96 SCIENCE PROGRESS 



than the ordinary, in order that the results may be independent 

 of temperature ; and measurements of coefficients of thermal 

 expansion have also been carried out for the three principal 

 directions within the crystals. For the optical and thermal 

 work no less than 600 section plates, parallel-faced blocks, 

 and 6o° prisms have been cut or ground out of the crystals, 

 accurately orientated to 2' of arc as regards the theoretically 

 desirable positions within the crystals. 



Improvements in Experimental Methods 



The meagreness and inaccuracy of previous optical work 

 on the crystals of artificial chemical preparations was largely 

 due to the limitations imposed by the only available method 

 of preparing plates and prisms — namely, by holding the small 

 crystal between the finger and thumb and grinding the re- 

 quired surfaces on a plate of ground glass. It was necessary, 

 therefore, to devise an instrument for the absolutely accurate 

 preparation of such surfaces before any real progress could 

 be made. After having had several constructed, the final form 

 of the cutting and grinding goniometer is shown in fig. 1. It 

 has proved all that can be desired, after several years' hard 

 work with it, and without it the investigation would have been 

 impossible. 



The crystal is attached to an adjustable holder by means 

 of hard optician's wax, or is held in a grip holder, and then 

 suspended from a delicate apparatus, which serves not only 

 for the adjustment of a zone of the crystal's faces to the 

 vertical axis of the instrument, as determined by observation 

 through the telescope of the reflected images of the collimator 

 signal slit, but also, as the movements are graduated, for its 

 setting to any position with respect to the axis. Separate 

 and interchangeable cutting and grinding gear is provided, 

 and also a delicate means of varying the pressure of the crystal 

 on the grinding disc, so that the most fragile crystals can be 

 manipulated without danger of fracture. 



Another original piece of apparatus devised for these re- 

 searches is a spectroscopic monochromatic illuminator, which 

 is represented in fig. 2. This has not only enabled the optical 

 observations to be made for six different spectrum wave-lengths 

 of light in as short a time as was formerly occupied in taking 



