Tokeshi & Schmid-Araya, 2000 , Steadet al., 2005). A commonly observed phe-
nomenon within such distributions is the presence of distinct multiple modes
within the size spectrum of a given food web: these separate peaks usually
describe, respectively, the meiofauna (metazoans passing through a 500mm
sieve), the invertebrate macrofauna (e.g. stoneflies, shrimps), the vertebrate
macrofauna (e.g. fishes) or the megafauna (e.g. cetaceans), if present (Warwick,
this volume). There is often a degree of blurring between these boundaries,
with some of the smaller macroinvertebrates passing temporarily through the
meiofauna in their early life stages, but in general the peaks and their distribu-
tions are usually relatively discrete (Warwick, this volume). Unfortunately, very
few studies of food webs have included all portions of the size spectrum, largely
due to the difficulties in sampling the feeding links of such a vast array of
organisms and in identifying many of the more cryptic, but highly speciose,
groups of smaller taxa. Even in the few studies where the meiofauna has
been included together with the macrofauna, the microbial fauna has been
omitted, so even these highly resolved food webs are still incomplete (Schmid-
Arayaet al., 2002a, b). Nonetheless, these limited data do suggest some form of
dietary size partitioning is operating in response to the distribution of peaks and
troughs in the underlying size spectrum. For example, within the ‘simple’
Broadstone Stream community, the meiofaunal component has been resolved
to yield a summary food web that contains over 130 species and more than 800
links (Schmid-Arayaet al., 2002a; Woodwardet al., 2005c). Some intriguing
results emerge from these data. For instance, the slope of the relationship
between log number of links and log species richness of a set of these food
webs (b¼1.3) did not conform to the predictions of either the constant con-
nectance hypothesis (b¼2) or the link-scaling law (b¼1); this intermediate
slope suggests some form of size partitioning between the macrofaunal and
meiofaunal webs (Schimd-Arayaet al., 2002b; Woodwardet al., 2005b, c). This
(^0) –6 –4 –2 0 2 4
50
100
150
200
250
log 10 body-size ratio
Frequency
Figure 6.7Distribution of feeding links
within the Skipwith Pond food web across a
gradient of potential consumer:resource
body-mass ratios. The clear bars show
potential values, as derived from the full
community matrix; the dark bars show the
frequency of consumer-resource pairs for
which feeding links are realized.
BODY SIZE AND PREDATORY INTERACTIONS IN FRESHWATERS 111