Transparent Bodies by the action of Magnetism. 487 



analyser, and since this circumstance might modify the sensibi- 

 lity of the eye, I did not take the mean of these two observations 

 as corresponding to the position of the plane of polarization. I 

 made at least four observations, oftener six or eight; these ob- 

 servations differed usually by ten minutes, but often the differ- 

 ences were as much as thirty minutes. 



The observation of the tint of passage has given me generally 

 more accuracy than the observation of the extinction of homo- 

 geneous light, and I have employed it more frequently. In fact, 

 although the tint of passage results from the extinction of the 

 most intense rays of the spectrum, i. e. of the mean yellow rays 

 whose plane of polarization is much less deflected than the plane 

 of polarization of the indigo rays, yet this smallness of deflection 

 is more than compensated by the exactness with which the eye 

 appreciates the variations of colour in the neighbourhood of the 

 tint of passage. 



Again, I had to employ solar light in this case, in consequence 

 of the conditions peculiar to the phenomena which I studied. 

 The rotation of the planes of polarization of the different colours 

 being always very small, their dispersion was very small also, 

 and consequently when the analyser had a position convenient 

 for completely extinguishing the mean yellow rays, it extinguished 

 also in a great measure the other rays of the spectrum, so that 

 the tint of passage was only produced by a very small part of 

 the incident light. If, then, this light had not been extremely 

 intense, it would have been impossible to perceive the tint of 

 passage, and the eye would have only perceived so feeble a mini- 

 mum of light that all coloration would have escaped detection. 

 It is superfluous to add, that it would have been necessary to 

 operate quite otherwise in order to measure large rotations, such 

 as those produced by great thicknesses of quartz or organic 

 liquids. In this case the use of solar light would have only 

 dazzled the eye and rendered all observation of the tint of pass- 

 age quite inexact. I always determined the azimuth of the tint 

 of passage four times ; twice setting out from the red, and twice 

 from the violet. There was generally never more than five 

 minutes' difference between these four observations. The mean 

 could therefore be regarded as correct to two or three minutes. 



As to the estimation of the magnetic action, it was only after 

 numerous unfruitful attempts that I found a satisfactory process. 

 The first idea was to cause a steel needle, strongly tempered, 

 and magnetized to saturation, to oscillate in the intermediate 

 space between the two armatures. The square of the number 

 of oscillations in a given time would have served as a measure 

 of the magnetic action. 



This process would have been sufficiently exact if the actions 



