Advisor

Brown, Richard L.

Date of Degree

1-1-2006

Document Type

Graduate Thesis - Open Access

Abstract

The relationship of ant community composition to various habitat characteristics is compared across four habitat types and 12 environmental variables in Mississippi. The four habitat types include pasture, prairie, and oak-hickory forests in the Black Belt and forests in the Flatwoods physiographic region. Ants were sampled using pitfall traps, litter sampling, baiting and hand collecting. A total of 20,916 ants representing 68 species were collected. NMS and ANCOVA both revealed three distinct ant communities (pasture, prairie, and ?forests?) based on species composition and mean ant abundance per habitat type between the four habitat types. Principal component analysis (PCA) partitioned the 12 environmental variation into four axes with eigenvalues >1. Axis 1 differentiated open grass-dominated habitats from woodlands. In contrast axis two mainly separated pastures from prairie remnants. Multiple regression models using the four significant PCA axes revealed that total species richness was significantly affected by variation in the first two PCA axes. Forested sites supported approximately nine more species of ants than prairies and 21 more than pastures. Comparisons of the abundance of ant functional groups were also made between the four habitat types with multiple regression models to investigate how the environmental variables affected certain groups of ants. Annotated notes are included for each ant species encountered during this study.

URI

https://hdl.handle.net/11668/20634

Share

COinS