Ant Species Diversity in a Costa Rican Rainforest


Lee Ann Acker


Anthony Kirkendall

|Rancho Mastatal|La Cangreja|




One of the most indicative representations of the health of an ecosystem is the activity and prevalence of its ant species. The following report undertakes such a study in a wet-premontane rainforest of Costa Rica known as Rancho Mastatal, an area bordering a nationally protected wildlife preserve. Our findings indicate the highest level of species richness in the areas we defined as old growth and disturbed whereas the lowest level of such figures were found in an environment landscaped by human settlement. This neither supports nor disproves the intermediate disturbance hypothesis, which posits that species diversity will be greatest in areas between the extremes of ‘old growth’ and ‘severely disturbed’ habitats. Although evidence here may indicate the greatest loss to hymenoptera diversity in areas most directly affected by development, a more extensive endeavor into the greater pattern of ant occupation between old growth and recovering tropical rainforests is needed.


Rancho Mastatal in Puriscal County of Costa Rica is an environmental learning and sustainable living center that includes 300+ acres of land. The land transitions between very wet and pre-montane rainforest. The property borders the La Cangreja Reserve, a protected area since 1984. Altitude at Rancho Mastatel ranges from 400-600 meters and rainfall exceeds 4000 mm annually. The area is described as having high humidity and poor soils, which are defining factors of a tropical rain forest.


The purpose of this study was to compare the number of morphospecies in three levels of habitat disturbance in Central American tropical rain forest. We hypothesized that there would be a greater number of morphospecies in the disturbed landscape that was in the process of recovering and had the most vegetation diversity than in the primary rain forest or residential area that was not allowed to recover. If this intermediate disturbance hypothesis were true, then our data would show a strong positive correlation between the numbers of morphospecies of ants in the disturbed habitat.

Disturbed site, Old growth site and Residential site

We collected ants for three days a total of six times using bait that consisted of palm lard and honey. Preliminary tests demonstrated that ants preferred a 50% mixture of palm lard and honey over honey or palm lard alone.


We placed bait on Solo cup lids and set them 6.1 m apart along one transect in each treatment area at compass readings of 266°, 210°, and 70°. All of the lids were kept along the edge of a footpath on bare soil to control for leaf litter. Lids were picked up after 30 minutes and placed in sealed containers (.47 l Solo Cups) containing 70% isopropyl alcohol. We measured soil pH, moisture and took densiometer readings at every site on the first day of data collection. We measured temperature at each treatment every time we sampled.

Ants were separated into morphospecies by visual examination and stored in #2 drams in 70% alcohol for transport to the United States for identification and cardboard point pinning.


We found the highest number of total morphospecies in the old growth forest and the disturbed forest. The residential area had only four morphospecies compared to the six we found at the other two treatments. There was a strong positive logarithmic correlation between ground temperature and number of different species found per sample at each treatment. The highest mean number of species was at the disturbed forest site with the mean number of species caught per sampling event 3.16 compared to old growth forest and residential area that both produced a mean number of two morphospecies per sampling event. The ant species were presented as total number of ant morphospecies per location in all traps (Table 1). Additional data presented are ground temperature per morphospecies number per treatment (Figure 1) and mean number of morphospecies per treatment per sampling event.. Our data collection was limited due to time and resources.

Table 1. Total number of ant morphospecies per disturbance level in Costa Rica


Total Number of Morphospecies

Old Growth








Table 2. Initial readings taken at three treatment types for moisture, densiometer % daylight, site description, compass reading, and soil pH.




Site Description



% Daylight

Compass Reading

Soil pH

Old Growth



266 °





210 °





270 °



Figure 1. Logarithmic relationship between ground temperature in celsius and number of ant species found per sampling event.


ant and temperature graph


The highest number of morphospecies was found in both the old growth forest and the disturbed forest. This does not necessarily support the intermediate disturbance hypothesis where the increase of plant diversity may enhance niche specialists by increasing habitat heterogeneity or microclimate favorability.

Even though there were as many morphospecies in the disturbed area as in the old growth, we found a higher mean number of ant morphospecies per sampling event in the disturbed area. One possible reason for the same number of morphospecies in the disturbed area as in the old growth forest could be interferences with chemical trails left by ants to the bait by the more complex vegetation of the disturbed forest. There was a positive logarithmic correlation between ground temperature and number of species per sample. This result was not found in other studies of temperature and foraging activity in ants




We would like to thank Professor John Banks for all of his guidance and support, Tom McDonald for his advice on pre-tests for experimental design, and Tim O’Hara and Robin Nunes for being our gracious hosts at Rancho Mastatal.