In forecasting the likelihood of BYDV infection, information on infectivity index is necessary. The infectivity index can be used to guide decisions for insecticidal sprays. However, to improve on BYDV forecasting and hence aid epidemiological studies, improved prediction of secondary virus spread is also desirable. These experiments sought to show this by measuring primary infection and quantifying factors affecting aphid movement. Using two species of aphids Sitobion avenae (Fabricius) and Rhopalosiphum padi (L.) and two isolates of virus (MAV and PAV), the proportion of plants infected for six inoculation access periods (2,6,12,24,48 and 72 hrs) at four temperatures (5^oC, 12^oC, 18^oC and 25^oC) were compared. There was no further increase in the proportion of plants with virus after a 24 hour inoculation period for each of the MAV or PAV like isolate. Also using four aphid population densities (2,8,16 and 24 aphids per plant) at four temperatures (5^oC,12^oC,18^oC and 25^oC), the probability of an aphid leaving a plant and distance it subsequently moved was studied. There was very little movement of aphid at densities of 2 per plant. For all other densities, the proportion of aphids that moved increased with temperature. Methods for improvement on the forecasting system at Rothamsted Experimental Station, United Kingdom are suggested.

Barley yellow dwarf viruses (BYDV) are a group of luteoviruses which can cause serious losses to barley (and other small cereals especially oats and wheat) in the United Kingdom, the extent of which varies annually and spatially. A forecasting system is being developed to prevent prophylactic spraying of autumn sown cereal crops to control the main aphid vectors, Sitobion avenae (Fabricius) and Rhopalosiphum padi (L.).

Interest in BYDV at Rothamsted Experimental Station continues from an early forecasting technique termed the `infectivity index' (Plumb et al., 1981). This is calculated from the total number of each species of cereal aphid caught in the Rothamsted 12.2m suction trap (Macauley et al., 1988; Taylor, 1986) multiplied by the proportion of these species, caught alive in a 1.5m trap in the same area, that subsequently transmit the virus to test cereal seedlings. The index was highly correlated with virus incidence in autumn cereal crops in the south-east of England but failed to predict BYDV outbreak in south-west Scotland (Forster et al., 1993). A seven-year study at four sites throughout England showed variable correlations between the infectivity index and virus incidence (Tatchell et al., 1994). However, a problem associated with the infectivity index is that it fails to allow for secondary spread of the virus incidence (Forster, et al., 1993).

This paper describes experiments which will lead to improved predictions of secondary spread within the crop as the ultimate objective of the research programme is to provide a decision support system for the use of insecticides to control the aphid vectors.

Transmission of the virus can be separated into three periods; a) the acquisition access period, during which the vector has access to the virus source plant although is not necessarily feeding; b) the latent period i.e. the time after the vector has acquired the virus but before it becomes infective; and c) the inoculation threshold period, that is the minimum time necessary for a vector to feed on a healthy plant before inoculation occurs (Federation of British Plant Pathologist, 1973). Work is being done to quantify these different periods using alate aphids, at temperatures common in the autumn in the U.K.

Two species of aphids, S. avenae transmitting the MAV-like isolate, and R. padi transmitting the PAV-like isolate to oats (cv. Dula) were used. Oats (cv. Dula) was used since it is widely believed to be the most susceptible cereal to BYDV. The proportion of plants infected for six inoculation access periods (2,6,12,24, 48 and 72 hours) at each of four temperatures; 6^oC, 12^oC, 18^oC, and 23^oC were compared. Each species/isolate combination was tested separately and each experiment was replicated five times. In each experiment a batch of ten viruliferous aphids were used at each temperature/inoculation access period combination. Uninfected aphids were placed on ten control plants at each temperature and removed after 72 hours.

After a 72-hour acquisition feed, single winged aphids were placed on the oat seedlings for the required time. The aphids and any nymphs were then squashed to prevent further inoculation and the plants removed to a glasshouse. The plants were assessed for visual symptoms of virus at 17 and 24 days, and any plants with indistinct symptoms were tested by double antibody sandwich ELISA (Lister and Rochow, 1979).

There was no further increase in the proportion of plants with virus after a 24 hour inoculation period for either the MAV-or PAV-like isolates. For the PAV-like isolate there was no difference between the proportion of plants infected at 12^oC, 18^oC or 23^oC, but at 6^oC the proportion was significantly smaller. For the MAV-like isolate, there is a weak positive relation between the proportion of plants infected and temperature.

