Right on time as always! Over the month of May, I read a digital stack of Oikos papers, but since they are all published already, I get to pick one that I like best here to promote it a bit. My favourite was ‘Dioecy, hermaphrodites and pathogen load in plants’ by Williams, Antonovics and Rolff (doi: 10.1111/j.1600-0706.2011.19287.x). There were many superb papers, but this was the winner for me for several reasons. This paper was very novel (not just in the weird and wonderful way) yet synthetic. It was novel in that predictive set and application of ideas to plants was unique and extremely well executed. It was synthetic in that two large databases were used to test the hypothesis and since sexual dimorphism in plants is such a cool phenomenon – yet relatively untapped as a means to explore evolutionary ecology. A corollary of Bateman’s principle was tested in that males most likely invest more in mating whilst females invest more in offspring. I know we have heard this a million times in ecology and evolution but a nice implication or twist is tested here. If the above is true, then in plants we have a unique opportunity since the reproductive bits can be spread out in various permutations.
Here, dioecious plants (male or female) are compared to hermaphrodites (both) to determine if investment in defence differs by gender. From Bateman, females should invest more in defence and should thus have lower pathogen loads whilst hermaphrodites with both genders should have higher loads. The second prediction tested was that dioecious plant species should have a higher diversity of pathogens since they are more likely to be outcrossed (and since the two genders are separated). As an aside, I think it would also be interesting to test if dioecious plant species are more outcrossed in a similar fashion with large plant databases. However, this does seem like a reasonable assumption for this study. Hence, the predictions are clear, well derived, and adequately test the overarching hypothesis. I really enjoy a paper when it is set up like this.
The statistics and data were also really interesting. The statistics included both GLMs and a neat test I have not seen to explore differences in diversity. Also, a clear test of pathogen load was done using a clumping factor. This is a nice treatment of the data with limited assumptions. The data was also clearly described including the limitations which I always find useful, and to my surprise, citation data was included to control for sampling effort. Wow, neat idea. I still don’t quite get it, but it seems like an interesting control for effort within study when comparing across studies. This should be explored in more detail using other datasets.
The authors found that the first prediction was clearly supported, i.e that the mean pathogen load was higher on hermaphrodites. The second prediction was not however supported in that diversity of pathogens was not higher on average for dioecious plant species. I agree with the authors in that the data were limited to some extent in its capacity to test the latter prediction however the family-level comparisons certainly also reduce power. It would be interesting to see the data re-tested using species (I know a dream) or formal phylogenetic contrasts or application of some of the community-genetic tools now prevalent. Also, I am very curious how other pathogens might be responding to the genders of plants or if above versus belowground patterns are different. Great paper in that the future research on this topic and the method of testing could be expanded. Back to the big picture, I think Bateman’s principle (corollary) is well applied here and very suggestive of similar differences in plants. I wonder how many other crossovers there are between the dominant hypotheses in plants and animals and if they could be tested using large online databases.