RESULTS AND CONJECTURES 413 



feature which certain Cumberland terriers seem to reveal. In 

 general, however, the parcels in the larger township fields which 

 lay in runrig, especially those in Northumberland, doubtless 

 remained widely dispersed. 



Why, then, it may be asked, was the custom of allotting strips 

 in runrig incompatible with a three-field system ? The answer 

 is that it was not necessarily incompatible, since runrig might 

 under certain conditions develop into the system in question. 

 To understand what these conditions were we must turn to an- 

 other aspect of Celtic field arrangements still visible in eight- 

 eenth-century Scotland. This was the practice, appropriate to 

 primitive agriculture, of improving successively different parts of 

 the waste and allowing each part in turn to revert to fallow for 

 a series of years. Traces of such a custom are also perceptible in 

 documents from Cumberland, but are more apparent in others 

 from Northumberland, a fact that has led us to formulate the 

 hypothesis that the English border counties originally had the 

 same field system as Scotland but developed it differently. 



In both regions, we may surmise, field arrangements were 

 based upon runrig, a device that assigned to all tenants within 

 the township strips in any tract of waste brought under transient 

 tillage. As agriculture advanced, however, the two regions ex- 

 panded this system in different ways. Scottish husbandry turned 

 to an intensive tillage of the arable which lay nearest the home- 

 stead, the so-called " infield," and by the aid of manure took from 

 it an annual crop, the remaining " outfield " being treated in the 

 old manner. In the English border counties, on the other hand, 

 no permanent differentiation was made between infield and out- 

 field; but, as the demand for a greater return from the soil grew, 

 the period of fallow which had been allowed to the transiently 

 improved parcels of waste was shortened. Eventually, we may 

 suppose, it was reduced to an interval of one year in three, as it 

 appears in fourteenth-century Northumberland extents. When 

 this stage was reached, transition to a three-field system was 

 feasible, involving only such regrouping of the parcels of the 

 holdings as would render compact the area left fallow each year. 

 Since the advantage of a fallow field of this sort lay in its utility 



