34 BUREAU OF AMERICAN ETHNOLOGY [Bull. 193 



end of the pit near Mound III, the stratigraphic results may be dis- 

 torted by the presence of Mound III refuse. However, excavated as 

 it was in 25 cm. levels, this pit probably offers the most reliable strati- 

 graphic evidence from the site. 



Trench 9, a 5X2 meters cut on a north-south axis, was located in 

 the flat or ''courtyard" about 20 meters north of Mound II. The 

 upper 30 cms. of the pit consisted of humus with a few sherds underlain 

 immediately by sterile gray-white clay rock and granite nodules. 

 Level 1, 0-25 cms., the only level with cultural material from the pit, 

 yielded 1 milling stone and 11 sherds of Red-buff ware. No distinc- 

 tive features were noted. 



Trench 10, a 7X2 meters cut on a north-south axis in the crown of 

 the north ridge about 25 meters north of Trench 9, was also exca- 

 vated in 25 cm. levels and jdelded many sherds and at least three 

 burials. The trench was carried down to 1.25 meters in most of its 

 length, to 1.6 meters in the southern end and to 2.6 meters in the 

 northern end. Excavations were carried down to the basal clay rock 

 through an apparent midden accumulation consisting of layered bands 

 of dry refuse honeycombed with ant passages. Beneath this midden 

 a number of pits filled with loose dust penetrated into the clay rock; 

 two of these contained burials (Find Units 370 and 374). These pits 

 appear to have been made prior to the deposition of the concentrated 

 gray refuse stratum, since the latter was undisturbed. 



Besides the layermg and middenlike appearance of the upper fill 

 areas of Trench 10, additional evidence of the occupational nature 

 of the fill is suggested by the very high proportion of monochrome 

 to polychrome wares recovered from the uppermost meter of the 

 trench, the primary refuse strata. In Level 5 (1.00-1.25 m.) a large 

 lump of burned floor or waU plaster with a weU-smoothed surface on 

 one side was recovered. Of the three finds encountered in the trench, 

 only one contained polychrome pottery (Find 374) and this was 

 entirely of the Macaracas type, Cuipo and Higo varieties. If this 

 bm"ial was laid down prior to the layers above, as seems evident 

 from the field observations, and if the layers above are true occupation 

 midden rather than purposeful fill, the Macaracas type would be 

 placed earlier in time than the El Hatillo or Parita types which are 

 more heavily represented in the refuse layers. It should be noted 

 that in this case Macaracas sherds are represented in the upper four 

 levels by only 18 sherds. 



CHRONOLOGY OF TYPES AND VARIETIES 



The stratigraphic evidence at He-4 is scanty, and in many cases 

 untrustworthy, because some of the mounds with the greatest number 

 of sherds appear to have been completely unstratified and those 



