832 
Telopea Vol. 6(4): 1996 
monophyletic species. Based on the frequency of metaspecies lacking evident 
autapomorphies, they may comprise a high proportion of all recognised species. 
Our partial survey of two angiosperm families suggests that 21% of species and 
subspecies are potentially paraphyletic. From a broad survey of speciation patterns 
in animals, Lynch (1989) concludes that 21% of animal species arose by either 
sympatric speciation or peripheral isolation, and thus are likely to have a 
complementary paraspecies (above). Ackery & Vane-Wright (1984) estimate the 
proportion of paraspecies to be 50% in the milkweed butterflies. All these estimates 
fall within the theoretical range (above), so are plausible. 
Biogeographic differences 
The empirical difference in the proportion of metaspecies between eastern (35%) 
and western (16%) Australia is intriguing, because it reflects variation within higher 
taxa that are common to both regions. Interspecific hybrids also appear to be more 
common in the east than in the west, e.g. within Daviesia (Crisp 1991). Perhaps both 
patterns are related to the same cause: a historical difference in mode of speciation 
in the two regions. By definition, a metaphyletic species has not diverged in 
morphology from its nearest ancestor. Therefore there is a good chance that it has 
not diverged reproductively from its relatives either, and it may be more likely to 
hybridise than a highly autapomorphic species. Further research should investigate 
what biogeographic differences may underly these east-west contrasts. 
Are paraphyletic species avoidable? 
Some cladists go out of their way to avoid recognising non-monophyletic taxa, even 
at species level. For instance, Thiele & Ladiges 0994) refrained from segregating the 
autapomorphic taxon aquilonia from Banksia integrifolia on the implicit grounds that 
the residual B. integrifolia would be paraphyletic (confirmed by K.R. Thiele, pers. 
comm.). Instead, they treated aqtiilonia as a subspecies, but this action probably only 
regressed the problem to a lower rank, that of subspecies. The remaining subspecies 
of B. integrifolia {compar, integrifolia and monticola) are not as distinct as aquilonia, and 
appear to be metaphyletic. This manner of avoiding paraphyletic species by regress 
does not solve the problem; it merely pushes it back to a lower rank. Another way 
of avoiding paraphyletic species would be not to recognise any taxon that leaves a 
paraphyletic residual. For example, no taxa might be recognised within B. integrifolia 
sensii lato, not even the autapomorphic However, this solution is unrealistic 
and unreasonable. Any well-corroborated, monophyletic taxon is worthy of formal 
recognition and sooner or later will be recognised. 
Some authors have suggested that paraphyletic taxa are an artifact of the Linnean 
system, because it prescribes mandatory categories (de Queiroz & Gauthier 1990; 
de Queiroz & Gauthier 1992). Under the rules of nomenclature, all organisms must 
be assigned to a species, genus and family. Therefore when a new family, genus or 
species is segregated from a taxon of the same rank, the rules require that the 
residual group be formally named at that rank too. If the segregant taxon is 
autapomorphic (i.e., monophyletic), then the residual is likely to be symplesiomorphic 
(i.e., paraphyletic). For example, the segregation of three autapomorphic species 
from the old Eucalyptus alpina (above) leaves a paraphyletic group of populations 
that must take a species name (in this case, E. baxteri). At higher ranks, say genus, 
the problem can be solved by splitting the residual group further into monophyletic 
taxa, each of which is named as a genus. However, at species level this solution is 
not possible — there are unlikely to be monophyletic groups within the residual. 
But even without the mandatory Linnean categories, paraphyletic taxa would be 
