One hundred and two virulent (V) strains of Cryphonectria parasitica were isolated from chestnut blight cankers on American chestnut trees (Castanea dentata) in western Massachusetts, and vegetative compatibility was investigated. The 102 strains fell into 54 vegetative compatibility groups (VCGs): 10 groups contained more than 2 strains in each, 6 groups had 2 strains per VCG, and 38 VCGs contained only one strain each. A wide diversity of VCGs may be due to the presence of C. parasitica in Massachusetts for at least 80 years. A representative strain from each of the 10 VCGs with more than two strains, was paired with 22 hypovirulent (H) strains not native to Massachusetts (4 strains with French double-stranded RNA (dsRNA), 17 strains with Italian dsRNA, and 1 strain with American dsRNA). Some of the V strain could be converted to hypovirulent by the French and/or Italian H strains, but none were converted by the American H strain. Ten V strains from different VCGs were compared with 25 H strains, and with 29 strains converted to hypovirulence in cultural characteristics, pathogenicity, and phenol oxidase activity. Fungal growth, colony color, and pycnidial development were compared on various media. Fungal growth could be accentuated by the addition of ground bark or wood tissue to the media. Cultural characteristics were consistent among the V strains, but much variability was noted with the H strains. A rapid method (bark/wood test) for testing the virulence was developed, and compared with living tree inoculations and excised stem sections. The bark/wood test gave the fastest and most consistent results. Phenol oxidase activity of V strains was much stronger than those of H strains on the modified Bavendamm's medium. Several H strains were cured by the treatments of glucose and cyclic nucleotides (cAMP, cGMP). When the H strains were cured, they took on the cultural characteristics of normal V strains and virulence was restored. DsRNA bands were not detectable by gel electrophoresis from the cured strains, while uncured strains retained dsRNAs. It is assumed that RNA-dependant RNA polymerase gene, which regulates RNA replications, was depressed by cyclic nucleotides. Subsequently, RNA replication was inhibited and virulence returns as all genes can be expressed.

We used the nuclear polyhedrosis virus (LdNPV) of the gypsy moth, Lymantria dispar (Lepidoptera: Lymantriidae), to test a basic assumption of most models of disease dynamics; that the rate of horizontal transmission is directly proportional to the product of the densities of healthy larvae and virus. We made measurements of virus transmission, using small-scale experiments in bags on red oak (Quercus rubra) and black oak (Q. velutina) and observed a decline in the transmission constant as the densities of both healthy larvae and pathogen increased. We hypothesized two possibilities for the non-linearity observed in this system: (1) pathogen inhibition via the effects of induced foliage chemistry and (2) the effect of spatial heterogeneity of LdNPV (pathogen clumping effects). Previous work has linked larval damage on oak foliage to subsequently higher levels of tannins in damaged leaves; and laboratory bioassays have linked higher levels of tannins to a decrease in mortality caused by LdNPV. We damaged LdNPV-contaminated oak foliage using larvae, then bioassayed foliage with test larvae and measured tannins in damaged leaves. None of the experiments showed significant effects of foliage damage on mortality of test larvae or tannin content of damaged leaves, and we concluded that induced foliage responses cannot explain non-linearity of transmission in our previous work, or transmission dynamics in the field. To evaluate the role of pathogen clumping in transmission dynamics, we redefined parameters of the host-parasitoid model of Nicholson and Bailey (1935) for use in the gypsy moth-LdNPV system: a relationship between the assumptions of this model and the continuous time Anderson-May model was described. Using a version of this discrete-time model incorporating spatial heterogeneity, we observed transmission dynamics similar to those in field experiments. To evaluate safety issues in use of genetically altered baculoviruses as insecticides, we field-tested LdNPV that had been genetically-engineered for non- persistence. The EV was released in a forest setting, and did not persist or spread past the first year of release. Concerns regarding movement of baculoviruses were also addressed in experiments exploring the action of rainfall on the translocation of LdNPV.