THE nineteenth century discarded horoscopy of the weather with the other appurtenances of astrology. Dr. Z. Kamerling I would have us reverse the verdict and employ the motions of the planets as the basis of long-range forecasting. His thesis is that widespread periodicities must have a cosmic origin, and sunspots having failed, there remain only the planets. Accordingly he investigates periodicities of the length of the ‘synodic rotations,’ that is, the intervals between the dates at which the various planets are nearest to the earth. (It appears that the theoretical basis of this planetary connexion has been given in a previous paper; one wonders what it can possibly be.) These give him in years: Venus 1460, Mars 2-13, Jupiter 109, Saturn 1-035, Uranus 1018. As material he has forty years of average monthly rainfall over east Java and over west Java, twenty years' rainfall at Pernambuco, and temperatures at Winnipeg, K6nigsberg, and Zwanenburg. The method is to write down the monthly data in sets corresponding with the ‘synodic year,’ so that, for example, all the months of perigee come in the same vertical column, and to plot the smoothed means of these columns. From his graphs Dr. Kamerling concludes that there are real periodicities corresponding with the synodic periods of each planet, giving maxima generally near perigee and secondary maxima near apogee.