As posted to Stack Overflow, please try this out: plot(1:10,t='n') text(1,1,c(intToUtf8(c(1,9684))),cex=3) text(2,2,c(intToUtf8(c(9684))),cex=3) text(3,3,c(intToUtf8(c(9679,9684))),cex=3) text(4,4,c(intToUtf8(c(9679))),cex=3) text(5,5,c(intToUtf8(c(9685,9684))),cex=3) text(6,6,c(intToUtf8(c(9682,9685))),cex=3) text(7,7,c(intToUtf8(c(9682,9675))),cex=3) text(8,8,c(intToUtf8(c(9682,9682))),cex=3) In theory, those 9600-series values should be interpreted as "Harvey Balls" symbols. In practice, what gets plotted seems to depend on the order of the numbers in the argument to intToUtf8, and varies significantly from computer to computer (and from OS to OS). Any ideas? (Note, there's an alternative solution posted to SO http://stackoverflow.com/questions/22225086/harvey-balls-in-r?lq=1 which uses the extrafont package to drive points(x,y, pch=[stuff] ) , but I'm interested in why the code above is so strange). thanks Carl -- Sent from a parallel universe almost, but not entirely, nothing at all like this one.