294 Joty— On the Volume Change of Rocks and Minerals attending Fusion. 
mineral fragment hanging downwards, the flame is then directed upon the mineral 
till it is brought to the desired state of fusion. The bead, finally ready for 
experiment, appears perfectly smooth and spherical in form, is not ‘strung’ 
upon the wire, but is attached to it at its extreme point. 
An alternative method is to fuse the substance entirely by radiation. This 
may be effected most conveniently as follows. After the attachment of the 
fragment to the wire, this is hung down into a narrow platinum tube, closed at 
the lower end. The diameter of the tube is about three millimetres, and its length 
a couple of centimetres. The oxylydrogen flame is now directed to the outside 
of the tube till it is perceived on looking down into the tube that the globule has 
been formed. This method avoids the risk of chemical alteration due to the 
flame. It will not serve to free from bubbles the more infusible silicates, however. 
It serves in the case of basalt and other basic material. It is also to be observed 
that the colouration of the flame for the few minutes during which the bead is 
being prepared by the first 
method represents probably 
but a very minute loss of 
matter. The adjoiming figure 
(fig. 6) represents the opera- 
tion of fusing the bead by 
radiation. 
Before considering the ex- 
periments effected by this 
method of determining the thermal expansion of minute quantities of substances it 
is necessary to consider the nature and importance of the errors to which it is 
exposed. 
The most serious source of error in effecting the measurements will arise if the 
substance is exposed to any marked inequality of temperature. I refer now to 
the case of dealing with beads at such a temperature that they owe their spherical 
form mainly to their surface tension and not to their rigidity. If the areas on the 
sphere facing the open ends of the tube still retained a certain amount of rigidity 
of surface while the circular band approximating to the heated platinum was 
completely liquefied, expansion would doubtless be attended with a certain amount 
of distortion of form, and the increase of diameter observed on the image would 
probably be unduly great. A  smail cooler area at the exposed faces of the sphere 
will not give rise to appreciable error in this manner as this small area will “ float”’ 
out on the expanding bead without serious distortion of the spherical form. The 
Fic. 6. 
error arising in this way will in fact depend upon the degree in which uniformity 
of temperature is departed from. The following observation appears to show that 
this departure from uniformity must be very small. A bead of potassium car- 
