The rates of movement of 11 families of transposable elements of Drosophila melanogaster were studied by means of in situ hybridization of probes to polytene chromosomes of larvae from a long-term mutation accumulation experiment. Replicate mutation-accumulation lines carrying second chromosomes derived from a single common ancestral chromosome were maintained by backcrosses of single males heterozygous for a balancer chromosome and a wild-type chromosome, and were scored after 116 generations. Twenty-seven transpositions and 1 excision were detected using homozygous viable and fertile second chromosomes, for a total of 235056 potential sources of transposition events and a potential 252880 excision events. The overall transposition rate per element per generation was 1·15×10−4 and the excision rate was 3·95×10−6. The single excision (of a roo element) was due to recombination between the element's long terminal repeats. A survey of the five most active elements among nine homozygous lethal lines revealed no significant difference in the estimates of transposition and excision rates from those from viable lines. The excess of transposition over excision events is in agreement with the results of other in situ hybridization experiments, and supports the conclusion that replicative increase in transposable element copy number is opposed by selection. These conclusions are compared with those from other studies, and with the conclusions from population surveys of element frequencies.