An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-sig-ecology/attachments/20120530/67e36f5d/attachment.pl>
species richness
2 messages · Linda Bürgi, Ivailo
On Wed, May 30, 2012 at 10:42 PM, Linda B?rgi <patili_buergi at hotmail.com> wrote:
Hi, I am interested in testing whether increasing the number of plant species sampled increases the number of herbivore species I find, irrespective of the number of herbivores collected. To do that, I was thinking of fixing the number of herbivores collected (e.g. 100) and randomly pulling 99 samples of 100 herbivores from all possible combinations of 2 plant species, 3 plant species, 4 plant species, etc. This should then yield a curve with number of plant species on the x axis and average number of herbivore species found on the y axis, always for a sample of 100 herbivores. Does such a function already exist? My data (see below) is in matrix form with columns representing herbivore species (14) and rows representing plant species (14), with the numbers in the cells representing number of specimens collected per herbivore and plant species combination. I?m not quite sure how to tackle this?. Thanks! Linda
Hi Linda, your question whether increasing the number of plant species sampled increases the number of herbivore species seems to require a contingency table and a corresponding test. I am not sure, however, how to treat the requirement "irrespective of the number of herbivores collected" as usually the number of species increases with the number of individuals sampled. Perhaps to test additionally if the proportion of herbivores sampled is independent of the proportion of hon-herbivores in the sample(s)? Hope this helps, Ivailo
UBUNTU: a person is a person through other persons.