Phylogenetic analysis is not "usually" based on molecular data. In fact, practically all such analyses were based on morphological characters prior to the 1960's. Even in the 1980's to early 1990's many analyses were still based on morphology alone. However, the ease with which DNA can be sequenced has made molecular systematics much more popular in recent years. The popularity of molecular systematics is understandable as well from the standpoint of reliability. Nevertheless, as I pointed out before, molecular data is not always available, especially in the case of fossils. Paleontologists therefore have to rely almost exclusively on morphological characters. That does not mean that such analyses are necessarily unreliable.

Dr. Michael Lee of the University of Queensland in Australia, for example, finds that snakes are closely related to varanoid lizards, and his morphological analysis has now been corroborated by Rest et al.'s (2003) mtDNA data. Dr. Lee is probably also correct that turtles are not diapsids in disguise because Rest et al.'s (2003) analysis showing turtles as diapsids is unreliable because of their choice of mammals as outgroup to the other amniotes. Mammals are amniotes too. Rest et al. assume (probably incorrectly) that mammals (actually synapsids) are the first group that branched off from the ancestor of the amniotes. Ironically, their tree then shows the placental and marsupials as two distinct lineages that form an unresolved polytomy with the lineage that is ancestral to all other remaining amniotes. That is of course anomalous because the placental-marsupial split occurred much more recently than the synapsid-anapsid split. Rest et al.'s tree shows the placentals and marsupials last shared a common ancestor at the same time as the synapsid-anapsid split!

What distinguishes Dr. Lee from most (but not all) other cladists is his willingness to analyze his characters. By doing so, he is able to achieve more reliable results than most cladists, who, in the words of Kurt Schwenk, "...cannot judge the quality of a character..." and must therefore merely "...assume that characters are independent and hope that...enough 'good' characters will outweigh ". "Hope" is not a good tool for phylogenetic analysis; character analysis (e.g. comparative anatomy) is far better.

Yes, a good systematist like Dr. Lee can still come up with reliable results using morphological characters only. However, it is next to impossible to construct a reliable phylogeny using morphological characters alone. That is because lineages evolve at different rates morphologically. A lineage (e.g. the whales) can diverge so much from their close relatives that virtually all signs of their ancestry have disappeared. In such cases, morphological data often fails not only to unite closely related groups in an analysis, but morphological data often fails to reveal the paraphyletic nature of a given group. No one, for example, knew that Artiodactyla was paraphyletic prior to the use of molecular data, which shows that whales are descended from an artiodactyl that is closely related to the hippo, the pig and cattle. In fact, hippos are more closely related to whales than they are to the camel, another member of the Artiodactyla. Despite the fact that Dr. Lee correctly identifies varanoid lizards as the closest relatives of snakes, his analysis fails to show that the anguimorphs are a paraphyletic group.

In sum, morphological characters can be a reliable source of phylogenetic information, especially if their goodness is carefully analyzed (for example by using detailed comparative anatomy). Where morphological analyses often fail is their inability to delimit holophyletic groups. Many and perhaps most of the groups that the cladistic morphologists think are holophyletic often turn out to be paraphyletic or even polyphyletic. Hence it is not a surprise that even morphologists are turning more and more to the use of molecular characters when they try to ascertain branching order. Morphology nevertheless retains an important role in systematics and taxonomy. Morphological characters serve as corroborative evidence for molecular analyses. In taxonomy, it is next to impossible to classify organisms without studying their morphological disparity. A branching diagram alone, especially if it is based entirely on molecular data, is practically useless for delimiting both species and higher taxa.

Reference: Schwenk, Kurt 1994. Systematics And Subjectivity: The Phylogeny And Classification Of Iguanian Lizards Revisited. Herpetol. Rev. 25(2):53-57

PS Character choice, as well as character coding is a subjective process; different systematists see different numbers of characters when looking at the same part of the animal and they of course more likely than not code their characters differently.

[ Reply To This Message ] [ Subscribe to this Thread ] [ Show Entire Thread ]