"axis" won't add x-axis to boxplot - r

Question

I'm trying to make a boxplot with custom axis labels, but I can't seem to add an x-axis to the plot.
For example:
test <- data.frame(year=as.integer(rep(1963:2014, each=10)),
response=rnorm(520))
boxplot(response~year, data=test, ylim=c(-3,3), xlab="", ylab="",
range=0, xaxt="n", yaxt="n")
responselabs <- as.numeric(c(-3:3, by=1))
yearlabs <- as.integer(seq(1965,2015, by=5))
axis(2, at=responselabs, tck=0.03, las=1)
axis(1, at=yearlabs, tck=0.03)
returns the boxplot, but no x-axis labels:
Trying to hack it the other way by making an empty plot first, I can get the axes, but it won't add the boxplot:
plot(NA, ylim=c(-3, 3), xlim=c(1962, 2015), xaxt="n", yaxt="n", ylab="", xlab="")
axis(2, at=responselabs, tck=0.03, las=1)
axis(1, at=yearlabs, tck=0.03)
boxplot(response~year, data=test, ylim=c(-3,3), xlab="", ylab="",
range=0, xaxt="n", yaxt="n", add=T)
What's going on here?

Answer 1

I think what's happening is that boxplot converts year to a factor. We can get around this by using the labels argument in axis:
boxplot(response~year, data=test, ylim=c(-3,3), xlab="", ylab="",
range=0, xaxt="n", yaxt="n")
responselabs <- as.numeric(c(-3:3, by=1))
yearlabs <- as.integer(seq(1965,2015, by=5))
axis(2, at=responselabs, tck=0.03, las=1)
axis(1, at = yearlabs - 1962, labels = yearlabs)

Answer 2

Why not use ggplot2?
library(ggplot2)
p<-ggplot(test,aes(x=year,y=response,group=year))+
geom_boxplot()+
scale_x_continuous(breaks=round(seq(min(test$year),max(test$year),by=5),1))
If you want to be rounded to the nearest 5 then the code is fairly easy to adjust in the scale_x_continuous() argument.
p<-ggplot(test,aes(x=year,y=response,group=year))+
geom_boxplot()+
scale_x_continuous(breaks = round(seq(round(min(test$year)/5,0)*5,round(max(test$year)/5,0)*5, by = 5),1))

Answer 3

Like others have stated, you may have confused the values underlying the x-axis. Playing around with abline() reveals that abline(v=2010) does not show up, but something like abline(v=50) will show up.
Here's a plot you're after, I believe (with some extra modifications):
year_vals <- as.integer(rep(1963:2014, each=10)) # pulling out of next line for easy reference
test <- data.frame(year=year_vals,response=rnorm(520))
boxplot(response~year, data=test, ylim=c(-3,3), xlab="", ylab="", range=0, xaxt="n", yaxt="n")
responselabs <- as.numeric(c(-3:3, by=1))
yearlabs <- year_vals[year_vals%%5==0] # alternate, more general definition; note that 2015 wasn't in the original 'test' data.frame #as.integer(seq(1965,2015, by=5))
axis(2, at=responselabs, tck=0.03, las=1)
# axis(1, at=1:length(yearlabs), tck=0.03)
axis(1, at=which(unique(year_vals)%in%yearlabs), labels=unique(yearlabs), tck=0.03)