The Fruitful Facets of Galton's Polyhedron 367
of their theory (see Gould, 1977b, Chapter 4). These two scientists proposed a new
mechanics for the biogenetic law, and this "principle of acceleration" forms the
core of their theory and the basis of its orthogenetic implications. Ontogeny will
recapitulate phylogeny provided that two necessary principles can be validated. All
supporters of the biogenetic law promoted some version of these principles. First,
new characters must arise in phylogeny as additions to the end of previous
ontogenies. This principle of "terminal addition" (Gould, 1977b) would, of itself,
engender recapitulation, as descendants pass through the adult stages of ancestors
before accreting their own novelties in ontogeny—except for a logical problem that
required the second principle for a full and coherent theory. The phylogeny of a
lineage unfolds through thousands of steps in geological immensity; new stages
cannot be added indefinitely to the unaltered ends of previous ontogenies, lest
growth to adulthood take untold years to reach completion. Some process—some
law of heredity—must produce a general speeding-up of development, so that
ancestral ontogenies can unfold more rapidly, leaving time at the end for addition
of novel features.
All recapitulationists necessarily defended some form of speeding up for
ancestral ontogenies through phyletic time. Haeckel himself, who thought a great
deal about genealogies but precious little about mechanisms, advocated a
differential dropping out of stages, with compression of the remaining steps to a
shortened ancestral ontogeny (Gould, 1977b). Cope and Hyatt, who both devised a
theory of recapitulation in 1866 (independently of each other and of Haeckel), first
proposed an ultimately more popular and plausible version of ontogenetic
quickening—the "principle of acceleration," or general increase in rate of
development (with no necessary excision of stages). This law of acceleration (as a
foundation for recapitulation) became the most important theoretical contribution
of the American school (Fig. 5-2).
The law of acceleration held that ancestral ontogenies unroll more and more
rapidly in successive descendants, thus making room for new stages in phylogeny.
(The gill slits of a human embryo can therefore represent the compressed and
accelerated adult form of our piscine ancestry.) With ontogeny thus depicted as a
quickening treadmill through time, attention could shift to the nature of stages that
accumulated in successive terminal additions—for the accreted stages form a series
that defines the lineage's phylogeny. If the sequence of accreted stages represents
the unfolding of an internal "program," then phylogeny can be justly called
orthogenetic—for the law of acceleration makes room on the treadmill, and the
next stage of a predictable sequence then struts its hour upon the adult stage.
Later versions of "Neo-Lamarckism" in this American style cannot be called
orthogenetic because they did not propose an internally programmed series of new
stages. These later accounts embody a view of causality every bit as functionalist
as Darwin's, though relying upon Lamarckian mechanisms rather than natural
selection. New features arise adaptively as organisms actively respond to needs
imposed by altered environments. The sequence of novel stages accretes as a
contingent series, mapping the functional requirements of a changing external
world.