This report describes the foraging pattern of the Florida harvester ant Pogonomyrmex badius in a high-density population of colonies. The foraging pattern has both promoted and been influenced by the colony distribution. Pogonomyrmex badius forages from short trails which extend into a surrounding foraging range. Direction of foraging trails is influenced by the location of a colony's near neighbors. Seasonal nest relocations always occur along a foraging trail, usually the main trail. Foraging ranges are not actively defended, but are used almost exclusively by foragers from a single colony. Foraging ranges will be extended into an area abandoned by neighboring foragers, indicating that forager presence may define each colony's range. Colony distribution has remained essentially the same for several years, despite seasonal nest relocations and addition of new colonies. Establishment of trails and exclusive foraging ranges by each colony minimizes encounters with neighboring foragers and guarantees access to available resources; this pattern also promotes maintenance of the existing colony distribution and partitioning of resources.

The bioenergetics of the southern harvester ant were studied on the Savannah River Project, Aiken, South Carolina. Excavation of hills revealed that density of ants ranged from 4,000 to 6,000 per hill. Labeling ants with P32 indicated that only 10% of the ants in a colony were active above ground during any 2—week period (the limit of recognition of the label); thus, short—term marking recapture estimates based on above—ground individuals greatly underestimate the size of the colony. There were 27 hills per hectare on the study area. The consumption of oxygen of workers at various temperatures was determined in a simple respirometer, and the temperature at which the ants were living in the field was determined during a year's observation at mounds. The energy expense of heat production was calculated from the oxygen consumption and activity temperature records. Total energy flow (heat production plus tissue growth) was highest in summer and lowest in winter. The annual range was from 14 to 48 kgcal/m2/yr. Tissue growth was estimated to amount to only 0.09 kgcal/m2/yr. Energy flow in this species was higher than that of 2 vertebrate granivores studies in the same field, the old—field mouse and the savannah sparrow.

The effect of predation (by the systematic removal of forager ants) on the behavior and population structure of Pogonomyrmex badius colonies was studied during the summer of 1965. For 23 wk, 50 workers per day were removed from one group of 10 colonies and 10 ants per day from another group of 10 colonies. Another 10—colony group served as a control. No effects of removal were observed in the 10/day group. Within a week after removal began, aboveground activity was reduced drastically in many of the 50/day colonies and the removal of a complete sample each day was impossible. Later, colonies underwent extended periods (3 wk or more) of complete inactivity aboveground. The mean dry weight of workers removed at the rate of 10 per day increased over the removal period. The mean dry weight of workers removed at the rate of 50 per day increased for the first 9 wk and then decreased rapidly over the rest of the removal period. As removal continued, craters became smaller and many disappeared; nest entrances were shifted under the overhang of grass clumps and were covered with dried leaves. Intercology conflicts became common at most 50/day colony sites. After the 23—wk removal period, five colonies from each treatment group were randomly selected for excavation. Significantly fewer ants were excavated from both the 10/day and 50/day colonies compared to the controls. Significantly more ants altogether were produced (removed plus excavated) by the 10/day and 50/day colonies. However, the increase in the number of ants produced could not immediately counteract the loss of the ants to aboveground removal. At excavation, workers in the 50/day colonies weighed less than those in the control and 10/day group. In each treatment group, workers collected from the top 20 cm of the nest system weighed less than those collected from the remainder of the nest system. Also, in the 10/day group, workers collected at the surface during the removal phase weighed less than those present in the top 20 cm of the nest at excavation. No such difference existed in the colonies of the 50/day group. By midsummer of the next year (1966), the remaining experimental colonies appeared to be back to normal strength. Although the forager population was greatly reduced, the colonies were able to survive until more foragers could be produced. Such an adaptive strategy is possible because of a highly developed social behavior which results in a well—defined division of labor within the ant colony.

Polish workers have proposed a method of estimating small rodent populations based on removal trapping. Using a linear regression method, they obtained reliable population estimates with five days of prebaiting followed by five days of removal trapping. The aim of our study was to compare the results obtained, using the same method at the Atomic Energy Commission's Savannah River Plant in the Southeastern United States, with the results of the Polish investigators. Our study area consisted of 5.8 ha over which a grid of 256 trap stations was arranged with two-snap traps at each station. Since by the fifth day of trapping, the number of daily captures had not declined, the trapping period was extended to cover 27 days. An increase in daily captures was observed between the 13th and 15th day in three of the four most common species. These fluctuations in capture rates were believed to be due to behavioral changes in the population. The results from this study indicated that at least nine days of trapping were necessary to obtain reliable estimates for the population as a whole. Even so, reliable estimates could not be calculated for two of the four most common species after 27 days of removal trapping.

We estimated the densities of termites in dead wood using a point-quarter technique in four habitats: pine plantations subject to control burning and unburned, lowland hardwood forest, and turkey oak woodland. The burned pine plantation had the highest estimated termite density, 13×10·ha unburned pine plantation and lowland hardwood had estimated densities of 2.6×10·ha and 2.2×10·ha, respectively, and the turkey oak woodland had an estimated densities of 61.9×10·ha. There were varying percentages of Reticulitermes flavipes and R. virginicus in the various habitats. There was nearly linear increase in percentage of pine blocks attacked by termites in the pine and hardwood forests and by the end of the growing season, nearly all had been channelized by termites. In the turkey oak habitat 70% of the pine blocks were channelized. Termites removed between 3% and 12% of the original mass of over one-fourth of the pine blocks during the growing season.Wood litter in fall was highest in the lowland hardwood forest: 2869 kg·ha·yr. Wood litter input in the long leaf pine plantation, 792 kg·ha·yr was nearly equivalent to the three year average wood litter in fall in a Danish oak forest.