research highlights

Here are some papers illustrating my research direction over the last few years. Papers towards the top of the page highlight our current research, whereas papers towards the bottom of the page show how we came to be where we are.

Acta Oecologica 75: 25-34 (2016)


Although structured by interspecific competition, ecological communities experience predation, mutualism and parasitism, and sometimes chance events. However, finding statistical evidence for specific biological interactions is extremely difficult. We set out to isolate the effects of interspecific competition from all other interactions, using ant communities nesting in epiphytic bird’s nest ferns as a model system. The Venn diagrams above show how the ant species overlapped between the rainforest canopy and the understory, and at different stages of ecological succession. By recording the frequencies with which each and every single ant species occurred together, rather than looking at overall levels of competition, we were able to reveal some of the deep patterns associated with interspecific competition. The results were unexpected. We found evidence for competition, but the co-occurrence patterns between ant species were the opposite of what we anticipated. Rather than detecting species segregation—the classical hallmark of competition—we found species aggregation. We suspect that these unexpected positive associations were linked with the displacement of certain species following asymmetric competition.

Sibbaldia 14: 55-69 (2016)

This paper introduces our latest research direction, as we study the links between invertebrates and microbes in the suspended soils of bird’s nest ferns. Linking the above and below ground communities within these ubiquitous epiphytes will reveal important insights into the ecosystem functioning (e.g. carbon and nitrogen cycles) of rainforest canopies. Our previous work has shown that bird’s nest ferns house a remarkable diversity of insects and other arthropods. However, almost nothing is known about the microbial communities associated with the suspended soils associated with the ferns. Access problems, and difficulties with microbial assays in the field, have meant that little work has been done on the functioning of canopy microbes. The Eden Project in Cornwall, U.K., is the world’s largest indoor tropical rainforest, and represents an ideal ‘half-way house’ between complex field experiments and controlled laboratory conditions. We introduce the fernarium (above), an adjustable canopy research platform for the standardisation, manipulation and detailed study of bird’s nest ferns. Using Phospholipid Fatty Acid Analysis (PLFA), we revealed striking similarities in microbial structure and function between ferns in the Eden Project, and those that we sampled in the rainforest canopy of lowland dipterocarp forest in Borneo (below). 

Ecological Modelling 258: 33-39 (2013)

Communities often contain large numbers of rare species. Why are some species common and others rare? Niche partitioning is the traditional explanation, with some species being adapted to a particular niche, and therefore able to multiply, while less adapted species struggle to increase their numbers. The alternative, neutral explanation is simply that competitively identical species drift from one environment to another. By simulating dispersal between local communities (above) we showed that for communities to develop, migration and niche width must act together. Only when migration increased (left to right) with the niche width of each species (upper versus lower panels) did our simulations contain species from red, green, blue and pink local communities (below).  

Biology Letters 7: 601-604 (2011)

Water behaves differently during evaporation and condensation. As water evaporates the lighter isotope of oxygen (O16) evaporates first, resulting in enrichment of the remaining heavier isotope (O18). Oppositely, the heavier isotope is the first to condense. We showed (above) that this process leaves an atmospheric imprint on the haemolymph of insects—much of which is water. As far as we know, this is the first time that a direct measure of the physical or abiotic niche of insects has been quantified in this way. We then tested our hypothesis that cockroaches from a range of habitats (below) would show isotopic enrichment or depletion, depending on how wet or dry the habitat was. We found that, while the O18 signature of the cockroaches did not correlate with precipitation, it correlated strongly with Relative Humidity, which of course is what affects respiratory water loss.  

Bird’s nest ferns (Asplenium nidus) occur throughout the old world tropics and contain very large numbers of insects and other arthropods, as well as earthworms, reptiles, amphibians and mammals (above).


Ferns grow at all heights, from the forest floor to the emergent tree crowns that make up the forest’s high canopy (above). Accessing ferns at these heights is not only physically challenging, but can be rather daunting; we are all accomplished tree climbers.  

Biotropica 34: 575–583 (2002)

This publication was the first to describe fern population densities throughout the lowland rainforest of Danum Valley, and quantify the numbers of animals within them. We showed that the ferns contained entire communities of insects and other arthropods, from the highly diverse beetles to the ecologically important ants and termites. It appears that—as one might imagine—ants often outcompete the less aggressive termites when engaged in direct competition. But once the ferns reach a certain size (above), there is enough room for ants and termites to co-exist. This observation led me to ponder the importance of space as a limiting resource for the fern’s invertebrate inhabitants; moreover, space is a resource that can be manipulated, quantified and tested. In many ways, this paper paved the way for later papers, including my latest paper on competition for space between ant species of different sizes. This paper has just been accepted and will be added to the site as soon as it is published.

Nature 429: 549-551 (2004)

To develop the bird’s nest fern into a model ecosystem we had to establish whether it is possible to predict how many animals were living in each fern. For example, would larger ferns support more animals than smaller ferns? We found a striking relationship (above) between the size of ferns (as measured by the number of leaves) and the amount of invertebrate biomass. We found that ferns contained around the same amount of invertebrates as the entire crown of their host tree. Adding these animals essentially doubled our estimate of the invertebrate biomass in the entire forest.

Ecology Letters 12: 277-284 (2009)

We had established that ferns are important reservoirs of invertebrate diversity, but how were these communities structured? The insects and other arthropods, as well as earthworms, living in pristine ferns appeared to be assembled by chance, or at least, not significantly different from random (black arrows above). This striking result suggested that an earthworm is essentially the same as a millipede, a cockroach or a woodlouse, even though these animals are from very different phyla. It was only when we combined communities from the high and low canopy that we observed the kind of niche partitioning that one would expect. But then, was this significant result (red arrows) genuine—many researchers sample whole trees, from top to bottom—or simply an artefact of introducing habitat heterogeneity into the analysis? At the level of individual ferns, communities really did appear to be assembled at random. 

 Ellwood Lab

Evolutionary Biology, Ecology, Biogeography, Entomology, Conservation