THE CELTIC SYSTEM 1 87 



they restrayned sowing of wynter corne but as nowe sythence the 

 use of gavelkinde is abohshed for these threescore yeares past [by 

 statute of 34 and 35 Henry VIII c. 26, sees. 36, 64] in many partes 

 the grounde is brought together by purchase & exchanges and 

 headging & enclosures much encreased, and now they fall to the 

 tillinge of this wynter corne in greater aboundance then before." ^ 



From this it is clear that parcels of holdings in the Welsh parts 

 of Pembrokeshire had once been intermixed and unenclosed but 

 that the abolition of transmission by gavelkind had encouraged 

 consolidation and enclosure. The reference to gavelkind sug- 

 gests that it was a determining principle in the Welsh field system, 

 and at once calls to mind the part played by co-tenants in Scottish 

 agrarian arrangements. Before following out these suggestions, 

 however, we shall profit by attending for a little to Irish condi- 

 tions; and we shall naturally inquire first whether Irish units of 

 settlement were, like those of Scotland and Wales, ^ of the hamlet 

 type surrounded by small arable fields. 



In Ireland the units of settlement are and long have been the 

 townlands, but in seventeenth-century surveys they assume 

 various names and are variously grouped into larger units. Since 

 many of these units were more or less artificial, subserving pur- 

 poses of rating or assessment, hke English hides or virgates, it is 

 always necessary to keep apart the actual from the artificial units. 

 The size and shape of an actual Irish townland of the nineteenth 

 century is illustrated by any section of the six-inch ordnance sur- 

 vey map ; and the areas of the eight towns which Seebohm has 

 reproduced from county Monaghan, and which range in size from 

 35 to 165 acres with an average of about 90 acres, are entirely 

 typical.^ 



Seebohm has gone farther, and identified these eight townlands 

 by means of their names with eight tales of a survey of 1607. 

 The tate was primarily a unit of rating, whereas the Latin term 

 for townland was villata. Sometimes, as in the instance cited 



1 George Owen, The Description of Penbrokshire, (ed. H. Owen, Cymmrodorion 

 Record Series, 3 pts., London, 1892-1906), i. 61. 



^ The places referred to in the survey of the lordship of Bromfield and Yale 

 were usually hamlets, the arable fields of which were inextensive. 



^ English Village Community, p. 224. 



