2.2.83
INTERPRETATION OF HOST SPECIFICITY STUDIES IN RELATION TO THE BIOLOGICAL CONTROL OF THE INVASIVE SHRUB LANTANA CAMARA WITH THE RUST PROSPODIUM TUBERCULATUM

CA ELLISON and SE THOMAS

CABI BIOSCIENCE, Silwood Park, Ascot, Berks. SL5 7TA, UK

Background and objectives
Lantana camara L. (Verbenaceae) is a spreading, woody shrub with showy flowers native to the Americas, which is now a major pantropical weed invading areas of both natural vegetation and pasture land. L. camara sensu lato in its exotic range is considered to be a complex of horticultural hybrids, with geographical origins obscured. Hence, populations have a broad genetic base, and consequently the selection of biological control agents with a sufficiently wide infectivity within the weedy populations becomes highly problematic. The rust Prospodium tuberculatum (Speg.) Arthur (Uredinales) was selected from a large number of fungal pathogens associated with the weed in its native range [1]. This study reports on host specificity screening designed to determine not only the range of the rust within populations of the target weed species, but also within related and unrelated species, as required for any classical biological control project prior to gaining permission for release.

Materials and methods
Prospodium tuberculatum (isolate W1241 ex pink flowered biotype of L. camara, Rio de Janeiro state, Brazil) was inoculated on to 30 different varieties/biotypes of L. camara and screened against 32 nontarget plant species. The latter were selected according to their relatedness to the weed, using a phylogenetic, centrifugal system. Urediniospores were formulated in talc (1:50), brushed onto the underside of the leaves (0.045 ;g of spores per leaf) and placed in a dew chamber set at 21C for 48 ;h. Test plants were monitored for 8 ;weeks and macroscopic symptoms scored. In compatible host-pathogen combinations an optimum dew period of 13 ;h is required and sporulation occurs after 3 ;weeks. However, to enable detection of any latent infection, extended dew periods and monitoring were carried-out. The Bruzzese and Hasan staining technique [2] was employed to enable microscopic analysis of the inoculated leaves.

Results and conclusions
The rust sporulated on five of the weedy varieties of L. camara, all with a pink flower colour. No macroscopic symptoms were observed on any other species/varieties that were tested. Microscopic analysis of the inoculated leaves revealed a wide range of interactions between non-host plants and the rust: complete inhibition of spore germination and occasional spore lysis; spore germination without recognition of stomata, often resulting in extended germ-tube growth over the leaf surface; recognition of leaf surface, with appressorial formation over stomata and no further development; penetration of the leaf with restricted internal mycelial growth. These results demonstrate the high level of host specificity of the rust, and confirm its suitability as a classical biological control agent against at least one of the most invasive weedy varieties of L. camara. In addition, the microscopic analysis forms an interesting study of how a biotrophic plant pathogen not only interacts with compatible and incompatible varieties at an intraspecies level, but also with other plant species of decreasing relatedness.

References
1. Barreto RW, Evans HC, Ellison CA, 1995. Mycoiogical Research 99, 769-782.
2. Bruzzese E, Hasan S, 1983. Plant Pathology 32, 335-338.