The cerebellum is a vital organ for postural control, equilibrium, and motor coordination; its increasing dysfunction as a result of normal aging can lead to life threatening accidents. The objective of this study was to identify and quantify all expressed cerebellar genes in the adolescent, adult, and aging mouse cerebellum using SAGE, a powerful tool for assessing the absolute expression levels of multiple genes simultaneously. The SAGE method is based on the isolation of short nucleotide sequence tags from a defined position within a transcript, followed by serial concatenation of tags. Sequencing of clones isolates individual tags, allowing identification and quantitation of cellular transcripts. SAGE libraries were constructed from P23, P92, P150, P810, and P840 C57BL/6 male mouse cerebella. A 'closest-neighbor' algorithm was used to differentiate low abundance tags from possible sequencing errors in each of the libraries. An online database, MmSAGEClass, was used to assign genes into functional categories. The percentage of genes within each functional class was virtually identical in each library. The most abundant genes identified in each library included ribosomal protein-L32, myelin basic protein, and tumor differentially-expressed-1. Comparison of young adult and aged cerebellar libraries revealed many genes that were differentially expressed, including growth hormone and prolactin (100% decrease in the aged cerebellum), and peroxisome-proliferator-activator receptor (100% increase). In addition, several tags differentially expressed in the aging cerebellum represented genes absent from the Unigene database, some of which are likely to represent novel genes. Our data show that comparison of transcriptomes at key stages can reveal candidate genes that may be involved in the aging process.