Correlations among network and community structure metrics

<p>Kerri Crawford, now a postdoc at Washington University St. Louis, and I were writing a paper about correlations among network and community structures - an attempt to think more explicitly about these two fields. </p> <p> </p> <p>We include our paper that we submitted to Oikos, and was rejected. Upon further thinking about this, I don't know if the paper even makes sense! </p> <p> </p> <p>Figure captions:</p> <p>Figure 1. Correlations among network structural properties (ordinate) and species richness (abscissa) for animals and plants. Plant-pollinator (empty circles) and plant-seed disperser (filled circles) mutualistic networks are represented. Network structures: number of links (a,b), links per species (c,d), nestedness (e,f), and connectance (g,h; see Table 1 for definitions). Values in each panel represent Pearson correlation coefficients, and p-values (H_o, <em>r</em> = 0). All variables were log₁₀ transformed prior to analysis. Note that all variables are displayed as log₁₀ transformed. Sample sizes: species richness (58-59 communities); evenness (20 communities). Data from the Interaction Web Database (http://www.nceas.ucsb.edu/interactionweb/), and Rezende <em>et al</em>. (2007). </p> <p> </p> <p>Figure 2. Correlations among network structural properties (ordinate) and evenness (abscissa) for animals and plants. Evenness (Pielou’s J) is constrained between zero and one. Larger values of J represent communities with abundance spread more equally among species, whereas small J values represent the presence of few very abundant species. See Figure 1 caption for details on figures. </p> <p> </p> <p>Figure 3. Correlations among network structural properties (ordinate) and mean and standard deviation of the log normal distribution fit to abundance data for each community (abscissa) for animals. See Figure 1 caption for details on figures, and Appendix 1 for abundance distribution model results. </p> <p> </p> <p>Figure 4. Correlations among network structural properties (ordinate) and mean and standard deviation of the log normal distribution fit to abundance data for each community (abscissa) for plants. See Figure 1 caption for details on figures, and Appendix 1 for abundance distribution model results.