OBSERVATIONS AND EXPERIMENTS ON JOINT PLANES 165 
Instead of beeswax and resin, mixtures of paraffin and resin were 
used in the experiments. Several precautions were necessary in 
order to obtain good results. Paraffin alone faulted with a slicken- 
sided surface along a plane making an angle of 45° with the pres- 
sure, but no fine cracks appeared. The addition of a little resin, 
however, brought the desired cracks. The amount of resin used 
varied from just enough to stain the paraffin yellow to enough to 
give it a brown color. No good results were obtained with the 
material above a freezing temperature and the best results were had 
at about o° F., or, rather, at the lowest temperatures available. 
Artificial cooling did not give such good effects as those obtained by 
allowing the material to harden by exposure to the air on the coldest 
winter days and subjecting it to pressure at the same temperature. 
The paraffin was cooled in tins and cut into convenient sizes while 
soft. It was found that the cracks came out better on the natural 
upper surface than on the smoother, glazed sides and bottom which 
had been in contact with the tin and better than on surfaces which 
had been planed smooth. This made it necessary to cool the 
paraffin carefully; otherwise, with its large contraction, the surface 
became badly wrinkled. It was also found best to cool the material 
rapidly and use it as soon as hard. The better results seemed 
to be connected with lack of uniformity in the material. Results 
were better on the less regular surfaces and on material so recently 
cooled that it probably was not equally hard throughout. Like- 
wise, since it was cooled by being placed on snow with the upper 
surface exposed to the air, the rates of cooling of the upper and lower 
surfaces were different, with a consequent variation in grain through 
the mass. Daubrée was careful to have his blocks regular in shape, 
with faces planed smooth, and probably his material was nearly 
homogeneous. ‘The experiments on paraffin showed that such care 
was more than wasted in preliminary experiments like these, where 
no attempt was made to use carefully regulated conditions for the 
purpose of obtaining mathematical results. 
In size the blocks used were from two or three to eight or ten 
centimeters in length and breadth and from less than a centimeter 
to three centimeters in depth. These dimensions were employed 
in all combinations of long and narrow to square, thick, or thin. 
