 164 (2), pp. 328-360 : 353

TARENTOLA PROTOGIGAS JOGER 1984 B

Diagnosis: Medium to large-sized gecko [maximum SVL 98.5 mm ( Schleich, 1987); 71.9 mm on average, see Appendix 2]; eye/ear opening ratio averages 1.69; ear–eye/eye–snout distance ratio averages 0.75. Eight to 12 supralabials; seven to nine infralabials; ten to 13 enlarged lamellae under the 4th finger; 144–181 midbody scales ( Joger, 1984b); conical to apical prominent dorsal tubercles with a narrow central keel ( Fig. 5D 4 ), especially on the sacral region, with 12–15 transverse rows and 15–21 longitudinal rows; several enlarged tubercles between the eye and the ear opening. Grey, brownish to yellowish dorsal pattern with a series of four (sometimes five) light middorsal patches, each preceded by a more indistinct and lighter W-shaped dark mark, usually connected by a light middorsal line ( Figs 6D 3–5 and 7D 3–5 ); goldenyellowish grey ventral parts; dark spots on the labials, sometimes creating an alternating light and dark pattern; eye iris grey with an indistinct broad horizontal dark area.

It differs from T. bocagei , T. fogoensis , T. darwini , T. substituta , T. raziana , T. caboverdiana , and T. nicolauensis by having prominent conical dorsal tubercles, enlarged tubercles between the eye and ear opening and a different dorsal pattern ( Fig. 6 ), and from T. gigas by the presence of a narrow wellmarked central keel, especially on the sacral region. It also differs from T. gigas by having important morphological, bioacustical, ecological, and behavioural differences. It differs from T. boavistensis , T. rudis , and Tarentola from Maio by its yellower ventral coloration. It also differs from T. rudis by a higher number of scales around midbody and interorbital scales [18–21 versus 16–19 ( Joger, 1984b)], by having four to five more indistinct and lighter W-shaped dorsal bands ( Fig. 6 ), fader spots on the labials and less contrasted eye iris coloration ( Fig. 7 ). It differs from Tarentola from Maio by a higher number of scales and lamellae under the fifth toe [22–26 versus 19–21, rarely 22 ( Joger, 1984b)] and interorbital scales [19–21 versus 16–18 ( Joger, 1984b)].

Distribution: The southern islands of Fogo, Brava, and Rombos Islets, Cape Verde.

Genetic and phylogeographical remarks: Tarentola protogigas is monophyletic ( Fig. 2 ) and presents a considerable level of genetic divergence from other sister taxa from clade D, as T. gigas , T. rudis , and Tarentola from Maio: D3–D1, D3–D2, and D3–D6 p- dist (cyt b) = 2.5 ± 1.2, 2.6 ± 0.9, and 5.3 ± 1.2%, respectively ( Table 5). The population from Fogo presents a considerable level of genetic divergence with the populations from Brava and Rombos: D3–D4 and D3–D5 p- dist (cyt b) = 2.1 ± 0.8 and 2.3 ± 0.8%, respectively. However, the Snn test values for PDC, ACM4, and MC1R are not significant between T. protogigas from Fogo versus Brava and Rombos (Appendix 5). The population from Brava presents very low values of genetic divergence with the population from Rombos: D4–D5 p- dist (cyt b) = 0.4 ± 0.3%. Therefore, only one of the three lines of evidence (morphology) differentiates the population of Fogo from Brava and Rombos. Consequently, according to the IPC protocol, T. p. protogigas and T. p. hartogi comb. nov. are considered only distinct subspecies ( Fig. 2 ). The lack of differentiation in at least two of the three lines of evidence precludes any further differentiation between the island populations from Brava and Rombos.

TARENTOLA PROTOGIGAS PROTOGIGAS JOGER,

Figures

Figure 2. Phylogenetic relationships of endemic Cape Verde Tarentola taxa and their relatives from the Canary Islands modified from Vasconcelos et al. (2010) based on cytochrome b and 12S rRNA genes. The tree was inferred using maximum likelihood (ML) and GTR+I+G model of sequence evolution (log likelihood = -6468.896) and was rooted using Tarentola americana. Bootstrap support values above 60% for the ML analysis are shown below nodes. Posterior probability (PP) values higher than 95% for the Bayesian analysis are represented by an asterisk (*) and are shown above nodes. Names in bold follow the new taxonomic proposal and non-bold ones the taxonomy accepted in previous recent papers (Carranza et al., 2000; Jesus et al., 2002; Vasconcelos et al., 2010). For further details see Vasconcelos et al. (2010). Characters immediately to the right of island names correspond to the 15 evolutionarily significant units (ESUs) of A, B, C, and D clades recognized in the present work and represented in split green bars. Lines of evidence (in grey): 1, mitochondrial DNA (independent cyt b parsimony networks with a connection limit of 95%; see Appendix 3); 2, nuclear DNA (absence of shared haplotypes in MC1R); 3, morphology (detection of any diagnostic morphological character or a set of a unique combination of characters). Integration approaches (in red) from the most conservative to the most inflationist: ITC stands for integration by total congruence (all lines of evidence should be congruent), IPC stands for integration by partial congruence, retained in the present study (at least two lines of evidence are necessary); IC stands for integration by cumulation (one line of evidence is sufficient). Species are represented in split red bars and subspecies in yellow.

Figure 3. Parsimony networks corresponding to the PDC, ACM4 and MC1R nDNA sequence variation in Tarentola from the Cape Verde Islands. Lines represent a mutational step, circles haplotypes and dots missing haplotypes. The size of circles is proportional to the number of haplotypes and colours to the number of individuals. The dotted circles represent the most probable ancestral haplotype. Samples from the same island are similarly coloured but with different tonalities for different taxa. For correspondences of sample and location codes see Appendix 1.

Figure 4. Discriminant analyses for males and females previously called T. darwini. The total contribution of each of the two Canonical Discriminant Functions (CDF1 and CDF2) to explain the total morphological variation is also given. See Material and methods for details.

Figure 5. Magnified dorsal tubercles of Tarentola species of the Cape Verde Islands.

Figure 6. Typical dorsal patterns of Tarentola species of the Cape Verde Islands (adapted from Joger, 1993).

Figure 7. Photographs of the dorsal and lateral views of Tarentola of the Cape Verde Islands. A1, T. boavistensis; A2, T. bocagei; A3, T. fogoensis; A4, T. darwini; B1, T. substituta; B2, T. raziana; B3, T. caboverdiana; C, T. nicolauensis; D1, T. gigas (T. gigas brancoensis on the left and T. gigas gigas on the right); D2, T. rudis; D3, T. protogigas protogigas; D4, T. p. hartogi from Brava Island; D5, T. p. hartogi from Rombos Islets; D6, T. maioensis.