Comparison of the potential rate of population increase of brown and green color morphs of Sitobion avenae (Homoptera: Aphididae) on barley infected and uninfected with Barley yellow dwarf virus

Life tables of brown and green color morphs of the English grain aphid, Sitobion avenae (Fabricius) reared on barley under laboratory conditions at 20 ± 1°C, 65% ± 5% relative humidity and a photoperiod of 16 : 8 h (L : D) were compared. The plants were either: (i) infected with the Barley yellow dwarf virus (BYDV); (ii) not infected with virus but previously infested with aphids; or (iii) healthy barley plants, which were not previously infested with aphids. Generally, both color morphs of S. avenae performed significantly better when fed on BYDV‐infected plants than on plants that were virus free but had either not been or had been previously infested with aphids. Furthermore, when fed on BYDV‐infected plants, green S. avenae developed significantly faster and had a significantly shorter reproductive period than the brown color morph. There were no significant differences in this respect between the two color morphs of S. avenae when they were reared on virus‐free plants that either had been or not been previously infested with aphids. These results indicate that barley infected with BYDV is a more favorable host plant than uninfected barley for both the color morphs of S. avenae tested, particularly the green color morph.     

Serological and biological characterisation of isolates of barley yellow dwarf virus in Sweden in 1983

In 1983, cereal plants showing symptoms of barley yellow dwarf virus (BYDV), collected from 15 localities in Sweden, were tested for BYDV using enzyme-linked immunosorbent assay (ELISA). Antisera against two Swedish isolates of BYDV were used, a mild isolate (27/77) transmitted specifically by Sitobion avenae and a severe one (39/78) transmitted mainly by Rhopalosiphum padi. No virus was detected in 57 of 607 plants of oats and barley tested. Of the 550 plants in which virus was detected, 366 were infected with viruses similar to isolate 27/77, 116 with viruses similar to 39/78 and the remaining 68 reacted strongly with both antisera. When tested, the latter isolates were shown to be mixtures. Thirty-nine selected samples were also tested with antisera against the USA isolates RPV, RMV, MAV and PAV, and for transmission by S. avenae and R. padi. Twenty-six of these samples were transmitted specifically by S. avenae, one was transmitted only by R. padi and the remaining 12 samples were shown to be infected with a mixture of an S. avenae-specific isolate and one transmitted mainly by R. padi. Antisera against PAV and MAV each detected all isolates tested and the results were very similar to those with the antisera to the 39/78 and 27/77 isolates, respectively. None of the field isolates reacted with antisera against RMV or RPV. It was concluded that 1983 was an epidemic year for BYDV in Sweden and that isolates specifically transmitted by S. avenae predominated. Symptoms of infection by these isolates on oat plants ranged from mild to severe.     

Barley yellow dwarf virus, wheat, and Sitobion avenae: a case of trilateral interactions

We analysed interactions in the system of two Barley Yellow Dwarf Virus (BYDV) strains (MAV and PAV), and wheat (cv. Tinos) as host plant for the virus, and the cereal aphid Sitobion avenae (F.) as vector, in particular whether or not infection by the virus might alter host plant suitability in favour of vector development. By measuring the amino acid and sugar content in the phloem sap of infected and non‐infected wheat plants we found a significant reduction in the concentration of the total amount of amino acids on BYDV‐infected plants. Qualitative and quantitative analysis of honeydew and honeydew excretion indicated a lower efficiency of phloem sap utilisation by S. avenae on infected plants. In addition, S. avenae excreted less honeydew on infected plants. Both BYDV strains significantly affected aphid development by a reduction in the intrinsic rate of natural increase. Hence, infection by the virus reduced the host suitability in terms of aphid population growth potential on BYDV‐infected plants. However, more alate morphs developed on virus‐infected plants. These findings are discussed in relation to the population dynamics of S. avenae, and, as a consequence, the spread of BYDV.     

The effect of barley yellow dwarf virus on field populations of the cereal aphids, Sitobion avenae and Metopolophium dirhodum

The numbers of cereal aphids, especially Metopolophium dirhodum in 1979, and Sitobion avenae in 1980, were significantly increased on BYDV infected wheat and oats in 1979, and wheat, barley and oats in 1980. The differences were probably caused by attraction of alates of each species to virus infected plants which had changed colour as a result of their infection. Significantly more alates of M. dirhodum were found on virus infected oats in 1979, and of S. avenae on oats and barley in 1980, although not on wheat in either year. probably because the colour contrast in wheat was less intense than in the other crops. Flight chamber experiments with alates of both species confirmed their visual attraction to virus-infected leaves. The interaction between virus, vector and host plants is discussed with reference to the ecology of virus spread.     

