An improvised drag-plate apparatus, on the principle of that used by Sheppard (1947) on a concrete surface, but suitably modified in design, has been used for exploratory measurements of the aerodynamic drag of grassland. The grass cover was variable (1 to 15 cm. in height), and measurements were made at a number of positions (not simultaneously) in order to obtain an approximate representative value of the drag over a considerable area. Wind velocities were in the region of 500 cm./sec. and, judged in terms of the Richardson number, effectively adiabatic conditions of flow prevailed. The drag (r
0
) and the simultaneous vertical distribution of wind velocity (u
z
) up to a height of 2 m. were found to be expressible in terms of the law well established in the laboratory, i.e. with
k
(von Kármán’s constant) = 0.37,
d
(the zero plane displacement) = 8 cm. and z
0
(the roughness parameter) = 0.66 cm. For reasons which are discussed the drag measurements are regarded as approximate, and the close agreement of the numerical value of
k
with the laboratory value of 0.4 is probably fortuitous. However, the general consistency achieved in this preliminary application suggests that the technique could be profitably developed for a critical investigation of the relation between drag and wind profile and its dependence on atmospheric stability. In a brief discussion of previous work some evidence is now provided for the validity of the Reynolds formulation of the turbulent shearing stress. Attention is drawn to the application of the present results in treatments of the diffusion of matter in the lower atmosphere.