IN connexion with the note by L. W. Tilton on the relation between the sunspot number and refractivity of dry air1, it is of interest to examine whether such a relation may be deduced from the astronomical observations. For this purpose we have selected twenty independent determinations of the refraction constant () and reduced them to 0° C., 760 mm. pressure, 6 mm. vapour pressure (), = 45° and sea-level. The corresponding values of the index of air (n) as well as the relative sunspot numbers for the approximate mean epochs of the determinations of are shown in the Y and X axes of Fig. 1. The diagram does not show any marked dependence of index of air on sunspot number. Assuming y = ax + b, we have: a = + 0·10, b = 1·00029171, the correlation coefficient being r = + 0·15. All the above values of are within the limits of 60·10 and 60·22. The corresponding values of n are 1·0002915 and 1·0002921, the range of variation being only 6 × 107. The mean of n for = 0 mm. is 1·00029208, that given by Tilton being 1·00029237. These values agree well considering that this latter is referred to the D line while in the astronomical observations the bisections of a stellar spectrum are usually made between C and D lines.