The effect of barley yellow dwarf virus on honeydew production by the cereal aphids, Sitobion avenae and Metopolophium dirhodum

Individual S. avenae and M. dirhodum excreted significantly fewer droplets of honeydew on plants infected with BYDV than on healthy plants. S. avenae excreted less honeydew on the ears than on the leaves of wheat. M. dirhodum excreted less than S. avenae on the leaves. The size of honeydew droplets increased with the age of aphids but was not affected by BYDV infection. Possible reasons for the observed effects of BYDV on honeydew excretion are discussed.     

Spread of barley yellow dwarf virus and relative importance of local aphid vectors in central Washington

Data from bioassays of field collected aphids, barley indicator plants exposed to natural conditions, and various types of aphid traps were used to describe the spread of barley yellow dwarf virus (BYDV) in wheat and barley near Prosser, Washington. Bioassays were also used to assess the relative importance of local vector species. Of alate aphids collected from grain in the 1982 and 1983 fall migration seasons, 3.4–14–5% transmitted BYDV. Data from concurrent and post-migration assays of resident aphids (apterae and nymphs) reflected an increase in the proportion of infected plants in the field. Maximum increase in the percentage of viruliferous aphids occurred in late November and December of 1982 and November of 1983. The 1982 increase occurred after aphid flights had ceased for the year, suggesting active secondary spread. Collections in pitfall traps and infected trap plants from November to February confirmed aphid activity and virus spread. Rhopalosiphum padi was the most important vector in central Washington in 1982 and 1983 because of its abundance and relative BYDV transmission efficiency. Metopolophium dirhodum was more winter-hardy than R. padi and equal to R. padi in its efficiency as a vector; however, it was not as abundant as R. padi except during the mild winter of 1982–83, when it was a major contributor to secondary spread. Sitobion avenae may be important in years when it is abundant, but it was only a quarter as efficient as R. padi. Rhopalosiphum maidis was a much less efficient vector than R. padi and it only reached high populations in late autumn barley.     

Identification, Distribution and Vector Population Dynamics of Barley Yellow Dwarf Virus in Three Cereal-Producing Areas of Spain

ELISA-based surveys during 1985–87 in three major cereal-growing areas of Spain confirmed the presence of barley yellow dwarf virus (BYDV). Samples of small grain cereals and grasses with and without BYDV-like symptoms were collected in the central, southwestern, and northeastern Spain. Infections were found in all cereal species sampled and in some grasses. About 37 % of the samples collected in 1985 were infacted with isolates of the PAV serotype. Isolates of the RPV serotype were less common, and were detected only in samples from the central region at El Encin, Madrid. Only a single sample, collected from El Encin in 1987, was unequivocally diagnosed as containing an isolate of the MAV serotype. Aphid vector population dynamics was monitored during fall and winter of 1984–87 in the central region. Rhopalosiphum padi L. appeared to be the most abundant species during fall and winter months, infesting grasses and volunteer wheat. Other species present were Sitobion avenae (F.), Metopolophium dirhodum (Walker) and Rhopalosiphum maidis (Fitch). Both R. padi and S. avenae seem to be anholocyclic in the central region of Spain, and are able to remain and reproduce on wheat volunteers and grasses until the beginning of spring. S, avenae populations increase quickly on wheat volunteers in April, while populations of R. padi remain low. Therefore, spread of S. avenae-transmitted BYDV types to neighbouring cereal fields seem more likely to occur than spread of other types. Other possible virus reservoirs, such as maize, also need investigation for a better understanding of BYDV epidemiology in the central and other cercal-growing areas of Spain.     

Interaction between barley yellow dwarf virus and rust in wheat, barley and oats, and the effects on grain yield and quality

