Hubert Herz

Host Plant Selection

Generally, host plant selection of leaf-cutting ants and particularly the mechanisms of selective foraging in Atta ants have been extensively studied with laboratory colonies as well as with natural colonies in the field using pick-up assays as method for both (e.g., Cherrett 1972a,b; Hubbell and Wiemer 1983; Howard 1987, 1988; Vasconcelos and Cherrett 1996; Salatino et al. 1998). However, long-term descriptive field studies are rather scarce (Rockwood 1975, 1976; Shepherd 1985; Wirth et al. 1997). Thus, for a better understanding of the factors influencing food choice under natural conditions throughout the year, it was necessary to investigate host plant selection and to determine the degree of polyphagy of A. colombica throughout an entire year.

The Natural History of Leaf-Cutting Ants

No one can give us an exact number of animal species living on earth today, but all biologists agree that millions more species exist than the approximately 1.5 million that have been described so far. Quantitative faunistic investigations in many habitats suggest about 8 million extant species; other assessments claim 30 million species or even more (Wilson 1992 from Erwin 1982, but see Novotny et al. 2002). Most of these species that share mother earth with us are still unknown to science, and sadly, may never become known, because of ongoing man-made habitat destruction and ensuing species extinction.

The Trail System

Foraging trails make it easier for workers of social insects to find resources after leaving the nest (Holldobler 1977; Shepherd 1982; Fowler and Stiles 1980). The way in which the food sites are used is thus a consequence of the distribution of the food in space (Carroll and Janzen 1973). Apart from this partitioning effect (i.e., the location of the harvesting sites within the colony territory), trails have been attributed to an aggression-reducing function between neighboring colonies whose resource-containing areas overlap (Vilela and Howse 1986; Farji Brener and Sierra 1993). According to Holldobler and Lumsden (1980), the gradually shifting trunk trails and the area immediately around the nest are regarded as a “core area” which is the actual territory to be defended. Thus, the trails would obviously reduce the chance of aggressive confrontation between adjacent colonies. However, this hypothesis has not yet been conclusively proven for leaf-cutting ants.

The Study Area — Barro Colorado Island

The present study was performed on Barro Colorado Island (BCI) in the Republic of Panama (Plate 1; Fig. 7). Located about mid-way (9°09′N, 79°51′W) between the Atlantic and the Pacific Oceans, the island lies within an artificial lake (Lake Gatun) created by impounding the Rio Chagres during construction of the Panama Canal between 1911 and 1914. Like various other islands, BCI is high land which was isolated from the surrounding terrain as the lake filled. The island covers an area of 1564 ha. and, due to the numerous bays, a shoreline of 65 km. Located on a central plateau, the highest point on the island is 171 m above mean sea level and 140 m above the shoreline. The relief of the island appears very irregular, consisting of gentle hills traversed by gorges and steep slopes. The island is well drained by small watercourses and creeks. BCI was declared a nature reserve in 1923 shortly after its isolation and since 1946 it has been under the administration of the Smithsonian Tropical Research Institute (STRI) which operates a modern field station on the northeastern shore. Due to decades of research activities, BCI is probably the most thoroughly studied piece of tropical rainforest in the world (Leigh and Wright 1990).

Species Composition of the Forest

One of the most striking features of tropical forests is their extraordinary species diversity (MacArthur 1969; Gentry 1982). The diversity of tropical trees supports a remarkable diversity of vertebrate and insect herbivores that have been shown to inflict a more intense pest pressure than in temperate settings (Landsberg and Ohmart 1989; Coley and Barone 1996). As a consequence, both discussions on plant diversity influencing (e.g., Erwin and Scott 1980) and being influenced by herbivore diversity (e.g., Janzen 1970; Connell 1971) have been of far-reaching importance in tropical ecology (for reviews see Crawley 1997; Leigh 1999).

