Parque Nacional del Manu plants vascular plants Spermatophytes angiosperms dicots animals invertebrates arthropods insects chordates vertebrates ant gardens inventories plant-animal interactions ant-plant interactions Cocha Cashu small spatial scales ecology epiphytes ants experiments Azteca Camponotus femoratus Crematogaster Inga Fabaceae Calyptranthes Myrtaceae Cordia nodosa Boraginaceae Madre de Dios Bibliography
In a census taken in Peru's Manu National Park, ten epiphytic angiosperms from seven plant families established principally on arboreal carton-ant nests. These "ant gardens" (AGs) were most often inhabited by parabiotic ants, Camponotus femoratus and Crematogaster cf. limata parabiotica, whose polygynous and polydomous colonies fissioned to form extensive AG aggregations. AGs tended by polydomous but probably monogynous Azteca cf. traili occurred on average in smaller isolates. All three ant species enriched nest gardens with vertebrate feces, but frequencies of occurrence of most AG epiphytes were lower on the less organic carton of Azteca AGs. Interspecific differences in epiphyte abundance and distribution were related to light requirements of plants and to colonizing abilities, as influenced by differences in allocational preferenda and life history. AG aggregations occupied 16-39% of five forest habitat types present and were especially common in frequently flooded habitats and areas of high light intensity. Patchy distribution was explained partly by overrepresentation on resource trees, such as Inga and Calyptranthes (parabiotic ants) and Cordia nodosa (Azteca). Habitat associations did not result from reduction of the terrestrial ant fauna in flooded forests. Other arboreal ants, but not terrestrial ants, were markedly lower in AG aggregations than in areas that lacked AGs, perhaps due to competition from aggressive and dominant AG ants. AGs formed principally by directed dispersal of epiphyte seeds to ant nests, where larvae fed on seed attachments without damaging seeds. AG ants also recognized and retrieved seeds of at least one AG epiphyte from feces of vertebrate fruit dispersers. The preference ranking of epiphyte seeds by Ca. femoratus was not correlated with either obvious differences in quality of seed appendages or long-term resource potential of plants. Seeds of AG epiphytes were rejected by three ants that do not tend AGs but were collected by a fourth such species. Seed attractiveness may depend in part on nonnutritional cues. Preadaptations of plants and ants appear to have been very important to the origin of AGs. Evidence for evolutionary specialization and coadaptation is circumstantial but suggestive.