In three separate experiments, the upper leaf surface of the fifth formed leaf of wheat cv. Highbury, the fourth and fifth leaves of barley cv. Julia and the third and fourth leaves of oat cv. Mostyn were inoculated in a spore settling tower with wheat brown rust (Puccinia recondita f. sp. tritici), barley brown rust (P. hordei) or oat crown rust (P. coronata f. sp. avenae), respectively. Fewer pustules developed on distal portions of leaves of plants infected with barley yellow dwarf virus (BYDV) than on similar portions of leaves from virus-free plants. There were no significant differences in the number of pustules on proximal leaf portions. In barley and oats, the number of pustules on distal leaf portions was negatively correlated with the amount of yellowing of the leaf areas scored. In wheat, symptoms of BYDV were mild and leaves were little affected by yellowing. The latent period of rust on wheat and oats was not affected by BYDV. In barley, BYDV reduced the latent period of rust on leaf 5, but not on leaf 4, and reduced it on proximal, but not distal, leaf portions. In other experiments, BYDV reduced the yield of wheat and oats by 44% and 66%, respectively, while BYDV-infected barley was almost sterile. The appropriate rust reduced the yield of wheat, barley and oats by 33%, 13% and 86%, respectively. When infected with both BYDV and rust, yield of wheat and oats was reduced by 63% and 91%, respectively. Neither BYDV nor rust affected the percentage crude protein content of wheat grain, nor did rust affect that of barley. In oats, BYDV and rust each significantly increased crude protein of grain, but rust infection of BYDV-infected plants tended to reduce it.     

The effect of different virus isolates on the expression of tolerance to barley yellow dwarf virus in barley

A barley variety of Ethiopian origin, with a single Mendelian gene con-fering tolerance to barley yellow dwarf virus (BYDV), was equally tolerant to a number of isolates of the virus, whereas a susceptible European barley was more susceptible to isolates transmitted by Rhopalosiphum padi L. than to those transmitted by Macrosiphum (Sitobion) avenae (Fab). However, hybrids between these two varieties homozygous for the Ethiopian tolerance gene were more tolerant to ‘mild’ than to ‘severe’ isolates, irrespective of the vector specificity. The European variety was damaged more severely by all isolates when infected early than when infected late in its development, but the hybrids were damaged more severely by M. awraae-transmitted isolates when infected late. It is suggested that in susceptible plants the concentration, rather than the virulence, of the virus determines disease severity, whereas the reverse is true in plants possessing a gene which reduces virus multiplication. Virus concentration appears to determine the severity of R. padi-transmitted isolates, while virulence determines the severity of M. avenae-transmitted isolates. The latter would also seem to be adapted towards late infection.     

Barley yellow dwarf virus in aphids caught in suction traps, 1969-73

Suction traps operating at low level (1 5 m) were used to catch live alate Rhopalosiphum padi, Macrosiphum (Sitobion) avenae and Metopolophium dirhodum which were tested for transmission of barley yellow dwarf virus (BYDV). The first species caught and infective was R. padi, followed by M. (S.) avenae infective some 2–3 wk later and M. dirhodum 3–4 wk later still. Never more than 11-5% of the annual catch of any species transmitted BYDV and the proportion fluctuated from week to week and between seasons in different years. The relative abundance of infective vectors of ths three species varied; annual numbers of infective M. (S.) avenae and M. dirhodum varied inversely with infective R. padi, the latter also usually transmitted severer virus. The results of the infectivity tests have been compared with the catches of these aphids by the Rothamsted Insect Survey and show that numbers of alate aphids do not necessarily indicate the likely incidence of BYDV.     

Virus infection stimulates phyto-oestrogen production in pasture legume plants growing in grazed swards

Four field experiments were sown with AMV‐infected or healthy seed of burr medic (Medicago polymorpha) and grazed by sheep; two were sown with cv. Circle Valley and two with cv. Santiago. Seed‐infected plants acted as primary sources for virus spread by naturally occurring aphids. Insecticides and admixture with annual ryegrass (Lolium rigidum), a non‐host of AMV, suppressed virus spread to different extents in the plots sown with infected seed. Effects of the different amounts of virus spread obtained on overall concentration of the oestrogenic compound coumestrol (dihydroxycoumestan) in dry stems and pods, and on seed production were measured in the medic. With cv. Santiago, stem and pod coumestrol concentration values for plots sown with healthy seed were significantly smaller than those for all plots sown with infected seed regardless of whether they were sprayed. With cv. Circle Valley, the coumestrol values for stems from plots sown with healthy seed were significantly smaller than those for unsprayed plots sown with infected seed but not than those for sprayed plots or ones with grass admixtures, and there were no significant differences with pods. There was always a significant positive relationship between concentration of coumestrol in medic stems and percentage AMV infection of swards; this was also so with pods in two experiments. A linear model best fitted this relationship with cv. Circle Valley but a logarithmic model did so with cv Santiago. In glasshouse grown plants, the coumestrol content of dried medic shoots was increased 11 (cv. Circle Valley) and five (cv. Santiago) times by AMV infection. AMV increased mean coumestrol concentrations up to 256 ppm (field) and 237 ppm (glasshouse) in stems and 223 ppm in pods (field). Sowing healthy seed in new pasture swards was an effective strategy for minimising coumestrol accumulation in burr medic swards. Two, but not single, applications of a newer generation pyrethroid insecticide to swards in which AMV was spreading significantly diminished coumestrol accumulation but applying organophosphorus insecticide twice and carbamate insecticide repeatedly did not. Medic seed yields and individual seed weights were sometimes significantly increased by the treatments, suppression of AMV spread by regular carbamate sprays being sufficient to increase seed yield by 55%.     

