124 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 146 
lowest levels of such caves as Eucaliptus, Samambaia, and Limeira, is 
characterized by a high percentage of stemmed bone projectiles and 
completely polished stone axes (Evans, 1950). On the other hand, 
Hurt (1960, p. 587), after excavating six different caves, decided 
that "although differences were noted in the vertical distribution of 
artifacts from one meter section to another there were no variations 
that could not be attributed to random distribution or size of sample. 
For this reason all the preceramic artifacts collected in the 1956 proj- 
ect were assigned to a single complex." This complex was composed 
predominantly of hundreds of flakes and fragments of quartz crystal, 
which Hurt considers to be tools and not rejects. The complex also 
includes several varieties of axes and celts made by percussion flak- 
ing and polishing, pitted hammerstones and choppers. The relatively 
rare bone artifacts include projectile points made from the hollow 
cannon bone of birds and a few splinter awls. Beads and ornaments 
were made from drilled olivella shells (op. cit., p. 583) . 
More progress has been made toward evaluation of the antiquity of 
sambaqui sites, which occur along the seacoast from the Amazon to 
Rio Grande do Sul, although many problems of interpretation remain. 
A series of carbon-14 dates from the southern sites gives them a time 
range of between 7,500 and 1,500 years ago (Laming and Emperaire, 
1958) . This long span suggests that efforts to classify sites on the basis 
of differences in artifact content and typology may have a chronologi- 
cal basis. 
Among the first to attempt a subdivision was Serrano, who pro- 
posed four phases with both areal and temporal implications: "the 
southern; the middle; that of the sambaquis with Azara prisca (ar- 
chaic) ; and the Amazonian" (1946, p. 404). The southern phase, 
which includes the sites from Rio Grande do Sul, Santa Catarina, 
Parana, and southern Sao Paulo (fig. 15), is characterized by zooliths 
and round stones or bolas. Pottery may occur in the upper level. The 
archaic phase, which he correlates with the Lagoa Santa complex, is 
distinguished by stone axes of trianguloid form, chipped stone knives 
and scrapers, and hammerstones. Little information is provided on the 
middle and Amazonian phases (op. cit., pp. 405-407) . 
Recent fieldwork in sambaquis of Santa Catarina, Parana, and Sao 
Paulo States by Orssich (1956), Bigarella (1954) and Laming 
(1958) generally supports Serrano's point of view. Laming and Em- 
peraire distinguished two complexes. The oldest one, corresponding 
to Serrano's "archaic" phase, is represented by crudely chipped bi- 
face artifacts, hammerstones, and crude flaked knives. The second 
complex, much more recent and corresponding to Serrano's "south- 
