HISTORICAL NARRATIVES AND FIELD WORK 31 



A trench wus dug from the outer edge of the midden nearest 

 the water's edge, toward the center of the deposit. This trench was 

 4 feet wide and was carried down to the bottom of the midden. 

 Scattered through the midden were fragments of shells, principally 

 conch and clam ; fragmentary shards of undecorated pottery ; flaked 

 and slightly chipped stone implements; caches of spherical pebbles 

 and of coral; polishing stones and stones obviously used as hammers. 

 An unusually large amount of partially burned or roasted turtle 

 carapaces and leg bones was included in the kitchen middens of 

 San Gabriel cave, although here as in other cave middens moUusk 

 shells far outnumbered mammal and bird bones. 



From the Samana caves were obtained three distinct types of 

 pottery — a thin brown ware, well fired but undecorated; a coarser, 

 poorly fired, but decorated terra-cotta ware; and a third type con- 

 sisting of fragments of crude, brick-red potter}^, similar to the 

 thick-walled globose bowl shapes of Samana Peninsula, but also 

 fragments of the typical earthenware cassava griddles are of fre- 

 quent occurrence. The poorly fired, decorated terra-cotta ware con- 

 sisted of fragments of large shallow bowls with zoomorphic figurine 

 handles. Shards of decorated pottery covered with rectilinear 

 incised lines were outnumbered by innumerable fragments of 

 shallow bowls and of potsherds, plain for the most part, but belong- 

 ing to a thin- walled, well-fired type of brown ware. 



An interesting discovery was that of a many faceted tool of sand- 

 stone with each dimensional diameter approximately 2 inches. This 

 object was dug up from the bottom of the midden at the Cueva del 

 Templo. A similar faceted stone object has been designated as a 

 celt polisher by M. R. Harrington,^^ supposedly characteristic of 

 Tainan stone culture in Cuba and elsewhere in the Greater Antilles. 

 The implement is undoubtedly a polishing implement, but its pres- 

 ence at the bottom of a midden where other characteristic Arawakan 

 objects are lacking is anomalous. A similar multiple-faceted stone 

 was recently excavated by G. A. Duncan in Haiti from a shell 

 midden on the large sisal plantation near Fort Liberte on the north 

 coast. It is more than likely that this and other faceted polishing 

 implements of stone were also of use to the primitive Arawak potter. 

 Improvised implements of shell or stone were found in the middens 

 of each of the caves visited. Hammerstones, abrading tools, and 

 flaked stone implements were, however, markedly similar and uni- 

 form in type. Stone knives, scrapers, picks, and hammerstones 

 were of the same type as those recovered later from open village 

 and habitation sites on the north coast of Samana Peninsula and in 

 the Province of Monte Cristi. 



1^ Cuba before Columbus, Indian notes and monographs, Mus. Amer. Indian, Heye 

 Found., vol. 2, pi. 108, 1921. 



