THIRTEENTH AND FOURTEENTH CENTURIES. 59 



I do not rely, as many antiquaries do, on the size and num- 

 ber of churches as positive evidence of abundant population. 

 The church in medieval time was a familiar place of resort. 

 It generally contained many sacella or chapels, in which mass 

 was constantly said for founders and benefactors. It had what 

 private houses had not generally, abundant warmth and light. 

 It was plainly often a refuge from violence, and occasionally 

 formed a store room for valuable produce. The chapel of 

 Merton College is a gigantic building when compared with the 

 number of fellows which ever were or could have been sup- 

 porte<i|2&y:he foundation. The same may be said of the 

 monajy marches, the size of which is wholly disproportionate 

 to the numbers which the domestic buildings could have shel- 

 tered, however closely packed the inhabitants of the cells may 

 have been. So in the ancient buildings of New College, one 

 side of the single quadrangle which was built by the founder, 

 is occupied by chapel and hall, part of another by offices and 

 library, and the rest of the building, only two stories high, was 

 inhabited by a warden, seventy fellows, ten chaplains, besides 

 choristers and servants. Private buildings were small and 

 inconvenient, while churches were large, and provided, as far 

 as the state of industrial knowledge allowed, with every con- 

 venience possible. And, as we well know, these churches 

 were treated with a strangely mixed feeling of reverence 

 and familiarity. Business was transacted in them, mys- 

 teries were played, and this perhaps, when in some other 

 part of the structure the most solemn offices were being 

 performed. The church, in short, was the common hall of 

 the parish, and was as much needed for secular as for sacred 

 purposes. 



I should be disposed to place absolutely no reliance on the 

 numbers of the population contained in contemporary historians 

 or chroniclers. In the absence of any exact statistical infor- 

 mation, calculations as to the numbers of people invariably err 

 largely on the side of exaggeration. In the case of alarm the 

 exaggeration is still more gross. The most striking instance 



