Biosystematic Studies of the Native Clematis Species of New Zealand with Particular Emphasis on Clematis Forsteri Sensu Lato

Abstract

The morphological and allozyme variation within and between New Zealand's native Clematis species was surveyed in this study, particularly with respect to the Biological. Phylogenetic, and Evolutionary Species Concepts. Using the isozyme data, the relationships between populations were also analyzed using two different phenetic clustering techniques and a cladistic technique. Using the Phylogenetic Species Concept and morphological data, it was possible to differentiate all of the native Clematis species (sensu Garnock-Jones, 1988), except perhaps C. forsteri sensu lato and C paniculata. However, using isozyme data and the Phylogenetic Species Concept, only populations of C. paniculata, C quadribracteolata and C vitalba were each separated. Similarly, only populations of these three species could be clearly distinguished as separate species in the phenetie analyses. That much morphological diversity was contained in the species that were difficult to distinguish using isozymes suggests that many of the native Clematis species may have diverged relatively recently, possibly through adaptive radiation. However, alternative explanations are that occasional hybridization between various species may have prevented divergence, or the sample of loci analyzed may have been conservative. The isozyme data did strongly support the treatment of C marata and C. quadribracteolata as separate species rather than members of a variable complex, which is in concordance with the delimitation of these two species using morphological data. The identity of two putative semi-juvenile C. foetida populations could not be confirmed using either morphological or isozyme data as they shared similarities with both C. cunninghamii and C. foetida, and few fixed differences were detected between these two species (although fulvous pubescence on the abaxial surfaces of the leaves appears to be diagnostic for C. cunninghamii). It seems likely that C. cunninghamii and C. foetida have only diverged recently. The allozymes present in the only known C. afoliata population in the local Wellington region (Waipapa Stream) were very similar to those of the South Island populations, which is reassuring given the regionally threatened status of the former population. The introduced species Clematis vitalba, was genetically diverse from all of the native species. An analysis of the Clematis herbarium specimens held at WELT indicated that two discordant entities appear to have been included in Allan's (1961) description of C.petriei: an entity including the nomenclatural type and corresponding to C. petriei sensu H. & C and another entity (comprising populations from Marlborough with predominantly entire leaflets). Thus, it appears that the C. forsteri sensu lato complex is actually made up of five major entities rather than four (C australis. C. forsteri sensu stricto, C. hookeriana, C. petriei sensu H. & C and C. sp. (= C petriei Allan pro parte)). Much morphological variation was detected within and between populations of C. forsteri sensu lato. To an extent, this variation correlated to the differences described by Allan (I961). and Heenan and Cartman (1998) between various entities (or groups of entities). However, much overlap in variation occurred between populations and it was not possible to delimit any particular entity or group of populations using morphological data and the Phylogenetic Species Concept. This is with the possible exception of C. petriei sensu H. &C. which may be distinct in having glabrous achenes (although only limited data about this character was analyzed in this study), and the sepal colour and flower scent may also be somewhat different (although these characters are difficult to quantify). Other proposed morphological differences between C. petriei sensu H. & C. and the other members of the complex (e.g. sepal base staining, adaxial sepal pubescence, and bracteole size and shape) were not found to be consistently diagnostic characters, although in the Canterbury region they were often useful. Although fixed allozyme differences were not detected between any of the morphologically identified entities (or other groups of populations) in the C. forsteri sensu lato complex (apart from one which involved a locus for which the genetic basis of the observed banding patterns was uncertain), geographical patterns in the distribution of particular alleles at eight loci were detected. These patterns often corresponded with the geographical distribution of each entity, and some reproductive isolation between sympatric populations of different entities also seemed apparent. Rare alleles also tended to be restricted to either side of the Southern Alps, and/or Cook Strait, and thus it is proposed that these geological features act as barriers to gene flow between C.forsteri sensu lato populations. These isolation mechanisms appear to be incomplete though and thus conferring species status to most of the entities of the C.forsteri sensu lato complex is not fully supported. However, some genetic divergence of the C. sp populations sampled was found (and this was considerable in some analyses, although the results were not consistent when all analyses were considered), and so further research on populations of this entity is recommended. Further analysis of other Marlborough populations is also recommended, as two populations in this study showed morphological characteristics of both C. petriei sensu H. & C. and C. sp. but whether this represented natural polymorphism within either population or was caused by hybridization or introgression could not be resolved using the isozyme data achieved in this study. No association between the genotype of an individual within a population and its morphology was detected. Indeed, in populations in which two extreme leaf forms were detected, individuals with both leaf types often had exactly the same genotype.