TERRA-COTTA ROOFINU-TILES. 7 



ther north in Europe it is usually to be seen on the older 

 buildings and is the tile most often seen depicted in mediae- 

 val paintings of places outside of Belgium and Holland. 



If this form of tile really represent the earliest type, 

 one might readily believe that its form was derived from 

 sections of bark which must have come early into use as 

 a roof-covering. In lapping the sections of bark from the 

 eaves to the ridge, the concave as well as smooth sur- 

 face, would be placed uppermost as form ing the best water 

 gutters. Other sections of bark, perhaps from smaller 

 trees, would have been used to cover the joints of the 

 larger pieces and these would have been placed with their 

 convex surfaces uppermost. Such surmises are quite 

 justifiable when one sees so many forms of pottery whose 

 shapes have been derived from natural objects, as shown 

 in the Pitt-Rivers collection in the Ethnological Museum 

 at Oxford. [Professor Tylor, its director, has brought 

 out in a striking manner similar relations in other depart- 

 ments of the collection.] In other museums, notably the 

 museums in Stockholm and Copenhagen, the change from 

 stone to bronze and iron shows successive derivations of 

 form from objects first made in a ruder material or from 

 natural objects. 



As the origin of roofing-tiles is probably not lost in a 

 very dim past, philology may throw some light on the sub- 

 ject. The material of which they are made is among the 

 most enduring of man's fabrications and the earliest form 

 must sometimes be found. 



The arrangement of feathers on a bird in shedding the 

 rain would have given a sufficient hint for the proper ar- 

 rangement of material on a sloping roof. From the rough 

 natural substances used in the prehistoric roof there came, 

 not only slabs of wood, flat pieces of stone, terra-cotta 

 tiles of many kinds, but worked marble tiles (620 B. c.) 



