Travel Reference
In-Depth Information
collected in the censuses. But I was able to use the locations of many trees
that were not counted in one census but appeared in the next. Such a sud-
den appearance shows that the tree had exceeded the one-centimeter cutof ,
which allowed it to be recruited into the census data. Ecologists refer to such
arriviste
trees as recruits.
I could count the numbers of such recruits in dif erent parts of the forest,
including parts where there are lots of adults of the same species and other
parts where there are few or none. Are there more recruits where there are
few adults? If so, then I would have found the frequency-dependence pre-
dicted by the Janzen-Connell and the N-C models.
The analysis was complicated by the fact that trees tend to shed their
fruit and seeds in a decidedly non-random way. Parental trees tend to shower
most of their seeds around their immediate vicinity. Fewer of their seeds
land further away. Even if the seeds that land near the parental trees have
dii culty surviving, because of all those pathogens and predators waiting to
devour them, their sheer numbers mean that more of them will survive than
in places where both the adult trees and the seeds of that species are less
plentiful. This excess of seedlings may not be enough to maintain the cluster
of trees over time, but it will be enough to confuse the analysis.
To circumvent this bias in the data, I asked what would happen if recruit-
ment were to take place completely at random. I did this by building a series
of “pretend” forests in the computer, and then comparing results from them
with the way recruitment happens in the real forest.
In these pretend forests I put all the same trees into the same positions
as in the real forest, but then I went to each species in turn and shul ed the
recruits around at random. In the pretend forests, unlike the real forest, the
chance that a given tree would be a recruit was always the average chance for
that species, regardless of whether that tree was surrounded by lots of other
trees of the same species or by trees of dif erent species.
When I did this, I found that the pretend forests were clearly dif erent
from the real forest. For almost all of the commonest species, when the
real forest was compared to the pretend forests there were more than the
expected number of recruits in places where parental trees were sparse,
and fewer in places where the parental trees were plentiful. Even some rare
