742 THE STRUCTURE OF EVOLUTIONARY THEORY
features that diminish its influence, and others that enhance its power. Among
factors that weaken the potential of species selection, we may mention:
(1) The generally low population size of species in clades, and the generally
long life of species-individuals, both factors limiting the amount of variation
usually available for a process of selection.
(2) Unlike the organism, the species-individual does not actively suppress
selection at lower levels within itself. Since the individuals of lower levels, by their
shorter cycle times, present much more variation for selection (per given unit of
time), this unsuppressed lower-level selection may overwhelm the operation of
species selection.
(3) Species selection, as the analog of asexual reproduction at the organismal
level, becomes subject to the same important limit that favorable traits arising in
one individual cannot be transferred laterally (for mixing and matching) to other
individuals, but only vertically to direct descendants.
(4) Species selection is limited by particular structural constraints,
encountered only at this higher level, most notably the apparently unbreakable
correlation between origination and extinction rates, thus tying together by
negative interaction the very two phenomena that, if positively associated—that is,
high speciation with low extinction—could so powerfully accelerate any trend
produced by species selection.
But several features that grant potential strength to species selection will
counteract these negative forces and limits:
(1) Species selection may be theoretically weak relative to the power of
transformation by continuous selection of lower-level individuals (organisms in
this case) within species. But, in fact, such transformation by anagenesis rarely
occurs in nature, as the great majority of species exhibit stasis during their
geological lifetimes. With general anagenesis usually weak or inoperative, and
with effective organismal selection concentrated at the origin of new species and
their differentia (and thus also limited to the cycle time of species themselves),
species selection can become a predominant process.
(2) The population size of species in clades may be low, but each event of
speciation must produce difference from parental traits (at least enough to yield
reproductive isolation)—whereas events of organismal birth need not add any new
variation to the population. The amount of change per speciational event may be
large, even providing a potential macroevolutionary analog to macromutation. (At
such a point, however, we must also allow some possible weakening of selection's
power as well, for macromutation, by producing a completed form of change in
one step, deprives selection of its creative role in building adaptation gradually—
see Chapter 2.)
(3) At the species level, not only does each birth of a new individual include
novel variation that may be substantial, but the variation also arises in an adaptive
context (whereas mutation, the source of variation at the organismic level, will
usually be detrimental to the organism). Of course, the adaptive component in the
production of a new species-individual need only exist at the level of its own
causal origin—often the organismal level, rather than the