THE ROMAN VILLA AT HEMSWORTH. 9 



angle of 45 degrees and meeting at the points. A very perfect 

 water bath, nearly 6ft. square by 30 inches deep, has a mosaic 

 floor of the axe pattern in black and white. An exactly similar 

 bath, but paved with tiles, was found at Hartlip, Kent, and is 

 figured in Wright's Celt, Roman, and Saxon. 



Some of the spaces of plain mortar or rammed chalk must 

 have had boarded floors above them, for traces were found of 

 sleeper walls to take timber joists. Wooden floors were probably 

 common in these houses. In the small villa at Clanville, Hants, 

 I noticed three coins lying in a perfectly straight line across a 

 mortar floor showing a brown film of decayed wood. The room 

 had sleeper walls, and the coins, no doubt, had fallen through 

 the space between two badly fitting boards. The large hoard of 

 Roman pewter vessels found by myself at Appleshaw, Hants, 

 was buried in a pit sunk through a hard mortar floor. Such a 

 disturbance would certainly have drawn the attention of raiders, 

 unless boards had been lifted and replaced over the spot. 



Vastly more valuable than the recovery of coloured pavements, 

 or objects however interesting or curious, are any clues that may 

 lead us to a clearer historical view of these houses and the 

 persons who built and lived in them. How much more should we 

 know if every Roman site in England, or even in one county, had 

 been examined" with the methodical pains of your great Dorset 

 example, General Pitt- Rivers ! The unavoidably partial explor- 

 ation of the Hemsworth building has not, so far, added much to 

 our knowledge. As to its individual history, we may think it 

 was built and stood during a longer or shorter possibly only a 

 short period of the first three-quarters of the fourth century. 

 The unusually few and poor coins range from Constantine the 

 Great, who succeeded in 306, to Gratian, who died in 383. A 

 solitary coin of Tetricus, 267-272, is much worn, and most of the 

 coins may well have been in circulation many years before the 

 house was built. It is, of course, possible that earlier or later 

 coins may have been or will be found on the site. There are 

 some slight tokens that the house was comparatively new when 

 destroyed. None of the pavements look worn by treading, and 



