The majority of the explained variation in our finescale
survey is attributed to the covariance between
distance and environment—in other words, to spatial
structure in the environment. Taken alone, this finescale
survey could be interpreted as providing support
for models positing either dispersal or environmental
determinism as controls on community assembly. Because
the covariance between distance and environment
cannot be broken at the fine scale, it is impossible
to know the independent effects of either factor. Specifically,
in the range of 4–140 m at our site, it is
impossible to separate the effects of dispersal from
those of soil moisture, thus undercutting any ‘‘pure
environment’’ signal of this important environmental
variable. Nonetheless, the combined results from surveys
at both spatial scales clearly indicate that these
ferns do consistently segregate along soil moisture gradients.
If stochastic dispersal processes alone were determining
fern distribution, we would not expect any
particular environmental gradient to structure fern
communities at multiple scales, least of all a single
environmental gradient.