ABBOT: SOLAR RADIATION MEASUREMENTS 233 



new empirical method of "solar constant" determination, based 

 upon data obtained by Langley's methods, but independent of 

 gradual changes of transparency of the atmosphere during ob- 

 servations. Hitherto it has required several hours of uniform 

 atmospheric transparency to enable us to make the proper ob- 

 servations. If the sky was growing clearer our result was too 

 high, and if more hazy, too low. The new method of "solar 

 constant" determination is based on the fact that the atmos- 

 pheric transparency varies in an opposite sense to the varia- 

 tions of the brightness of the sky. Increased haziness means 

 more reflecting surface to scatter the solar rays indirectly to the 

 earth. At the same time it means more obstructing surface to 

 cut off the direct solar beam. We have found accordingly that 

 from measurements of the brightness of the sky near the sun it 

 is possible to infer the atmosphere transmission coefficients at 

 all wave lengths. In our new process all the observations can 

 be made in 15 minutes, and the "solar constant" value can be 

 completely worked out in a couple of hours. The following great 

 advantages may be claimed for the new method: i. Great sav- 

 ing of labor. 2. Possibility of making several independent de- 

 terminations each day. 3. Greater accuracy because independ- 

 ent of the variability of the atmospheric transparency. 4. Avail- 

 ability on partly cloudy days. On the other hand the new 

 method is empirical, and must be frequently checked against the 

 old to make sure that no new atmospheric conditions have 

 arisen to invalidate it. 



In letters just received from Mr. Clayton, he states that his 

 most recent studies have but increased his enthusiasm for the 

 value of solar radiation observations in forecasting. He en- 

 closes curves showing a striking direct correspondence between 

 the temperature departure for Buenos Aires and other South 

 American cities in November and December of 1919, and the 

 slightly antedating solar radiation changes. The direct tem- 

 perature effects lag from two to three days behind the solar 

 fluctuations. 



We now come to the most interesting and puzzling feature of 

 these new discoveries. It is that the dependence between solar 



