HOLMES] METHODS OF DECORATING 65 



have two classes of features oapalile of giving rise to ornament; ttiese 

 are constructional and functional. Ttiose of tiie former class are 

 represented by such features as the coil employed in building, and the 

 stitch, the plait, and the twist employed in textile fabrics. Those of 

 the latter are repi'esented by handles, legs, bands, perfoi'atious, etc. 

 Suggestions incidental to manufacture, such as finger markings, 

 imprints of implements, and markings of molds, are fruitful sources 

 of nonideographic decorations. 



In the primitive stages of the art simple nonideographic elements 

 seem to predominate, but it is difficult to draw a line sejjarating them 

 from the ideographic, for an idea may at any time become associated 

 with even the most elementary design. When, however, we encounter 

 delineative elements or subjects emjiloj'ed in ornamental offices, we 

 may reasonably assume that ideas were associated with them, that they 

 were symbolic. It is pi"etty generally conceded that life forms were 

 not employed in early art save when they had a peculiar siguilicauce 

 and applicability in the connection in which they wei'e used, and it is 

 pi'obable that the associated idea was often retained even though the 

 representation became so conventionalized and formal that the ordinarj- 

 observer would no longer recognize the semblance of nature. This 

 topic was examined in detail in a recent study of the art of ancient 

 Chiriqui,'' and is presented in equally definite form in the section of 

 this paper devoted to Gulf Coast ware. 



The range of imitative subjects employed in surface decoration is 

 not large. AVithin the whole area studied, no representation of a plant 

 has been found; birds and the human figure were rarel}- delineated, 

 and even quadrupeds, so generally employed in modeling, do not 

 appear with frequency in other forms of expression. Ceramic decoi'a- 

 tion is proliably late in taking up the graphic and ideographic art of a 

 people. This conservatism may be due to the fact that in early stages 

 the art is purely domestic, and such delineations would have little 

 appropriateness. It is probably not until the fictile products come to 

 take a prominent place in superstitious usages that- significant designs 

 are demanded and emploj'cd. 



Methods ok Decoratixi; 



The decoration of earthenware was accomplished in a numlter of 

 ways which are classified by fonii characters as relieved, flat, and 

 depressed. The processes employed are modeling with the fingers and 

 with tools, molding in baskets or other vessels having ornamented 

 surfaces, and stamping, paddling, impressing, puncturing, carving, 

 incising, polishing, and painting with such tools as were most conven- 

 ient. A brief review of the decorating processes has already been 

 given under the head Manufacture. 



rt Holmes, W. H., Ancient art of the province of Chiriqui, in Sixlli Annual Report of the Bureau of 

 Ethnology. Washington. 1888. 



20 ETH— U3 5 



