IV NOTE 



the first place, the observations have been taken mainly on Mt. Wilson 

 during the season from May until November, and although continued 

 for many years, there have been in each year many gaps due to the 

 interference of cloudiness and other causes. Not only are there gaps, 

 but many of the days which are included in the record are days in 

 which observations have been made despite conditions which were not 

 perfectly favorable. It may have been that the sky was growing 

 clearer or more hazy, due to the approach or recession of cloudy con- 

 ditions, or it may have been that untoward conditions affected the 

 instrumental equipment. On these accounts, I think it probable that 

 at least one-third of the Mt. Wilson observations are Hkely to be one 

 or more per cent in error. If uniformly excellent solar values had 

 been available, I think it is fair to suppose that much higher correla- 

 tions would have been found by Mr. Clayton. 



His whole paper deserves careful attention, but in order to fix in a 

 striking manner in the reader's mind the strength of his case for a 

 real correlation between solar radiation and terrestrial temperature, 

 I would draw attention to tables i and 2 of Mr. Clayton's main paper 

 and to the little table in the Appendix. Part of the data in table 2, 

 changed to the Fahrenheit scale, forms the frontispiece. 



Having drawn attention to these, which are but samples of Mr. 

 Clayton's results, I now anticipate the question of the reader : Is it 

 not more probable that these apparent correlations between solar 

 changes and terrestrial changes are really altogether of a terrestrial 

 origin ? In other words, is it possible that the apparent variations of 

 radiation were not truly solar, but were caused by changes in the 

 transmissibility or other properties of the air which affected the solar 

 radiation measurements in one way and the temperature and rainfall 

 of the earth in another? In answer to this possible objection, I draw 

 attention to the fact that the correlations which Mr. Clayton has 

 found between the weather of Argentina and the results of solar 

 investigation have been based not only on the results obtained at the 

 near-by station of Calama, in Chile, but much more on the results 

 obtained at the very distant station of Mt. Wilson, in California. 



^In table 2 and the little table of the Appendix note the gradual change 

 which is indicated from the results of high maxima of radiation at the top of 

 the tables to low minima of radiation at the bottom. Table 2 is indeed particu- 

 larly striking. It shows that, in the mean, all large deviations of temperature 

 at Buenos Aires are definitely correlated with large deviations of solar radia- 

 tion, while no large deviations of temperature are correlated with mean values 

 of solar radiation. 



