394 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1960 



some domains to grow at the expense of others having a less favorable 

 orientation, and also in some cases the magnetization direction within 

 a domain was thought to rotate to a more favored direction. This 

 remarkable insight antedated the observation of domains by about 25 

 years. 



In addition to the magnetic rotation of the plane of polarization, 

 we can distinguish between some domains lying in the plane per- 

 pendicular to the line of sight by a magnetic birefringence. A thin 

 section of a crystal can be seen to be birefringont if, when it is placed 

 between crossed polarizers, the light is no longer extinguished except 

 at certain angular settings of the crystal. This happens because the 

 crystal transmits light polarized along one crystal direction at a 

 velocity different from that at which it transmits light polarized 

 along a perpendicular direction. If light entering the crystal is 

 plane polarized exactly along one of these axes, the emergent light is 

 plane polarized and can be extinguished by a properly set analyzer. 

 If the light entering the crystal is plane polarized but not along either 

 of these axes, the emergent light is elliptically polarized and cannot 

 be completely extinguished by an analyzer. 



In the present case it is found that domains whose magnetization 

 lies in the plane show a birefringence. This is called a magnetic 

 birefringence, since it is associated with the magnetization. If the 

 direction of polarization is parallel or perpendicular to the magneti- 

 zation, the emergent light can be extinguished by the analyzer. If 

 the plane of polarization is neither parallel nor perpendicular but at 

 some other angle to the magnetization, the light passing through can- 

 not to be quite extinguished by the analyzer. The effect is a very feeble 

 one and the distinction can be made only if domains with two head- 

 ings of the magnetization are present. Figure 5 illustrates the man- 

 ner in which a contrast between the two domains is achieved. Note 

 that light and dark can be interchanged by rotating the crystal 

 through an angle equal to that between the magnetization directions. 

 In the case of domains which we can distinguish by virtue of a dif- 

 ference in the rotation of the plane of polarization, light and dark 

 can be interchanged by moving the analyzer to the opposite side of the 

 original extinction position ; rotation of the crystal between polarizers 

 has no effect on the contrast. Plate 4 shows a domain structure seen 

 by virtue of the birefringence. 



DOMAIN STRUCTURE 



It is a general physical principle that a system assumes the lowest 

 free energy state accessible to it [12, 13, 14]. The magnetization dis- 

 tributions which we encounter here or in any other sample are the re- 

 sult of minimizing the total magnetic energy of the crystal. It is 