Cereal aphids and the infectivity index for barley yellow dwarf virus (BYDV) in northern England

Suction traps at Leeds University Farm, N. Yorkshire, monitored aerial populations of cereal aphids over three autumns. Different migration patterns were observed between the four main species, Sitobion avenae, Metopolophium dirhodum, Rhopalosiphum padi and R. insertum. The relevance of these patterns to the epidemiology of barley yellow dwarf virus (BYDV) is discussed. Transmission tests revealed S. avenae to be the major vector of BYDV, rather than R. padi, which is responsible for disease outbreaks in the south and west of Britain. An Infectivity Index (II) of 50 has been advocated for R. padi-transmitted BYDV, above which economic damage is likely to occur. This value is shown not to be applicable to the Vale of York, and methods of adapting the data are proposed. Such adjusted II values depend on the behaviour and reproduction of the aphids during the transmission tests, and produce II values that correlate well with levels of field infection in the area.     

Aphid Vectors of Barley Yellow Dwarf Virus in West-Central Morocco

The ability of seven aphid species, collected in west-central Morocco, to transmit barley yellow dwarf virus (BYDV) was determined. Aphids were either collected from grasses showing symptoms of BYDV infection or were allowed acquisition access to plants infected with a PAV-like isolate of BYDV before transfer to oat test plants. BYDV transmission by six of the seven aphid species was confirmed by ELISA test; only Melanaphis donacis failed to transmit. The six newly defined BYDV vector species brings the total known to occur in Morocco to ten.     

Assessment of Sublethal and Transgenerational Effects of Pirimicarb on the Wheat Aphids Rhopalosiphum padi and Sitobion avenae

The wheat aphids, Rhopalosiphum padi (Linnaeus) and Sitobion avenae (Fabricius), are key pests on wheat crops worldwide. Management practices rely primarily on insecticides. The pirimicarb (carbamate) is used extensively as an effective insecticide to control these two aphids. In addition to the mortality caused by pirimicarb, various sublethal effects may occur in aphids when exposed to low lethal or sublethal doses. Understanding the general effect of pirimicarb on aphids could help increasing rational use of this insecticide. Under laboratory conditions, we assessed the sublethal effects of a low lethal concentration of pirimicarb (LC₂₅) on biological traits and acetylcholinesterase (AChE) activity of R. padi and S. avenae. Both direct and transgenerational effects, i.e. on parent and the F1 generations were assessed, respectively. We found that R. padi and S. avenae responded differentially to the LC₂₅ of pirimicarb. The parent generation of R. padi showed a 39% decrease in fecundity and multiple transgenerational effects were observed in the F1 generation; overall juvenile development, reproductive period, adult longevity and lifespan were longer than those of the control group. By contrast, LC₂₅ of pirimicarb showed almost no effects on S. avenae biological traits in both the parent and F1 generations; only the pre-reproductive duration was reduced in F1 generations. Demographic parameter estimates (e.g. r_m) showed similar trend, i.e. significant negative effect on R. padi population growth and no effect on S. avenae. However, AChE activity decreased in both R. padi and S. avenae treated by the LC₂₅ of pirimicarb. We demonstrated sublethal and transgenerational effects of pirimicarb in the two wheat aphid species; it hinted at the importance of considering sublethal effects (including hormesis) of pirimicarb for optimizing Integrated Pest Management (IPM) of wheat aphids.     

