The results of a survey of 50 O stars in the J, H, K, L, and M bands are described. The observations are described, and the fitting of reddening relations to them is discussed. The zero points in the reddening relations are related to the intrinsic colors of the samples of normal stars, and these are compared with theoretical colors. The comparison reveals anomalies at L and M, which are discussed. The residuals for an individual star from the reddening relations are a measure of its infrared excesses, if indeed it has any. How the excesses can be fitted to a stellar wind model to derive a parameter that involves the rate of mass loss in the wind, the stellar radius and temperature, and wind terminal velocity is considered. The results for a number of individual stars are discussed and compared with other mass loss measurements, and it is concluded that no single model is capable of reconciling the IR, UV, radio, and H-alpha observations for the stars.