frutescens INTRODUCTION
Bocconia L. (Papaveraceae), or plume poppy, is an invasive shrub listed as a noxious weed by the State of Hawai`i (HDOA 1992). Bocconia is native to Central and
South America and the West Indies and was introduced to Hawai`i as an ornamental (Wagner et al. 1999, Starr et al. 2003). This species is characterized by a soft-wooded trunk, large lobed leaves, fast growth and copious seed production. Bocconia is known to occur on the Hawaiian islands of Maui and Hawai`i. A single broad infestation is documented from Maui, in primarily dry forests on the south and west slopes of Haleakal between 490 and 1,200 m elevation (Starr et al. 2003). Three distinct populations were documented on Hawai`i Island in late 2002 by the Big Island Invasive Species Committee (BIISC); however, population densities and range extents were unknown at that time. This report summarizes survey work initiated in early 2003 to document the extent of Bocconia invasion on Hawai`i Island. The purpose of this study is to provide data for development of a strategy to manage Bocconia infestations on the island, and to provide baseline data for future Bocconia studies.
METHODS
Surveying was conducted from April through July 2003, to document the distribution and population density of Bocconia frutescens in Wood Valley (Ka` District), Honomolino and Manuk (South Kona District) on the Island of Hawai`i (Figure 1). The study areas were selected based on known infestations identified from prior field data (Benitez 2002, BIISC 2002). Initial ground surveys were conducted along roadsides near known infestations, and presence/absence of Bocconia was used to direct further sampling. Seventy-nine linear transects, totaling 30.7 km in length, were established at Wood Valley and Manuk. These transects ranged from 40 to 1,600 m in length and were placed parcels in of uniform vegetation, generally bounded by roads, or other conspicuous features such as gulches or forest. Transect width varied between three and 50 m between transects, but was uniform within individual transects. The start of each transect was determined randomly along the most accessible edge of the parcel to be sampled, using a GIS. Within sampled parcels, between one and eight transects were established and monitored. Counts of mature plants (>1.25 m or fertile) and juveniles (<1.25 m and sterile) were recorded along each transect. Population densities (plants per hectare) were calculated by dividing the number of plants found along a transect by the area sampled within the transect. These values were extrapolated across the parcel. Where multiple transects were sampled within a parcel, the counts and transect area were summed, and these values were used to calculate the parcel's density. Transects were generally representative of the parcels with respect to Bocconia abundance and vegetation type. One 4.3 km roadside transect was sampled in Wood Valley in mixed vegeta...