E

One common, but contentious (Nixon and Carpenter 1993), method for determining polarity is outgroup comparison, with outgroup defined as a taxon that is outside the taxa being analysed (the ingroup). A character state in the outgroup is likely to have been ancestral relative to the taxa under investigation. Technically, any taxon can serve as an outgroup, but the more closely it is related to the ingroup taxa, the greater the number of character states whose polarity

might be clarified. Selecting an outgroup roots the tree below the intersection of the outgroup and ingroup; ingroup taxa are then arranged in terms of their best fit relative to the character states examined. Although in the simple examples presented thus far all character states evolved only once, such is rarely encountered in real-world situations. More likely a tree will contain multiple character states that show up in lines not related directly through one common ancestor. These are homoplasies. One kind of homoplasy results from character state reversals - meaning that character state A changed to state A' and then at some later point in the set of related lineages reverted to state A. This kind of homoplasy is a classification problem (O'Brien et al 2001,2002) because rarely will precisely the same character state re-emerge after it disappears (Dollo's Law). More likely, classification makes it appear as if the character state has reappeared. Another kind of homoplasy results from parallelism or convergence - organisms, perhaps because of anatomical and/or environmental constraints (the first the result of common history, the second because of similar environments), independently evolve the same character state (Figure 6.4). All but the simplest cladograms contain homoplasy, and the analytical task is to reduce its influence on phylogenetic hypotheses - probably the most difficult problems in cladistic analysis. Homoplasy leads to multiple solutions to arranging taxa, and it is up to the analyst to sort through the solutions and defend why one of them was chosen. Various computer programmes (eg, Swofford 1998) simplify this task and produce indices by which to judge the overall strength of the ordering.