Adjusting the Collimator of a Spectroscope. 97 



until the line in question, either dark or bright, is distinctly 

 seen. 



2. Move the prism into position B, and focus the collimator 

 until the same line is distinctly seen. 



3. Repeat the operation, always focusing the telescope 

 when the prism is in position A, and the collimator when the 

 prism is in position B. After three or four trials no change 

 of focus is required ; both collimator and telescope will then 

 be adjusted for parallel rays. I find that it is by no means 

 necessary to work much out of the position of minimum de- 

 viation in order to gain a delicate adjustment. If the adjust- 

 ment is made in the centre of the field, then I usually put the 

 telescope into such a position that the line, when the prism is 

 placed at maximum deviation, should just be out of the field 

 of view ; this gives quite a sufficient change of focus if the 

 rays are not parallel on entering the prism. 



The following measurements, which were purposely made 

 without special care, will show the accuracy of which the me- 

 thod is capable. The sliding tube of the collimator was divided 

 into millimetres. Two different adjustments for the sodium- 

 line, made in the way described above, gave the readings 5*0 

 and 4*0. The prism was now turned round so as to deflect 

 the ray to the other side. Two adjustments now gave 4*1 and 

 5*0. The mean of the four readings is 4*5. The adjustment 

 was then made according to the well-known method of first 

 focusing the telescope on a distant object and focusing the col- 

 limator to the telescope afterwards : the reading was 4*2. As 

 the focal length of the collimator was 300 millimetres, the two 

 results differ only by a thousandth part of the focal length. 

 Whether this difference is due to errors of observation, or 

 whether it is produced by a difference in the focus of the yel- 

 low rays and the mean visible rays, I cannot say ; but I believe, 

 with a little precaution, the method can be adapted to the 

 study of the achromatism of a lens. 



I have assumed that the faces of the prism are perfectly 

 plane. Practically it is difficult to get a prism in which this 

 condition is accurately fulfilled ; and it may be questioned 

 whether the curvature of the prism may not seriously interfere 

 with the accuracy of the method. To this I reply : — 



1. That a prism which is known to be good may always be 

 set aside to do this work. 



2. That the reason of having the rays strictly parallel on 

 entering the prism is based on the supposition that the faces 

 of the prism are plane. It is by no means evident that parallel 

 rays will give the best definition when the faces of the prism 

 are curved. 



