Variation in Communities of Seed-Dispersing Ants in Habitats with Different Disturbance in Knox County, Ohio

Disturbance may disrupt mutualisms between plants and seed dispersers in a variety of ways including changes in the community of dispersers present. We tested the hypothesis that habitat alteration disrupts the seed-dispersal mutualism between ants and spring-flowering, perennial herbs. Fourteen study sites in Knox County, OH, represented a range of disturbance including mature woodland, woods recovering from partial logging up to 30 years prior to the study, recently logged woods, early successional field, and corn field. Baits were distributed among 480 positions in 48 randomly located transects. Ground-foraging ants were collected and vegetation characteristics of the transects were recorded. In 315 encounters of ants with baits, locations with tuna and seeds of the myrmecochore, Sanguinaria canadensis, were visited at equal rates. Visitation rates to the baits were higher in the more disturbed sites. There was little evidence for depressed ant diversity in disturbed habitats, though ant community composition varied. Seed dispersal distances were likely to be depressed only in agroecosystems, due to the smaller average size of ants, but even this difference is moderated by the tendency for myrmecochores in disturbed sites to have small seeds relative to plants in forest habitats. Where sites have extensive tree canopies the ant communities and dispersal rates are likely to be very resistant to change. Thus, habitat alteration may only affect the ant-plant seed dispersal mutualism in highly and continually disturbed habitats. OHIO J SCI 103 (4):89-97, 2003 INTRODUCTION Disturbance to ecosystems can change the nature of interactions between species as well as influence the composition of communities. Mutualism where seeds are dispersed by animals is one kind of interaction that might be particularly vulnerable to disturbance, whether that disturbance is caused by invasive species (Bond and Slingsby 1985; Zettler and others 2001; Christian 2001) or by habitat alteration that shifts the composition of species (Andersen and Morrison 1998). Dispersal of seeds by ants (myrmecochory) involves many species of plants and ants in mesic habitats of North America (Beattie 1985; Holldobler and Wilson 1990). Up to 30% of the ramets of spring-flowering, perennial herbs in these habitats are specialized for having their seeds carried by ground-foraging ants. Typical genera of plants with species using this dispersal syndrome include Asarum, Claytonia, Dentaria, Dicentra, Erythronium, Jeffersonia, Sanguinaria, Trillium, and Viola. Specialization for this mutualism involves plants producing small seeds that have a fat-rich aril (elaiosome). Although ants are not attracted from a distance to these seeds (Sheridan and others 1996), upon contact with the elaiosome, ants are stimulated to carry the elaiosome (and secondarily the attached seed) to their nests. There, the elaiosome is fed to larvae. Undamaged seeds can be deposited outside the nest or may become buried in the nest. Seed dispersal for these plants has both demographic (Hanzawa and others 1988; Kjellsson 1991) and genetic consequences (Kalisz and others 1999). Normally, ground-foraging ants collect 'Manuscript received 20 November 2001 and in revised form 1 July 2002 (#01-31). insects from the leaf litter and ants are not obligately dependent on myrmecochory, but adding the food resource of elaiosomes can increase a nest's production of gynes (virgin queens) (Morales and Heithaus 1998). Producing more gynes is required for founding new colonies, so this is an important consequence. Pudlo and others (1980) suggest that disturbance of a forest habitat decreases the diversity of ant fauna, including the numbers of seed-dispersing species, which eventually leads to a reduction in myrmecochorous plant species because of the loss of their seed-dispersing mutualists. In this study, we examined the hypothesis that disturbance disrupts the mutualistic system of myrmecochory. We sampled the ground-foraging ant community and the rate of discovery of baits in a series of habitats that represent a range of disturbance. We asked whether disturbance reduces the probability that antdispersed seeds are located and transported. MATERIALS AND METHODS Site Selection: Field work for this study was carried out from 23 May 1986 through 8 July 1986 in Knox County, OH. All sites were located within 8.0 km of Gambier, OH, USA. Potential sites were first located on recent aerial photographs obtained from the Ohio State Agricultural Extension Office. We used Sigma-ScanTM to digitize and measure areas of each site. Fifty potential sites were then visited, and fourteen sites were selected for study based on accessibility, availability of historical information, and representation across a range of disturbance. Three woodlots were sub-sampled to test for within-site variability, and samples within these different subsites were analyzed separately. All sites were located on ground moraine deposited by the Wisconsin or Illinoisan glaciers, on upland habitats. 90 DISTURBANCE AND SEED-DISPERSING ANTS VOL. 103 Historically, these sites would have been occupied by mixed, deciduous forest (especially Quercus, Acer, Fagus, and Caryci). All sites had been modified by some degree of timbering, but they varied in the time since the last harvest, the amount of canopy removed, and the use of the land after removal of trees. Sites were characterized for "degree of disturbance" using parameters that literature suggests are important for ant communities (canopy closure, litter coverage, and understory coverage). The categories were "Least Disturbed," "Moderately Disturbed," and "Open or Mostly Open Canopy." Site descriptions follow and site characteristics are compared in Table 1. Low Disturbance All sites had closed canopy on the 1940 aerial photographs. The understory was comprised of perennial herbs and scattered saplings, with a well-developed layer of leaf litter. Myrmecochores comprised about one-quarter of the species of understory perennial herbs. "Burns" This woodlot had remained undisturbed for 30 yr (neither grazing nor selective cutting). Prior to that time, the site had been grazed and selectively cut for oak. The canopy over transects was closed and dominated by Acer saccharum and Quercus alba, with large Fagus grandifolia near but not over the transects. "Sapp" The canopy was closed and dominated by Acer saccharum, Carya ovata, and Fraxinus americana; the understory included six species of myrmecochorous plants, and the density of myrmecochores was high. "Crouch" The canopy was closed at the time of sampling. Acer saccharum and Quercus rubra dominated trees. Moderate Disturbance These sites had varying history, but generally there was evidence of more extensive or more recent timber removal than for the Least Disturbed sites. The proportion of myrmecochores in the understory varied, but low frequencies were interpreted as another indication of historical disturbance (see Matlack 1994; Jessen 2000). "Baube" — This site may have been cleared for row crops prior to 1940, but the canopy was continuous at the time of sampling. Trees tended to be even-aged and dominated by Fraxinus americana and Ulmus rubra "Heithaus" This closed-canopy site had been selectively cut at a high intensity prior to 1940, and then selectively cut at a low intensity about fifteen years prior to sampling. Trees were co-dominated by Fraxinus americana, Ulmus rubra, and Acer saccharum. "Backbone" Located on the Brown Family Environmental Center at Kenyon College, this woodlot in 1940 consisted of clumps of trees with many open areas. Many of the largest trees (dominated by Quercus alba and Carya glabrd) were planted in a forest management program in 1912, and the canopy was continuous at the time of sampling. "Biology" A narrow section of closed-canopy woods between the main campus of Kenyon College and State Route 229. The site is more exposed than most other sites to trampling. The trees were dominated by