Pratt, Linda W., Gregory L. Santos, and Charles P. Stone. 1994. A test of four herbicides for use on strawberry guava (Psidium cattleianum Sabine) in Kipahulu Valley, Haleakala National Park. Cooperative National Park Resources Studies Unit Technical Report #90. University of Hawaii at Manoa. Honolulu.
Strawberry guava or waiawi (Psidium cattleianum Sabine), a tree native to Brazil, was intentionally introduced to the Hawaiian Islands in the early nineteenth century. It is now naturalized at low to middle elevations on all the main Hawaiian Islands except Niihau and Kahoolawe. A fast-growing species capable of spreading by seed or vegetative suckers, strawberry guava forms dense stands, which shade out native species and prevent their reproduction. Strawberry guava is the most serious invasive alien plant in the wet koa/ohia (Acacia koa/ Metrosideros polymorpha) forests of Kipahulu Valley, Haleakala National Park, Maui. In 1987-88 ten herbicide treatments were, tested on strawberry guava trees at two mid-elevation wet forest sites in KIpahulu Valley. Six treatments involved ap lication of Garlon 3A (triclopyr, triethylamine salt) or Garlon 4 (tricP o pyr, butoxyethyl ester) to cut stumps of guava (undiluted, 50%, and 5% dilutions in either water or citrus oil). Control treatments were cut stumps sprayed with either water or citrus oil. One treatment was undiluted Garlon 3A ap lied to continuous frill cuts on guava trunks, and another was 50% Garlon S in citrus oil applied as a thinline to the basal bark of large trees. Additionally, at one of the wet forest sites, two soil application methods (spot or broadcast) were tested with two different herbicides: Spike 20P (tebuthiuron) or Velpar (hexazinone). Systematically selected untreated trees served as a control for the four soil treatments. The most effective treatment for strawberry guava tested in Kipahulu Valley was undiluted Garlon 4 applied to cut stum s; 21 months after herbicide application, this treatment had killed 80% of P. guava trees. Trees 5 cm or larger in diameter were more likely to survive treatment. The undiluted Garlon 4 treatment had the lowest number of resprouting stumps and the fewest resprouts per stump. All cut-stump treatments were better at inhibiting resprouts than were the water or citrus oil controls. Garlon 3A, 5% dilution in water, was significantly less effective than other treatments. The application of 50% Garlon 3A to continuous frill cuts on guava trunks was also moderately effective, resulting in the death of 55% of strawberry guava after 21 months. The thinline application of 50% Garlon 4 in citrus oil to the basal bark of guava was ineffective; no trees were killed by this treatment. Garlon 3A and Garlon 4 applied to cut stumps, frill cuts, or the basal bark of guava trees did not adversely affect native plants in plots surrounding treated strawberry guava. No significant losses of 14 native plant species were observed during 21 months of monitoring. Soil treatments with Velpar and Spike 20P were not effective in controlling strawberry guava. Native plants surrounding guavas in treated plots were adversely affected in Spike treatment plots. The recommended treatment for strawberry guava in Kipahulu Valley is the cut-stump application of undiluted Garlon 4. Because of potential re-rooting of cut slash and the less than 100% kill achieved with Garlon 4, further research with other application techniques and herbicides is suggested.