Geoscience Reference
In-Depth Information
et al., 2011
), but in the other four streams, the fish assemblages were not
sampled. The Tadnoll Brook is the only stream food web that includes both
fishes and invertebrates, so in this sense, it is intermediate between the other
streams and the Celtic Sea in terms of the absolute size range of organisms it
contains. Nonetheless, it seems that many of the patterns we have observed
here are consistent despite these differences, even among food webs with no
species in common or with very different biogeographical histories, which
gives us some confidence when generalizing or extrapolating to other sys-
tems. One response variable we should be perhaps be particularly cautious
about relying on is TH, as this was calculated from the partial food webs and
hence included no autotrophic species, which will inevitably lead to some
uncertainty in the estimates.
All of these food webs focus almost solely on animal-animal interactions in
which the predators engulf their prey in its entirety. This is largely due to our
desire to standardize across systems where possible, but care should be taken to
remember that predatory interactions that include truly carnivorous predators
or filter feeders, as well as detritivorous and herbivorous interactions, are not
included here. These interactions can be problematic when assessing size struc-
ture, as the resources are often modular, amorphous, or damaged beyond
recognition, so no clear individual size can be assigned. It is less problematic
to consider herbivory in aquatic systems where single-cell primary producers
are consumed (e.g.
Cohen et al., 2003; Layer et al., 2010a,b
) than is the case for
terrestrial systems where substantial supporting structures are often necessary
and modularity is common. As such, the individual and size-based approaches
based on gut content analysis used here may simply be inappropriate to apply in
all circumstances. That is not to say, however, that trophic size structure will be
absent from these types of interactions (e.g.
Humphries, 2007; McCoy et al.,
2011; Novotny and Wilson, 1997; Rall et al., 2011
).
Another caveat when comparing among our study systems is that the
results where size classes have been used are to an extent dependent on the
number of size classes, which differ among the seven food webs. However,
our principal objective was comparing different approaches within rather
than across systems, which is why the number of size classes has been
standardized to equal the number of species in each food web. Nonetheless,
the potential effect of the number of size classes used should be considered
when interpreting the results.
E. Future Directions
In short, we have shown that many predator-prey body mass relationships in
aquatic food webs (e.g. prey mass, PPMR, prey body mass range, TH and
out-degree as functions of predator or species body mass) scale differently
