178 



tal so as to split the boule to favor the- correct cutting, but the crack 

 tends to turn around to the other side of the boule. I suspect that this 

 is due to the following fact : As was said above, the flattened face of 

 the boule is almost always parallel to the longer axis of the oval cross- 

 section of the boule. Now natural ruby tends to cleave parallel to the 

 basal planes, but only feebly so. The boules are under internal strain, 

 somewhat as Prince Rupert's drops are, and when tiie surface of the 

 Itoule is abraded even slightly, cracking follows aTid the strain is relieved. 



FiQurell B. iSuku 

 A/ieujedon top. 



Hy cracking along tlic Iciigdi of tlie boule at right angles to (lie plane of 

 the flattened surface a smaller cross-section is made than would he made 

 werc! Ilie ])onle to si>lil parallel lo the basal jilanes. The natural cleavage 

 is so feeble that I suspect the s]tlit, follows the otliei' (liieciicm .is a line 

 of least resistance. .\t any rat" the bonles split contrary to the natural 

 cleavage and (lisa(l\an(ageously as regards cutting to ])roduce th<' best 

 color, r.v slitting with the diamond saw alter the break b;is occurred 



