endophyte,but large numbers may build up in endophyte-free pastures in late summer in dryland conditions (C.G.Pennell,unpub.cited by Easton,1999).Serious infestation is associated with poor ryegrass growth and persistence,but cause and effect are difficult to establish. insect that is sensitive to endophyte (Dymock et al., Evid ence that ryegrass shw greater growth (Latch et al.,1985)or ith d fo estoera when 1993 cabinet and glasshouse-based e eriments of a few ks duration have not sho any en dophvte effects on stress tolerance (Barker et al.,1997:Eerens et al.,1998c:Hume et al.,1993 Easton Rolston,unpub.data cited by Easton 1999;Easton Hume,unpub.data cited by Easton 1999).Field experiments are difficult to protect from all invertebrate pressure factors,but a series of row trials run over 2-4 years,where no serious invertebrate pressure was observed,has indicated significant advantage to ryegrass naturally infected with endophyte or infected with certain strains (Hume Latch,unpub.data cited by Easton 1999). Other field experiments have indicated no such effects (Eerens et al., 1998a Widdup Ryan,1992). Likewise,results in France were inconsistent,but endophyte conferred some advantage to ryegra absence of apparent invertebrate attack.in ,1995 n endoph 1993 be less ed ards (F by lives swards clos ser to the g nd further into the leaf sheath horizon.especially after the first few days Similar differences have been docur mented in non-choice grazing experiment by R.H.Watson (cited by Easton,1999). If such differ es in a razino pressure were sustained over time,they could,in the absence of any other factors.lead to a more rapid decline in the endophyte-free sward. Whether due to pest pressure,drought tolerance or any other mechanism,endophyte- infected ryegrass pastures produce more dry matter than endophyte-free pastures. In a recent series of trials at six sites throughout NZ(Popay et al.,1999),endophyte-infected ryegrass consistently out-yielded endophyte- -tree ryegrass. Averaged over three years and the k sites,th difterence was more than 20%in summe and mo e than 0%in autumn.At some sites and in some years,the difference was much greater than this be a 20:Su d for cted pasture 1993).The effect is not alway rens et al..1998a).and is mostly an inv e effect of rven igour (Prestidge Direct inhibitor(allelopathic)effects on white clover have been documented (Sutherland et al..1999).but apparent stimulatory effects have also been noted (Eerens et al.1998d).Livestock may preferentially araze clover in a mixed sward(Milne et al.,1982).This preference is partial (Parsons et al. 1994),and more active avoidance of ryegrass if infected may intensify it.Alternatively,if clover is growing evenly through a grass sward,rather than in a mosaic,endophyte by discouraging close grazing by stock may protect the clover stolons.Poorer clover content in a ryegrass pasture will exacerbate any negative effects of endophyte on livestock performance,and these two factors have been confounded in some experiments PAGE6endophyte, but large numbers may build up in endophyte-free pastures in late summer in dryland conditions (C.G. Pennell, unpub. cited by Easton, 1999). Serious infestation is associated with poor ryegrass growth and persistence, but cause and effect are difficult to establish. Cutworm (Graphania mutans) is another insect that is sensitive to endophyte (Dymock et al., 1989b). It is found throughout NZ but is of unknown significance. Evidence that ryegrass shows greater growth (Latch et al., 1985) or stress tolerance when infected with endophyte, as has been widely reported for tall fescue (West & Gwinn, 1993), has proved to be inconsistent. Intensive growth cabinet and glasshouse-based experiments of a few weeks duration have not shown any endophyte effects on growth or stress tolerance (Barker et al., 1997; Eerens et al., 1998c; Hume et al., 1993; Easton & Rolston, unpub. data cited by Easton 1999; Easton & Hume, unpub. data cited by Easton, 1999). Field experiments are difficult to protect from all invertebrate pressure factors, but a series of row trials run over 2-4 years, where no serious invertebrate pressure was observed, has indicated significant advantage to ryegrass naturally infected with endophyte or infected with certain strains (Hume & Latch, unpub. data cited by Easton 1999). Other field experiments have indicated no such effects (Eerens et al., 1998a; Widdup & Ryan, 1992). Likewise, results in France were inconsistent, but endophyte conferred some advantage to ryegrass, in the absence of apparent invertebrate attack, in plots at more stressful sites (Ravel et al., 1995). Ryegrass infected with endophyte may be less severely grazed by livestock than uninfected swards (Edwards et al., 1993). Sheep offered a choice, grazed uninfected swards closer to the ground, and further into the leaf sheath horizon, especially after the first few days. Similar differences have been documented in a non-choice grazing experiment by R.H. Watson (cited by Easton, 1999). If such differences in grazing pressure were sustained over time, they could, in the absence of any other factors, lead to a more rapid decline in the endophyte-free sward. Whether due to pest pressure, drought tolerance or any other mechanism, endophyte￾infected ryegrass pastures produce more dry matter than endophyte-free pastures. In a recent series of trials at six sites throughout NZ (Popay et al., 1999), endophyte-infected ryegrass consistently out-yielded endophyte-free ryegrass. Averaged over three years and the six sites, the difference was more than 20% in summer, and more than 30% in autumn. At some sites and in some years, the difference was much greater than this. To the list of adverse effects can be added a trend for poorer clover growth in endophyte￾infected pasture (Stevens & Hickey, 1990; Sutherland & Hoglund, 1989; Valentine et al., 1993). The effect is not always observed (Eerens et al., 1998a), and is mostly an inverse effect of ryegrass vigour (Prestidge et al., 1992). Direct inhibitor (allelopathic) effects on white clover have been documented (Sutherland et al., 1999), but apparent stimulatory effects have also been noted (Eerens et al., 1998d). Livestock may preferentially graze clover in a mixed sward (Milne et al., 1982). This preference is partial (Parsons et al., 1994), and more active avoidance of ryegrass if infected may intensify it. Alternatively, if clover is growing evenly through a grass sward, rather than in a mosaic, endophyte by discouraging close grazing by stock may protect the clover stolons. Poorer clover content in a ryegrass pasture will exacerbate any negative effects of endophyte on livestock performance, and these two factors have been confounded in some experiments. PAGE 6