Statistical analysesnext section
Fig. 1. Ordination of species A and species B along the first two principal components, based on a PCA of the characters VENT, SUBC, TAIL and EYED.
TwoStep Cluster analysis split the sample into two clusters and validated the initial assignment of specimens to the two species (henceforth: species A and species B). The results of a PCA are shown in Figure 1. A plot of the number of ventral scales against eye-diameter is shown in Figure 2. Both figures demonstrate a clear separation between the two species.
The most prominent results of univariate confirmatory analyses are summarized in Table 2. Species A has a higher number of ventral scales, a higher average number of subcaudal scales and a smaller average eye-diameter. The relation between SVL and eye-diameter for the two species is depicted in Figure 3. With regard to coloration, a conspicuous character of species A is a bright interparietal spot which is absent in species B. However, several specimens constituted exceptions to this rule. In one specimen of species A (from Bombay, India) and two specimens of species B (from Sri Lanka), an interparietal spot was present but in a rudimentary (i.e. barely visible) form. Finally, species A has a substantially shorter vertebral stripe than species B.
Table 2. Descriptive statistics with regard to the most diagnostic characters of species A and B. Eye-diameter is SVL-adjusted (see Materials and Methods). Significance-levels are provided in the last column.
Sexual dimorphism in the diagnostic characters was not detected in either species, a finding that is largely in line with results regarding several congeneric species (How et al., 1996; Vogel and Van Rooijen, 2007; Van Rooijen and Vogel, 2008).
Eye-diameter, number of ventrals and number of subcaudals were subjected to analysis of geographical variation. This analysis was restricted to OTU’s from which both species had been sampled, thus excluding OTU 3 and OTU 4. Only the number of ventral scales was found to exhibit significant geographic variation (P = 0.01), although this variation is negligible in the context of interspecific differences (P < 0.000001). With regard to this character, the two species were found to differ slightly in their patterns of geographic variation based on a significant interaction OTU x species (P = 0.02).
The analyses demonstrate unequivocally that two species are currently referred to the binomial Dendrelaphis tristis. Dendrelaphis tristis was described by Daudin (1803) on the basis of information on a single specimen published by Russell (1796). Daudin did not deposit a type-specimen in a collection as was usual at that time. He mentioned 181 ventrals and 130 subcaudals, enlarged vertebral scales and a whitish-grey first dorsal scale row. The enlarged vertebral scales and the light first dorsal row are consistent with both species but the high ventral and subcaudal counts are only consistent with species A. Therefore, we refer species A to the name Dendrelaphis tristis (Daudin, 1803). The next available name according to the rule of priority is Dipsas schokari Kuhl, 1820. This name was synonymized with D. tristis by Boulenger (1894). Kuhl also did not assign a type-specimen, nor a type locality, but he provided ventral and subcaudal counts for five syntypes. The range of ventrals and subcaudals (168-183 and 111-131 respectively) given by Kuhl indicate that both species were included in his type-series. In his general description Kuhl mentioned a vertebral stripe, a light ventrolateral line, a postocular stripe and 5-6 temporal scales. These characters are also consistent with both species. Despite the fact that Kuhl’s name Dipsas schokari refers to both taxa and a type-specimen was not designated explicitly, the name is available according to the International Code of Zoological Nomenclature (ICZN, 1999). The fifth syntype mentioned by Kuhl agrees beyond doubt with species B. In Figure 4, the number of ventrals and subcaudals of the specimens examined for this study are plotted together with those of Daudin’s type of D. tristis and Kuhl’s fifth syntype of D. schokari. Figure 4 shows that Daudin’s type is situated within the area occupied by species A whereas Kuhl’s 5th syntype is positioned within the area occupied by species B. Indeed, Daudin’s type differs significantly from species B as the ventral and subcaudal counts differ more than 2.5 standard deviations from the respective means of species B. In a similar vein, Kuhl’s 5th syntype differs significantly from species A as its ventral count is situated further than 3 standard deviations from the mean of species A. As such, there is no doubt that Dendrelaphis tristis (Daudin, 1803) is the correct name for species A whereas Dipsas schokari Kuhl, 1820, in the combination Dendrelaphis schokari (Kuhl, 1820), should be adopted for species B.
Fig. 4. Plot of the number of ventrals against the number of subcaudals for species A and species B. Ventral and subcaudal counts of the holotype of D. tristis and the 5th syntype of D. schokari are included.
An excellent overview of the life and work of Heinrich Kuhl is provided by Klaver (2007). Heinrich Kuhl (1797-1821) studied natural sciences at the University of Groningen, The Netherlands from 1816 to 1819. In 1820 Kuhl was employed by the Dutch government and was sent to Java to study the natural history of this island. After reaching Java he died within a year.
Kuhl described Dipsas schokari while he studied at the University of Groningen. The types mentioned by Kuhl were part of his personal collection. He donated this collection to the museum of natural history of the University of Groningen when he departed to Java. Unfortunately, the collection of the museum of natural history of the University of Groningen was destroyed in a fire in 1906. The syntypes of Dendrelaphis schokari should be considered lost. In order to stabilize the names D. tristis and D. schokari as well as designate a type locality for the latter, we designate RMNH 842 as the neotype of Dendrelaphis schokari (Kuhl, 1820) and SMF 58442 as the neotype of Dendrelaphis tristis (Daudin, 1803) on the basis of article 75 of the ICZN. The type locality of Dendrelaphis schokari becomes Ceylon (now: Sri Lanka). The type locality of Dendrelaphis tristis becomes Calcutta, India.