Fróðskaparrit - 31.12.2000, Side 112

116
DOMINANT SPECIES ABUNDANCE RELATED TO ENVIRONMENTAL
FACTORS ON ROCKY SHORES IN THE FAROE ISLANDS
the ordination diagram, thus failing to de-
monstrate any large negative effect of Litto-
rina on any of the included species living at
the same station, but not necessarily at the
same level. The apparent large effect of Lit-
torina was therefore probably not due to a
negative effect of grazing, but rather to it
often occuiTÍng together with certain speci-
es. However, grazing may be a secondary
explanation for this coexistence, as L. obtu-
sata is known to use Ascophyllum nodosum
as the main food source (Watson and Nor-
ton, 1987), probably without having a sig-
nificant negative effect on the population
level of the host. It is also probable that
juvenile L. obtusata use microalgae on the
algal surface as a food source (Williams,
1990), thus reducing the epiphytic growth
on the host plant. Such interactions might
possibly explain some of the variation
extracted by the first axis in the partial CA,
and consequently also along the first DCA
axis.
As seen by the low amount of variance
explained, grazing by Patella was not
shown to have any large effect. Semibal-
anus and Mytilus, which are predated on by
Nucella, had their centroids near the zero
value of the Nucella vector, thus failing to
demonstrate any predation effect.
When interpreting these results, two
points should be kept in rnind. First, the
scaling: Significant effects of grazing and
predation might have been detected more
locally than on 8 m stretches of shoreline.
Second, the sampling method, by which
only the abundance of a species was record-
ed for the horizontal zone where it was
most abundant: Recording of total abun-
dance might have been more sensitive to
grazing and predation effects. The results
suggest, however, that on an intermediate
scale the maximal abundances of the inves-
tigated species were not strongly affected
by grazing by Littorina or Patella or preda-
tion by Nucella, except possibly indirectly
by Littorina grazing.
Species abundance curves
Considering the earlier discussion, the
species abundance curves along the first
DCA axis may be assumed to represent the
species responses to the main environmen-
tal factor(s). All species abundance curves,
except those for Mytilus edulis, Palmaria
palmata and Patella vulgata, were signifi-
cant at a 5% level, but the test is only sug-
gestive, as the ordinal abundance data hard-
ly fit a Gaussian distribution pattern. As
Fig. 4 shows, the species may be divided
into three main groups. The first group in-
creased in abundance with increasing site
scores on the first DCA axis. The second
group, as evidenced by the curves, seemed
to have the highest abundance at intermedi-
ate scores on the first DCA axis. The plots
for the individual species revealed, howev-
er, that several of these species were found
at variable abundance at all, or nearly all,
axis scores. The group, therefore, includes
species with no clear response to the factors
underlying the axis. The shapes of the
curves for these species appear somewhat
arbitrary. The third group decreased in
abundance with increasing site scores on
the first DCA axis. The groups were nearly
the same as those identified by the CCA
plot (see above).