Often we are interested in the joint behaviour of a number of random variables and the relationships between them. This chapter describes methods of dealing with these multivariate distributions. We first consider a motivating data set which illustrates why, in many applications, it is not enough to consider the random variables individually.