Ant/orchid Associations in the Barro Colorado National Monument, Panama

Of 23 orchid species studied at Barra Colorado National Monument in central Panama, 16 were found to produce extrafloral nectar. Extrafloral nectaries were found primarily on reproductive structures and developing shoots. Only one species (Caularthron bilamellatum) showed evidence of ant occupation, indicating that most orchids are not likely to derive nutritional benefit from their associations with ants. Observations of ants visiting inflorescences of Aspasia principissa and occupying C. bilamellatum showed that the ant assemblages were species-rich, indicating that these are general rather than species-specific mutual isms. The generalized pattern of ant occupation of C. bilamellatum was not different from that observed for host trees in which orchids did not occur. The results suggest that ant associations are common in lowland Neotropical representatives of the Orchidaceae. ASSOCIATIONS BETWEEN ants and plants have been described for many groups of plants, and it has been shown frequently that plants will benefit from the presence of ants (Bentley, 1977; Beattie, 1985). These associations appear to be abundant in epiphytes (Beattie, 1985; Davidson and Epstein, in press), where plants were often shown to benefit nutritionally from the presence of ant colonies in roots or in specialized structures that house the ants (Huxley, 1980). In the Orchidaceae, there are many reports of ant associations (Jeffrey, Arditti, and Koopowitz, 1970; Bentley, 1977; Davidson and Epstein, in press), often involving the occupation of some part of the plant by ants. In their study, Jeffrey, Arditti, and Koopowitz (1970) emphasized the importance that these ant associations may have for protection of orchids against insect herbivores. Because of the large number (over 19,000) of representatives in the Orchidaceae (Atwood, 1986), it is difficult to determine if reports of anU orchid associations represent a few noteworthy cases or a common theme in the family. Because no 'The authors thank J.D. Ackerman, T. M. Aide, and R. L. Dressler for helpful comments on the manuscript and R. Snelling for ant identifications. They also thank the Smithsonian Tropical Research Institute for the excellent facilities provided for this study. Jess K. Zimmerman would like to thank S. Madriiian for having the wherewithal to climb trees. Brian Lee Fisher is especially grateful to H. F. Howe for encouraging tropical studies. Fisher was supported by NSF grant BSR8604687 to H. F. Howe. Zimmerman was supported by a Predoctoral Fellowship from the Smithsonian Institution and NSF Dissertation Improvement Grant BSR-8501246. 0rder of authorship determined alphabetically. 3 Address for reprint requests. study has considered possible ant associations of a number of orchid species in one locality, it is difficult to determine how general these relationships might be. We have investigated 23 species of orchids occurring in the Barro Colorado National Monument in central Panama to determine how many are involved in ant associations. This was largely done by surveying these species for the production of extrafloral nectar and investigating plants for evidences of ant occupation. 12 Additional information is provided on the nature of anUorchid associations. For two orchid species, we describe the ant species involved in the associations and the frequency with which orchids are attended or occupied by ants. For one of these orchids, we indicate the degree to which the distribution of ant occupation among plants is similar to that in the general environment. METHODS-Barro Colorado National Monument (BCNM) is a protected area of lowland, semideciduous tropical forest, located in the Panama Canal in central Panama, consisting of a central island, Barro Colorado Island (BCI), and three adjacent mainland peninsulas. See Croat (1978) and Leigh, Rand, and Windsor (1982) for detailed descriptions of BCI. Croat ( 1978) lists 90 orchid species that occur on BCI, but well over half of these are listed as rare. Our data for 23 orchid species thus represents a substantial portion of the orchid flora in the area. Orchid nomenclature follows Croat ( 1978), except Brassavola nodosa (L.) Lindley, which occurs on one of the mainland peninsulas of the BCNM but not on BCI. Production of extrafloral nectar by orchids from the BCNM was studied in several ways. Most inFISHER AND ZIMMERMANANT/ORCHID ASSOCIATIONS formation was collected from plants in live collections on BCI, with additional observations of orchids at nearby Summit Gardens and in the field. Usually, extrafloral nectar production was observed directly by excluding ants, but the presence of ants was sometimes used as an indication of extrafloral nectar production when the ants were observed feeding from the plants. Ants attending or occupying two orchid species, Aspasia principissa and Caularthron bilamellatum, were studied in detail. Ants attending extrafloral nectaries of 14 plants of Aspasia principissa were censused once every five weeks during the day using plants occurring near ground level along or near trails on BCI. On the last census date, ants were collected from 13 remaining, active inflorescences. More extensive data were collected for the ants occupying pseudobulbs of Caularthron bilamellatum. Ants were sampled over a five-month period from 573 orchids occurring on 87 trees and from 44 trees unoccupied by orchids. All sampled trees were Annona glabra L. occurring along the shoreline of the BCNM. Frequencies of ant occurrence among orchids and trees were compared using the log-likelihood ratio, G, an analog to the chisquare test (Sokal and Rohlf, 1981). RESULTS AND DISCUSSION-Of the 23 species of orchids investigated, 16 species were observed to produce extrafloral nectar (Table 1). All16 species were found to produce nectar on reproductive structures; nectaries were always found on developing inflorescences (Fig. 1-3) and often found on mature flowers (Fig. 4) and fruits. Nectar was most often produced at the base of pedicels (Fig. 3, 4) or at the bases of the dorsal sepal (Fig. 4) or lateral sepals (Fig. 1). Of the 16 species producing extrafloral nectar on reproductive structures, 10 also produced nectar on · developing shoots (Table 1, Fig. 5). However, only one species, Caularthron bilamellatum, produced extrafloral nectar on mature shoots. Interestingly, this is the only species in the study in which the ants occupy any portion of the plant (Dressler, 1981). Although we were unable to collect information for Epidendrum imatophyllum or Coryanthes speciosa, these might also be expected to produce extrafloral nectar on mature shoots, since ants commonly build nests among the roots of these two species (Jeffrey, Arditti, and Koopowitz, 1970; Croat, 1978). In most cases, no specialized structures were observed to be associated with the production of nectar. Exceptions were noted for Caularthron bilamellatum, which commonly has two or three small, nectar-producing "pits" at the base of the leaves, and in Aspasia principissa, which has quite obvious nectaries at the bases of the lateral sepals (seen in the developing bud, Fig. 1). 13 TABLE I. Orchid species from Barro Colorado National Monument, Panama, investigated for the production of extra· floral nectar. The symbol + denotes presence of extrafloral nectar and denotes absence of extrafloral nectar.