Page 3 - Pardi&al_2006
P. 3

Pardi, Piazzi, Balata, Papi, Cinelli & Benedetti-Cecchi                                                   Posidonia oceanica epiphytes of Sicily

Vanderklift & Lavery 2000; Buia et al. 2003). The first                             considered were: Geographic Area (three levels, fixed),
10 cm of the internal face from the tip of the leaves was                          Habitat (small islands versus mainland coasts, fixed and
sampled under a dissecting microscope supplied with a                              crossed with Geographic Area), Meadow (three levels,
grid of 100 squares. The abundance of each taxon was                               random, nested in the interaction Geographic
obtained by examining five field views of 1 cm2 and                                  Area · Habitat), Site (five levels, random, nested in Mea-
recording the presence or absence of that taxon within                             dow) and Plot (five levels, random, nested in Site). Five
nine squares per field view, for a total of 45 (9 · 5)                              leaves from five different shoots were considered for each
squares examined (Piazzi et al. 2004a). The abundance of                           Plot. Homogeneity of variances was tested by Cochran’s
each taxon was determined as the ratio between the                                 C-test (Underwood 1997) and data were ln(x + 1) trans-
squares in which the taxa were present versus the total of                         formed when necessary. Student–Newman–Keuls (SNK)
the squares (45) and then expressed as percentage. The                             test was performed to check for a posteriori comparisons
macroalgal epiphytes were lumped into morphological                                of mean values after significant effects in ANOVA
groups (filamentous, red crustose, brown crustose, green                            (Underwood 1997). In addition to hypothesis testing,
crustose, red corticated, brown corticated, foliose and arti-                      anova was also used to estimate the variance compo-
culated calcareous algae) according to Steneck & Dethier                           nents associated with the random factors (Meadow, Site,
(1994); animals were analysed at the level of phylum.                              Plot, Leaf) in the model of analysis (Sokal & Rohlf 1981;
Bryozoans were distinguished as erect and crustose forms.                          Underwood 1997). These data were used for qualitative
                                                                                   comparisons of patterns of variability across scales.
   The percent cover of the most abundant groups and
taxa was analysed by five-way ANOVA. The factors                                       Six separated ANOVA for the most abundant groups
                                                                                   were conducted to evaluate variability among meadows in
               Encrusting red algae                                                each habitat and region, in order to determine possible
                                                                                   patterns of variability related to the different distances
Cover (%)100                                         Island                        among meadows in the regions studied.
          80                                         Mainland
          60                                                                       Results
          40           West  South
          20                                                                       In Table 1, the most frequent species and/or genus
                                                                                   included in the morphological groups are reported.
            0                                                                      Seven taxa were sufficiently abundant to be used in un-
                                                                                   ivariate analyses (Fig. 2) (Table 2). Significant differ-
                North                                                              ences were detected among the geographic areas for
                                                                                   encrusting brown algae, filamentous algae, encrusting
               Encrusting brown algae                Erect Bryozoans               bryozoans and Foraminifera. SNK tests indicated that
                                                                                   encrusting brown algae, filamentous algae and Fora-
           40                                     8                                minifera were significantly more abundant on the west-
                                                                                   ern coast of Sicily when compared with the northern
Cover (%)  30                          Cover (%)  6                                and southern coasts that did not differ significantly (SE
                                                                                   for comparisons: encrusting brown algae ¼ 4.027; fila-
           20                                     4                                mentous algae ¼ 0.151; Foraminifera ¼ 0.760). A signi-
                                                                                   ficant geographic Area · Habitat interaction was
           10                                     2                                detected for hydroids (Table 2, SNK test). Multiple
                                                                                   comparison tests within this interaction indicated that
           0           West  South                0            West       South    hydroids were more abundant on islands compared
                                                                                   with the mainland only on the southern coast of Sicily,
               North                                   North                       while the two habitats did not differ in the other geo-
                                                                                   graphic areas (SE ¼ 0.166). In contrast, the SNK test
               Filamentous algae                                Hydroids           was not able to discriminate any alternative to the null
                                                                                   hypothesis for encrusting bryozoans. With the exception
           15                                     8                                of hydroids, there was no significant effect of habitat
                                                                                   for any of the other taxa considered. Spatial variability
Cover (%)  12                          Cover (%)  6                                within habitats was, in contrast, large and significant at
                                                                                   all the scales investigated (meadow, site and plot) and
           9                                      4                                for all the taxa examined (Table 2).

           6

           3                                      2

           0           West  South                0            West         South

                North                                  North

                 Encrusting Bryozoans                         Foraminifera

           20                                     9

Cover (%)  15                          Cover (%)  6

           10

           5                                      3

           0           West  South                0            West         South

               North                                    North

Fig. 2. Percent cover (mean ± SE, n ¼ 3) of the most abundant
groups of epiphytes of Posidonia oceanica leaves.

Marine Ecology 27 (2006) 397–403 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd  399
   1   2   3   4   5   6   7