To understand the potential impact of exposure misclassification on water fluoridation studies in England, this paper aims to describe the long-term variation in water fluoride concentrations in both artificially and naturally fluoridated water supplies. Water fluoridation dose monitoring data were requested from all five English public water suppliers who artificially fluoridate their water, as well as from one water company that supplies naturally fluoridated water. Descriptive statistics were calculated, including annual means, standard deviations, minimum-maximum and absolute and relative frequencies. Data were made available by two of the five English water companies who supply artificially fluoridated water and one water company that supplies naturally fluoridated water (40 398 individual samples). The data for fluoridated water spanned 18-35 years, whilst the data on naturally fluoridated water spanned 14 years. The artificially fluoridated samples showed wide variation in fluoride dose control, both between different water treatment works and over time. Mean fluoride concentrations in the artificially fluoridated supplies ranged from 0.53 (SD 0.47) to 0.93 (SD 0.22) mg F/L and were within the optimal range of 0.7-1.0 mg F/L in 27.7%-77.8% of samples. The naturally fluoridated supplies had a higher mean fluoride concentration of 1.06 (SD 0.18) and 1.15 (SD 0.16) mg F/L than the artificially fluoridated supplies, with lower variation over time. The naturally fluoridated supplies were above the optimal range in 75.5% and 53% of samples. Assumptions that populations living in areas with a water fluoridation scheme have received optimally fluoridated water (0.7-1.0 mg F/L) are invalid. To support future research endeavours, as well as to provide 'external control' and facilitation of optimal dosing, it is recommended that a quarterly record of water fluoride concentrations (mean, standard deviation and minimum and maximum) are made available for every water supply in England, in a format that can be mapped against residential postcodes.