A reconciliation of allopatric speciation with long-term trends can be formulated ... We envision multiple . . . invasions, on a stochastic basis, of new environments by peripheral isolates. There is nothing inherently directional about these invasions. However, a subset of these new environments might... lead to new and improved efficiency ... The overall effect would then be one of net, apparently directional change: but, as with the case of selection upon mutations, the initial variations [species] would be stochastic with respect to the change [trend].
Several paleontologists groped towards a generalization during the next few years, but Stanley (1975, 1979) made the greatest headway in appreciating the full generality of such an analogistic procedure for macroevolutionary theory: “In this higher-level process species become analogous to individuals, and speciation replaces reproduction. The random aspects of speciation take the place of mutation. Whereas, natural selection operates upon individuals within populations, a process that can be termed species selection operates upon species within higher taxa, determining statistical trends” (Stanley, 1975, p. 648).
Stanley preceded this statement with a claim that I regard as fully justified and prescient, but that became a lightning rod for unfair criticism: “Macroevolution is decoupled from microevolution, and we must envision the process governing its course as being analogous to natural selection but operating at a higher level of organization” (1975, p. 648). Largely on the basis of this claim about “decoupling,” Stanley, Eldredge and I, and others, were often accused of trying to scuttle Darwinism, and to invent an entirely new (and fatuously speculative) causal apparatus for evolutionary change (meaning, and explicitly so stated in this reductionistic critique, a new genetics).
We made no such claim, and the words quoted above speak for themselves. We were trying to explore the different workings of selection on individuals at levels of the evolutionary hierarchy higher than the conventional Darwinian focus upon organisms. Not only do I continue to regard this procedure as [Page 716] fruitful and fully justified, but I would also defend such an effort as the basis for an independent macroevolutionary theory that can harmoniously expand our conventional and exclusive focus on organisms to yield a more satisfactory general account of life's workings and history.
I also continue to regard the individuality of species as the central proposition of such an expanded theory. If organisms are the traditional units of selection in classical Darwinian microevolution within populations, then species operate in the same manner as basic units of macroevolutionary change. This perspective establishes an irreducible hierarchical structure in nature, precluding the smooth upward extrapolation of microevolutionary change within populations to explain evolution at all scales, particularly phenotypic trends and patterns of diversity displayed in geological time — the proposition that true devotees of microevolutionary exclusivism rightly feared. If species, as stable units and genuine evolutionary individuals, interpose themselves between populational anagenesis and trends within clades, then the lower-level process cannot smoothly encompass the higher-level phenomenon. For this fundamental (and excellent) reason — and not because any “new” genetics or anti-Darwinian forces reign in a threatening world of macroevolution — Stanley introduced his key notion of “decoupling.”
The levels become decoupled because macroevolution must employ species as “atoms,” or stable and basic units of change. Decoupling then becomes intensified because higher levels exhibit allometric properties that distinguish their phenomenology from the workings of lower levels. Thus, macroevolution with species as individuals must differ, in deep and interesting ways, from microevolution with organisms as individuals. These differences, and not any fatuous claims about “new genetics,” express the uniqueness of macroevolution, and the validity of our argument for decoupling.
An extensive analogy — “the grand analogy,” if you will (see Gould and Eldredge, 1977, p. 142) — between organismal microevolution and speciational macroevolution provides a good tool for assessing the differences imposed by scaling among the levels. Stanley (1975, p. 649) and Gould and Eldredge (1977, pp. 142-145) proposed some partial and preliminary schemes, and several others have added components along the way (Stanley, 1979; Vrba, 1980; Grantham, 1995, for example). I present this grand analogy below, largely in the form of a chart contrasting the key features of organic structure and evolution in their organismal and speciational manifestations. For each major category, I list the most important differences between the levels. A fuller explication of all items on the chart follows.
