Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

SECONDARY SEED REMOVAL BY ANTS, BEETLES, AND RODENTS
IN A NEOTROPICAL MOIST FOREST

By
Matthew B. Jones
April, 1994
Chair: Douglas Levey
Major Department: Zoology

I examined two aspects of secondary seed removal for four plant species in two Panamanian tropical moist forests. First, I examined the effect of seed manipulation by primary dispersers on secondary seed removal, fungal attack, and germination for Ficus maxima (Moraceae) and Ossaea quinquenervia (Melastomataceae). For F. maxima, I examined removal of seeds from three primary dispersal substrates: feces, fruit, and no substrate. Beetles removed F. maxima seeds from howler monkey feces and ants removed seeds from fruit and seeds not in a substrate. Significantly more seeds were removed from feces than from fruit and seed without substrate. Beetles separated dung from seeds whether or not they moved seeds. Because fungi infected defecations more than other substrates, fungal growth on treatments with beetle access was much lower than on treatments from which beetles were excluded, regardless of whether seeds were moved by beetles. Thus, seeds that remained in feces had significantly lower viability than those that were separated from dung by beetles. For O. quinquenervia, ants removed an external seed body without moving the seeds. Seeds missing these external bodies almost never germinated, while intact seeds and control seeds freshly removed from fruits showed greater than 90% germination success. In summary, the effect of plant-insect interactions after primary dispersal depended on the plant species involved, the particular seed harvesting insects, and the substrate in which seeds were dispersed initially.

Second, I studied secondary seed removal by ants and rodents for two plants with intermediate-sized seeds in two forests that probably differed in their rodent assemblages. Groups of five seeds of Virola nobilis and Psychotria marginata were placed on the forest floor at one site on the Gigante peninsula and one site on Barro Colorado Island (BCI). Ectatomma ruidum, a common litter ant, removed P. marginata seeds from both Gigante and BCI sites approximately equally. Rodents did not remove P. marginata. Removal by ants was rapid during the first day of the experiments but quickly declined thereafter. Rodents removed V. nobilis seeds but ants did not. Removal was significantly greater on the Gigante peninsula (96%) than on BCI (38%). Removal in rodent-accessible stations was similar to rodent- and ant-accessible stations on Gigante but not on BCI, which suggests a difference in seed harvester assemblages at the two sites.

[Chapter 1] [Chapter 2] [Chapter 3] [References] [My Home Page]