Field studies of competition

For most ecologists, the question of how much interspecific competition is experi­enced by species must be settled by controlled field manipulations. The review by Adler et al. (2010), quoted in the first paragraph of this chapter, suggested that these studies had not been conclusive.

7.3.1 A historical review

There is little doubt that the first prominent experimental field study of competition, Connell's (1961) experimental work on two barnacle species, would not have been carried out had there not been highly suggestive evidence of a large effect of at least one of the species on the other. Connell's article was followed by many field studies on competition over the next two decades. Most of these involved addition or removal of a single species from a pair. The two comprehensive reviews of field studies of com­petition, which both appeared 22 years later (Schoener 1983; Connell 1983), found sizable competitive effects in a large majority of the studies. Schoener's review includ­ed 150 separate studies, and he remarked that the number seemed to be increasing faster than exponentially. Most of the studies measured the change in population size in one species following addition or removal of a second species. In most studies, the species were closely related taxonomically. However, the finding of many large effects was likely due to a selection bias for studying systems in which there was already some evidence suggesting large effects. In addition, experimental studies are almost always carried out on a small spatial scale, and these have uncertain applicability to effects on larger spatial scales (where effects are likely to be smaller).

Regardless of the large body of experimental studies published by 1983, many still argued against the importance of interspecific competition (many chapters in Strong et al.

MacArthur's own fieldwork primarily involved competition between bird species for food, but was purely observational, and did not provide estimates of competi­tive effects. Later research by his student, Martin Cody (1973), greatly expanded this observational approach to other bird communities. Unfortunately, long-term manip­ulative experiments in the field were impractical for the species he studied, so the applications largely involved the assumption that there was a single limiting category of resources, and that simple measures of overlap described competition. One of the more rigorous early observational studies involved different sets of species of seed­eating sparrows in different habitats (Pulliam 1975), and was published in Science. It appeared to support MacArthur's theory, but was later shown by the same author not to be consistent with it (Pulliam 1983).

Terrestrial plant communities have been favourite subjects for those studying competition (Keddy 2000; Kraft et al. 2015; and many others). Large interspecific effects have frequently been described in plant communities (Levine et al. 2017).

However, a recent meta-analysis (Adler et al. 2018) found that intraspecific competi­tion exceeded interspecific on average by a factor of four to five. Unfortunately, most field studies have been very limited in their spatial and temporal scales, so have shed little light on the strength of interactions at more biologically relevant scales. Adler et al. (2018) was consistent with this generalization, in that the relative magnitude of intraspecific competition was greater in observational than experimental studies, and in studies that quantified growth over the full life cycle.

The literature on plant competition is too extensive to provide a short summary of it here. The immobility of the adult stage and the observable effects of one individual on other nearby individuals make competition especially evident in terrestrial plants. However, the involvement of spatial resource segregation at moderate to large spatial scales calls into question estimates of population-wide effects based on experiments involving much smaller spatial scales; i.e. the vast majority of studies. The spatial proximity of interacting individuals in terrestrial plant communities also allows a variety of positive effect pathways to co-occur with negative effects via other mecha­nisms (Bertness and Callaway 1992). As a result, estimating interactions at a proper spatial scale and studying all potential pathways of effects both are very difficult and seldom occur. Hart et al. (2017) have recently emphasized the importance of spatial scale in assessing the causes of coexistence, and Clark et al. (2010) describe several neglected aspects of niche partitioning in plant communities.

Space is not the only complicating feature of terrestrial plant competition. Polli­nation, particularly in insect-pollinated plants, provides another coexistence mech­anism, with recent work suggesting that it often provides a rare-species advantage (Wei et al. 2021). Interactions of various sorts with soil-based fungi and microbes are increasingly being recognized as having important and varied roles in determining the outcome(s) of competition (Klironomos 2002; Cardinaux et al. 2018; Ke and Wan 2020). Lekberg et al. (2018) have recently carried out a meta-analysis suggesting that plant-soil feedback could promote coexistence, but its extent is usually not known. Plants are also subject to herbivory by a variety of organisms, which may often have a decisive effect on coexistence (Levi et al. 2019).

In the aquatic realm, phytoplankton species are more easily studied, and have a much shorter generation time than most terrestrial plants.

7.3.2 An illustrative example: competition between hermit crabs

Among animals, most studies of competition have failed to explore resource dynam­ics in spatially extended systems, and the effects that this has on competition. Such an approach is particularly necessary in many marine species, which share with plants the characteristic of dispersal during a life history stage during which competition is unlikely to occur. Hermit crabs present a case where resource use is easy to observe and manipulate, and the resources are not self-reproducing. Early in my graduate studies, I decided that I would study competition for shells between hermit crab species for my Ph.D. Recent work by Richard Vance (1972) had documented that empty shells were a major limiting resource for a set of three species, and because crabs ceased growing when their shell was too small, there seemed to be little pos­sibility of other major limiting resources. The number of species present in hermit crab communities varied on a large spatial scale, but the same communities occurred over large geographic zones, and all had sufficiently few species that community-wide studies seemed possible. The outer coast of the Pacific Northwest had similar species from Central California to Alaska.

