kuento@kuhub.cc.ukans.edu (02/02/91)
In article <6781@harrier.ukc.ac.uk>, sss3@ukc.ac.uk (S.S.Sturrock) writes: > > While we are at it, the reason cladistics does not work is because it relies > on similarities of morphology, what about homomorphy? Also, how can it This isn't really quite that much of a problem with cladistics. In fact, it poses more of a problem for *other* taxonomic methods than it does for cladistics - which is why cladistics is still the method of choice, among other reasons. The biggest problem with cladistics is that it assumes all speciation to be vicariant (splitting off of populations) and dichotomous (a species gives rise to only one new lineage besides itself) - other modes of speciation will lead to cladograms which, to a practicing cladist, would be considered poor and/or unresolved. If there was one widespread hominid species which gave rise to *different* new species in different parts of its range, cladistics (assuming it gets everything right) would generate a polytomy, and no self-respecting cladist tolerates polytomies. Many other similar plausible evolutionary scenarios can be developed which would be rejected by cladists because of problems with the trees they result in. Cladistics also can be thrown off if there are taxa that belong in the tree and have not been included. > be applied to fragmentary remains of hominids? Cladistics relies on The problem is not how fragmentary the remains are, but how many *characters* can be gleaned from these fragments for inclusion into the analysis, and whether these characters are independent of one another. I know nothing of hominid cladograms, so I don't know how many characters they use, and whether they are informative characters or not. > primitive and advanced features to produce a tree, the problem today is that > all species are equally evolved, they are at the same point in time. Some > species may have a primitive feature (according to us) but may also be This is why cladistics relies on outgroups - to *AVOID* problems with arbitrary designations of primitive or derived characters. It sounds to me like you've never studied cladistic methodology...what is of critical importance is the presence of *shared, derived* features. The retention of primitive characters contibutes nothing to a cladistic hypothesis - only the distribution of *derived* traits that occur in more than one lineage. If you use an appropriate outgroup, then there is no problem deciding what is derived and what isn't. > advanced in other ways. Also, a similar external feature may evolve > more than once. If so, a proper cladistic analysis will *reveal* this fact, as long as there are enough *other* characters to rely on for a proper "structure" in the tree. The fewer the number of informative characters, the less likely you are to recognize parallelisms and convergences for what they are. If your only character was general body shape, you'd lump whales, ichthyosaurs, and fish together, but when you increase the data set, it eventually becomes obvious by the weight of the evidence that this grouping is based on a convergence rather than a shared evolutionary history. Cladistics is only as good as the data set, and the underlying assumptions. It ain't perfect but it IS the best we've got!! (I oughta know - I argue with cladists *all* the time). Cheers, -------(please include "DY" in subj header of mail to this user)-------- Doug "Speaker-To-Insects" Yanega "UT!" Bitnet: KUENTO@UKANVAX My card: 0 The Fool (Snow Museum, Univ. of KS, Lawrence, KS 66045) "Bobby, jiggle Grandpa's rat so it looks alive." "Roota! Voota! ZOOT!"