number of exceptions?² Proponents can always furnish their lists as well. And since each list must include a ridiculously small percentage of all possible cases, how can a theory of natural history be rejected by simple enumeration? We cannot know whether cenogenesis is " really" more common than palingenesis. What we can test is the validity of proposed laws—terminal addition and condensation—that entail the dominance of palingenesis.

In this chapter, I will treat the unsuccessful attempts of empirical cataloguers to refute Haeckel's theory of recapitulation. I will then argue that the biogenetic law fell only when it became unfashionable in approach (due to the rise of experimental embryology) and finally untenable in theory (when the establishment of Mendelian genetics con- verted previous exceptions into new expectations). The biogenetic law was not disproved by a direct scrutiny of its supposed operation;* it fell because research in related fields refuted its necessary mechanism. If these arguments offend some scientists' beliefs about the way science should operate, they reflect, nonetheless, the way it does operate.

An Empirical Critique

Many critics tried to refute recapitulation by using it unsuccessfully —by working within its confines to obtain confusion rather than phylogeny. This empirical approach engendered three main objections to Haeckel's doctrine:

     1. Acceleration is not general or equal for all organs. Each ontogenetic stage is an inseparable mixture of organs in different stages of ancestral repetition.

     2. Larvae and embryos have evolved many features as adaptations to their own mode of life. New characters can be introduced at any stage of ontogeny, not only as additions to the end of ancestral growth.

     3. Development can be retarded as well as accelerated. Embryonic or larval stages of ancestors can become the adult stages of descendants—a phenomenon directly opposite to recapitulation.

* I do not mean to imply that this direct scrutiny only served to buttress recapitulation. In fact, it produced interminable arguments that could not be resolved in the absence of firm criteria for distinguishing palingenetic from cenogenetic features. I am only arguing that there was nothing in these debates to compel believers to question the premises of the theory itself (Marshall, 1891, and MacBride, 1914, for example, express sorrow at the continued wrangling but are confident that further factual inquiry will resolve all the issues). The debates did, however, discourage many young scientists from working with the theory and led them to experimental work, with its promise of true testability. Both Roux and Driesch, after all, had studied with Haeckel.