In freshwater invertebrates, dispersal is poorly understood and
difficult to study directly. Consequently, indirect estimates of
gene flow are often used as surrogates for dispersal, but the
accuracy of these estimates is generally unknown. In this
dissertation, I investigated dispersal and gene flow in fairy shrimp
(Branchinecta coloradensis) and water mites (Arrenurus spp.)
through a number of approaches. To establish the general
relationship between gene flow estimates and dispersal, I
summarized genetic data for 27 groups of animals in which
dispersal potential can be ranked from high to low based on
criteria such as behavior or morphology. I found that increased
dispersal is almost always associated with less genetic
differentiation between populations. However, allozyme variation
in Arrenurus water mites did not follow this pattern. In 11
species of Arrenurus collected throughout northeastern North
America, patterns of variation were unrelated to dispersal
potential, and many species were essentially panmictic. Although
this may be indicative of pronounced historical effects, contrasts
between closely related sister species also suggested that ongoing
dispersal makes a small but measurable contribution to population
differentiation.

In populations of the fairy shrimp Branchinecta coloradensis from
Colorado, the dominant dispersal vector can be studied directly.
Metamorphic salamanders (Ambystoma tigrinum nebulosum) eat
egg-bearing females, move between ponds with full guts, and
defecate once they reach their destination. By combining
behavioral information for the salamanders with
experimentally-obtained hatching data for B. coloradensis, I found
that salamanders move thousands of fairy shrimp eggs at this site
annually. Gene flow for B. coloradensis was calculated to be 1-3
individuals per pond each generation. This number is very close to
the estimate of salamander-mediated dispersal, and also to
estimates obtained for other fairy shrimp species. Thus,
allozyme-based gene flow estimates appear to be accurate on a
local scale for these organisms.

Finally, I examined the ecological importance of dispersal through
mathematical modeling. By extending metapopulation models to
include refuges free from extinction pressures, I quantified the
relationship between dispersal, extinction and species coexistence.
The model suggests that in many systems, creating even a very
small number of refuges can profoundly impact the ability of
competing species to coexist.