Gregg Henderson, Russel O. Wagner, and Robert L. Jeanne.
Prairie ant colony longevity and mound growth.
Psyche 96:257-268, 1989.

Full text (searchable PDF, 1228K)
Durable link: http://psyche.entclub.org/96/96-257.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

PRAIRIE ANT COLONY LONGEVITY
AND MOUND GROWTH
Of the social Hymenoptera (ants, some bees, and some wasps), colonies of ants appear to live longest. However, little information bearing on ant colony longevity exists (Wilson, 1971; Tschinkel, 1988), and much of it is anecdotal. Darwin (1859), for example, commented that some ant colonies must live as long as a human generation, and Fore1 (1928), fascinated by a colony of Formica pratensis in his childhood, found the nest still active 56 years later. Donisthorpe (191 5) knew of a rufa nest in Weybridge, England, that was active for at least 20 years.
More recently, long-term monitoring of mound growth has been conducted in an attempt to understand colony population dynamics (Andrews, 1925; Dreyer and Park, 1932; Dreyer, 1942; Scherba, 1958; Talbot, 1961). Dreyer (1942) projected a 25- to 30-year maxi- mum colony lifetime for Formica ulkei Emery, based on average changes in mound size for a population of colonies. Andrews (1925) took mound measurements for 19 years on a single nest of Formica exsectoides Forel, showing a clear exponential growth in mound size in the first half of the colony's life. Recently Tschinkel (1988) found that colony growth in Solenopsis invicta Buren fits a logistic curve, supporting the hypothesis that the growth of social insect colonies is regulated by density-dependent negative feedback mech- anisms (Wilson, 1971; Oster and Wilson, 1978; Levings and Tra- niello, 1981; Holldobler, 1986; Rissing, 1987; Tschinkel, 1988). Individual ant-mounds of the prairie ant, Formica montana Emery, discovered in a prairie remnant in southwestern Wisconsin, were mapped in 1956. Here we report on the survivorship of 69 nests during the next 33 years. We also present 31 years (1957-1987) of data on growth patterns of eight of the original mounds that were 1) Department of Entomology, University of Wisconsin-Madison, Madison, Wis- consin 53706.
2) N6706 Shorewood Hills Road, Lake Mills, Wisconsin 53551 Manuscript received by the editor August 6, 1989.



================================================================================

258 Psyche [vo~. 96
still active in 1988. This represents the most extensive longitudinal study on a population of social insect colonies to date. F. montana is a mound-building member of the fusca group that prefers sunny places and damp soil (Francoeur, 1973). Insolation of the relatively bare mound helps to regulate temperatures inside to levels that are optimal for brood development (Andrews, 1927; Scherba, 1958, 1961; Wilson, 1971; Bruskewitz, 1981). Gregg (1948) regarded the prairie ant, F. montana, a subdominant species, if not a true ecological dominant. F. montana builds mounds of excavated soil that may occupy up to 1.7% of the surface area of a prairie, and are considered to be one of the most important biotic forces in the prairie environment (Baxter and Hole, 1967; Fig. 1). Homopteran honeydew, a major dietary constituent for F. rnontana, is supplied by a complex of prairie plants on which the homopterans reside throughout the foraging season (Henderson, in preparation). Ipswich Prairie is a 1500 m by 10 m prairie remnant located along a railroad right-of-way on the Grant-LaFayette County line in southwestern Wisconsin. Over 120 native prairie plant species are located at Ipswich Prairie, and of Curtis' (1955) 50 prairie plant indicator species, 3 1 are present (Wagner, 1960). The Wisconsin Scientific Areas Preservation Council (Natural Areas now) initiated acquisition of Ipswich Prairie about 1979 and now has it listed as one of over 200 such areas.
In 1956, four 150 m X 1 m transects along the 1500 m length of Ipswich Prairie were established by one of us (ROW) and all nests of F. montana located within their boundaries were mapped. The pres- ence of ants entering or exiting nest holes and/or the condition of the nest (e.g., whether the nest showed signs of recent excavation) were used as indicators of nest activity. Abandoned nests lacked recently excavated entrances and other signs of ant activity.



