86 Caster ley Camp Excavations. 



baked on the outside of the vessel to a dull red colour. It is very 

 soft and sandy, and only occasionally mixed with grains of 

 pounded flint. Tt is imperfectly baked, and when wet very apt to 

 crumble to pieces, and even after it has been dried it breaks easily, 

 and can be cut with a knife. Both the inner and the outer surfaces 

 are often striated, the result apparently of tools used in smoothing 

 the surfaces before baking. A few pieces have been smoothly tooled 

 and bear a fair polish. Compared fragment with fragment, or taken 

 as a whole, this pit-pottery presents a striking contrast to the hard- 

 baked, wheel-turned wares of the two succeeding stages. The frag- 

 ments found seem all to have been parts of bowls or cooking pots, 

 with straight rims, rather straight sides, and flat bottoms, but the 

 pieces are mostly too small to determine the exact shapes of the 

 vessels. The two largest pieces are shown on PI. IV., Figs. 8 and 12. 1 



Scarcely any of the pit type of pottery was found except in the 

 pits themselves, and this perhaps suggests that there was no con- 

 siderable occupation of the site by these people. This soft badly 

 baked pottery, however, crushes into powder when trodden upon, 

 and a later occupation of the site would in this way destroy much 

 of it beyond recognition. 



(2) Between the pottery of the two later stages there is no such 

 distinct difference in quality, and it is only when the two are 

 compared as a whole that the difference becomes apparent. 



To demonstrate this, the pottery from all the lower strata of all j 

 the cuttings made through the ditches was arranged side by side ! 

 with that from the upper strata in the same ditches. Half the 

 depth was taken as the line of separation, because the deepest find I 

 of Samian ware was at just half the depth of ditch No. 1 ; it seems 

 that where, as in this case, an earlier merges without a break into 

 a later stage, some such arbitrary line must be drawn. 



As a result of this arrangement, it appeared at a glance that | 

 there was a very striking difference between the pottery from tliej 



1 Similar pottery, equally badly baked and roughly made, has been found 

 with other Early Iron Age remains in Wilts, as at a site near All Cannings 

 Cross {Wilts Arch. Mag., xxxvii., 526) and in pits on Wilsford Down, near, 

 Avebury, etc. 



