Thermometric observations over a row crop (cotton, Gossypium hirsutum L.) with 20 per cent cover using hand-held radiometers were made during several clear days near Maricopa, a town in south central Arizona. A ground sampling routine developed for estimating a composite temperature for the cotton field compared favourably with surface temperatures taken for one day from an aircraft flying at an altitude of approximately 150 m. Both ground and airborne radiometric measurements were from nadir angle traversing the row crop. It was found that by considering an equation for the radiative balance at the surface, a composite surface temperature could be calculated which was within ± 1.5°C of the value given by the labour-intensive ground sampling method or the aircraft. The data requirements were temperatures of sunlit and shaded soil and the sunlit canopy temperature and their respective fractional areas. The fractional areas of the sunlit and shaded soil as a function of time and the vegetative cover were estimated with extensive agronomic measurements and nadir photographs. A model similar to that of Kimes (1983) requiring only information on row orientation and the height and width of the vegetation produced estimates of the fractional areas close to the labour-intensive ground measurements. Therefore, it is possible to obtain estimates of the composite surface temperature comparable to low-flying aircraft observations over row crops with sparse cover without having to devise a labour-intensive ground-sampling routine and extensive agronomic measurements and photographs to determine fractional areas of the composite scene. However, there is some uncertainty as to the number of thermometric observations needed to obtain reliable estimates of sunlit and shaded soil and sunlit canopy temperatures.