A major problem with the infectivity index is its failure to account for secondary spread during winter months. This is of particular importance under mild winter conditions, when transmission is more efficient and increased aphid survival and movement enhances spread of the virus through the crop. Sitobion avenae which overwinters more successfully than R. padi (Dewar and Carter, 1984; Tatchell et al., 1994) is thought to be the prime contributor to secondary spread of the virus. An accurate prediction of disease progression during the winter relies on an understanding of how biotic and abiotic factors determine vector movement. Some of these including temperature and aphid density were investigated in a series of laboratory experiments.

To determine the probability of an aphid leaving a plant and the distance it subsequently moved, S. avenae were released on the central plant of a 9 x 9 grid of oat seedlings at growth stage 12, planted at a spacing of 6cm. Densities of 2, 8, 16 and 24 aphids per plant and temperatures of 5, 12, 18 and 25^oC were compared. Aphids were contained on the plant by clip-cages for 24 hours before being released. After a further 48 hours all plants were thoroughly searched and aphid positions noted. The proportion of aphids which had left the central plant and the mean distance they moved were calculated.

There was very little movement of aphids away from the central plant at densities of 2 per plant. For all other densities the proportion of aphids that moved increased with temperature and for any given temperature, increased with density. There was no significant interaction between densities of aphids and temperature. The average distance moved also increased significantly with temperature and with density although the latter effect was not consistent for all temperatures at low densities, whereas at temperatures of 18^oC and 25^oC there was little difference among densities.

The results obtained for the determination of primary infection in the autumn, and those for the acquisition and latent period experiments yet to be completed, will be used to estimate the time taken at a certain temperature for an infective aphid to transmit the virus and thereby improve future prediction models. Results for virus transmission are more useful when combined with data on aphid movement. The time an aphid stays on a plant at a given temperature, when combined with data on transmission of the virus by apterous aphids, facilitates understanding of the movement of aphids and its relation to the spread of virus within a crop over a given period.

For the experiment on quantifying factors affecting aphid movement, it was observed that high aphid densities influence their dispersal but at low aphid densities (2 per plant or less) as might occur during winter months, aphid density was not important. Temperature had a clear effect, with movement increasing with temperature. At 5^oC very little movement occurred, indicating that the threshold for movement may be around that point. Data presented will be useful in predicting aphid movement during the winter based on ambient temperatures, particularly when combined with results on virus transmission.

At Rothamsted Experimental Station, United Kingdom, trap catches in the autumn have been combined with the proportion of infective aphids to give the infectivity index that can be used to guide decisions on the need for aphicidal sprays. However, for better BYDV forecasting results, improved predictions of the secondary spread of the virus is also desirable. While suitable methods for determining the amount and distribution of primary inoculum are still being assessed, a more appropriate forecasting model should therefore include the effects of temperature on aphid development, reproduction and survival: the effects of temperature, wind, rain, plant density, plant age, aphid density and aphid age on aphid movement; and the effects of temperature, plant age and aphid age on virus transmission characteristics.

Dewar, A.M. and Carter, N. 1984. Decision tree to assess the risk of cereal aphid (Hemiptera: Aphididae) outbreaks in summer in England. Bulletin of Entomological Research 74: 387 - 398.

Federation of British Plant Pathologists 1973. A guide to the use of terms in plant

pathology. Phytopathology papers No. 17.

Forster, G.N., Holmes, S.J. and Bone, S.F. 1993. Ten years' experience of infectivity indexing as a method of predicting the risk of barley yellow dwarf virus outbreaks in autumn sown cereals in the west of Scotland. Proceedings Crop Protection in Northern Britain 1993: 97 - 101.

Lister, R.M. and Rochow, W.F. 1979. Detection of barley yellow dwarf virus by enzyme linked immunosorbent assay. Phytopathology 69:649-654.

Macauley, E.D.M., Tatchell, G.M. and Taylor, L.R. 1988. The Rothamsted insect survey "12-meter" suction trap. Bulletin of Entomological Research 78:121-129.

Plumb, R.T., Lennon, E.A. and Gutteridge, R.A. 1981. Barley yellow dwarf virus (BYDV). Report of Rothamsted Experimental Station for 1981. Part 1, 195-198.

Tatchell, G.M., Hill. S.A., Tones, S., Barker, I. and Smith, A.J. 1994. Monitoring aphids

and virus to improve forcecast of barley yellow dwarf virus. Project report 87 of the

Home Grown Cereals Authority.

Taylor, L.R. 1986. Synopic dynamics and the Rothamsted insect survey. Journal of Animal Ecology 55: 1-38.