128 //. E. Merwim — Density of Minerals. 



accuracy of the refractometer be tested, for errors in the 

 adjustment of the cross-hairs and of the vernier are not compen- 

 sated as they are when both parts of the circle are used. The 

 testing may be done by observing the error of the instrument 

 for &>Na of quartz, 1*54:125. A drop of the concentrated solu- 

 tion may be used between the cell and the hemisphere. The 

 trouble from diffraction bands may be largely avoided by plac- 

 ing a triangle of thread between the cell and the hemisphere. 



Before finding the density of grains of a material the grains 

 should be examined under the microscope in a watch glass of 

 xylol or alcohol to make sure of freedom from foreign matter. 

 So little attention has been given to such an examination that 

 many of the recorded densities of minerals which have been 

 carefully analyzed are in error "5 per cent or more. Greater 

 error is apt to come from lack of care in selecting material 

 than from other causes. 



If only the approximate density of a substance is desired, it 

 may be found quickly by matching its density approximately 

 with that of the heavy solution, placing a drop of the solution 

 on the refractometer, covering with a glass slip (to prevent 

 evaporation) and determining the index of refraction. 



Geophysical Laboratory, 



Carnegie Institution of Washington, 

 Washington, D. C, Oct. 2, 1911. 



