OF AETS AND SCIENCES. 27 



polarization probably more than making uj? for the loss of light and 

 distortion induced by the increased obliquity of the incident rays. 



The want of perfect transparency of the glass would also tend to 

 increase the polarization by enfeebling the secondary reflection, and 

 dirt or grease on the surfoce of the glass would produce the same effect. 

 With eight surfaces, these disturbing causes are much less marked, 

 except for large angles of incidence, and hence the' agreement with 

 theory much better. Beyond 60°, however, it becomes perceptible, 

 producing the increased polarization noticed above. Even with a single 

 plate of glass, this disturbing cause becomes perceptible, which probably 

 accounts for the divergence for angles greater than 80°. 



