120 



MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Table I.t 



t This is Table IV of the article cited, altered to give percentages and distances from the limb. 



The first column, under rf, gives tbe angular distance from the limb of the sun ; the next, 

 under r", gives the corresponding percentages of polarization for an atmosphere of unvarying- 

 density, while the other columns give the values for different laws of atmospheric density ; e. (/., the 

 fifth column contains the results of calculation under the assumption that the density decreases as the 

 the sixth power of the distance from the sun's center. It will be seen that it makes little difference 

 what laws of distribution be assumed for the neighborhood of the limb ; the polarization is there 

 feeble, not far from 12 per cent., and increases continuously outwaixls. Since we have seen, how- 

 ever, that not more than one-fourth or one-fifth of the light of the lower corona cau be diffused 

 sunlight, even eight times the polarization theoretically possible would not yield the percentage 

 observed. 



Again, the polarization must, if due to reflection by an atmosphere, increase continuously with 

 increasing distance from the sun. That it does not do so is established beyond question. The 

 only inuiginable explanation for this is the admixture of increasing quantities of nonpolarized 

 light at higher altitudes, which, as we have just seen, cannot be admitted near the sun. But if 

 "■we decline to accept any values for polarization other than those yielded by direct measurements, 

 we have enough in Professor Weight's observations of 1878 to demonstrate that the assumed 

 explanation is untenable, notwithstanding that his measures did not come nearer than C from 

 the moon. 



It is easy to calculate from the last table what must be the relative quantity of mixed light 

 from other sources, in order to i educe the theoretical numbers to correspond with his observations. 

 If we represent this quantity for any point in the corona by A, the quantity of diffused light for 

 the same poiut being unity, then the values for A are given in the following table : 



Table II. 



Here Professor Weight's " trace " at 22' is assumed to be equal to 0.02. 



