Experimental Zoogeography of Islands: Defauantion and Monitoring Techniques by Wilson and Simberloff, 1969

This paper provided an extensive account of the methodology used by Wilson and Simberloff, to remove fauna on small islands. The purpose was to develop an effective technique of defaunation that would allow experimentation of colonization and provide data to test the hypotheses of MacArthur and Wilson (1963 and 1967). These hypotheses make suggestions on the equilibrium number of species on an island and the predicted survival times of recolonization, but are limited in what they can explain and the data to support them were scarce (at least at the time). Empirical evidence was needed to further explore these ideas.

The authors decided that in order to accurately determine appropriate methods for defaunation of small island a few criteria were needed:

1. There needed to be enough islands for replication and variation.
2. There needed to be sufficiently large animal diversity and organisms that were large enough to accurately find and identify.
3. Small island size was needed to compensate for the close distance they were to the mainland source.

Small islands along the Overseas Highway of southern Florida fit all of the study criteria. A total of nine islands of varying size (11-25m diameter) and distance from the mainland source (2-1188m) were censused (25-43 initial numbers of arthropod species on each island). Island flora consisted of mangrove trees and fauna consisted of arthropods, mostly arboreal, with about 75 species of insects, about 15 species of spiders and various other arthropod species. A few vertebrates including birds, snakes and raccoons were also found but were not included in the census as these were usually on the island to forage and were not considered inhabitants.

Of the nine islands censused, two were controls and not subjected to defaunation, one was an island used to test defaunation method and six were experimental islands were defaunation was performed. Five islands censused (four experimental, one control) were located in Great White Heron National Wildlife Refuge and three islands (two experimental, one control) were located in the Everglades National Park.

Two methods of defaunation were tested to determine overall effectiveness. The first attempt used a spray application of parathion and diazinon (insecticides). However, it was found that the spray couldn’t penetrate hollow twigs and although it did kill many arthropods species, it did not completely defaunate the island.

The second attempt was to use a fumigation method to remove arthropods from the island. Field and lab tests were conducted on mainland Florida to test four insoluble fumigants. Of the four tested, methyl bromide was found to have the lowest impact on plants and was the most effective at killing arthropods, including eggs and pupae of the more resilient species. Fumigation was conducted on two experimental islands, however heat damage occurred to the plants during the daytime fumigations and all other islands were fumigated at night.

Once the chemical and application technique was determined the next experimental issue was how to construct a fumigation tent without damaging the trees. The first tent constructed was supported by a temporary full frame that was constructed at the island site and used to raise the tent. Methyl bromide was introduced through the tent wall and an electric fan was used to disperse the gas. After fumigation, the gas was released through a seam in the tent for 45 minutes and the tent was removed. Although a success, the few live insects that were found in the six man hours spend exploring the island resulted in an increased concentration of methyl bromide on all subsequent islands.

Due to the difficulty in erecting scaffolding framework to hold the tent at islands of farther distance from a land source, a new method was devised. A tower was created at the island center and used to support most of the tent weight. This proved to be easier to construct on islands farther out, but was also more vulnerable to wind.

All islands were fumigated under a tent and examined after tent removal. Of the six experimental islands, there were no live insects on four, one live beetle on one island and one live millipede on another island. The authors concluded that their fumigation technique was efficient at removing all arthropods species, with the possible exception of deep boring species.

During monitoring, care was taken not to contaminate islands and also to prevent destruction of possible habitats. Dead insects were collected after fumigation to determine species composition on the island before defaunation. Species census was taken for two days, every 18 days after defaunation. Discovery of new species was very high until about 14 man hours post defaunation, at which point the new species discovery declined to near zero. Variation in island size and distance, and weather conditions were found to cause minor differences in the times in which species were discovered, however the cumulative species count curves were very similar in shape for the three islands that were plotted in species accumulation graphs.

This paper presents an interesting account of the author’s persistent attempts to refine and perfect their methodology of small island defaunation. Although I appreciate their perseverance and thorough account of their trials and errors, I wonder if all of their “failures” would be published in present-day articles, or if only the successful methods and subsequent results would be publishable and thus not needing a separate methods paper?