532 



SCIENCE 



LN. S. Vol. XXV. No. 640 



that Lambert's cosine law can be true only 

 for a black body, but no satisfactory ex- 

 periments have been made, >u) for as the 

 writer knows, to test the law empirically in 

 its application to glowing surfaces. 



These two laws, the inverse-square law 

 and the cosine law, are applicable to the 

 infinitesimal elements of a radiating sur- 

 face, and large errors may result if they 

 are assumed to apply to an extended source 

 as a whole. Particularly is this so in the 

 ease of the inverse-square law, which un- 

 derlies the great majority of photometric 

 measurements. 



In each of the above two cases the ex- 

 pression is deduced for the illumination at 

 different distances in a single plane or 

 along a single line symmetrical with re- 

 spect to the radiating surface. In the case 

 of a uniformly bright strip of infinite 

 length but of finite width it is not difiieult 

 to derive the expression which will give 

 the illumination at any point in space. 

 Instead of illumination, however, the more 

 general term specific luminous flux is sub- 

 stituted. At every point in space there is 

 some definite direction in which the flux of 

 luminous energy is a maximum. The 

 quantity of luminous energy which in one 

 second flows normally across a surface of 

 unit area placed perpendicular to the direc- 

 tion of maximum flux is defined as the 

 specific luminous flux at the point. It is a 

 vector quantity, and the component in any 

 direction equals numerically the difference 

 in illumination on the two sides of an in- 

 finitely thin material screen placed perpen- 

 dicular to the direction. 



Two examples are given of the errors 

 incident to assuming for a finite surface 

 the inverse square law, which only applies 

 to the elements of the surface; the above 

 case of an infinite strip is used to show the 

 errors incident to assuming for the strip as 

 a whole the cosine law, which is true only 

 for the elements of surface of the strip. 



The value of ^^ at different distances in 

 a direction normal to the strip at its middle 

 point, and in a direction making an angle 

 of 45° with the normal at the middle point, 

 are calculated. For any definite distance 

 the former multiplied by cos 45° would 

 equal the latter if the cosine law, which 

 has been assumed for the elements of sur- 

 face, applied to the surface as a whole. 

 The difference between the two values gives 

 the errors for the distance used, and by 

 plotting the errors for different distances 

 a curve of deviation from the cosine law is 

 obtained. 



On the Magnetic Properties of Eeusler's 



Alloys: J. C. McLennan. 



In this paper the author describes some 

 experiments made during the past year by 

 Messrs. Dawes, McTaggart and Robertson, 

 and Mr. L. B. Johnston, on the magneto- 

 striction and permeability of Heusler's 

 alloys of varying composition. 



From measurements on both phenomena 

 the alloys are shown to be in an exceedingly 

 unstable condition when freshly made, and 

 from observed changes in their magnetic 

 behavior the conclusion is drawn that pro- 

 found modifications are made in their 

 structure through the lapse of time as well 

 as by their being subjected to changes in 

 temperature and to repeated magnetiza- 

 tions and demagnetizations. 



In connection with the phenomenon of 

 magnetostriction the gradual shortening 

 observed by Austin with rods of the alloys 

 subjected to long-continued high magnetic 

 fields is shown to disappear when the rods 

 have reached a stable condition after being 

 repeatedly magnetized. 



In a series of rods containing the 

 same amount of manganese with varying 

 amounts of aluminium the greatest elonga- 

 tion was observed with rods in which the 

 manganese and aluminium were present in 

 the ratio of their atomic weights. 



