1886.] Light by Reflection from Iceland Spar. 



177 



of carbon, its reflecting power was so much, diminished tliat a more 

 intense source of light was necessary, and a magic lantern (a 

 " sciopticon ") was used, a black card with a slit 3'5 mm. wide 

 being placed in the slide-holder, and focussed on the surface of the 

 crystal, care being taken in both cases that the direction of the inci- 

 dent light should coincide as nearly as possible with the axis of the 

 collimator tube. 



The measurements were made by altering the angle of incidence 

 and the azimuth of the observing Nicol until the light was reduced to 

 a minimum, the position of the crystal remaining fixed. 



In order to obtain anything like accurate results with observations 

 of this kind it is necessary to make a large number of determinations 

 and take their mean: it was obvious that there were two ways in 

 which any given number of observations might be grouped, either by 

 making a good many separate determinations for a few positions of 

 the crystal, or by making a few observations at a number of different 

 azimuths ; the latter alternative being the one adopted, two readings 

 were made at seventy-two different azimuths of the crystal. 



In the first series the observations started from one of the edges 

 of the crystal, the tube containing it being turned through 10 J after 

 each pair of readings ; after thirty-six pairs of readings the crystal 

 was turned through 6° 20', and then thirty-six more double readings 

 made at intervals of 10° from each other. 



In the second series the observations started from the principal 

 section, and were also made at intervals of 10° ; the crystal was turned 

 through 5° after thirty-six observations had been made, and then 

 thirty-six more were made, also at intervals of 10°. 



It was thought that by working in this way the results would be 

 more independent of each other, and therefore more trustworthy 

 than if the readings had been made continuously round the whole 

 circle. The position of the crystal was determined by placing a 

 square on the horizontal stage of the goniometer, and rotating the 

 tube carrying the crystal until the edge of the crystal appeared to 

 coincide with the vertical edge of the square, and noting the reading 

 of the divided ring attached to the tube; several such readings were 

 made, and the tube and crystal turned through 180° and several more 

 observations made, and the mean of these taken as the position in 

 w r hich one of the sides of the crystal was vertical (i.e., perpendicular 

 to the plane of incidence) ; the position of the principal section could 

 then be readily determined as it bisects the obtuse angle, and there- 

 fore, that angle being one of 101° 55', or nearly 102°, forms an angle 

 of about 51° with the adjacent edges. 



The position of the crystal in which the principal section was in 

 the plane of incidence and the obtuse summit nearest the observer 

 was considered the zero position ; when the principal section was in 