THE PARTICULARS OF MACROEVOLUTIONARY EXPLANATION
The structural basis
The first category of structural differences seems straightforward enough. In order to construct the analogy, we ratchet the focal level of individuality up from the organism to the species, thus redefining both lower components and higher contexts in the structural triad of part-individual-collectivity (see page
[Page 717]
Table 8-1. The Grand Analogy
Feature
Organismal Level
Species Level
I The Triad of Structure
1. Individual
Organism
Species
2. Part
Gene, cell
Organism, deme
3. Collectivity
Deme, species
Clade
2a. Usual result of
proliferation of one pan to
crowd out others
Cancer
Immediately adaptive
anagenesis
II The Criteria of Individuality
1. Production of
new individuals
Birth
Speciation
2. Elimination of individuals
Death
Extinction
3. Sources of cohesion
i) Stability of individual
Physiological homeostasis
in ontogeny
Sources of stasis in punc-
tuated equilibrium
ii) Boundaries against
invasion
Skin to delineate; immune
system to police
Reproductive isolating
mechanisms
iii) “Glue” of subparts
Functional integration &
division of labor
Social structure & behav-
ioral interaction among
parts (organisms); re-
combination in sexual
reproduction to mix
parts in their replica-
tion
4. Inheritance
Asexual by budding from
one individual, or sex-
from one individual
ual by mixture of two
individuals
“Asexual” by budding
from one individual
5. Source of new variation in
newborn individuals
Mutation
Geographic (or some
other form of) isolation
& selection (mecha-
nisms), causing differ-
ences that break repro-
ductive integrity
4a. Spread of new variation to
other individuals in the
collectivity
Recombination in sexual
reproduction
Generally absent except
for hybridization be-
tween species in some
clades
5a. Frequency of new variation
in replicated individuals
Very rare for any single
trait
Inherent in birth process
and always present
III Modes of Change in the Collectivity
A. Drives, or Directional Variation Within or Between Individuals
1. Heritable ontogenetic
&n
bsp; change within the
individual=ontogenetic
drive
Lamarckism—powerful if
it occurred, but pre-
cluded by nature of he-
redity
Anagenesis (gradualism
within species); rare by
punctuated equilibrium
[Page 718]
Table 8-1 (continued)
Feature
Organismal Level
Species Level
2. Biased production of new
individuals=reproductive
drive
Mutation pressure
Directional speciation
2a. Frequency of biased
production
Very low (if harmful to
organism) because
organismal selection
effectively suppresses
lower levels
Potentially common for
two reasons: 1) species
processes don't strongly
suppress lower-level se-
lection; 2) new individ-
uals must originate
with change from par-
ent
B. Selection, or Differential Proliferation Due to Traits of Interactors
1. Name of process
Natural (organismal)
selection
Species selection
2. Basis in birth
Differential birth
Differential speciation
3. Basis in death
Differential death
Differential extinction
2a. Reason for non-
directionality of variation
as precondition of
selection's power
Inherent in nature of
mutation as unrelated
to needs of organism
No necessary reason;
benefits of organism &
species frequently coin-
cide. No relation if im-
mediate adaptive con-
texts of new species
uncorrelated with direc-
tion of trend. Testable
as Wright's Rule
2b. Distinctive feature of birth
bias
Usually internal to organ-
ism; need not lead to
adaptation to environ-
ment
Usually irreducible as
based on traits of pop-
ulations, not organisms
3a. Distinctive feature of death
bias
Usually yields adaptation
to local environment
Often reducible as simple
summation of organism
deaths
C. Drift, or Random Differential Proliferation
1. Within the collectivity
Genetic drift
Species drift
2. In founding of new
collectivities
Founder effect
Founder drift
1a. Frequency
Rare except in special cir-
cumstances of small
populations, or neutral-
ity of many genie sites
Common because most
clades have low N;
intensified by reduction
of N in mass extinction
2a. Frequency
Common; depends on N
of founding population
Very common for two
reasons: 1) necessary
(and often large) differ-
ence from ancestor at
each founding; 2)
greatly different poten-
tials in allopatric re-
gions
[Page 719]
Table 8-1 (continued)
Feature
Organismal Level
Species Level
IV External and Internal Environments
A. Competition and the External Environment
1. In direct contact
Most often biotic
More likely to produce
an effect by differential
elimination
2. Not in direct contact; often
allopatric
More often abiotic
More likely to produce
an effect by differential
birth
1a. Main feature
Major source of occa-
sional biomechanical or
general progress
Often reducible to
organismal level
2a. Main feature
Adaptation to local cir-
cumstances; no general
vector
