Consider the following two ways of stair steps plotting: plo <- function(ty = 2, steps = 200){ old.par <- par(mfrow = c(2, 1)) x <- seq(1, 100, length = steps) y <- seq(1, 10, length = steps) plot(x, y, type = "s", lty = ty, main = "Stair steps") x <- rep(x, each = 2)[-1] y <- rep(y, each = 2)[-2*steps] plot(x, y, type = "l", lty = ty, main = "Line steps") par(old.par) } This shows a problem with 'type = "s"' combined with "lty = "; the lines look solid no matter what lty is set to, if the steps are small enough. Is this intentional, and if so, why? This bit me when I tried to plot empirical cdfs with the 'type = "s"' method; some curves were impossible to distinguish. Btw, I checked 'plot.survfit', and it uses the second method, probably of this reason. G?ran --- G?ran Brostr?m tel: +46 90 786 5223 Department of Statistics fax: +46 90 786 6614 Ume? University http://www.stat.umu.se/egna/gb/ SE-90187 Ume?, Sweden e-mail: gb at stat.umu.se