Below is a short summary of the results we've found so far. Statistics were done in R, and all results presented occured with p-values of .05 or smaller (due to risk of death-by-boredom, model details are omitted)
We found 15 species of ants on campus! Specific 5x5m plots had anywhere from 0 to 5 species of ants. The top 3 species in terms of abundance (percent of specimens collected) were T. caespitum (the pavement ant), P. pennsylvanica, and S. molesta (making up fractions of .67, .13, and .035 of specimens collected respectively). These were also the top 3 in terms of prevalence (proportion of sites that a species was found at) with values of .91, .53, .10 for T. caespitum, P. pennsylvanica, and S. molesta. This means that T. caespitum was found at 91% of sites surveyed - it occurs practically everywhere! Below are graphs of this data for comparison:
Performing multiple surveys of a subset of sites shows that species richness is a relatively replicable measure. This means that surveying a specific site on different days and at different times during midsummer will yield a similar species richness finding for that site. Besides validating our surveying technique, this result shows that the richness we found at a site is primarily a result of site attributes, not an attribute of temporally variable things like temperature, weather, ground moisture, etc.
Mulched garden sites and grassy sites (we call these good sites) had significantly more species in them than bare soil sites and evergreen sites covered with needles. This may be a result of gardening practices (many bare soil garden sites are regularly disturbed with hand trowels) or resource availability at these sites. Species richness was positively influenced by the number of trees in a 20x20m plot around the site. More trees = more species of ants. The presence of trees offers habitat for arboreal ants (such as Camponotus species) and ants that tend aphids or use nectar resources (Formica species).
Our results show no significant effect of our measured variables and T. caespitum's abundance. This could mean that we didn't measure the right stuff, or that T. caespitum has the ability to numerically dominate a site regardless of specific site attributes.
We were surprised to find P. pennsylvanica as the second most abundant and prevalent ant on campus. As shown in the prevalence plot above, P. pennsylvanica occured at about half of all places surveyed - though the abundance plot indicates that it was usually found in much smaller numbers. The native range of this ant is supposed to be "west to the limits of the eastern deciduous forest (Fisher and Cover, 2007)", however we obviously found plenty of them in an urban environment in Utah.