Crisp & Chandler, Paraphyletic species 
827 
For example, trigonous leaves appear to be an autapomorphy for Daviesia idicifolia 
when it is assumed to be monophyletic as traditionally circumscribed (above). 
However, cladistic analysis using geographic forms as terminals nests three 
recognised species within D. ulicifolia: D. acicularis, D. aremria and D. microcarpa. 
None of these has trigonous leaves, so the character shows three homoplastic changes 
within the D. ulicifolia clade, and D. ulicifolia is shown to be paraphyletic. These two 
sources of error have opposite effects, respectively overestimation and 
underestimation of the number of paraphyletic species, although this is not to say 
that they will cancel out one another. 
With these caveats in mind, we have tabulated all known species and subspecies 
(206) from eight genera in two distantly related families (Appendix 1). These are 
Brachi/scma, Chorizcma and Daviesia (Fabaceae) and Alloxylon, Embothrium, Lomatia, 
Oreocallis and Telopea (Proteaceae). All have been subject to recent taxonomic revision 
and cladistic analysis at species level (Pate et al. 1989; Crisp 1990; Crisp 1991; Taylor 
& Crisp 1992; Weston & Crisp 1994; Crisp 1995a; Crisp 1995b; Crisp & Weston 1995). 
In this sample, methodological bias between taxonomic workers is minimised because 
all treatments involve Crisp as an author. Appendix 1 lists the hypothesized 
autapomorphies for every species and .subspecies. Where a taxon appears to lack an 
autapomorphy, we have listed its putative sister-group(s) whose formal recognition 
may render the metataxon paraphyletic. These data are summarised in Table 4. 
Telopea speciosissima (Proteaceae) is a good example of a metaspecies from Appendix 
1. As originally circumscribed, it possessed obvious autapomorphies in the enlarged, 
bright red involucral bracts of the conflorescence (see the cover of this journal), as 
well as toothed leaves, raised venation and pollen sculpturing. However, with the 
segregation of T. aspera on the basis of its autapomorphic harsh texture and rusty 
indumentum of the leaves (Crisp & Weston 1993; Crisp & Weston 1995), the 
autapomorphies of the old T. speciosissima became synapomorphies for a monophyletic 
group comprising both species (Weston & Crisp 1994). In its newer, restricted 
circumscription, T. speciosissima lacks an evident autapomorphy. Moreover, those 
populations of T. speciosissima which are geographically closest to T. aspera have a 
low density of rusty hairs on the leaves (Crisp & Weston 1993), and thus may be 
more closely related to T. aspera than to more distant populations of T. speciosissima. 
This hypothesis, though plausible, should be tested by cladistic analysis at the 
population level before T. speciosissima can be declared paraphyletic with confidence. 
Until then it should be considered a metaspecies. 
Within the same genus, T. oreades may also be paraphyletic. Its sister taxon is 
T. mongaensis, which has an autapomorphic absence of leaf sclereids (Weston & 
Crisp 1994). Absence of lobing in the early intermediate leaves is an apparent 
autapomorphy for T. oreades; however, this feature is very homoplastic within the 
subtribe Embothriinae (Weston & Crisp 1994) and recent morphometric studies have 
revealed a sporadic occurrence of lobes in some populations of T. oreades (Parrish & 
Crisp, unpublished). The geographic distribution of T. oreades populations also 
suggests paraphyly. Most populations occur in the east Gippsland region of Victoria; 
however, one highly disjunct population occurs sympatrically with the southernmost 
population of T. mongaensis, in New South Wales (Crisp & Weston 1987; Crisp & 
Weston 1993). In the area of sympatry, morphometric and molecular (RAPD) evidence 
indicate hybridization between the two species, and this site falls outside the main 
bioclimatic envelope of T. oreades (Parrish & Crisp, unpublished). Thus it seems 
likely that the disjunct population is more closely related to T. mongaensis than to the 
Gippsland populations of T. oreades, in which case, T. oreades would be paraphyletic. 
This hypothesis is currently being tested by molecular phylogenetic analysis of 
populations (Parrish & Crisp, unpublished). 
