rageena ARCHEOLOGY IN BRITISH GUIANA 319 
pean trade materials, the correlation with the seriated position derived 
from pottery type classification is very good (table M). Only two 
TABLE M.—Correlation between dated and seriated position of Rupununi Phase 
sites with European trade materials 
Dated sequence from European Seriated Dated sequence from European Seriated 
trade goods sequence from trade goods sequence from 
Went TOS Tee JISH aborigimhlOol) Oise too COO Gort OO || fabonginal 
pottery pottery 
Date Site types Date Site types 
1880-1910... -......-.s Res R-12 Pre IRhi) 2 = soe oe eee IRoGr cases R-6 
Late 19th century -____- R-36 B__..| R-8 Prerilssu ces BOA rit R26222.5.2 R-17 
ISS Rages R-36 B AYS0s1850 oe eS Rei7 22-5 R-3 
1820-40. BE. DATs WS R-12____-- R-19 
sites are out of order. R-3 is third in the European dating and 
earliest in the pottery seriation, but this is not serious contradiction 
because the trade goods at R-3 give only a terminal date, and the 
period of occupation might well be several decades earlier, in which 
case the site could be placed lower down in the sequence. R-12 is more 
seriously out of line, being in the middle of the dated sites but last 
in the pottery seriation. Here, the discrepancy is more difficult to 
explain. The ceramic seriation and the characteristics of the site both 
establish R-12 as very recent, but the glass and earthenware are both 
dated as about 1820-40. The only explanation that suggests itself 
is that the European trade materials were used longer than would 
normally be expected or were received by the Indians after they were 
out-dated among the Europeans. Since the other identifications all 
confirm the ceramic seriation, this single exception cannot be con- 
sidered a significant argument against the validity of the sequence. 
The ceramic history of the Rupununi Phase is characterized by the 
decreasing popularity of Kanuku Plain, with an orange paste, and a 
corresponding increase in frequency of Rupununi Plain, with a gray 
core. During the period of time represented by the site survey, 
Rupununi Plain increases from 11.6 percent to 79.2 percent of the 
sherd total. At the same time, Kanuku Plain declines from an initial 
high of 88.4 percent to 20.8 percent. In all but five of the sites, these 
two pottery types comprise 98 percent or more of the total sherds, with 
the result that they show a smooth sequence of change. Only one 
site, R-40, fails to fit well because of an unusually high frequency of 
decorated sherds. 
A small number of sherds with unusual temper or some form of 
decoration appear sporadically, one or more kinds occurring at only 
12 of the sites. In several cases, origin by trade is demonstrable or at 
least suspect. Cariapé tempering was detected at five sites, with the 
majority of the sherds from each site attributable to one or two 
vessels. The distinctive nature of this temper and of the process by 