Nutrient Cycling and Water Relations

Nutrient content of the vegetation in the humid tropics does not necessarily reflect the nutrient paucity of tropical soils. Acquired nutrients are retained very tightly within the vegetation and reacquisition from decomposition of dead biomass is very effective (for reviews see Vitousek and Sanford 1986; Bruijnzeel 1990; Grubb 1995). In undisturbed primary forests, the amount of nutrients lost through leaching from soil is typically less or equal to amounts added by precipitation. By far the greatest proportion of the annual nutrient requirement of the forest is satisfied by nutrients that are cycled within the forest. Because of the limited availability of free nutrients, nutrient cycling is a key process in tropical rainforests.

Forest Light Regimes

Light availability is both a cause and an effect of forest dynamics. It is a major environmental factor limiting growth and survival of many forest species (Chazdon et al. 1996; Montgomery and Chazdon 2002) and its heterogeneity is thought to be crucial for the maintenance of species diversity in old growth forests (e.g., Denslow 1987; Canham et al. 1994). As a potential consequence of foliage loss caused by leaf-cutting ants, changes in the penetrability of the canopy for light may be expected. In order to demonstrate such effects, it is necessary to analyze the spatial and temporal light variability patterns in the forest understory as well as within the canopy. In this study, relative irradiance (RI; i.e., the fraction of photosynthetic photon flux density, PPFD, reaching a sensor somewhere in or below the canopy relative to the PPFD flux above the canopy) was chosen as a measure for light penetrability (Wirth et al. 200 lb).

Canopy Structure of the Forest

In order to assess herbivory rates (see Chap. 11) and also to calculate the effect of foliage removal on canopy carbon and water relations (see Chaps. 12 and 14), it is necessary to gain sufficient information on the amount of foliage present and its arrangement in space. This chapter contains such an analysis in which the vertical and horizontal distribution of foliage density (leaf area index, LAI) and foliage angles were investigated. Additionally, due to the semideciduous nature of the forest, the seasonal variations of LAI were also analyzed.

Herbivory and Light

In Chapter 5, the light climate within the rainforest was assessed within the vicinity of a leaf-cutting ant colony. These measurements highlighted the enormously variable light conditions both vertically through the canopy and near the forest floor. Since the cutting of leaves by ants affects canopy structure and subsequent light penetration (Plates 12, 13), the significance of this foliage removal to the light climate and primary production of the canopy becomes an interesting issue. As mentioned earlier, light availability may limit growth and survival of many plant species by limiting photosynthesis, and heterogeneity in the light environment may be an important functional attribute of the rainforest ecosystem. This chapter focuses on the measurable consequences of leaf harvesting by ants on the canopy light climate and potential primary production.

Conclusions: Ecosystem Perspectives

Because of their multifarious effects on the vegetation, leaf-cutting ants have been denoted as key species of Central American rainforest ecosystems (Fowler et al. 1989; Perfecto and Vandermeer 1993). First, they have long been identified as important herbivores in tropical rainforests. As discussed previously (Chap. 12), foliage removal by ants undoubtedly has direct effects on individual plants with some losing up to 40% of their leaves. However, when scaling this patchily distributed herbivory up to the whole ecosystem (Fig. 52), and comparing it with the estimated 15% consumption of annual leaf production by all herbivores on BCI (Leigh and Windsor 1982), one is tempted to conclude that the effect of leaf-cutting ants at the ecosystem level is practically negligible. On the other hand, much of the information on the activities of leaf-cutting ants in tropical rainforests, including the material in this book, points in the direction that the overall effects of leaf-cutting ant activity in tropical rainforests may go well beyond the simple removal of foliage. These effects include enhancement of nutrient availability through the enrichment of soil from nest refuse dumps (Haines 1975; Farji Brener and Silva 1995) and the transfer of nutrients to upper soil layers during nest construction (Weber 1972a,b; Perfecto and Vandermeer 1993). Their nests can contain several thousand chambers comprising a total volume of up to 20 m3 (Weber 1966). However, there are also more direct effects of these ants on the vegetation: understory vegetation growing on or overhanging the immediate nest surface is constantly cleared, frequently resulting in understory gaps near nest sites. Moreover, the ants can directly affect vegetation succession through the destruction of numerous flowers of forest tree species (Haines 1975), and can also significantly contribute to seed dispersal of certain forest plants (Roberts and Heithaus 1986; Kaspari 1993, 1996; Dalling and Wirth 1997).