摘要：The assessment of modality or \bumps" in distributions is of interestto scientists in many areas. We compare the performance of fourstatistical methods to test for departures from unimodality in simulations,and further illustrate the four methods using well-known ecological datasetson body mass published by Holling in 1992 to illustrate their advantagesand disadvantages. Silverman's kernel density method was found to be veryconservative. The excess mass test and a Bayesian mixture model approachshowed agreement among the data sets, whereas Hall and York's test providedstrong evidence for the existence of two or more modes in all datasets. The Bayesian mixture model also provided a way to quantify the uncertaintyassociated with the number of modes. This work demonstrates theinherent richness of animal body mass distributions but also the dicultiesfor characterizing it, and ultimately understanding the processes underlyingthem.
关键词：Bayesian; body-size data; excess mass test; kernel density estimate;mixture model.