================================================================================

19891 Henderson el al. - Prairie ant colonies 259 Fig. I. Formica montana mounds in Green Prairie (University of Wisconsin Arboretum, Madison, WI), after a prairie fire. Photograph taken in 1940s. As a non-destructive estimator of colony size, mound measure- ments were taken in mid-summer or early fall every year between 1957 and 1962, and intermittently thereafter through 1987. The volume of each mound was estimated by using its length, width and height to compute the volume of a cylinder. To establish the correlation between mound volume and colony size in F. mantma, forty ant colonies were dug in 1957 and 1958 (outside of the four transects) and the ants counted. Colonies having mounds of rather uniform size (15 cm X 15 crn X 13 cm, length, width and height) were selected, and the dry weight and volume of soil in each mound was determined. The number of workers inhabit- ing a colony positively correlated with mound volume at the 1% level (J. T. Medler, unpublished data). Mound size for other ant



================================================================================

260 Psyche [vo~. 96
species is also considered a good approximation of colony popula- tion size (Andrews, 1925; Weber, 1935; Scherba, 1958; Talbot, 1961; Wilson, 197 1 ; Tschinkel, 1988; but see Porter and Jorgensen, 1988). Multiple regression analyses (via Minitab) were used to evaluate the pattern of mound volume changes for the eight longest-lived nests. We used indicator variables (X2 = 1 if a mound had not yet reached maximum volume, X2 = 0 for each subsequent year after reaching maximum volume) to determine if the data points for each nest best fit a line that showed 1) a significant growth phase and decline phase, 2) a growth phase with no subsequent decline, or 3) no significant change in volume throughout the 31-year period. Regression equation: y = BO + BIXI + B2X2 + Bs(XlXl), where Y = mound volume, Bo, Bl, B2, B3, = constants, XI = year, X2 = indicator variable; Minitab automatically eliminated the (XIX~) combination if it was highly correlated with the other X variables. From observations of nests outside of the transects it was appar- ent that prairie ants initiate two kinds of new colonies, permanent and temporary, or food-source mounds. Food-source mounds were often located about the stems of prairie plants, found most com- monly associated with the prairie rose, Rosa arkansana and the cone flower, Ratibida pinnata. Excavations of these nests showed that about two dozen workers and a species of aphid they were tending were in residence. Food-source mounds were most often initiated in June, but did not grow in size and often disappeared by late season.
Of 58 F. montana colonies in the four 150 m X 1 m transects in 1956, survivorship was 40 (69%) after 10 years, 28 (48%) after 23 years, and 8 (14%) after 33 years (Fig. 2). Three of the 18 nests that disappeared between 1956 and 1966 lost their identities through merging with closely adjacent colonies, indicating that polydomy occurs in this species. Most of the 58 mounds were already of mod- erate size (mean = 38 cm X 39 cm X 25 cm) in 1957 and had probably been founded several years before 1956. Between 1956 and 1959 mounds ranged in size from the newly initiated (measuring less than 5 cm X 5 cm X 5 cm) to some of the largest (1 m X 1 m X 23 cm) .




================================================================================

Henderson et al. - Prairie ant colonies
Fig. 2.
Comparison of survivorship curves of established F. montana mounds with incipient mounds. Lines are interpolated from data points. See text for details. In 1957,21 new mounds were recorded. Of these, 12 (57%) were active in 1958, 10 (47%) in 1959, three (14%) in 1979, and none were active in 1988 (Fig. 2). In contrast, of 40 food-source mounds recorded in June 1959,38 (95%) disappeared by the end of August 1959.
The relatively large volumes of six of the eight mounds in 1957 (all except nests c and g) indicated that they were initiated some years before recordings started (Fig. 3; Appendix I). Nest mound h grew very little over the 31-year census (Fig. 3h; Table I). Volumes of nest mounds a, b, c, and f increased rapidly and then declined, the decline starting between 1973 and 1979. Part of colony decline was coincident with the beginning of a reduced burning schedule for the prairie. Following the last regular burning of Ipswich Prairie in 1964 woody species (especially sumac and black cherry) began to invade



