GEOGRAPHIC AND TAXONOMIC DISTRIBUTION OF A POSITIVE INTERACTION - ANT-TENDED HOMOPTERANS INDIRECTLY BENEFIT FIGS ACROSS SOUTHERN AFRICA

Although species pairs and assemblages often occur across geographic r egions, ecologists know very little about the outcome of their interac tions on such large spatial scales. Here, we assess the geographic dis tribution and taxonomic diversity of a positive interaction involving ant-tended homopterans and fig trees in the genus Ficus. Previous expe rimental studies ar a few locations in South Africa indicated that Fic us sur indirectly benefited from the presence of a homopteran (Hilda p atruelis) because it attracted ants (primarily Pheidole megacephala) t hat reduced the effects of both pre-dispersal ovule gallers and parasi toids of pollinating wasps. Based on this work, we evaluated three con ditions that must be met in order to support the hypothesis that this indirect interaction involves many fig species and occurs throughout m uch of southern Africa and Madagascar. Data on 429 trees distributed a mong five countries indicated that 20 of 38 Ficus species, and 46% of all trees sampled, had ants on their figs. Members of the Sycomorus su bgenus were significantly more likely to attract ants than those in th e Urostigma subgenus, and ant-colonization levels on these species wer e significantly greater than for Urostigma species. On average, each a nt-occupied F, sur tree had 37% of its fig crop colonized by ants, whe reas the value was 24% for other Ficus species. H. patruelis was the m ost common source for attracting ants, although figs were also attacke d by a range of other ant-tended homopterans. P. megacephala was signi ficantly more common on figs than other ant species, bei:ng present on 58% of sampled trees. Ant densities commonly exceeded 4.5 per fig, wh ich a field experiment indicated was sufficient to provide protection from ovule gallers and parasitoids of pollinators. Forty-nine percent of all colonized F. sui trees sampled had ant densities equal to or gr eater than 4.5 per fig, whereas this value was 23% for other Ficus spe cies. We conclude that there is considerable evidence to suggest that this indirect interaction occurs across four southern African countrie s and Madagascar, and involves many Ficus species.