APPENDIX TO MEMOIR OF PELTIER. 191 



more perfect is the polarization and tlie more intense the blue color of the ordinary- 

 ray. For a pile of eight glasses, the polarization may be considered as perfect 

 when the ray of light reaches it under an incidence of 10 degrees; it is at that 

 point, therefore, that we should have the most intense blue, and it is evidently 

 this angle which might serve as a point of departure, were it not for the circum- 

 stance which I am about to mention. The pile of glasses transmits, in effect, 

 only a part more or less considerable of the incident light, and reflects the rest. 

 Now, when the incidence of the ray is too oblique, the quantity of reflected 

 light is so much augmented that more is lost in vivacity of color by the reflec- 

 tion, than is gained by the perfection of the polarization. There is, therefore, 

 an angle at which the polarized ray gives a maximum of the image ; that point 

 passed, the ray still gains in polarization, but loses considerably in brightness. 

 It is thought by most authors that this maximum is obtained when the pile of 

 glasses makes with a perpendicular to the ray an angle of 55°; in other words, 

 when the ray reaches the pile under an incidence of 35°* It is in fact under 

 this angle that we obtain the maximum of absolute, but not of relative polari- 

 zation. 



This is the angle also that Peltier has taken as being that which gives the 

 maximum blue. I confess, however, that it has appeared to me that we should 

 still gain by continuing to incline the pile. We lose, it is true, a little in light, 

 but to me it has seemed that the blue tint became more pronounced. I think 

 that the angle which gives the maximum of coloration is rather between 25° and 

 30° than at 35°; it may be, however, that this would vary according to the indi- 

 vidual. 



Optical principles of polarimetry . — We now pass to polarimetry. In researches 

 on this subject, the observer is always supposed at the centre of a sphere of 

 which the sun is one of the poles and the anti-sun the othei*. This sphere has 

 its meridian and its equator, endowed with the properties which characterize 

 those great circles. 



We will suppose, then, the axis of the objective tube to be in the plane of 

 the meridian, the pile also, and moreover rectangular with the incident ray ; we 

 will suppose, in fine, that the index of the ocular points to the zero of the 

 graduated circle. If now, by means of the vertical joint, the objective tube be 

 carried successively to all points of the meridian of the optical sphere which Ave 

 are considering, the following is what we observe: the rays proceeding directly 

 from the sun and those little distant from them give no signs of polarization, 

 and consequently no coloration in the images, but in proportion as the angle of 

 the radius vector with the direct rays of the sun is enlarged, the signs of polar- 

 ization supervene and coloration makes its appearance. The extraordinary 

 image takes the blue color, and the ordinary image assumes the orange-yellow 

 complementary tint. 



Ttie intensity of the tints increases up to about 90°, that is to say, to about 

 the point of intersection of the meridian and equator; thence it decreases till 

 about 150°. This number attained, we find the neutral point for whose discov- 

 ery we are indebted to M. Arago. IBeyond this, polarization is again reproduced, 

 but in an opposite direction; that is, the plane of polarization of these new 

 polarized rays is perpendicular to the plane of polarization of the preceding; 

 consequently it is no longer the extraordinary image which is colored blue; it is 

 the ordinary image. 



This singular change in the plane of polarization of the reflected rays results 

 from the circumstance that that portion of the sky no longer reflects the rays 

 proceeding directly from the sun in so great quantity as the rays proceeding 

 irom the different illuminated points at the horizon. Consequently the neutral 

 of M. Arago evidently results from the union of equal rays polarized rectanga- 



* Peclet, Traite de Physique, § 1439, p. 447. 



