the proper criterion for identifying higher-level selection (Lloyd and Gould, 1993;
Gould and Lloyd, 1999), I think that we finally reached, by a circuitous route
around many stumbling blocks of my previous stupidity, a consistent and truly
operational theory of hierarchical selection (see Chapter 8).
I must also confess to some preconditioning beyond punctuated equilibrium. I
had admired Wynne-Edwards's pluck (1962) from the start, even though I agreed
with Williams's (1966) trenchant criticisms of his particular defenses for group
selection, rooted in the ability of populations to regulate their own numbers in the
interests of group advantage. Still, I felt, for no reason beyond vague intuition, that
group selection made logical sense and might well find other domains and
formulations of greater validity—a feeling that has now been cashed out by
modern reformulations of evolutionary theory (see especially Wilson and Sober,
1998, and Chapters 8 and 9 herein).
My odyssey on the second branch of balancing internal constraint with
external adaptation in understanding the patterning and creative population of
novel places in evolutionary morphospace followed a much more complex,
meandering and diverse set of pathways. As an undergraduate, I loved D'Arcy
Thompson's Growth and Form (1917; see Gould, 1971b, for my first "literary"
paper), and wrote a senior thesis on his theory of morphology. But I thought that I
admired the book only for its incomparable prose, and I attacked the anti-
Darwinian (and structuralist) components of his theory unmercifully. I then took
up allometry for my first empirical studies, somehow fascinated by structural
constraint and correlation of growth, but thinking all the while that my task must
center on a restoration of adaptationist themes to this "holdout" bastion of formalist
thought—particularly the achievement of biomechanical optima consistent with the
Galilean principle of decreasing surface/volume ratios with increasing size in
isometric forms. I remain proud of my first review article, dedicated to this subject
(Gould, 1966), written when I was still a graduate student, but I am now
embarrassed by the fervor of my adaptationist convictions.
I emphasized allometric analysis, now in a directly multivariate
reformulation, in my first set of empirical studies on the Bermudian pulmonate
snail Poecilozonites (see especially Gould, 1969—the published version of my
Ph.D. dissertation). And yet, of all the long and largely adaptationist treatises in
this series, and for some reason that I could not identify at the time, the conclusion
that I reached with most satisfaction, and that I somehow regarded as most
theoretically innovative (without knowing why), resided in a short, and otherwise
insignificant, article that I wrote for a specialized pale-ontological journal on a case
of convergence produced by structural necessity, given modes of coiling and
allometry in this genus, rather than by selectionist honing (for some cases rested
upon ecophenotypic expression, others on paedomorphosis, and still others on
gradual change that could be read as conventionally adaptive): "Precise but
fortuitous convergence in Pleistocene land snails" (Gould, 1971c).
Five disparate reasons underlie my more explicit recognition, during the 1970's and
early 1980's, of the importance and theoretical interest (and icon-
Defining and Revising the Structure of Evolutionary Theory 43