Spirally wound 12-V valve-regulated lead-acid batteries were subjected to conventional-charge and rapid-charge cycling tests. The cycle life was 250 cycles for the conventional-charge regime and 1000 cycles for the rapid-charge regime. In conventional-charge cycling, the positive active material quickly expanded and developed a coralloid structure in association with lowered utilisation and integrity. In rapid-charge cycling, no coralloid structure developed and the expansion was smaller and much slower. Correspondingly, the particle size of the negative active material grew in both cycling tests, but at a much slower rate in rapid-charge cycling. With the expansion of the positive active material, the negative active material was compressed. In the failed batteries, about one-third of the negative active material in the centre of the electrode was compressed almost into a solid non-porous mass. This densification process also occurred at a much slower rate in rapid-charge cycling. At the point of failure, the discharge capacity of all test batteries was limited by the negative electrode, although it was limited by the positive electrode at the beginning of the cycling tests. The cause of failure for most of the batteries, regardless of the charging regime, was the occurrence of "soak-through" shorts caused by numerous minute lead dendrites formed in the separator. This might have been encouraged by the formation of shorter distances between the two electrodes, created by the compression of the separator as a result of the expansion of the positive active material.