PHYLOGENETIC RECONSTRUCTION OF RECENT SHORT-LOOP TEREBRATULIDE BRACHIOPODS: COMPARING SOURCES OF VARIABILITY

How do different sources of heritable data compare in their reconstruction of phylogenetic relationships? Brachiopoda provide excellent case studies to answer this question. Our previous analyses of extant Rhynchonellida generated conflicting results from morphological and molecular data. We chose to compare these results with investigations of phylogenetic relationships among extant short-looped Terebratulida (Terebratulidina; Brachiopoda). Parsimony analyses of 23 terebratulidines plus five outgroup genera for 70 qualitative morphological characters result in relatively well-resolved paraphyly or polyphyly of all three superfamilies; adding 15 quantitative characters of loop morphology actually reduces resolution; adding 18S rDNA sequence data, available only for 13 of the 23 genera, provides considerably greater resolution. Bayesian analyses of qualitative morphological characters yielded almost no resolution in tree topology; adding quantitative characters of loop morphology did not improve resolution; adding molecular sequence data again produced greater resolution. Analyzing only those 13 taxa for which all three sources of data are available yielded a well-resolved topology that separates one of the three extant superfamilies clearly from the other two superfamilies, which were intermixed.

Of what broader significance are these kinds of comparative analyses? First, it is not possible to test morphological patterns in extinct taxa with molecular data, which increases the importance of understanding the nature and reliability of the morphological data in hand. Second, when investigating phylogenetic patterns that can only be detected morphologically (e.g., body size evolution), independent molecular data can provide a critical test. Very small adult size is often considered to be a secondary paedomorphic adaptation among extant articulated brachiopods, possibly related to advantages gained by the earlier acquisition of sexual maturity. Our preliminary results suggest that very small size is a basal feature rather than a derived feature among short-looped terebratulidines. When studying macroevolutionary patterns and processes, combining all relevant sources of information is very likely to yield the most robust phylogenetic results.