Comment 2 for bug 137035

This is still reproducible in Hardy. I am not going to upgrade to Intrepid just yet but it should be easy to check for anyone by pasting and copying the following in a console (you r-base and littler, perhaps more R packages)

r << EOF
 ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
 trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
 group <- gl(2,10,20, labels=c("Ctl","Trt"))
 weight <- c(ctl, trt)
 lm.D9 <- lm(weight ~ group)
 postscript("anova-h.ps", height = 4.5, width = 6,
             horizontal=T, paper="special", pointsize=11,
           family="NimbusSan")
 plot(lm.D9, las = 1) # Residuals, Fitted, ...
 dev.off()
 postscript("anova.ps", height = 4.5, width = 6,
             horizontal=F, paper="special", pointsize=11,
           family="NimbusSan")
plot(lm.D9, las = 1) # Residuals, Fitted, ...
dev.off()
EOF
grep -F "%%Orientation: Landscape" anova-h.ps