THE CELTIC SYSTEM 203 



devoted to pasturage, closes, more or less irregular in shape, 

 would appear. In the Donegal townland the pasture was " held 

 in lots in common." The map shows that the plots of pasture 

 in the occupation of a tenant were about as much severed one 

 from another as were the strips. Yet there was less reason for 

 their being so and remaining so, since pasture is not so diverse 

 in its quahties as arable and there was no question of common 

 ploughing. One can, therefore, imagine co-heirs subdividing a 

 pasturage township on broader Knes than they would have 

 thought applicable to one largely arable. For these reasons it is 

 not improbable that such a township sometimes broke apart into 

 closes which may have been to some extent consolidated. 



Probably this is what happened at times in Wales. There in 

 the sixteenth century township after township consisted of closes,^ 

 those of a holding being frequently contiguous. The principality 

 seems at that time to have been much more of a stranger to run- 

 rig than was Ireland or Scotland, a circumstance best explained 

 by the Tudor prohibition of transmission by gavelkind and by 

 the hypothesis of an early predominance of pasture. In Scot- 

 land, as we know, runrig prevailed in the first half of the eighteenth 

 century, and the situation in Ireland was without doubt similar. 

 The reason must have been that arable was, or had been, rela- 

 tively more extensive in these countries than in Wales. If this 

 supposition be correct, the different aspects assumed by the fields 

 of Celtic countries are only natural developments of a flexible field 

 system. 



We are left, accordingly, with four distinctive characteristics of 

 the Celtic field system. In the first place, the open arable fields 

 were small, a necessary corollary of the small size of the town- 

 ships; in the second place, they frequently consisted of the 

 intermixed strips of several tenants, but this intermixture was 

 variable, originating with and depending upon the extent to 

 which subdivision among co-heirs or co-tenants had proceeded; 

 in the third place, the rotation of crops, so far as we know it, was 

 not winter corn, spring corn, fallow, but something quite different, 

 viz., a succession of spring crops followed by several fallow years, 



* Archaeologia Camhrensis, Supplement of Original Documents, vol. i passim. 



