Communities of Harvestmen 2: A Range of Communities A consideration of communities from diverse locations worldwide.

So how do these diversity measures, considered in §1, compare with each other in describing opilionid communities? 

To explore this, data have been taken from 48 sources in the literature for 261 sampled locations (Table 1) and used to describe those communities in terms of species diversity; 122 species were involved. The samples have been labelled in terms of region (continental Europe, Great Britain, Japan, North America, and South America) and ten broad environmental categories: built, cave, cultivated, grassland, montane, open (mostly moor/heath), other (three samples of unspecified substratum), rocky, wetland (mostly peat-bogs, but also marshes, fens), and woodland (deciduous, coniferous and mixed, including scrub). The main trapping method was by means of pitfall traps (43 data sets) often supplemented by other techniques. Eighteen samples are only presence/absence data and it also has to be recognised that the data sources vary in terms of intensity and duration of sampling. Two of them have species richness probably elevated because the data comprise records over a relatively wide area (Eastern Alps and Brazilian valley woodlands), while at the other extreme some perhaps have only single species recorded in casual observations, as some of the Québec data points from Koponen (1995). These defects are largely overcome by the use of proportional values in calculating the diversity indices, and in the multivariate analysis below. Most samples have quantitative values, but thirteen are only presence/absence data, for which S is valid, but not the other diversity measures. In spite of this heterogeneity in sample quality, the data set provides us with an opportunity to see how these indices behave, and also to make some tentative comparisons between the harvestman fauna in different kinds of environment across geographical regions. Table 2 lists the species included and their frequency of occurrence over the 261 samples. Note the relatively low percentage occurrence values (maximum of 44% for Mitopus morio).

Table 3 lists the values for the diversity indicators: species richness (S), species diversity (B and H) and evenness (E(B) and E(H) ), summarised as average values for each of the ten environmental categories in each of the regions. Over the quantitative samples, the parameters S, B and H range from minima of 1, 1, and 0 respectively to maxima of 19, 8.92 and 3.54 (not all in the same sample). The minima are single-species samples in a variety of non-wooded environments in several regions, the maxima in Austrian woodland. Other samples (in Austrian and Belgian woodlands) have higher species richness (up to 19), but lower values of B and H because of greater dominance by one or two species. The higher values of diversity show in samples from European mountains and woodlands, including British. A Brazilian valley community (probably amalgamated survey samples) shows the highest species richness of 27, but as the data here are only presence/absence B and H are not valid. These actually compute to B = 27 and H = log2(27) = 4.75, with evenness of 1 for both indices as each species is given a nominal value of 1. The overall average values for these three parameters are S = 5, B = 2.55 and H = 1.34. The values for B and H are plotted against each other in Figure 1, which shows a curvature resulting from the logarithmic operation used in calculating H. The greater spread of values for B is apparent, suggesting better discrimination between samples in terms of their species diversity.

The data used for Figure 1 show some variation in diversity between different environmental categories and this is presented more clearly in Figure 2, for S and my preferred diversity index, B, as well as the probability of interspecific encounter (PIE; Table 4). The average values (medians) varied significantly according to Kruskal-Wallis test between the geographical regions (p < 0.001 for all three indices) as well as between habitats (p = 0.001 for S, 0.027 for B, and 0.032 for PIE); tests on B and PIE excluded samples with presence/absence only data. A more precise test of inter-habitat differences was restricted to the 211 European (including British; EUR + GBR) samples, yielding p<0.001 for S, p=0.035 for B, and p=0.020 for PIE, confirming differences in community complexity between the environmental categories. The median values are indicated in Figure 2, and means in Table 3 and Table 4.

But what is it like inside these communities? What do these data mean for any random individual animal living in these environments? The probability of interspecific encounter (PIE) gives us an intuitive glimpse - as an individual harvestman, how likely is the next harvestman we meet to belong to a different species? Values of PIE, calculated for each of the 261 locally sampled communities are presented in  Table 4. It is interesting to note that the overall average value is 0.50; in other words, over all of these communities there is a 50:50 chance that two random individuals (drawn from any single collection) will belong to different species. It needs to be recognised that the quality of the data varies over these samples and it is, therefore, not straightforward to interpret these PIE values. However, across a set of coherent samples (in terms of sampling effort and data quality) PIE can help us develop some understanding of the structure of communities. The most consistent samples in this data set (quantitative data from thorough trapping at single sites) are mainly in the European and British natural environments. Among these PIE average values do not differ greatly from 0.5 although as you can see from the median values there are wide variations between communities as described above. Notice that (1 - PIE) gives you a value for the probability of intraspecific encounter, i.e. likelihood of meeting another member of your own species, which might be worth knowing from a reproductive point of view!

Back to Opilionid Communities

or

Back to Arachnologia

Back to Home page