HOLMESJ PROCESSES OF FLAKING 59 



produced and the ratlier haphazard arrangeiueut of the percussion 

 points prechide the idea that any process capable of accurately adjusting 

 the point of contact between the tool used and the article shaped could 

 have been employed. At best such a method would certainly not be 

 readily applicable to a stone of the refractory nature of quartzite. 

 Though the manner of delivering the stroke seems sufficiently deter- 

 mined, the precise method of holding the stone shaped is left to con- 

 jecture. My own exj)eriments have been conducted on the assumption 

 that it was held in the hand. The account of tlaking processes given 

 in the following paragraphs is based on the belief that free-hand per- 

 cussion with hammers of stone or other hard and heavy material was 

 the exclusive or principal quarry-shop process. 



Eeferring to the series of graded rejects illustrated in plate xvii, we 

 observe that the process of manufacture and the steps of development 



Fig. 10— First step iu bowlder fi:ikiiig. 



were essentially as follows: Grasping a bowlder in either hand (sup- 

 jjosing bowlder hammers to have been used), the tirst movement was to 

 strike the edge of one against that of tlie other at the proper angle to 

 detach a Hake (figure 10). The second movement and the third were 

 similar, and so on until the circuit was completed. If no false stroke 

 was made and the stone had the right fracture, these few blows, occu- 

 pying but as many seconds, gave as a result a typical tnrtleback — a 

 bowlder with one side faceted by artificial fiaking, the other side, save 

 through accident, remaining smooth. If the removal of a single row 

 of flakes was not sufficient, the work was continued until the one side 

 was reduced to the proper degree of convexity, and the availability of 

 the stone for further elaboration was made apparent. A type iirofile 



