354 The Flint Tools of North Devon. 



smaller implements from the valley of the Somme, although 

 none of the large hatchets, so abundant in France, have as yet 

 been discovered in North Devon. Some also are analogous in 

 form to the ruder weapons found by M. Lartet and the late 

 Mr. Christy, in the caves of the Dordogne, others are clearly 

 the cores from which the thin knife-like flakes have been 

 struck. The chief number, however, are chippings, and the 

 larger fragments, although they show traces of design on the 

 part of the maker, have been rejected as utter failures to pro- 

 duce anything serviceable, either for domestic use or for pur- 

 poses of war. Some few are mere pieces of flint knocked off, 

 apparently, in the very earliest stage of the manufacture, 

 showing on one side the conchoidal fracture, and on the other 

 the original external coating, or crust, of the nodule. 



After many excavations, numbering altogether ten or eleven, 

 since August, 1863, I lately determined to make one on a more 

 extensive scale, in order to ascertain exactly the relative propor- 

 tion which the arrow-heads and knives bear to the failures, and 

 therefore, on the 19th of July last, took advantage of the kind 

 offer of a gentleman staying in the neighbourhood to supply 

 me with two or three men, so as to make a thorough examina- 

 tion of the spot which appeared most fruitful in tools. A pit 

 was dug through the alluvium down to the drift. Its length 

 was 12 feet, and the width at the narrowest end 7 feet, while at 

 the south end it widened to 14 feet. All the earth was thrown 

 up for me to examine, and. the result of a careful search (I may 

 say sifting) was the discovery of 366 flints, counting fragments 

 of all sorts, shapes, and sizes, which, on my return home, I 

 classed under the following heads. The first column of figures 

 gives the actual number of specimens of each type, and the 

 second the proportion that number bears to a thousand. The 

 latter will consequently be found the most convenient for pur- 

 poses of general reference, as from it the exact percentage can 

 so readily be obtained. 



Flint flakes shaped carefully by chipping, they 

 all approximate in form to one of three 

 types (see Plates), and are imperfect only 



as regards deficiency in point 72 196 



Ditto, badly shaped and imperfect .... 60 162 

 * Attempts of the rudest description . . .114 316 

 Cores, generally of large size, from off which 

 flakes have been struck . 53 144 



• Among the attempts and failures I havo counted many which were acci- 

 dentally broken by the workmen. 



