Date of Completion

4-22-2016

Embargo Period

10-18-2016

Major Advisor

Eldridge Adams

Associate Advisor

Paul Lewis

Associate Advisor

John Silander

Field of Study

Ecology and Evolutionary Biology

Degree

Doctor of Philosophy

Open Access

Open Access

Abstract

The Eurasian ant Myrmica rubra L. was first discovered in North America in the early 1900s in Boston and Woods Hole, Massachusetts, and is now established across large parts of the northern United States and Canada. Its aggressive behavior, painful sting and high local population densities make it a potentially serious pest to local residents and a threat to biodiversity. I studied the distribution, habitat affinities, and population structure of this species across southern New England.

A systematic search showed that M. rubra is widespread in the Greater Boston area and the Berkshires range, with smaller populations elsewhere in Massachusetts and Rhode Island. The ant was not detected in Connecticut or across large expanses of central Massachusetts and northern Rhode Island, despite the presence of suitable habitat. This distribution suggests a combination of long-distance dispersal mediated by humans coupled with slow local spread. Analysis of mitochondrial DNA revealed multiple haplotypes that are shared with populations in central Europe.

Invasive species distribution modeling is complicated by the fact that many introduced species are not yet in equilibrium with their environment. To address this challenge, a Bayesian hierarchical approach was employed that combines two components: a spatial model, allowing estimation of the geographic extent of the invasion, and a habitat suitability model, quantifying the association between several environmental predictors and the presence of the ant. The spatial component describes uncertainty in where the ant has arrived and can account for spatial dependence in outcomes at nearby locations. This approach discounts the contribution of sites where the species has not arrived to estimation of the habitat suitability model. M. rubra was more likely to be found at sites with high amounts of wetland, water, and developed habitat within 300 m, at sites with wetland plants and stands of the invasive Fallopia japonica Houtt (= Reynoutria japonica Houtt or Polygonum cusidatum; Japanese knotweed), and at sites with low maximum temperatures in the hottest month of the year. The ant was most prevalent in marshes, wet meadows, and wet forests and was rarely found in drier forests or in open areas. These results suggest that, within southern New England, M. rubra is limited to places where the soil does not dry out during the summer. Its restriction to these habitats may have contributed to its slow dispersal during the last 115 years.

Several widespread and damaging species of invasive ants have populations that lack colony boundaries, or that are characterized by large supercolonies with many nests and queens, such as Linepithma humile (Argentine ant), Wasmania auropunctata (little fire ant) and Anoplepis gracilipes (yellow crazy ant; Holway et al. 2002, Lach and Hooper-Bui 2010). Bioassays of aggression were used to determine colony and population structure of M. rubra at eight sites spanning two major invasions in Massachusetts. Populations at all sites consisted of multiple colonies that were mutually aggressive. The largest colonies can be considered supercolonies, expandable networks of nests distributed across more than a hectare. Hierarchical models, accounting for variation in aggressiveness among colonies and variation in outcomes across replicate assays, were used to analyze patterns of aggression. There was no evidence that aggressive responses depended on distance of separation between colonies or whether colonies were neighbors. Supercolonies were constrained by intraspecific competition and habitat variation and did not achieve the vast sizes seen in introduced populations of some other invasive ants.

Download
COinS