Dr. C. A. Abbot, in Smithsonian Miscellaneous Collections, Vol. 107, No. 4, announces that “there is a regular period of 6·6456 days in solar variation, and that terrestrial temperatures respond with changes ranging from 2° to 20° F. in exactly the same average period of recurrence. While the sun's variation appears to be perfectly regular in phase, always recurring on the day predicted, the terrestrial responses come sometimes for a month or more in succession from 1 to 3 days early or late. This, which by mechanical analogy we might call backlash, is doubtless the circumstance which hitherto has prevented meteorologists from recognizing the nature of this large temperature variation.” It is to be feared that “meteorologists, physicists and astronomers” who, as Dr. Abbot remarks, “with few exceptions, have remained to this day skeptical as to whether the Smithsonian observations of the solar constant really discover solar variations from day to day”, will manifest the same attitude to this new announcement. It is greatly to be regretted that Dr. Abbot, who has done so much to promote our knowledge of the solar ‘constant’, should so persistently refrain from using standard statistical methods for the discussion of his data, and proper statistical criteria for testing the significance of his results. The use of the word ‘period’ in connexion with an interval given to five significant figures (or 0·0001 day), when its phase is liable to be from one to three days too early or late, will generally be regarded as inappropriate, and neither the temperatures nor the solar data appear to warrant this pretention to accuracy in the period. The amplitude of the variation in both sets of data changes irregularly from one ‘period’ to the next, the average range in the solar variation being 0·13 per cent of the solar constant. The discussion is commendably sensible in its claims as to the practical value of this work for meteorological prediction.