Interactions of crop density of field beans, abundance of Aphis fabae Scop., virus incidence and aphid control by chemicals

The number of Aphis fabae Scop. per plant and per acre developing on field beans (Vicia faba L.) was inversely related to seeding rate (i.e. plant density) except sometimes at very low rates; with equal numbers of plants per acre, fewer aphids developed on plants in rows 11 in. than 22 in. apart. Plots sown in mid-March with more than about 150,000 plants per acre were more attractive than less dense stands to colonizing alate A. fabae, but established colonies multiplied most on the sparsest and least on the densest plots. The number of plants per acre infected by pea leaf-roll virus was inversely related to planting density. There were more virus-infected plants on II in. than on 22 in. spaced rows-in contrast to the numbers of A. fabae. A single spray with demeton-methyl, timed to control A. fabae, did not significantly decrease virus incidence. Grain yields of sprayed plots were little altered by increasing the seed rate above a critical minimum, except in one year when the densest crops lodged. Increased yields from spraying were closely related to the numbers of A. fabae on unsprayed plots. Dense planting (more than 400,000 plants per acre) prevented or greatly decreased losses caused by A. fabae in unsprayed plots except in one year when the aphids were exceptionally abundant.     

Vector specificity of barley yellow dwarf virus serotypes and variants in southwestern Idaho

Plants with symptoms of barley yellow dwarf virus (BYDV) obtained in infection feeding assays of aphids collected in the field in Idaho between 1986 and 1988 were tested for virus transmissibility by possible aphid vectors. Isolates obtained during 1987–1988 were also tested with a range of polyclonal antisera which distinguished PAV, MAV, SGV, RPV and RMV serotypes. In 1989 some Idaho (ID) BYDV isolates, maintained as standards for comparison, were serotyped and tested for aphid transmissibility, using 11 species of aphids. There was not always the expected correspondence between serotype and vector specificity for ID isolates. For isolates obtained from field-collected Rhopalosiphum padi, vector transmissibility and serotype corresponded with previous reports; however, 44% of isolates which were serotyped as RMV were also transmissible by species other than Rhopalosiphum maidis. Similarly, the transmissibility of the ID laboratory standards did not always conform to the reported vector specificity of serotypes. The laboratory ID-MAV culture was transmitted by Metopolophium dirhodum and Myzus persicae as well as by Sitobion avenae. The laboratory ID-SGV culture was transmitted by R. padi and 5. avenae as well as by Schizaphis graminum. The ID-RPV culture was transmitted by S. graminum and Rhopalosiphum insertum as well as R. padi. Both of two laboratory ID-RMV cultures were transmissible by R. insertum and R. padi transmitted one of them. The results indicate that, for isolates collected in Idaho, vector specificity cannot be assumed from their serotypes.     

High Levels of Resistance in Agropyron Species to Barley Yellow Dwarf and Wheat Streak Mosaic Viruses

Several Agropyron species were tested for new sources of resistance to barley yellow dwarf virus (Bydv ) and wheat streak mosaic virus (WSMV). With BYDV strain PAV, 11 of the 17 Agropyron species showed no virus transmission when plants were given access feed by viruliferous Rhopalosiphum padi. Similar trials with BYDV strain RMV (vectored by R. maidis) indicated that all plants, except susceptible control plants, remained virus free. Virus status was confirmed by enzyme-linked immunosorbent assays. When plants were mechanically inoculated with WSMV, 11 Agropyron species failed to express symptoms, while five other species showed a segregating response or had some accessions segregating and some resistant. Test results suggest that resistance to BYDV and WSMV in Agropyron species does not appear to be correlated with any specific genome of Agropyron species although most of the Agropyron species containing S genome were resistant to BYDV and WSMV.     

Barley yellow dwarf virus transmission and feeding behaviour of Rhopalosiphum padi on Hordeum bulbosum clones

A Hordeum bulbosum L. (Poaceae) clone A17 was identified, which showed complete resistance to Barley yellow dwarf virus (BYDV) and Cereal yellow dwarf virus (CYDV). It was not possible to infect plants of A17 with BYDV‐PAV, ‐MAV, or with CYDV‐RPV by the aphid vectors Rhopalosiphum padi (L.) or Sitobion avenae (Fabricius) (both Hemiptera: Aphididae). Plants of the A17 clone and of the BYDV‐susceptible H. bulbosum clone A21 revealed some resistance to R. padi compared to the susceptible winter barley cultivar Rubina [Hordeum vulgare L. (Poaceae)]. The development time to the imago was longer and the number of nymphs was reduced on both clones compared with cv. Rubina. The probing and feeding behaviour of R. padi on plants of the H. bulbosum clones was studied over 12 h and compared with that on plants of the barley cv. Rubina. Principal component analysis of the results of the feeding behaviour revealed a clear separation of the H. bulbosum genotypes from Rubina. On H. bulbosum the number of penetrations was higher but total feeding time was shorter. Significant differences were mainly found in the phloem feeding parameters for plants of both clones in comparison to Rubina, with the virus resistant A17 clone having the strongest effect and the susceptible A21 clone being intermediate. Most significant differences were found in parameters of the phloem salivation phase. On A17, an average of less than one (0.9) E1 phase per plant was observed (3.3 on A21 and 5.7 on Rubina) and its duration was reduced to less than 1 min (0.9 min) in comparison to 2.4 min on A21 and 5.7 min on Rubina. Also, the phloem feeding (E2) phase was clearly reduced on A17 plants with 0.5 E2 phases per test and a mean duration of 1.1 min in contrast with 2.9 and 3.5 E2 phases per test and 34.1 and 421.3 min for A21 and Rubina, respectively. These results point towards a phloem‐localized factor for aphid resistance in H. bulbosum, i.e., on A17 plants the phloem salivation time is too short for a successful infection by BYDV leading to vector resistance.     