The fact that hermit crabs do not kill snails to obtain shells means that the resources (shells of different shapes and sizes) had purely abiotic dynamics. The ability to fol­low marked shells made it possible to experimentally study the process of exploitative competition, and exchanges of shells between species via ‘shell fighting’. The system seemed much more amenable to estimating magnitudes of competitive effects than systems where competition was based on resources whose ‘consumption’ could not be so easily observed or whose dynamics were much more complicated.

While my Ph.D. thesis ended up being purely theoretical, most of my early aca­demic career was largely devoted to working on competition between hermit crab species for empty shells. The majority of the work was done with little grant support, and this contributed to several limitations to the results. The large spatial ranges of all the species made it impossible to have random range-wide sampling, and the extent of shell exchanges (due to ‘shell fighting’) in the field was difficult to determine over broad spatial scales. Nevertheless, it seemed clear that intraspecific competition was by far the predominant form in all sets of species that were studied in detail. Shell use could be followed by introducing marked shells, both in the field and in lab­oratory tanks. Intertidal communities studied ranged in species number from two (the southern end of the Great Barrier Reef (Abrams 1981a) to nine in Micronesia— studied primarily in Guam and Enewetak (Abrams 1981b). It was also possible to compare the subset of three species found in relatively sheltered intertidal zones in the Pacific Northwest to the set of six species found in outer coast locations (Abrams 1987e, h), which included the three protected-water species. Communities having more species tended to have somewhat larger estimates of average pairwise effects, particularly comparing the Southern Great Barrier Reef to Micronesia. In the Pacific Northwest intertidal species assemblages, the two species with relatively high total interspecific effects in the more speciose outer coast community were low intertidal species that also occurred in the subtidal.

Resource use and the strength of interspecific competition • 159 range was not sampled as extensively, which likely increased the estimates of relative size of interspecific to intraspecific effects.

Estimates of relative strengths of inter- and intraspecific competition produced from analysis of the occupancy of released marked shells (performed for some, but not all of the communities studied) closely matched that based on overlap in habi­tat and shell type. The most thoroughly studied system consisted of three intertidal species found in relatively protected shorelines in the Pacific Northwest region of North America (Abrams 1987e). This was one of several communities in which com­petition was studied by marked shell releases as well as size-specific resource overlap measures. The others were the intertidal species on the Pacific coast of Panama (Abrams 1980c, 1981c) and a pair of species found at the southern end of the Great Barrier Reef (Abrams, 1981a). All species experienced close to or more than an order of magnitude more intraspecific than interspecific competition in those size classes constituting the greatest part of the reproductive output of the population. For most species in other communities that I studied the estimated intraspecific fraction of all competitive effects ranged from several-fold to more than an order of magnitude greater than the summed effects of interspecific competition from all other species in the community. The largest estimates of interspecific competition were for a sub- tidal community (Abrams et al. 1986), where most of the sampling was carried out by trawling (Nyblade 1974). This blurs small-scale spatial segregation, and probably led to overestimates of interspecific competition.

There are many unknowns about the population dynamics of every species of her­mit crab involved in the studies discussed above. The long pelagic larval stage of most hermit crab species makes it difficult to determine the movement rates between spatial locations. It is possible that parasites play a significant role in population dynamics for some species of hermit crabs. Shell exchange between crabs (usually called ‘shell fighting') is another complicating factor. Bertness (1981) argued that this had a significant effect on the overall interaction of the Panamanian intertidal species. Shell fighting does have the potential to make interactions asymmetric, but is unlikely to have greatly changed estimates of interaction strength in this community (Abrams 1981c). Hazlett (1978, 2013) has repeatedly found that shell exchanges due to shell fighting between conspecifics are usually mutually beneficial. Even if exchanges were strictly competitive, the mean level of competition would still be limited by the high degree of spatial segregation documented in the studies reviewed above. None of the unknowns in these studies seems likely to significantly change the general conclusion that interspecific competition is much less than intraspecific.

7.3.3 Current status of field studies of competition

Little attention has been paid to quantifying interspecific competition in hermit crabs since the 1980s. It may have been regarded as too small and obscure a group of organ­isms to produce generalities. However, the 1980s was also the time when attempts to measure competition in any manner other than long-term manipulations became

unpopular, as documented in Strong et al. (1984). Interest in hermit crabs may have been a victim of this trend.

Notwithstanding many decades of research on competition, only a small fraction of the extant species in any higher-level taxonomic group have been studied with the goal of quantifying the interspecific competition they experience. While we can hope to have a more unbiased sample of species in future field studies, the ecological research community is too small to provide a comprehensive answer in the next cou­ple of decades. This means that theory is likely to provide our best indication of when and where competition is expected to be strong for some time to come. One aspect of such theory must be a better understanding of how overlap in resource use is related to the population-level effects of competition. It is unlikely that theory focused on the Lotka-Volterra or simplified MacArthur models will be sufficient for these tasks.

7.4