The discovery of genetically differentiated groups of populations within species raises the question of how inﬂuential genetic isolation has been in generating phenotypic diversity. Since morphological traits are products of both genetics and the environment (see e.g. Schluter, 2000) it is relevant, from an evolutionary perspective, to determine whether morphological variation reﬂects historical isolation or local adaptation despite recurrent gene ﬂow. Todistinguish between these non-exclusive hypotheses it is important to study a variety of traits. Differentiation arising from historical isolation of populations (i.e. with the stochastic effects of mutation and drift acting independently in isolated populations) will likely generate parallel patterns across characters of different nature, e.g. genetics and morphology, whereas adaptation across environmental selection gradients will not, or not necessarily so. Organismal constraints involving plasticity and autopoiesis may however cause morphological stasis despite considerable levels of genetic and environmental change through evolutionary time, such as in Plethodontid salamanders (Wake et al., 1983; see also Larson, 1984. In the case of cryptic closely related lineages one way of testing hypotheses on the extent and nature of phenotypic diversity has been the multivariate statistical analysis of morphometric data (e.g. Good and Wake, 1992; Arntzen and Sket, 1997; Puorto et al., 2001).
The golden-striped salamander, Chioglossa lusitanica Bocage 1864, is a streamside species inhabiting low and medium elevation mountainous areas in the northwest of the Iberian Peninsula (Fig. 1).
Alexandrino et al. (1997, 2000, 2002) analyzed allozyme and mitochondrial DNA variation and uncovered two genetically distinct groups of populations geographically separated by the river Mondego in central Portugal (group 1 south of the river and group 2 north of the river). The two groups represent lineages that separated in the early Pleistocene, probably as a result of climate change in combination with local environmental conditions and formed a secondary contact zone in postglacial times (Sequeira et al., 2005). The northern part of the present range was colonized from a refugium located in between the Mondego and the Douro. Major rivers such as the Douro and, more to the north, the Minho acted as barriers to dispersal, lowering genetic diversity through sequential bottlenecking (Alexandrino et al., 2000).
Morphological variation in C. lusitanica was ﬁrst observed by Vences (1990), but with only three populations studied in Portugal, robust spatial patterns could not be discerned. We here describe morphological variation within C. lusitanica to investigate if i) the diversiﬁcation of C. lusitanica in two population groups and ii) the decrease of genetic diversity in recently colonized areas, have been paralleled by morphological variation. If genetic drift has been underlying morphological evolution we expect concordance between genetic and morphological patterns of variation. Conversely, non-concordance between genetics and morphology is expected if directional selection and epistasis have been major sources of morphological variation (see Reed and Frankham, 2001). We speciﬁcally test morphometric and colouration pattern variation across the species’ range against the causal hypotheses of i) vicariance of groups 1 and 2 and ii) isolation by distance. Latitudinal increase/decrease in morphological trait diversity is examined against causal hypothesis of i) genetic heterozygosity ii) hybrid origin of populations and iii) geographic distance between populations.