In studying the functional significance of the coiled shell, it is important to be able to analyze the types that do not occur in nature as well as those represented by actual species. Both digital and analog computers are useful in constructing accurate pictures of the types that do not occur.

The classic biometrical study of phyletic evolution inKosmoceras(Brinkmann 1929) is evaluated using unpublished raw data provided by Professor Brinkmann. Most morphological characters show statistically significant changes over time yet it is difficult to provide an unequivocal biological interpretation for these changes. In a few cases, runs tests indicate that evolution was nonrandom in the sense that fewer reversals in the direction of evolution occurred than would be predicted from a null hypothesis based on a random walk. These cases suggest persistence of natural selection regimes for fairly long periods of time. In other cases, and with other kinds of testing, the random walk model cannot be rejected although failure to reject the hypothesis does not justify its acceptance! Thus, the contribution of random factors (either genetic drift or selection in a randomly changing environment) cannot be assessed with confidence. It is problematical also whether theKosmocerasseries represent significant evolution by phyletic gradualism or just the natural temporal variation that characterizes stasis (sensu Eldredge and Gould 1972).