Digit reduction: via repatterning or developmental arrest?

Hamrick (2002) contested the conclusion of our recent article on evolutionary constraints on digit numbers that in tetrapods limb reduction usually occurs via arrest of the initial development followed by degeneration (Galis et al 2001). Firstly, we agree with Hamrick that we unfortunately illustrated this phenomenon with the wrong figure: the digits in the hand of Didelphis marsupialis are indeed not reduced. We could have chosen many good examples in mammals (e.g., Fig. 1), and we apologize for not having done so. Fortunately, our other two illustrations of the phenomenon are good examples of the transient presence of reduced digits in amphibians (Fig. 2). Furthermore, Hamrick (2002 in press) contested the embryonal presence of a digit I Anlage in the pig ( Sus scrofa ), because he could not find such Anlagen in his serial sections. However, Baur (1884), Emery (1891, in Braus 1906), and Schmidt-Ehrenberg (1942) all have documented digit I Anlagen in the pig. In addition, both Schmidt-Ehrenberg (1942) and Holmgren (1952) confirm the presence of digit I in another artiodactyl with even more reduced digits, the cow ( Bos taurus ), and illustrate this with particularly clear drawings of the digit I Anlage (Fig. 1). The transient presence of these digits shows that reduction occurred via developmental arrest followed by degeneration. How general is this evolutionary mechanism of digit reduction? Evolutionary digit reduction can occur via two processes, repatterning of the initial embryonal Anlage or developmental arrest possibly followed by degeneration. Lande (1978) in a review on limb reduction concluded that in amniotes, evolutionary limb reduction (including digit reduction) typically occurs by the slow continued evolution of earlier developmental arrest followed by degeneration. He documented how both in mammals and reptiles with reduced limbs the initial anlage is first developed and then regresses in absolute size (mammals: Phocaena communis , P. dalli. Laegenorhynchus acutus , and Megaptera nodosa ; reptiles: Python reticulatus , Anguis fragilis , O. apodus , Scelotes brevipes ). In the reptilian examples cell death was shown to cause the partial or entire degeneration of developed structures. Lande (1978) also documented developmental arrest followed by degeneration reduction of digits in perissodactyls (the horse, Equus caballus ), artiodactyls (the cow, B. taurus ), lizards ( Scelotes gronovii ), and in birds (the ostrich, Struthio camelus ). Reviewing the digit embryology studies of Holmgren (1952), Schmidt-Ehrenberg (1942), Krölling (1934), and Kindahl (1944, 1949), we found that in all mammals that still have digits, the pentadactyl state can still be traced early in ontogeny, even in horses. Reduced digits are either permanently present as rudiments (e.g., digit II and V in cows and digit V in the hand of the armadillo Dasypus and the golden mole Eremitalpa granti ) or only transiently (e.g., digit