Hierarchy in evolution (from Plantinga)

David Campbell (bivalve@mailserv0.isis.unc.edu)
Mon, 18 Jan 1999 14:59:50 -0400

>Darwin predicted that phyletic hierarchies would be formed from the bottom up
>(in Linnaean terms). He states this clearly in _Origin of Species_ (Everyman
>Library Edition, pp. 109-122). Varieties and species, being constantly
>modified by natural selection, would eventually form what could be called
>genera, these in turn by further modification would form families, which in
>turn would form orders. He stops at this point. I call this "bottom-up"
>because the process of evolution begins with the lowest categories, according
>to Darwin, and eventually forms the higher taxonomic categories through the
>process of natural selection.
>The fossil record, however, fails to confirm this bottom-up direction.
>Numerous studies indicate that the direction is from the top-down, that higher
>taxonomic categories appear before lower ones in the fossil record. Once the
>phyla became organized in the Cambrian explosion, they began to differentiate
>into ever lower taxonomic categories over geologic time, until today only
>species and varieties are formed.

This pattern is expected from either a gradualistic or punctuated
evolutionary system because we define the major groups from the viewpoint
of the present.

In traditional evolutionary systematics, every organism is assigned to a
species, genus, family, order, class, phylum, and kingdom, plus various
subgroups or supergroups. By this system, all taxonomic levels appear
simultaneously. As soon as a new phylum evolves, it has one class, one
order, etc. Later evolutionary events may lead to a new group splitting
off from the old one. A more gradualistic view would suggest that it takes
a long time for a group to become distinctive and prominent enough to be
recognized as a phylum, not quite so long for a group to be recognized as a
class, etc., so we should expect phyla to have very distant origins,
classes almost as distant, etc. The separate evolutionary paths of the
descendants of sibling worms 600 million years ago by now may be recognized
as the difference between two phyla, but after another half billion years
we might be able to look back and see that two identical worms of today are
going to evolve into two very distinct phyla given enough time.

Another way of looking at it is to consider that most phyla are made up of
several classes and so on down. The phylum will be as old as the oldest
class it contains, etc. so that the average ages of each smaller
subdivision will be less than the average ages of each larger taxon. For
example, Chordata includes sea squirts, lancelets, conodonts, jawless fish,
sharks, bony fish, amphibians, reptiles, mammals, and birds, each assigned
to its own class. Averaging the times of appearance of these classes gives
roughly 350 million years, but the phylum fossil record extends as far as
the Early Cambrian or late Precambrian, when the oldest known conodonts
occur (classification of some of which is uncertain), about 545 million

Another factor affecting the pattern of appearance of novel kinds is the
limiting of opportunity. Once one organism has evolved the ability to do
something well, it can generally outcompete another that evolves a similar
ability. Thus, the dramatic differences we recognize as phylum-level
differences may not have been possible to evolve after the Cambrian because
all the good ideas were already taken. Many of the classes that appear
after the early Paleozoic are terrestrial, a place where there was less of
a history of evolution already in place than in the sea.

More recently, some of those who want to redo classification of organisms
according to a cladistic scheme want to throw out all groups that do not
include all the descendants of a common ancestor. For example, "reptile"
in the usual sense would be rejected because mammals and birds are
descended from reptiles, and crocodiles are more closely related to birds
than they are to lizards or turtles. Some people want to use "crown-group"
definitions, in which a formally-recognized taxon is based on the last
common ancestor of the living representatives and all its descendants. For
example, in the class Crinoidea (sea lilies), the Permo/Triassic extinction
killed what are traditionally recognized as several orders, and all modern
crinoids represent a small group that apparently evolved in the earliest
Triassic. In a crown-group definition, if the last common ancestor of all
modern crinoids was in the earliest Triassic, all pre-Triassic crinoids are
not assigned to a class. Crown-group definitions may change any time there
is an extinction of a group, a new group is discovered, or new study
suggests that the classification was mistaken and generally seem like a bad
idea to me. A more practical cladistic revision is an apomorphy-based
definition, in which all descendants of the first organism to have a
certain feature are grouped together. Thus, the phylum Chordata consists
of the first organism to have a notochord and all its descendants.
However, this creature would have been generic from the cladist's
perspective. Eventually, one of its descendants would have evolved the
distinctive feature identifying itself as a sea squirt, whereupon the first
class of chordates would have appeared. However, the remaining forms would
still be generic chordates until a new split occurred into the lancelets
and generic non-lancelets.
Any of these cladistic methods of defining taxa obviously forces
all lower-level taxa to be younger than the higher-level taxa that contain

David C.