HOLMES] THE PINY BRANCH QUARRIES 45 



The refuse about the roots of the chestnut tree contaiued more tliau 

 the usual perceutage of partially shaped tools, aud several bushels of 

 these, showing rude leaf-shape outlines, were collected. A photograph 

 made shortly after beginning the excavation shows the inclosure of 

 worked stones in the base of the tree and their prevalence in the mass 

 of refuse (plate xiv). 



THE SEl OND TREXCII 



A second trench carried across the old quarry in tlie spring of 1S90 

 failedtofurnishfeaturesof especial interest aud added little to the fund 

 of information acquired from the trench made the previous year. It was 

 not expected, however, that this second excavation would expose exten- 

 sive deposits of refuse or well marked quarrying. The site was chosen 

 in a depression, or incipient gulch in the slope, where no marks of dis- 

 turbance could be detected, whereas the first trench was carried across 

 a convexity in the face of the hill, which convexity bore every indica- 

 tion of being the result of artificial disturbance and accumulation. Hav- 

 ing determined that surface appearances in the first case really indicated 

 the conditions beneath the surface, the second trench was made where 

 no indications of ai-tiflcial disturbance could be noted. This trench was 

 100 feet noi'th of the first. No well-defined shop sites were discovered, 

 and evidences of ancient quarrying were quite meager. Artificial refuse 

 was evenly distributed throughout the overplaced gravels to a depth 

 of about ;5 feet. These couditious would seem to indicate that the shal- 

 low depression in which the trench was dug had been filled from shops 

 and quarries at the right and left, or perhaps from random working at 

 higher points on the slope. 



Excavation was begun in the rivulet bank, here about (5 feet high. 

 The immediate bank was found to consist of a mass of refuse, well 

 filled with broken bowlders and rejects and chips which exhibited a 

 sort of rude bedding as if rearranged by the action of the rivulet 

 or as if deposited on its successi\e though very nairow flood plains. 

 Our trenching soon passed through these deposits. The gneiss which 

 formed the bed of the stream rose i-apidly beneath the loose mass 

 forming the bank, and at 10 feet from the stream approached within .3 

 feet of the surface. From the tenth to the tliirtieth foot the gneissic 

 surface followed the slope of the hill at a prettj' uniform depth of 3 

 feet; beyond this it passed horizontally beneath deposits of Potomac 

 bowlders. Overplaced gravels from the tenth foot to the end of our 

 trench contained but few artificial objects, and these did not occur at 

 a greater depth than about 3 feet. These gravels for the most part 

 were made up of a heterogeneous mixture of clay, sand, and pebbles, 

 with occasional bowlders. Near the bottom they consisted principally 

 of material derived directly from the disintegrating surface of the 

 Potomac bowlder beds. 



THE THIRD TRENCH 



The site for a third trench was chosen with the view of secur- 

 ing evidence ou two questions of especial interest. The first was the 



