28 BULLETIN 15 6, UNITED STATES NATIONAL MUSEUM 



face the sea, and several large openings in the roof caused by the 

 fall of rock masses loosened by water seepage. 



No fresh water was found in any of the caves explored, although 

 a small quantity of water saturated with calcium carbonate 

 dripped from the cave ceiling and readily formed stalactites and 

 stalagmites. Shards of aboriginal pottery coated with calcium 

 carbonate lay in small heaps near the stalactites. A few unbroken 

 bowls were recovered from under rock masses but the midden de- 

 posits generally contained no pottery refuse such as was to be seen 

 in the kitchen middens surrounding habitation sites on Samana 

 Peninsula. 



According to historical accounts, fish nets of cotton cord were 

 woven by the pre-Columbian Indians of Samana, but the sole re- 

 minder of this aboriginal fishing practice was the recovery of sev- 

 eral net weights of notched stone from the middens near the cave 

 entrances. During a subsequent visit to the Samana caves in 1929 

 the writer succeeded in recovering two fishhooks of shell from the 

 midden debris on the cave floor in the Boca del Infierno. 



The prehistoric cave dwellers of Samana subsisted principally 

 on the meat of the conch and of other shellfish. A careful search 

 by Doctor Miller and the writer failed to reveal a bed of live 

 conchs anywhere near the caves of the Pla3'^a Honda coast. The 

 absence of beds of live conchs is remarkable, as the bulk of the 

 midden material covering the cave floors is made of these shells. 

 Natives professing to know of conch beds never were able to locate 

 one, although a small number of recently dead conch shells of the same 

 species {Stronibus pugilis L.) as those of the cave deposits w^ere 

 found in one of the shallow coves near the keys. The much larger 

 Strombus gigas appeared in smaller quantities in the cave middens, 

 but was fairly plentiful in the kitchen middens of the open village 

 sites on the peninsula. 



Deposition of shells in the kitchen middens of the caves had pro- 

 duced heaps of varying thickness; but in those portions of the caves, 

 usually near the entrances, which were obviously devoted to culinary 

 purposes, the refuse heaps reached a thickness of 9 feet or more. 

 Where the deposits had not been disturbed, excavation was under- 

 taken. The floor of the cave is covered with a thick layer of reddish- 

 yellow soil, not at all sandy. The layer is of irregular depth, 

 greatest near the cave entrance, but sloping down to isolated heaps 

 at a considerable distance away. 



Cave middens contain conch, clam, and other species of shell, 

 crab, claws, mammal, fish, turtle, and bird bones cast there by the 

 pre-Columbian Indian cave dwellers. The bottom of the deposits 

 of shell is embedded in the yellowish soil, while tlie upper sections 

 are interspersed with deposits of ash, charcoal, and a small quantity 



