Anomalous results concerning the micrometeorological temperature field in the boundary layer over the ocean have been obtained in many recent experiments. These include lack of an inertial-convective subrange in temperature spectra, unusually large values for the scalar universal subrange constant, underestimation of the sensible heat flux by the bulk aerodynamic formula, gross imbalance of dissipation and production terms in the temperature variance budget equation, and dissimilarities of the temperature and humidity statistics and time traces. Empirically it has been observed that such results occur for unstable conditions when the temperature time series is characterized by a peculiar waveform, termed a “cold spike”, which has no counterpart in the humidity field and has not been observed over land. To explain these results, it is proposed that surfaces of the small temperature sensors (thermistors, thermocouples and resistance wires) commonly used in marine boundary layer experiments become contaminated with salt spray when used over the ocean. Under typical ocean conditions (relative humidity > 70%), the results of Twomey (1953) indicate that the spray will exist as saline drops on the probe surfaces. Water will evaporate from or condense on the saline drops as the humidity around the sensor decreases or increases, respectively. The latent heat of vaporization associated with the evaporation and condensation processes will cool and heat the sensor, and therefore generate erroneous temperature signals. Evidence is presented that most of the anomalous temperature results observed over the ocean, including “cold spikes”, may be due to the spray-induced humidity sensitivity of such temperature sensors.