DR. E. C. S. MEGAW1 opens his valuable contribution to this subject with a challenging remark. He says, “knowledge of the average properties of turbulence in the free atmosphere is almost totally lacking”. Actually, so many meteorologists have published observations that it would be impossible here to give an adequate list of them. As long ago as 1926, I summarized the available facts in the statement that the diffusivity for neighbours is roughly proportional to the four-thirds power of their separation2. That was an empirical induction from observations: the accompanying theory was an aid to description, and did not use any dynamical hypothesis. The separations ranged from 108 cm. (Defant3) to 30 cm. (Richardson4). The theory afterwards started by Kolmogoroff to which Dr. Megaw refers is an independent and fascinating intellectual creation, based on energy and the theory of physical dimensions. Recently, Batchelor5 has related it to the aforesaid summary of observations. The behaviour of sea water is, strange to say, also relevant6. These objections to Dr. Megaw's opening paragraph do not in the least detract from the merit of the rest of his publication.