THE CELTIC SYSTEM 199 



enough horses or oxen for a plough team. Cooperative ploughing 

 must, in short, have been a custom complementary to the sub- 

 division of holdings among heirs. Further, in so far as the parcels 

 were arable and ploughed with a common plough they would 

 tend to be, not block-shaped, but long and narrow, for such was 

 the shape of the unit ploughed by the heavy plough. Pasture 

 subdivided among heirs might fall into parcels of any shape; 

 arable would in its nature separate into strips hke those described 

 by the Scottish reporters to the Board of Agriculture. 



The appearance of runrig can thus be explained as due to the 

 custom of subdividing arable land of different qualities among 

 co-heirs. This custom and its effects constitute the second of 

 the distinctions which differentiate the field system of Celtic coun- 

 tries from that of midland England. The first difference we have 

 found in the markedly smaller size of Celtic townships. It has 

 now become clear that the intermixed holdings of central England 

 had one history, those of Scotland, Ireland, and Wales another. 

 In the English midlands, virgates consisting of scattered strips had 

 been fully formed when they were first described in the thirteenth 

 century; after that they underwent little change through sub- 

 division, the integrity of the virgate almost never admitting of 

 fission into more than four parts. In Celtic countries, on the other 

 hand, subdivision of a townland or a township sometimes first 

 arose as late as the eighteenth century, and no Hmits were set to 

 the lengths to which it might go. The distinction is fundamental 

 for the comprehension of runrig, and explains the greater flexi- 

 bility of its open-field arrangements. 



In a general way, however, the furlongs of open arable field 

 cultivated in accordance with Celtic runrig presented an aspect 

 not very different from that of an English midland township. 

 We must therefore hasten to note two other distinctions between 

 midland and Celtic arrangements, those, namely, which resided 

 in the methods of tillage employed and in the grouping of the 

 parcels of the tenants' holdings. 



Relative to Welsh tillage the Denbigh survey of 1334 twice 

 mentions a three-course rotation of crops; but in both instances 

 the reference is to demesne lands and the usage was apparently of 



