Monitored natural attenuation (MNA) is a preferred remedy for petroleum hydrocarbon contaminants in groundwater. It is expected that natural biodegradation will halt the expansion of plumes and, as source strength declines, will eventually cause their complete attenuation. This hypothesis is examined in a controlled field experiment at the Borden aquifer, CFB Borden, Canada. About 70 kg of coal tar creosote (CTC) was mixed at less than residual saturation with clean coarse sand. This mixture was installed below the water table. Chemicals were leached from the CTC and the lateral groundwater flow created chemical plumes that have been monitored for about nine years. Phenol essentially vanished by day 626; m-xylene was receding and plume mass declining by 626 days. Naphthalene plume mass increased for 1357 days, but appears to have declined thereafter. Dibenzofuran formed an apparently steady plume over 626, 1008 and 1357 days, but dibenzofuran mass in the plume had increased from 22 g to 65 g by days 2900 and 3300, due mainly to an unanticipated increase in the input of dibenzofuran from the source. The usual anticipation of plumes growing, becoming steady and then declining has not been demonstrated in this study. Vigilant monitoring when implementing MNA may be essential if receptors are nearby.