Copies of Alu elements (so-called because the sequence contains an Alu restriction site) are littered throughout primate genomes. These elements appear to be pseudogenes of the (highly-constrained) 7SL RNA-encoding gene. The Alu elements move around the genome through an mRNA intermediate (a common process for retrotransposons and retroviruses) but do not code for the required enzymes (reverse transcriptase, transposase) themselves, ``borrowing" the enzymes from active transposons. It is not clear if all Alu copies retain the capability to be transcribed (and thus, possibly, transposed through the processed mRNA), or if there are a few ``master genes" that have retained mobility, but it is clear that new copies of Alu have appeared in the genome in the recent past. In fact, there are several presence/absence polymorphisms that are diagnostic for different human populations. It is easy to assign directionality to these differences (absence of an insertion is almost certainly the ancestral state), and thus Alu presence/absence polymorphisms may be good characters for building phylogenies of human populations. Alternatively, if we can identify inert Alu elements, sequence-level changes (most usefully, point mutations) between homologous elements should be selectively neutral. A lot of theory has been developed to describe selectively neutral evolution. One of the successes of this theory is that it predicts a very simple relationship between within-population processes (changes in allele frequencies) and between-population processes (divergence in allele composition).