Comparisons of frogs, humans, and chimpanzees.

Evolutionary biology is in need of a yardstick or metric with which to measure morphological evolution in creatures as diverse as frogs and mammals. Two types of metric have been suggested to measure morphological distance, one based on quantal traits and the other on quantitative traits. Quantal traits are those whose state varies and is scored as 0 or 1 (presence or absence of a tongue). Quantitative traits are those whose state varies continuously (length of leg) and in this case the degree of difference between states is measured quantitatively. Findley (1) recommends that quantal traits be used for comparing the magnitude of the morphological difference between humans and chimpanzees with that between frogs. He chooses six quantal traits such that the state in Xenopus differs from that in Rana, whereas the state in humans is identical to that in chimpanzees (1). We present a counter list of six quantal traits, for which the opposite result is obtained. For these traits, chimpanzees and humans differ, whereas Xenopus and Rana are the same. Table I records the results of comparing additional pairs of species with respect to the 12 traits. Frogs belonging to different species, genera or families are identical for all 12 of the traits, confirming the hypothesis that phenotypically chimpanzees differ more from humans than one frog family differs from another. Most numerical taxonomists, however, would argue that 12 quantal traits are too few for an adequate test. As chimpanzees and humans are reported to differ by at least 312 quantal traits (2), it occurred to us to invite Findley to try to come up with a similar number of quantal traits by which Xenopus and Rana could be distinguished. An exercise like this could be entertaining, but it might not have scientific value. We are not convinced that it is valid to use quantal traits for estimating overall degree of morphological difference between species. Our skepticism stems chiefly from considering the problem of how to avoid bias in the picking of quantal traits. This problem is illustrated by the contrasting results obtained in the above example. Depending on which of the two sets of six traits one picks, the chimpanzee human difference seems either small or evolution (6) large relative to the Xenopus Rana difference. We feel that quantitative traits may be more appropriate than quantal traits for obtaining reliable estimates of morphological distance. It is well known from studies of quantitative genetics that such linear traits usually exhibit continuous variation in genetic tests. Furthermore, in our own morphological work with hundreds of species (3, 4), every quantitative trait examined varies in length both within and among species. Variability within a particular taxonomic group is therefore not a criterion for choosing among quantitative traits. Cherry et al. (1) have developed a morphological distance metric based on quantitative traits from all major parts of the body. The metric is monotonically related to traditional zoological estimates of phenotypic distance among frogs (3). Findley (1) seems to have overlooked the significance of this empirical demonstration of the utility of the metric based on quantitative traits. Regardless of the approach preferred, it seems from the information already available that, relative to differences among frog families the morphological difference between humans and chimpanzees is large. However, at the protein sequence level, the chimpanzee-human difference is very small by frog standards. As pointed out before, species within a genus of frogs usually differ far more from each other in their protein sequences than do humans from chimpanzees (3, 5). Thus morphological evolution and protein sequence evolution can proceed at contrasting rates. This contrast has important implications for our understanding of the mechanism of