Usually irreducible
B. Constraint and the Internal Environment
1. Limits on runaway change
by directional evolution of
parts
Lamarckian inheritance
doesn't occur
Punctuated equilibrium
suppresses anagenesis
by stasis
2. Structural brakes upon
change
Design limits of Bauplan
Positive correlation of
frequency of speciation
and extinction appar-
ently unbreakable
3. Variational brakes
Rarity of new mutation
allayed by recombina-
tion in sexuals; serious
in asexuals (allayed by
short generations in
many unicells)
Sufficient change per new
individual, but low N
of species in clades
4. Developmental brakes
Von Baer's laws of com-
plex ontogenesis
Hold of homology
5. Positive channeling by
structure
Heterochrony and pre-
ferred ontogenetic ex-
tensions
Differential ease &
permissibility of
Bauplan modifications
6. Positive channeling by
variation
Not important, given rar-
ity of directional varia-
tion
Frequent correlation of
directional speciation
with differential prolif-
eration
7. Size of exaptive pool
High in such crucial cir-
cumstances as genetic
redundancy usable in
evolution of complexity
Generally high because
lower levels not sup-
pressed and frequently
correlative
[Page 720]
673). But this basic ratcheting already reveals some pivotal differences between the evolution of organism-individuals and species-individuals. In Table 8-1, line I2a, for example, notes the profoundly different outcome that usually ensues when particular parts of the individual proliferate differentially and crowd out other parts. Such a process usually spells disaster for a complex multicellular organism — and we call the result cancer — because parts lack independent viability (and therefore harm both themselves and their collectivity, the organism, by unchecked proliferation), and because organisms build coherence (an important criterion of individuality) by functional integration and division of labor among parts. But species achieve equal coherence by other routes. The parts of a species — that is, its component organisms — do have independent viability; moreover, their interests in proliferation often coincide with the health of the enclosing species. Thus, in a species-individual, differential proliferation of some parts at the expense of other parts does not lead to death of the full entity, but usually to adaptation by anagenesis.
Criteria for individuality
Moving to the second category of criteria for individuality (see pp. 602–613 of this chapter), we may regard the species-level analogs of o
rganismal birth and death (lines III—2) — speciation and extinction — as both evident and well recognized. But the different causes of cohesion (line 113) are both fascinating and portentous throughout the chart. I only remind readers that the mechanisms used by species, while not clamping down so hard on lower levels, and therefore providing substantial “play” for interaction between organismal and species selection, provide species with as much coherence and stability as the “standard” devices of morphological boundaries, internal policing and functional integration among parts, do for organisms.
Important differences arise in the mode of production for novel variation in newborn individuals. Mutation supplies this attribute at the organismal level. (Following conventional usage, I consider recombination in sexual organisms as a device for spreading variation among individuals, although I recognize, of course, that novel combinations also arise thereby. In asexual organisms, a better analog for species in any case, mutation alone supplies new variation.) Speciation itself is not the proper analog of mutation at the species level (an error previously made both by me, in Gould and Eldredge, 1977, and by Stanley, 1975). Speciation, the production of a new species-individual by budding, is the analog of organismal birth, particularly the birth of asexual organisms. We made this error by inadequately interpreting one of the most interesting differences between organisms and species as evolutionary individuals. The birth of a new organism, particularly in asexuals, may or may not engender any substantial difference from parental form or genetics. But the birth of a new species necessarily includes the generation of enough difference from ancestors to preclude reproductive amalgation between the parts (organisms) of the two species. We therefore mistook a forced correlate of birth at the species level (change at speciation) with the process of [Page 721] birth itself (speciation) — and equated the correlate at one level with the phenomenon at the other. The proper analog of mutation, a source of variation for new individuals, is the change that insures reproductive isolation between species (with geographic isolation as a usual precondition, and drift and selection as mechanisms) — see line 115.