================================================================================

262 Psyche [vo~. 96
the prairie, and by 1979 these had completely displaced native prairie plants in some locations. In 1979, ROW noted that in one invaded area 76% of the colonies died out, whereas 39% die-out was recorded in a relatively little-affected portion. The dramatic mound volume increase by nest a in 1973 was due to its merger with a nearby mound. The next measurement year how- ever, mound volume dropped just as dramatically. Nest mounds d and e showed a positive growth at first, then leveled off. Nest g grew in mound volume throughout the observation period. 92% of the variance in nest g's volume can be explained by the year. Nest mounds showing a significant change in volume were collec- tively analyzed in two groups according to year of colony founding. Since nest mounds a and h had essentially no sustained growth phase (Table I; Fig. 3) they were omitted from the analysis. Nest mounds c and g, whose small initial mound volumes in 1957 indi- cated they were probably incipient colonies in 1956, were analyzed together (group I). Nest mounds b, d, e and f had larger mound volumes in 1957, suggesting they were all initiated at some earlier date than nests c and g. We analyzed these nest mounds as group 11. For each group, mound volume was regressed on year to deter- mine if the pattern growth was logistic (an S-shaped curve), or more like a truncated S curve (such that the bottom of the S was cut off). Data were transformed using the guidelines given by Chatterjee and Price (1977: 29):
Logistic Equation Transformation: Y = ln(Y / 1-Y); Truncated S Transformation: X = ln(X);
where In = natural logs, Y = volume, X = year. The truncated S transformation gave the better fit for both groups (Table I). However, only data for group I are valid for evaluating mound growth for newly initiated colonies, since start-up year for colonies of group I1 was not known. Nests of group I1 may have had a phase of exponential growth prior to 1956. Interpreting the growth pattern of F. montana colonies requires that we consider the constraints that help shape it. To do this we must know some of the biology of F. montana. F. montana is multicolonial: workers distinguish nestmates from alien conspecifics and they aggressively defend food-territories from nearby colonies



================================================================================

19891 Henderson et al. - Prairie ant colonies 263 Table 1. Best fit regression line for mound a-h growth patterns. Nest groups I and I1 show fit of data to a logistic growth model and a truncated S growth model. Y = mound volume, X = year, I = indicator variable, In = natural logs, P = significance level, R2(%) = Percent of variance in mound volume explained by the X- azis variables.
Nest Regression equation R2(%) P
a y = 0.103 - 0.0001 X - 0.061 + O.O05(XI) 45 .10 b y = 0.27 - 0.005X - 0.241 + O.Ol(X1) 86 <.OO 1 c y = 0.09 - 0.000IX - 0.1051 + O.O06(XI) 96 <.OO 1 d y = 0.02 + 0.003X + 0.0131 80 <.001
e y = -0.007 + 0.003X + 0.0381 84 >.05<.10 f y = 0.08 + 0.00002X - 0.0461 + O.O05(XI) 68 <.05 g y = 0.00065 + 0.0036X 92 <.OO 1
h y = 0.06 + 0.0003X - 0.01 1 + O.O02(XI) 5 1 >.I0 Group I { WX) 1 y = -0.045 + 0.039X 85 >.05<.10 { ln(y/ l -y) 1 y = -5.88 + 0.15X 52 >. 10 Group 11 { ln(X) } y = 0.012 + 0.029X 64 >.05<.10 { ln(y/ l -y) } y - 3.09 + 0.036X 54 .10 of conspecifics (Henderson, in preparation). Nevertheless, several other attributes of F. montana are normally associated with a unico- lonial species (sensu Wilson, 1971; Holldobler and Wilson, 1977; but see Rissing and Pollock, 1988): they occur in patchily distrib- uted habitats, monopolize the habitat by excluding other ant spe- cies, build large, aggregated populations of colonies through budding, and are polygynous (Henderson, in preparation). Ant colony growth most often appears governed by the food limitations of the habitat, a density-dependent controlling mecha- nism (Brian, 1953, 1957; Wilson, 197 1 ; Kissing, 1978; Tschinkel, 1988). This may be going on in F. montana as well. That is, the behavior of territoriality between F. montana colonies limits a col- ony's access to food and thus governs the density-dependent nega- tive feedback that regulates colony growth. This provides a means by which a single species may populate a habitat and not over- exploit its food supply.
While the carrying capacity of the prairie habitat for F. montana may be set by food resources accessible to the territorial colonies, density-independent factors affecting habitat quality also appear to influence colony size and density in this population. The correlation of the decline of F. montana colonies with the cessation of annual



