A PREHISTORIC ROULETTE WEDEL AXU TROWBRIDGE 585 



While we incline to the view that this interpretation of certain 

 characteristic decorative techniques in the Kansas City locale may be 

 applicable over a very much wider area in the Mississippi-Ohio region, 

 there is no attempt to assert here that all pottery showing indentations 

 or rocker marks was necessarily worked over with a tool of identical 

 type. Simple toothed stamps, both straight and curved, were used to 

 a limited extent at the Renner and Trowbridge sites for roughening 

 portions of the vessel surfaces, and compound stamp impressions occur 

 on sherds from neighboring sites. It is possible that still other types 

 of implements with curved edges such as the shells already mentioned 

 served occasionally to produce rocker impressions. Forms without 

 side handles, however, must have been much more awkward to use or 

 would at any rate seem to involve more tiring movements of the wrist 

 and hand. The roulette from the Trowbridge site, requiring little 

 more than a simple rolling motion of the fingers, operates with an ease 

 and effectiveness that would seem difficult or impossible to equal with 

 flat end-notched objects, sticks, gougelike forms, or shells unless they 

 were in some way provided with a handle. We concur in Holmes' 

 observation (1892b, p. 150) that- 

 Mounting upon a handle is essential to the free and proper use of this tool [i. e. 

 the curved-edge stamp]. The step from the use of the curved edge to the employ- 

 ment of a wheel is a slight one, although the advantage gained is very great. 

 Mounted upon a handle the notched wheel * * * j^^^y j-,g revolved at will 

 encircling the vessel or giving lines or filling spaces of any length. 



It may be noted that sherds from the Renner and Trowbridge sites, 

 where ornamented with vertical rocker marks, characteristically 

 appear to have the convex side of the marks to the right as one views 

 the upright pot or potsherd. The Trowbridge roulette, however, 

 when held in the right hand, produces curves convex to the left. Does 

 this mean that the ancient potters here were left-handed? Or, if 

 right-handed, did they invert or lay the pots on the side to apply the 

 ornamentation? Or, again, did they lean over the upright pot and 

 work on the far side? 



It is unfortunate that the handle of this implement is broken oif. 

 Were the degree of taper manifested by the remaining stub continued 

 3 to 5 cm. farther, one would suspect that the tip could have been 

 used in making the punched bosses frequently found on local rim 

 sherds (Wedel, 1938, pi. 3, F-I). 



In our opinion, tliis specimen fully vindicates Holmes' theory that 

 a type of roulette was part of the material equipment of the native 

 potters who produced the "rouletted" wares of the upper Mississippi 

 Basin. It is immaterial that the technique often used was rocking 

 or partial rolling, since this can be done quite as effectively and 

 easily^perhaps more so— with a notched disk as with any other 

 primitive tool. As for the long indented lines sometimes found on 



