That is easier to do with the ggplot package.
pVals <- c(1/2, 0.75, 0.25)
set.seed(1)
biases <- vector("numeric", length = 9 * length(pVals))
StdErr <- vector("numeric", length = 9 * length(pVals))
estimators <- vector("numeric", length = 1000)
for (j in seq_along(pVals)) {
for (n1 in seq(1,9)) {
indx <- n1 + (j-1) * 9
for (i in 1:1000) {
x <- rcauchy(10 * n1, 0)
estimators[i] = quantile(x,pVals[j]) - tan(pi * (pVals[j]-1/2))
}
biases[indx] <- mean(estimators)
StdErr[indx] <- sd(estimators)/sqrt(n1)
}
}
nVal <- seq(10, 90 , by = 10)
DF <- data.frame(n = rep(nVal, length(pVals)),
Bias = biases,
StandErr =StdErr,
P = factor(rep(pVals, each = 9)))
library(ggplot2)
#> Warning: package 'ggplot2' was built under R version 3.5.3
ggplot(DF, aes(x = n, y = Bias, color = P, group = P)) + geom_point() + geom_line()

Created on 2019-12-06 by the reprex package (v0.3.0.9000)