Ecologie et évolution des jardins de fourmis en Guyane Française

Résumé En

Ant gardens were followed intermittently, from 1986 to 1989, in a Citrus grandis plantation and in a lowland rain forest, at the Piste de St Elie study site, French Guiana. This plant-ant asociation most likely originates in a different way in open and closed environments. Whereas the plant partner settles first on a tree in a plantation, the ants secondarily entering the root-ball to build their nest, the ant partner is the first settler on rainf orest trees, epiphytes later establishing themselves on their nests. The number of plant partners in an ant garden ranged from 1 to 5 species in the hundreds of gardens studied. They belonged mostly to a limited number of epiphytic families : Araceae, Bromeliaceae, Gesneriaceae, and Peperomiaceae. The ant partners were less numerous in species (1 to 3), belonging to genera Camponotus and Crematogaster. Ant gardens in tree plantations, forest edges, or river sides were composed of more plant species than those found in rain forests. The number of plant partners also increased with the size and age of the ant garden. No specific plant-ant association was ever found in the ant gardens examined. When Anthurium gracile was the only plant partner, there was also a single ant associate, but it was seldom the same, strongly suggesting that the association is initiated at random. In 50 % of the ant gardens studied, whatever the environment, there were two different ant partners per garden. Seven categories of plant-ant associations were recognized. A single tree could harbour as many as 10 ant gardens of different composition. The growth of a new aggregation of ant gardens was studied from 1986 to 1989 in a Citrus plantation, after destruction of all pre-existing associations. The establishment of new ant-gardens was very rapid, and took place within a single year. The major factors influencing their establishment were a sufficient amount of light, and the structure of the bark of the supporting tree. Too much shade and an abundance of epiphytic moss were obviously unsuitable conditions. In the Citrus plantation, ant gardens increased more rapidly in numbers and size on the trees grown on well drained soils ; conversely, they were less abundant on stunted, moss covered, trees planted in hydromorphic soils. Under favourable conditions, the number of ant gardens in Citrus plantations can become impressive : 1015 were enumerated on 456 trees in 1989, three years after complete clearance of all previous gardens ! Both plants and ants take advantage of their association in an ant garden. The plant partner benefits from the nutrients provided by the humus of the root-ball and the decomposing detritus of the ant nest. They also take advantage of the active protection afforded by their aggressive ant partners. Their seeds may also be dispersed by the ants within and around the garden. The ants also greatly benefit throughout the year from the secretions of the floral and extra-floral nectaries of their plant associates, and from the pulp of their fleshy fruits. The intricate root system of the epiphytes also provides shelter and reinf orcement to the brood chambers. An ant garden largely functions in isolation. Contrary to most true myrme-cophytes still deriving most of their necessary nutrients from the soil, ant associated epiphytes no longer depend on the soil substrate. All the necessary nutrients are provided by the enriched humus surrounding the arboreal ant nest(s).