242 BUREAU OF AMERICAN ETHNOLOGY [Bull, 177 
to 74 percent; calculated at the same rate, this gives 809.7 years. 
These figures are close to the lower total derived from refuse accumu- 
lation calculations, and suggest it may be the more accurate of the two. 
However, this rate of change does not give valid results for the 
Koriabo Phase. Here there is a change of less than 30 percent in the 
pottery type frequency in the two major plain wares giving a dura- 
tion of under 100 years, much less than is implied by other evidence. 
In this case, however, the total duration computed from the sherd re- 
fuse accumulation formula is also too low, suggesting that the dura- 
tion of the Phase in the area is not adequately represented by the 
sites included in the survey. Aside from this instance the results of 
these various calculations suggest that the rate of pottery evolution, 
measured by alteration in plain wares, may be relatively constant un- 
der undisturbed Tropical Forest conditions. However, this impres- 
sion, as well as the problems of interpretation and adequacy of site 
sampling, needs more investigation before the method can be consid- 
ered a reliable means of estimating the passage of time. 
An opportunity to check the rate of accumulation expressed in the 
sherd formula with ethnographic data came during the investiga- 
tions of the Wai Wai. Unfortunately, only one of the two former 
Wai Wai villages could be used because the other (K-11) produced 
too few sherds to be representative of a normal rate of refuse ac- 
cumulation (see pp. 198-199). At E-2, sherds were scattered on the 
surface of the ground inside and around the collapsing house. The 
number of sherds from a 2- by 2-meter area on the north side of the 
house was tabulated, and conversion of the total into the dimensions 
specified in the formula gave 177 sherds in a 1.5- by 1.5-meter area, 
the equivalent of 6.82 years (Meggers and Evans, 1957, p. 257). The 
actual duration of E-2 was reported by Wai Wai informants and 
resident American missionaries as about 6 years. This is remarkable 
agreement, but again the result cannot be considered proof of the 
validity of the formula without corroborating evidence from other 
ethnographic sites. A single instance could be just a lucky coinci- 
dence. It is our hope that these various attempts to derive formulas 
for determining rate of refuse accumulation, duration of Phases, 
and rate of evolution of plain pottery types will prove challenging 
enough to other anthropologists so that they will attempt similar 
studies with their data. Only thorough examination will make it 
possible to determine the reliability of this kind of approach. 
ETHNOGRAPHIC EVIDENCE 
The most. recent enthnographic account of the Taruma Indians is 
by William Curtis Farabee (1918) who traveled up the Essequibo 
River from the junction of the Kuyuwini River to the headwaters 
in the latter part of 1914. En route, he passed two Taruma villages, - 
