Linking Specialisation and Stability of Plant ... - OPUS Würzburg

94 the importance of being synchronous
bers compared to the other two, we multiplied all visitor numbers
by a correction factor to account for this inter-observer
difference and repeated the analyses presented in this paper.
Using the correction did not qualitatively change the results of
the analyses.
5.4 results
5.4.1 Overall patterns
In total, we observed 10504 interactions between 166 flowering
plant species and 444 insect species. Flower-visiting insects included
326 species of flies, 45 bees and 73 butterflies and moths.
The data set comprised 67 interaction networks, of which 59
were significantly specialised (mean H2’ = 0.554, SD = 0.149).
5.4.2 Local specialisation
Local specialization (d')
1.0
0.8
0.6
0.4
0.2
0.0
**
n = 166 n = 29 n = 27
**
Flies Bees Butterflies
Insect group
Figure 5.2: Local specialisation of insect taxonomic groups. Sample sizes
given above each box are species numbers. Each data point is
the mean d’ value over all networks in which a species was observed
weighted by number of observations. Single observations
were excluded from the calculation. Overall effect of taxonomic
group: F 2,219 = 4.399, p = 0.013. Asterisks indicate a significant
difference (p < 0.01) determined by Tukey’s post-hoc test
for pairwise comparisons.
At the species level, flower visitors were on average moderately
specialised (overall mean of species’ weighted mean d’
values = 0.332, SD = 0.208). Bees were significantly more spe-