Effects of sowing date and volunteers on the infectivity of soil infested with Gaeumannomyces graminis var. tritici and on take-all disease in successive crops of winter wheat

In a field experiment on winter wheat, take‐all on plants and the infectivity of the soil were studied in crop sequences with different combinations of sowing dates. Take‐all was negligible in the first wheat crop, but thereafter the mean disease intensity (measured using a take‐all rating, TAR, with a maximum of 300) was 108, 190, 118 and 251 in the second to fifth successive crops. In each growing season, the disease differed amongst sequences and built up more rapidly and was more intense on plants sown in mid‐September than on plants sown in mid‐October. In late‐sown plots, where volunteers had been present during the mid‐September to mid‐October period, take‐all reached an intensity intermediate between that in early‐sown plots and that in late‐sown plots that had been kept free of volunteers. Volunteers, therefore, partially offset the expected beneficial effect of decreased disease with later sowing. Differences in take‐all amongst sequences were most pronounced in the second wheat crop and early sowing of the previous wheat increased intensity of disease. In the following (third) crop, differences in disease intensity amongst sequences were smaller. Soil infectivity (measured by seedling bioassay after harvest) built up progressively from a low level after the first crop to peak after the third crop. In this build‐up phase, soil infectivity estimates were always numerically greater after harvest of early‐sown treatments than after later‐sown treatments, although never significant at P= 0.05. The greatest difference (P= 0.06) was recorded in October before sowing of the third crop, where the comparison was between soil after two previous early sowings and soil after two previous later sowings and control of volunteers. In the same autumn, presence of green cover (i.e. volunteers) was associated with a smaller loss of soil infectivity between harvest and later sowing than occurred in an absence of green cover. In 2^nd–4^th crops, where comparisons were available and mean TARs indicated moderate levels of take‐all, sowing later had no yield benefit, despite more take‐all and greater soil infectivity associated with early sowing. Important considerations for the management of crops at risk of take‐all are 1) choosing appropriate sowing dates to minimize take‐all or to encourage take‐all decline and 2) controlling volunteers and weed hosts where crops are sown late to minimise take‐all.     

Migration of alate morphs of the bird cherry aphid (Rhopalosiphum padi) and implications for the epidemiology of barley yellow dwarf virus

Suction trapping data indicate three periods of migration of Rhopalosiphum padi in spring, summer and autumn. Four alate morphs are present at different times during the year. A comparison of data from suction traps operating at 12·2 and 1·5 m suggests a different behaviour of females in autumn with more being recorded at 12·2 than 1·5 m. Males, which are only present in autumn, were also more numerous at 12·2 m. During tests to measure barley yellow dwarf virus (BYDV) infectivity, only 9% of female R. padi reproduced on oat seedlings in autumn compared with 74% in summer. Tests on alate female R. padi trapped alive showed that in summer all were exules, but during the first half of September these were largely replaced by gynoparae so that in autumn only 5% of all R. padi trapped at 12·2 m were alate exules. The aerial densities of gynoparae and males were 10 times greater at 12·2 than 1·5 m while densities of alate exules were similar at both heights. It is suggested that gynoparae and males fly higher to increase the chance of finding a taller dispersed host plant. The implications for BYDV epidemiology of the behaviour and presence of the various R. padi alate morphs indicate that autumn-sown cereals emerging before mid-September are particularly at risk from colonisation by alate exules before the transition to a mainly sexual migrant population is complete. Alate exules introduce BYDV from comparatively local sources. The ratio of total R. padi to Sitobion avenae in suction trap samples in autumn usually exceeds 100: 1, but on crops it was only 10: 1. The ratio of alate exule R. padi to S. avenae in suction traps in autumn was only 12: 1, similar to that observed on crops.