Page 5 - FATTORINI2011

Williams et al. (2009), residuals were examined with Shapiro-
Wilk statistics to determine which distributional assumption was
the most appropriate for modelling the dataset. These tests indi-
cated a marginally significant lack-of-fit for the log-transformed dominant gradients in species assemblages with geographical/
model (W = 0.916, P = 0.04), but not for the power function
fitted using non-linear regression (W = 0.970, P = 0.65). Thus,
preference was given to the untransformed model.
The resulting z-value (z | 0.25) matched values typically
found for island systems (Rosenzweig, 1995; Drakare et al.,
2006). The SAR is a consistent phenomenon in insular ecosys-
tems, and the best way to consider other sources of variation in
species numbers is through the analysis of residuals from
species-area regressions (Crowell, 1986; Rosenzweig, 1995;
Price, 2004). Thus, residuals from the SAR were correlated with
other geographical variables and environmental heterogeneity
indices using Spearman correlation tests, which simply assume
monotonic relationships without any reference to particular
functions.
Analysis of variation in species composition
Variation in species composition between the islands was
analysed with CCA using the CANOCO program, version 4.5A
(Ter Braak & Šmilauer, 2002). A Detrended Correspondence
Analysis with the option “detrending-by-segments” (Hill &
Gauch, 1980) produced a first axis of 9.239 SD, which is more
than 2 SD units and hence indicates that CCA is suitable for this
data set (Ter Braak & Prentice, 1988).
Significance of individual environmental parameters (geo-
graphical and environmental variables) was tested using a for-
ward selection with 999 Monte Carlo permutations (see Fatto-
rini, 2011 for details). The influence of geographical and land-
scape variables were tested separately. Both the extent of dif-
ferent landscape categories and their proportions can be impor-
tant characteristics of the landscape of a given island. Thus,
separate CCAs were performed using alternatively the raw and
proportional extent of landscape categories.
Inter-island biogeographical similarity
Canonical Correlation Analysis (CANCOR) was used to ana-
lyse the influence of geographical and landscape variables on
inter-island biogeographical similarity. The original presence/
absence matrix was then subject to a non-metric multidimen-
sional scaling (NMDS) using the Kulczynski 2 coefficient to
construct a dissimilarity matrix (for a discussion of the use of
this coefficient in biogeographical analyses, see Hausdorf &
Hennig, 2005). This technique is designed to construct a “map”
showing the relationships between a number of objects, given
only a table of distances or similarity between them, and is often A 0.851
best at capturing patterns in community data when similarity
coefficients are used (Legendre & Legendre, 1998). The good-
ness of results obtained by NMDS was measured as stress val-
ues. On the basis of the increase in stress values when the
number of dimensions was decreased (Shi, 1993) the retention
of three dimensions was considered to be sufficiently represen-
tative. These three dimensions were used as dependent variables
in CANCORs. Separate CANCORs were performed for geo-
graphical and environmental variables to meet “rule C” of
McGarigal et al. (2000). Inter-island faunal similarity was also
investigated by cluster analysis using the Kulczynski 2 coeffi-
cient as a measure of distance and the UPGMA (Unweighted
pair-group method, arithmetic average) amalgamation rule.
NMDS and CANCORs were performed using Statistica 6.0 soft-
ware.
Both CCA and CANCOR investigate the effects of
geographical/environmental variables on species composition
from complementary points of view and cannot be considered
alternative approaches (McGarigal et al., 2000). CCA identifies
geographical/environmental gradients which mainly influence
species composition of sites, whereas CANCOR correlates
environmental gradients. Thus, CCA is particularly useful for
investigating the influence of geographical/environmental gradi-
ents on beta diversity (second level of abstraction sensu
Tuomisto & Ruokolainen, 2006) whereas CANCOR investi-
gates the importance of geographical/environmental gradients in
regulating variation in beta diversity, i.e. “variation in variation
in community composition data” — third level of abstraction
sensu Tuomisto & Ruokolainen (2006).
RESULTS
Species richness
Area was an important correlate of species richness and
the species-area relationship (SAR) was well modelled by
a power function (S = 11.5 A0.23; R2 = 0.84). When
residuals of the SAR were plotted against other geo-
graphical variables, no relationship was found (Table 3).
Species richness was also tightly correlated with all meas-
ures of landscape heterogeneity (Table 3). When residuals
of the SAR were correlated with landscape heterogeneity
indices, significant correlations were found for Pielou
equitability and, possibly, for Simpson dominance and
Berger-Parker dominance (Table 3). These results suggest
that relationships between species richness and landscape
heterogeneity were mainly through area. When the
stronger effect of area was removed, the influence of
landscape heterogeneity was less evident, although there
is an indication that richness tends to increase with land-
scape diversity and equitability and decrease with land-
scape homogeneity (Table 3).
TABLE 3. Values of Spearman correlation coefficients of spe-
cies richness and residuals from the species-area relationship
(SAR) with geographical and landscape parameters. Abbrevia-
tions are the same as those in Tables 1 and 2. Residuals from
SAR were calculated using the power function. Values in bold
are significant at P < 0.05.
Correlation coefficients Correlation coefficients
between number of spe- between residuals from SAR
cies and environmental and environmental parame-
parameters ters
0.082
Ds –0.040 –0.194
Da –0.250 0.101
Di 0.409 0.188
SDs –0.024 0.281
SDa –0.114 0.133
SDi 0.232 0.081
N 0.873 0.314
C –0.840 –0.403
H 0.831 0.364
eH/N –0.628 –0.090
J 0.737 0.569
DMg 0.828 0.358
d –0.874 –0.504
663