Evolution can be inferred from studying extant populations of 
organisms. The process of evolution can be directly observed. The
inference of common descent can be made from this and an examination
of how traits are distributed within the living world. If all lineages
coalesce back to a common ancestor, then biological traits should
be distributed across groups in a nested fashion. This is because a
trait would evolve and be passed onto all its descendent species.
Within the descendent species new traits would be evolved and
passed onto their descendent species and so forth yielding the
following pattern:

g     h       i       j       k       l       m       n       
c     c       d       d       e       e       f       f       
a     a       a       a       b       b       b       b

Imagine each three letter "word" above is a species and each
letter designated a trait. Species could be grouped into
two broad catagories: those with a and those with b. Within a, 
species could be grouped into those with c and those with d
and so on. This is the pattern that would arise if all a
species had a because their common ancestor had a. Then, the a
species gave rise to two descendent lineages -- one that evolved c
and one that evolved d. 

Deviations from this pattern can arise for several reasons, some
real and some simply artifact. 

REAL
1.) If genetically based traits can move "horizontally", that is, can 
be transferred from one lineage to another, then a non-nested 
distribution of traits will be seen. (Imagine that the species
acg (above) donated a c to one of the b species.)

2.) If species did not actually share a common ancestor and an
intelligent designer used an assembly line process to
generate all extant species. (Take a look at the box of SimLife
sometime for one artist's rendition of what a designed ecosystem
would look like.)

ARTIFACT
1.) A trait can be lost in some lineages. (Of course, loss of a trait
can be counted as a new trait and then nesting is re-established.) 

2.) A trait can be present, but not identified as the trait
      it can be highly modified (this isn't all that different from loss) 
      it can be "hidden" in some stage of ontogeny (all vertebrates
              have gill slits, but most (besides fish) have them 
              "covered over" in early development)

3.) Two independent traits can be mistakenly thought to be homologous
      (the trait "wing" would be non-nested if insect wings, bird
      wings, pterosaur wings, flying fish wings and bat wings were
      thought to be the same trait)
      if convergent or parallel evolution is prevalent, this can
      be a problem
 
Traits that can be examined for this purpose are: gene sequences,
protein sequences, morphological traits and developmental patterns.

In addition, non-random patterns of biogeography would be expected 
if all organisms shared a common ancestor. There should be a
relationship between spatial distribution and phylogenetic
distance for groups of organisms. This relationship will be
modified by the time since divergence of species and dispersal
capabilities of the species. (You'd expect to find a neater
pattern amongst plants than bats, for instance.) This can 
be misleading if a once wide ranging species is reduced to relic
populations. Let's say a species of beetle covered all the US, but
now is only present in Oregon and Florida. (This is where a fossil
record can help out.) On a planet where everything had evolved,
but was newly transferred there (as per your question) you
wouldn't see this. Prior to the acceptance of plate tectonics,
some species' distributions were quite puzzling. 

Finally, if you posit natural selection as your mechanism of
evolution; further evidence can be gained from evidence of
modification with descent (i.e. "jury-rigged" design).


Chris Colby   ---      email: colby@bu-bio.bu.edu     ---