================================================================================

Psyche [vo~. 96
YEAR YEAR
Fig. 3.
Mound volume (m3) as a function of year for eight Formica montana colonies censused over a 3 1 -year period.



================================================================================

19891 Henderson et al. - Prairie ant colonies 265 burning and the subsequent invasion of woody plant species at Ips- wich Prairie suggests that the one was the effect of the other. The increasing shade imposed by the invading species could well have reduced the quality of the habitat, either directly by its negative effects on nest mound temperature, or indirectly via its influence on the growth of the prairie plant species harboring the homopterans whose honeydew is an important food source for this ant. Nest mound growth of a population of F. montana colonies was systematically measured over a 3 1-year period. Some nests were still active 33 years later and one nest was still in the growth phase in 1987. Regression analysis revealed that mounds grew quickly in the first years of life, suggesting that colony initiation through budding provides an efficient means of rapid growth at the point when col- ony survival is most in jeopardy. It is hypothesized that colony growth is mediated by intraspecific territoriality which limits a col- ony's access to food. Decline of the colony is correlated with density-independent factors.
We are especially grateful to J. T. Medler for his expert contribu- tion and interest in the earlier phases of this study. Joan M. Herbers, Sean O'Donnell, Edward 0. Wilson and two anonymous reviewers are thanked for critically reading earlier drafts of this paper. The College of Agriculture and Life Sciences (CALS) Statis- tical Consulting Service provided us with advice on statistical procedures. This work was supported by: the Research Committee of the Graduate School of the University of Wisconsin and funds from the National Science Foundation (G-13320) to ROW and J. T. Medler (principal investigator), Federal Hatch Support Project No. 2588 to RLJ, grants from the Friends of the Arboretum, University of Wisconsin, to GH, and a grant from the Roger Tory Peterson Institute of Natural History to GH. The writers express their appre- ciation to W. L. Brown Jr. and A. Francouer for helpful determina- tion of the ant species.




================================================================================

Appendix I. Volume measurement (m3) of eight Formica montana mounds over a 3 1-year census. Year
nest 1957 1958 1959 1960 1961 1962 1966 1972 1973 1975 1979 1987



================================================================================




================================================================================

268 Psyche [vo~. 96
RISSING, S. W. AND G. B. POLLOCK.
1988. Pleometrosis and polygyny in ants. In: Interindividual behavioral variability in social insects (ed. R. L. Jeanne). West- view Press, Boulder, CO.
SCHERBA, G. 1958. Reproduction, nest orientation and population structure of an aggregation of mound nests of Formica ulkei Emery (Formicidae). Insectes Sociaux 5: 20 1-2 13.
SCHERBA, G.
1961. Nest structure and reproduction in the mound building ant Formica opaciventris Emery in Wyoming. J. New York Entomol. Soc. 69: 71-87.
TALBOT, M. 1961. Mounds of the ant Formica ulkei at the Edwin George Reserve, Livington County, Michigan. Ecology 42: 202-205. TSCHINKEL, W. R.
1988.
Colony growth and the ontogeny of worker polymor- phism in the fire ant, Solenopsis invicta. Behav. Ecol. and Sociobiol. 22: 103- 1 15.
WAGNER, R. 0. 1960.
Distribution patterns of some prairie plants and their rela- tionships to methods of reproduction. Ph.D. Thesis, Dept. Botany, University of Wisconsin, Madison, WI.
WEBER, N. A. 1935. The biology of the thatching ant, Formica rufa obscruipes Forel, in North Dakota. Ecol. Monogr. 5: 166-206. WILSON, E. 0. 1971. The insect societies. BelknapIHarvard Press. Cambridge, MA, USA.




================================================================================