The existence of a coherent film of lubricant between highly loaded machine elements has been recognized for many years. Over this period of time measurements of film thickness have gone hand in hand with theoretical analyses in the field now known as elastohydrodynamic lubrication. The experimental techniques of capacitance, electrical resistance and X-ray measurement have been supplemented by the use of optical interferometry while the analytical expressions obtained with the use of elegant simplifications have been superseded by those developed from extensive and comprehensive computational procedures. These developments in experimental techniques have yielded a substantial number of measurements of both minimum and central film thickness. Likewise, the advent of the digital computer has allowed the derivation of a large number of solutions to the problem of elastohydrodynamic lubrication of concentrated contacts. All these results, covering a wide range of geometrical conditions, are to be found in the literature, yet little attempt appears to have been made to assemble a representative set of experimental data to permit a detailed evaluation of the theoretical formulae for elliptical contacts. The second part of this paper therefore considers the correlation between a number of experimental studies covering a wide range of operating conditions and geometries, and the predictions of recent elastohydrodynamic theory. Some of the important aspects of each set of experimental results are then considered and examples are provided which illustrate the following points: 1. Good estimates of lubricant film thickness may be obtained from the theoretical expressions recently derived, even when the dimensionless parameters involved are outside the ranges considered in the derivation of the formulae. 2. The discrepancies which exist between theoretical predictions and some of the measured film thicknesses are nevertheless quite large, even when the dimensionless parameters are within their usual limits. On the whole there is good agreement between experiment and theory, while the general trend of the results indicates that theoretical predictions may underestimate the minimum film thickness by about 10 per cent and the central film thickness by about 25 per cent. This measure of agreement is quite remarkable when the extreme difficulty of interpreting the magnitudes of effective and very thin mean film thicknesses between machined components in various forms of experimental equipment is considered.