Are lungfishes the sister group of tetrapods?

It is generally accepted that rhipaidistian crossopterygians are the closest relatives of tetrapods. Rosen, Forey, Gardiner & Patterson (1981) challenge this view and contend that lungfishes are the sister group of tetrapods. They present a detailed cladistic analysis and claim to identify a large number of synapomorphies shared by lungfishes and tetrapods but not by rhipidistians. Their analysis is faulty. Although Rosen et al. (1981) correctly emphasize that cladistic relationships must be based on shared derived characters, they often fail to take intragroup variation into account in postulating synapomorphies. They also use evidence inconsistently by attributing greater significance to similarities between lungfishes and tetrapods than to even more detailed similarities between rhipidistians and tetrapods. They misinterpret the skeletal pattern of the paired appendages. The many synapomorphies that they claim to have identified are either invalid, irrelevant, or are characters involving reduction or loss (which have a high probability of convergence). Consequently, they make an unconvincing case for a sister-group relationship between lungfishes and tetrapods. On the other hand, Rosen et al. (1981) do show that evidence for the orthodox view of rhipidistian-tetrapod relationships is not as strong as generally believed. The uncertain interrelationships among rhipidistians is a major problem. Tetrapod-fish relationships need to be re-examined by means of a properly conducted cladistic analysis.