>
> >The feral population does not exist in a bell jar.


No forms of life do (unless they happen to be placed into a bell jar).

>
> >Here in Australia they appear to do just that.


And elsewhere, too (see Clinal geographic  variation in feral honey  bees
in California, USA , Apidologie 22, also data from Turkey).

It is pretty clear that feral populations can maintain independent genetic
identity despite the influx of managed populations.  The reason is simple:
nature selects mercilessly against alleles nonadaptive to the local weather,
seasons, flora, and parasites, and on the other hand incorporates without
bias any new adaptive alleles.

The end result is that once a feral population is established (remember that
honey bees are an invasive species in the U.S.), it will begin immediately
to evolve to be most successful under the local conditions.  The Calif study
found that the genotypes in the various climes in Calif matched the
genotypes of bees found in similar climates in their native Europe--no
surprise!

As Geoff says, drones and swarms are only released by managed colonies at
certain times of the year, so the feral population will not receive genetic
input from managed bees at other times.

But how about the swarms and drones that are released?  You might think of
them as being analogous to hatchery raised trout that are released into
streams in "put and take" stocking programs.  The chances of a released
hatchery trout surviving for more than a few days are minimal.  In a normal
stream, the existing native, or previously introduced feral, trout already
have filled all the best eddies (critical to survival) and already eat most
of the available food.  Plus they are adapted for life in the wild, as
opposed to the regular feedings in hatchery raceways.

The net result is that there is little genetic change in a stream's trout
population due to introduced hatchery fish that were raised in managed
raceways, fed artificial feeds, and enjoyed parasite suppression by
antibiotics.  Sound familiar?  However, the introduced trout are notorious
for introducing new parasites to wild populations.

Bee populations will quickly incorporate any new alleles that add to their
success, and lose any that don't, as evidenced by the gene flow from
European stock to African in the New World, and vice versa.

The swarms that do issue from managed bees have little chance of survival in
any environment that already has a healthy population of feral bees, and
therefore would be unlikely the next season to issue either swarms or
drones.  Any nonadaptive genes that do make it into the feral population
would be bred out in the next several generations.

However, when varroa depopulates a feral population, the influx of swarms
from managed bees can greatly slow the natural selection for a mite
resistant new population.  This is because the escaped managed bees face
little competition in this temporary condition of an environment depauperate
of bees, and can survive until varroa builds up to levels that cause the
colony to collapse (since most managed bees do not exhibit enough resistance
to varroa).  At that point, they than are extremely detrimental to the
recovering feral population, since they contribute a major mite load to the
ferals that rob them.

This situation was avoided in the Gottland Island experiment, since
mite-infested colonies died during freezing weather, which killed the mites
before they could be robbed.

In summary, I find the argument that feral populations are simply managed
bees that have moved to the trees uncompelling, and not supported by the
data that I've seen.

Randy Oliver, apologizing for the long post

*******************************************************
* Search the BEE-L archives at:                       *
* http://listserv.albany.edu:8080/cgi-bin/wa?S1=bee-l *
*******************************************************