Do mutualistic networks follow power distributions?

Species interactions are central to the structure and dynamics of biological communities. Recently, of particular interest, is the analysis of ecological networks where the number of species to species interactions vary depending on the degree of specialization. For mutualistic networks, species degree has been suggested to follows a power distribution or a truncated power distribution. In this paper, I discuss previously estimated parameters and associated model selection may be unreliable. I use the likelihood approach and compare the results from previous findings and show that the parameter estimates as well as model selection results are indeed very different. Furthermore, the likelihood approach reveals that many mutualistic networks do not follow either power distribution or truncated power distribution.