1University of Tasmania, GPO Box 252, Hobart, 7001, Australia; 2University of Hawaii, Honolulu 96822, USA; 3GPO Box 1O, Hobart 70, Australia

Background and objectives
Three viruses commonly infect hops (Humus lupulus ) in Australia, hop mosaic (HMV), hop latent (HLV) and prunus necrotic ringspot apple (PNRSV-A) and hop (PNRV-H) serotypes. In the past, all viruses have spread slowly in Australian farms allowing management through roguing of infected plants and the use of virus-free stocks. A recent shift in the Australian industry to higher yielding triploid varieties has coincided with an increased rate of spread of PNRSV (both serotypes) reducing the cost efficiency of current control methods. The objectives of this study were to document the rate of spread of viruses in a triploid cultivar, to identify associations between the viruses, to partition associations between virus pairs, and to assess the strength of these relationships.

Materials and methods
A plot of 1275 cv. Victoria plants, established from virus-free material in 1989 on a commercial hop farm at Bushy Park, Australia, were surveyed for virus incidence in 1990, 1996 and 1997. Leaf tissue from individual plants was tested by enzyme-linked immunosorbent assay using polyclonal antisera to HMV (WSU Prosser, UUSA), and chestnut mosaic virus (PNRSV-A isolate), D. Barbara, HRI, UK and, in 1996 and 1997, to HLV (D. Barbara). Extensive isolate studies allowed PNRSV-A serotypes to be distinguished by absorbances (A405) greater than 35 times the mean of the negative controls and PNRSV-H serotypes 15 to 25 times the mean of the negative controls. Data were analysed to determine whatever individual viruses were more likely to occur together or apart within plants or whether there was no association between viruses. Significant associations between viruses were determined by a W statistic which approximates to a chi-square distribution. Measures of strength of pairwise associations were computed by the Ochiai Index, equal to zero when there was no association and one at maxixum levels of association.

Result and conclusions
In 1990, HMV was found in 0.8% of plants and PNRSV in 2.1% of plants. In 1996, HMV was detected in 60.0%, PNRSV in 66.0% (27.0% PNRSV-A, 39.0% PNRSV-H) and HLV in 64.7% of plants. In 1997, HLV infection increased to 77.4%, PNRSV to 91.0% (29.2% PNRSV-A, 61.8% PNRSV-H), and HMV to 88.5%. The spread rate of PNRSV observed in cv. Victoria is six times that recorded in traditionally grown cultivars [2], indicating that this cultivar is particularly susceptible to infection by PNRSV.

Significant positive associations were found in 1996 and 1997 between HLV and HMV (0.786; 0.876), and HLV and PNRSV (0.685; 0.907), while HMV and PNRSV were found to be independent. HLV and HMV have common vectors which may account for their positive association. Spatial analysis of epidemics of all viruses showed spread was more likely to occur up and down rows, following the path of cultural operations. This suggests the involvement of mechanical transmission and/or apterous aphid vectors for HLV and HMV. The association between HLV and PNRSV may also reflect a common mode of spread for these two viruses (e.g. mechanical) or that infection by one virus may increase the susceptibility of plants to infection by the other.