540 Prof. F. Rosetti's Experimental Researches 



very favourable for the experiments, we see, nevertheless, that 

 at an equal zenithal distance (i. e. with an equal thickness of 

 the atmospheric stratum for the rays to traverse) the observa- 

 tions of the morning are not altogether concordant with those 

 of the afternoon. Amongst the observations of 29th of Sep- 

 tember we find, for example, that the observations of 9 h. 35 m. 

 gave a deflection of 192 divisions, whilst that of 2 h. 5 m. only 

 gave 189 divisions; and in both cases the distance from the zenith 

 was equal to 56° 36'. In the same way, in the observations of 

 28th of September we obtain 191 divisions when the thickness 

 of the stratum of air traversed, at 9 in the morning, corre- 

 sponded to 60° 42 / distance from the zenith ; whilst at 2 h. 

 15 m. we had the same deflection of 191 divisions, notwith- 

 standing the decreased thickness of the atmospheric stratum. 

 In general I found that the observations of the morning gave 

 a somewhat smaller deflection than those of the afternoon, 

 although the hygrometer almost always indicated a less degree 

 of moisture in the afternoon than in the morning. But just this 

 coincidence gives the key to the explanation of the pheno- 

 menon. In the morning the lower strata of the atmosphere, 

 those which can influence the hygrometer, contain a large quan- 

 tity of vapour, while the upper strata contain a much smaller 

 quantity of it ; but as the sun rises, the vapour also rises, and 

 expands into the higher regions of the atmosphere. In the 

 morning the upper strata of the atmosphere are less saturated 

 with vapour, and consequently absorb less heat ; in the after- 

 noon, on the contrary, the absorption becomes greater on ac- 

 count of the preponderating action of the vapour expanded 

 into the higher strata, whilst at the same time the lower strata 

 are drier. 



On constructing curves with the data contained in the two 

 preceding Tables, taking for abscissae the values of e corre- 

 sponding to the distances from the zenith entered in the second 

 column, and for ordinates the values of the deflections of the 

 galvanometer, we obtain for each day two dissimilar curves. 

 I preferred to construct only the curve for the afternoon, as, 

 the points lying usually in a straight line, it could be drawn 

 easily and with great precision. In Table VIII. the values of 

 the coordinates obtained from the curves are given. 



