In 0.62.1 I found that code like t(mat/vec)/vec seems to drop the matrix attributes of the result. In this case the matrix is 1x1. Here is something tracked inside a debug-induced browser. Browse[1]> data.class(Var) [1] "matrix" Browse[1]> Var (Intercept) (Intercept) 31.662 Browse[1]> debug: stdDev <- sqrt(diag(Var)) Browse[1]> debug: names(stdDev) <- dimnames(Var)[[2]] Browse[1]> debug: value <- t(Var/stdDev)/stdDev Browse[1]> data.class(stdDev) [1] "numeric" Browse[1]> debug: attr(value, "stdDev") <- stdDev Browse[1]> data.class(value) [1] "numeric" (As you might guess this code takes a variance-covariance matrix and creates a correlation matrix with the standard deviations as an attribute.) -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._