So I am doing some work on statistical analysis of DNA microarray data, and one thing I notice is that the sample size is often quite small. For example, in Alon etc all they only have 62 tissue samples. I was just wondering why there is often a low sample size?

Well, there is no good reason for wanting low samples. The reasons were price of preps (though they have dropped enormously since the 90s). Other considerations are availability of samples (especially of clinical samples), and, sadly, time. Many just make the minimum of samples to survive simple statistical analyses and publish it.