108 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 146 
ences from the great center of this pottery tradition in the valleys of 
Hualfin and Santa Maria. 
The Pichalo III and IV ceramic tradition appears to be independent 
of those just mentioned, and to have moved down along the coast. 
Molle I resembles in general features, such as vessel shape and re- 
duced firing, the gray-black ware of the southern Puna (Laguna 
Blanca) and the Puna de Atacama. The gray incised pottery shows 
clear Cienaga influence. 
Molle II exhibits several features of great interest. One is the pres- 
ence of two distinctive vessel forms, the bridge handle and the stir- 
rup spout. Both traits are significant in the Central Andes, but ab- 
sent or highly unusual in northwestern Argentina, so that an origin 
from the latter direction can be ruled out. Another unusual feature is 
post-fired painting, otherwise known only at Tebenquiche. These 
traits are absent in Chile north of Molle, except for occasional stir- 
rup-spout vessels. Finally, negative painting is a Molle II characteris- 
tic absent in northwestern Argentina but with clear Andean deriva- 
tion, and particularly common on the coast of Peru. An explanation 
of the origin of these elements is difficult unless we postulate a mari- 
time connection with the Peruvian coast, which would be feasible even 
against the Humboldt current. 
Other elements of Molle II ceramics, such as the almost ex- 
clusive use of stepped motifs, polychrome painting, and highly pol- 
ished slips, are shared with Condorhuasi. Many features are identical, 
and it is difficult to determine which complex gave rise to the other. 
The final conclusion will be reached only when a number of absolute 
dates are available for the Molle complex, but there are some clues 
as a basis for conjecture. For example, grooved axes do not occur in 
the Molle area, although they are very common in northwestern 
Argentina. They are also unknown in Peru. If the direction of diffu- 
sion had been from Condorhuasi toward Molle, the grooved ax should 
have crossed the Cordillera. A movement in the opposite direction 
would explain more completely the evidence as it now exists. 
The postulation of the imposition of a polychrome tradition on a 
complex such as Candelaria would produce a result similar to that rep- 
resented by Condorhuasi, and would explain the close resemblances 
between these two complexes, which have even led some to consider 
them a single culture. The polychrome forms, however, are centered 
in the southern Valliserrana region, and do not occur in the Cande- 
laria area. An origin from the Selvas Occidentals can be ruled out. 
This leaves only two routes : From the altiplano, which would link 
