570 NATURAL SCIENCE. oct.. 



refuse, so tliat no trace of them could be seen at the surface, but 

 upwards of go were found in each village, and in both cases they were 

 drained by deep trenches. The excavations showed that one of the 

 chief concerns of the inhabitants in those days, was to carry off the 

 heavy rain, of the prevalence of which there is historical and other 

 evidence. Thus a well 188 feet deep was re-excavated, and the 

 Roman bucket found at the bottom, but no water, suggesting that the 

 water-line or ordinary plane of saturation must have lain somewhat 

 higher in the hill in former days than it does at present. The pits 

 were probably used to contain refuse, and they were subsequently 

 used as places of burial. 



Among the animal remains, those of horses, oxen, and sheep 

 were of small size, the horse rarely exceeding the size of the Exmoor 

 pony, while the sheep were of a breed the like of which is only to be 

 found at present in the island of St. Kilda. The pig, but slightly 

 removed from the wild boar, was of large size, and the dog varied 

 from the size of a mastiff to that of a terrier. The people themselves 

 were of small stature. 



Of the many interesting remains there were a number of iron 

 styli, which showed that the people were able to read and write ; and 

 one decorated tablet of Kimeridge shale appeared to be of the kind 

 used for writing upon with the stylus, by means of a coating of wax 

 spread over the surface. 



During the past three years (1888-1891) General Pitt-Rivers has 

 turned his attention to the Bokerly Dyke and Wansdyke, and the 

 evidence upon which the date of the latter has been to some extent 

 determined, has been derived chiefly from the careful record of dis- 

 coveries made in the two Romano-British villages before described. 

 Both earthworks, at the places where excavated, are Roman or post- 

 Roman, and the theory of the Belgic age of the Wansdyke has been 

 completely overturned. 



Bokerly Dyke, the present boundary-line between Dorset and 

 Wilts, is an entrenchment of high relief, nearly four miles in length, 

 running in a north-west and south-east direction, across the old 

 Roman road, which runs from Sarum to Badbury. 



The well-known Wansdyke runs, or probably did run, from the 

 fenny country in the neighbourhood of the Severn at Portishead, by 

 Bath, passing to the north of Devizes into Savernake Forest, and on 

 to Chisbury Camp, where it turns and runs southward in the direction 

 of Andover. 



Full accounts of the excavations are given, and also of the Pottery, 

 Coins, and implements of Bronze, Iron, and Bone that have been 

 obtained. There are notes on the Human Remains from Woodyates 

 on the Bokerly Dyke, l)y Dr. J. G. Garson ; and notes, for comparison, 

 on human skulls from Ilunsbui-y Camp, Northampton, and from the 

 Roman Villa at Llantwit Major near Cardiff. Numerous plates serve 

 to illustrate the characters of these remains and of the other objects 



