Predicting community structure of ground-foraging ant assemblages with Markov models of behavioral dominance

Abstract

Although interference competition is a conspicuous component of many animal communities, it is still uncertain whether the competitive ability of a species determines its relative abundance and patterns of association with other species. We used replicated arena tests to quantify behavioral dominance of eight common species of co-occurring ground-foraging ants in the Siskiyou Mountains of southern Oregon. We found that behavior recorded in laboratory assays was an accurate representation of a colony's ability to monopolize resources in the field. We used interaction frequencies from the behavioral tests to estimate transition probabilities in a simple Markov chain model to predict patterns of relative abundance in a metacommunity that is dominated by behavioral interactions. We also tested whether behavioral interactions between each pair of species could be used to predict patterns of species co-occurrence. We found that the Markov model did not accurately predict patterns of observed relative abundance on either the local or the regional scale. However, we did detect a significant negative correlation at the local scale in which behaviorally dominant species occupied relatively few baits. Pairwise behavioral data also did not predict species co-occurrence in any site. Although interference competition is a conspicuous process in ant communities, our results suggest that it may not contribute much to patterns of relative abundance and species co-occurrence in the system studied here. However, the negative correlation between behavioral dominance and bait occupancy at the local scale suggests that competition-colonization trade-offs may be important in resource acquisition and persistence of behaviorally subordinate species.

Climatic warming destabilizes forest ant communities

“…To explicitly model the associations among species and link the outcomes of these associations to warming, we constructed a Markov transition matrix model ( 14 , 15 ) for the data from each individual chamber. For a community of n interacting species, we constructed an ( n + 1) × ( n + 1) transition matrix, which includes a state for each species and one state for an empty patch.…”

Ant dominance hierarchy determines the nested pattern in ant-plant networks

“…We chose this method because numerical and behaviour dominance are extremely closely related, and therefore such a method has been widely accepted and used in the ant literature (Dejean & Corbara, ; Santini et al ., ; Parr, ; Parr & Gibb, ). Moreover, we observed few competitive interactions among ants in the field, making it difficult to build a valid model for all ant species; also, due to differences in physiological and ontogenetic conditions at each colony, one‐on‐one interactions may not be the best way to quantify competition in ant communities (Hölldobler & Wilson, ; Wittman & Gotelli, ). Moreover, based on the ‘ghost of competition past’, if competition among ants is really strong in our study area, we would expect that competitive interactions are rarely seen in the field (Connel, ).…”

“…Therefore, we also tested whether there was a relationship between the ant's spatial abundance and its position in the nestedness ranking (Spearman correlation). In order not to overestimate the ant species with more efficient systems for recruiting, we calculated the spatial abundance of ants based on the frequency of species occurrence in the pitfall traps and not based on the number of individuals (Gotelli et al, 2011). We also used t-tests to test whether the numerical dominance index and the number of individuals recruited per bait is greater in ant species of the central core than in peripheral ant species.…”

The discovery–dominance trade‐off is the exception, rather than the rule

“…Although we acknowledge this is a contentious issue, in our opinion, it is not appropriate to look at individual abundance when dealing with a super-organism and indeed, frequently, it is not possible to scale up from individuals to species (e.g. Wittman & Gotelli 2011). Colony size is generally characteristic of a species (e.g.…”