|
GOULD--PRESIDENTIAL ADDRESS
|
321
|
interpretations are skewed perceptions born of a bias to
search for progress and direction in the history of lineages, and of an
ancient Platonic tendency to abstract varying systems by an "essential"
property. "Trends," in such cases, are byproducts or oddly-distorted
perceptions of changes in variance.
Many peculiar, even amusing, misinterpretations arise from the myopic
perspective that takes a macroevolutionary system powered by strong variation
in number of species through time, and reads this pulsation in items as
an anagenetic "trend" in one limited or peripheral part of the
system. Consider, for example, the supreme irony that the "standard"
evolutionary trends of textbooks must represent the history of unsuccessful
lineages I have dubbed this expression of our interpretive biases "life's
little joke"--see Gould, 1987). Clades are bushes that increase
or decrease in richness of branching through time. Each current twig,
of course, has its own labyrinthine pathway through the bush back to the
clade's common ancestor. But these pathways are stories for individual
twigs, not tendencies of entire clades. If a clade has been markedly unsuccessful
and now lies at the brink of extinction with but one surviving twig, then
our anagenetic biases click in, and we often read the single extant path
as an anagenetic trend. Thus, we celebrate little, many-toed eohippus
marching towards the large, noble, single-toed Equus. But Equus
is the sole survivor of a tree once lush and vibrant (in an early Tertiary
world with few artiodactyls and abundant perissodactyls). We speak of
the anagenesis of horses only because our biases abstract bushes as ladders,
and the clade of horses has been so depleted that only one lineage remains
to be misread as the terminus of a trend. All our textbooks cite horses
as the prototypical evolutionary trend, but there is no classical tale
about the evolutionary "trend" of antelopes, rodents, or bats-though
these are the true success stories of mammalian evolution by the more
appropriate criterion of increasing representation. I need hardly remind
my colleagues that one other prominent mammalian lineage, iconographically
stereotyped both in textbooks and in pop culture as a march to big heads
and less hairy, more upright bodies, shares with horses the precarious
status of sole survivor on a once more luxuriant bush.
Consider, also, life itself. We speak of growing complexity as a trend
from stromatolite to eukaryote, to grubby Tommotian cap, to larger and
complex invertebrate, to fishes, and finally to self-conscious mentality.
But we are only documenting the upper tail of a right-skewed distribution
that expands because the number of entities has grown -- not because
organisms on the right tail possess an intrinsic drive for further complexity.
The first fossil organism was a prokaryotic cell, and little (or no) room
exists to the left of this origin for any degeneration of complexity preservable
in the fossil record. Life's history has produced no trend at all in one
important sense. The modal organism on earth was, and has always been,
a prokaryotic cell. The number of E. coli in the gut of each human
being exceeds the sum total of Homo sapiens now alive.
Moreover, increase or decrease in number of species is not the
only mode of evolution that evokes a misreading of changes in variance
as anagenetic trends. Runnegar (1987) makes the valuable distinction between
diversity, or number of species, and disparity, or average
difference among species. If number remains constant and disparity increases
or decreases (as new kinds of species replace old), the same effects may
arise-for example, means and extremes may increase while modes remain
constant if the mode lies near one end of a potential range and average
disparity among a constant number of species grows.
This paper does not claim that all trends must be viewed as byproducts
of changing variation in numbers or disparity of species within clades.
Some classical anageneses are correctly read as directorial changes of
state within unbranching lineages
RELATIVE BRAIN SIZE (EQ)
FIGURE 2 -- Generalized frequecy distributions
for encephalization quotients for ungulates and carnivores during the
Tertiary. Note that distributions increase continually in range, but remain
anchored at their lower limit From Jerison (1973).
(though I suspect that this mode counts for rather little
of the directionality of life's history). But I will argue that many standard
"trends" of paleontology are distorted views of macro-evolutionary
systems with changing variance in numbers or disparity of species within
clades. This paper shall present examples of both "increase trends"
and "decrease trends" to support its central argument that trends
are often properly interpreted as changes in variance, not as anagenesis
of extreme values or central tendencies. Finally, I shall explore the
meaning of this altered interpretation in the light of the two most important
challenges now being offered against conventions of the Modern Synthesis
-- 1) constraints upon adaptation imposed by structure and development
(not only by history); and 2) hierarchical levels and causality.
INCREASE TRENDS AND THE STARTING
POINT PRINCIPLE
Consider two cases that rank among the most widely discussed trends
of this generation in paleontological research. Harry Jerison (1973, 1985)
has provided our best quantitative data in support of an old claim: that
brain size tends to increase within mammalian lineages (even after making
proper corrections for changes in body size). Jerison calculates an EQ
(encephalization quotient) for each mammalian species -- the ratio
of actual brain size to brain size expected at the organism's body size
for an average mammal on the mouse-to-elephant curve.
The mammalian data do show a tendency for general increase -- in
the sense that modes, as well as means and medians, shift to the right
through time (Figure 2). But the major phenomenon -- as Jerison has
always indicated -- is not linear increase, but expansion and flattening
of the distribution as numbers of taxa grow: in other words, an increase
in variance. The left ends of the frequency distributions remain anchored
at the small size of ancestral lineages, while the curve of EQ spreads
out, as numbers of species increase, into the previously unoccupied right
end of the range. In his major work, Jerison wrote (1973, p. 315-316):
Diversity evolved just as average size evolved .... Despite the general
trend toward increase in average brain size, there is an interesting
and important overlap in the region of low brain size which indicates
that there were at least some
|