There are three nearly ubiquitous genomic "imbalances" in prostate cancer cells: 1) loss of sequences from the short arm of chromosomes 8, 2) loss of sequences from the long arm of chromosome 13q, and 3) gain of sequences on the long arm of chromosome 8, particularly in advanced disease. Candidate tumor suppressor genes and oncogenes affected by this trio of consistent changes include the c-myc gene on chromosome 8q24, the RB gene at 13q14, and potentially multiple novel genes on the short arm of chromosome 8, with a gene located more proximally potentially involved in tumor initiation and a gene or genes located more distally involved in tumor progression. Loss of regions of chromosomes 2q, 5q, 6q, 7p and 7q, 9p, 10p and 10q, 16q, 17p and 17q, and 18q, and gain of regions of 1q, 2p, 3p and 3q, 7p and 7q, 11p, 17q, and Xq have also been detected in the range of 25-50% of tumors studied. Analysis of candidate tumor suppressor genes in these regions is still in its early stages. Similarly, potential oncogenes on a series of chromosomal arms which undergo frequent amplification remain essentially uncharacterized. The basic outline of the chromosomal aberrations in prostate cancer has been well established; the details of the story remain to be filled in. This paper reviews the advantages and disadvantages of various techniques for detection of genomic loss and gain in prostate cancer cells, and reviews published reports of loss and gain in prostate cancer, focusing primarily on reports using microsatellite analysis, Southern analysis, and comparative genomic hybridization. Fluorescence in situ hybridization (FISH) based analyses of selected regions are also reviewed.