NO. 5 



SOLAR VARIATION AND FORECASTING- — ABBOT 



13 



" To tell the truth, I have been examining your paper and I fear 

 you are mistaken." "How so?" I ask. "I have taken the mean 

 value of the heights of the object& as you measured them," he re- 

 plies, " and find that it is but 4 inches. And, therefore, according 

 to the theory of probability it is excessively unlikely that there are 

 tall objects in the field." *' But, Sir," I say, " there are flowering 

 shrubs there at least 6 or 8 feet high, and trees which look at least 

 30 or 40 feet high." " I regret to differ with you," he answers, 

 " but I am sure my averages are right, and so the mathematics are 

 against you." " But, really, Sir," I reply, " the Washington Monu- 

 ment is in that field. The fact that there are also 17 milhon blades 

 of grass there cannot shorten it any, though it brings down your 

 average to 4 inches." 



Fig. 9. — Bolograplis of the solar spectrum energy distribution. 



Similarly, as I see it. the small average scatter of our solar radia- 

 tion values of recent years about their mean does not preclude us 

 from admitting that some even of the larger deviations are really 

 of solar origin. 



5. This leads to our critic's most serious charge. It is suggested 

 that Harqua Hala and Montezuma have lost their characters as 

 independent solar observing stations owing to our methods of remov- 

 ing systematic errors from their data. 



]\Iy colleague, Mr. Fowle, and I have devoted a great deal of 

 thought and time to a conscientious efifort to free the observations 

 of both stations from all terrestrial errors. 



Referring to figure 9, our trouble is this: Owing to the greatness 

 of the water-vapor absorption bands in the infra-red spectrum whose 

 areas we have to determine, it is not possible to know just exactly 

 how to draw smooth curves over the bands in that region. We draw 

 as best we can, but we expect to find, and do find, when we examine 

 a large number of solar-constant values, that the results are not inde- 

 pendent of the quantity of water vapor prevailing. 



