That is an interesting question. I didn't know of any
transitional sequences proposed to connect two separate
animal phyla via some common ancestor. But
Jonathan Woolf pointed out something to me:
For an example of a phylum-level transitional, please
see THE CRUCIBLE OF CREATION, by Simon Conway Morris, pp.
185-95. Conway Morris uses details of micro-anatomy to
connect the Cambrian fossils called _Wiwaxia_ and
_Halkieria_ to each other and also to the modern phyla of
Mollusca, Brachiopoda, and Annelida. One of the most
impressive fossils in this series is a species of
halkieriid which looks sort of like a big worm with shells
covering the tops of its anterior (head) and posterior
(tail) ends. From what we currently know of genetics, it
probably wouldn't be very difficult to shorten the length
between the shells, then bend in the middle so that the
posterior shell is below the anterior shell, and finally
attach the two shells together. The final result of this
would be something very similar to a primitive
brachiopod.
Further evidence for this comes from a living species of
brachiopod, called _Neocrania_. _Neocrania_ starts life as
a mobile, wormlike animal. As it grows, it builds shells at
both ends, then folds in the middle and anchors itself to
become a proper, sessile, shelled brachiopod.
But this brings up an interesting second question: How
likely might it be that we would find some phylum-level
split among animals in the fossil record?
Obviously, a phylum-level split is simply another
speciation event when we get down to the last common
ancestor that links two phyla. We have some numbers for
estimates of how many species have been described from
fossils and estimates of total species numbers over life's
history on earth. Raup is commonly cited in this regard,
and gives about 250,000 described fossil species, as
compared to about 5 billion total species that have ever
lived on earth. I don't have a precise number for animal
phyla, but between living and extinct groups the number
probably does not exceed 100. If we considered it possible
that any two phyla could be linked (an obvious error, but
one which favors the anti-evolutionary interpretation if
anything), we could be seeking any of 10,000 speciation
events in order to find one such phylum-to-phylum
transition. But observations of species is not the same
thing as observations of fine-grained speciation events in
the fossil record. In this case, the anti-evolutionary
interpretation would be better off if we over-estimated the
proportion of fine-grained species-to-species transitions.
Let's say that 10% of observed fossil species include
fine-grained species-to-species transitions.
OK, we have a bunch of numbers biased toward, if
anything, the anti-evolutionary side of the ledger. What
about the kicker, our expected likelihood that we should
have seen a phylum-to-phylum transition? That would be the
proportion of observed fine-grained species-to-species
transitions to total species times the proportion of
possible species-to-species events that could represent
phylum-to-phylum events to total species. Or...
(Total fossil species * Fine-grained speciation
proportion) / Total species
*
(Number of possible transitions of interest / Total
species)
=
(250000 * 0.1 / 5000000000) * (10000 / 5000000000)
= 0.000005 * 0.000002
= 1x10-11
In other words, the odds that we might expect to have
already have an event representing a animal
phylum-to-phylum link among our already known fossils is
1E-11.
1E-11 is a very small number. An event meeting these
odds would be like winning a couple of lottery jackpots in
a row.
The fossil record is not like a videotape recording
every event in life's history.
Wesley