MILPA AGRICULTURE COOK. 319 
burial places, if not as dwellings, and even constructed artificial caves 
or tumuli. Excavation of an artificial mound of earth on the Sepacuite 
coffee estate in a heavily forested district between Senahu and Caja- 
bon revealed a core of inide stonework, roofed with large rocks. (PI. 8, 
fig. 1.) This form of construction is entirely different from other 
ancient walls in the same district, which are made of thin, flat stones, 
not shaped artificially, though carefully laid together. The present 
Indians refuse to credit the idea that the walls were built by ancient 
inhabitants of the country and ask, "Where could anybody find so 
many flat stones? " Their belief is that these ancient buildings, 
now buried deep in the forest were the original habitations of man- 
kind, prepared in advance by the Creator, "When man was bom, 
when daylight broke over the earth." 
The time needed to complete the process of reforestation in a 
district that has been denuded must depend, as in denudation, very 
largely upon the local conditions of climate, soil, and topography. 
Several decades are required for the growth and production of seed 
by the pioneer individuals, the pines, oaks, or other fire-resistant 
types, the first invaders of the grasslands, and other decades for 
the more abundant trees that must develop before the growth be- 
comes dense enough to exterminate the grass and permit the succes- 
sion of genuine forest types to begin. Many kinds of trees that are 
abundant in the new forests, such as Cecropia, Castilla, Heliocarpus, 
Ochroma, and Attalea, are only vanguard species and gradually give 
place to the more permanent types of slow-growing hardwood trees. 
Still more time is required for the flora of the undergrowth and 
the fauna of the humus layer of the soil to be fully restored after 
the necessary forest conditions have been established. 
By taking account of the succession of types of trees and other 
biological features it is possible to recognize the stage of develop- 
ment that any particular woodland may have reached, or even to 
gain an idea of the approximate age of a forest. From the open 
grasslands to the dense tropical forest, with its slow-growing hard- 
wood trees, is obviously a long sequence of biological events. A 
hundred years would be entirely insufficient, and 200 probably not 
enough, under the most favorable conditions. Even after 5 or 10 
centuries the effects of previous denudations might still be trace- 
able by sufficient study of a forest and what it contains. 
MAIZE PLANTINGS IN UNCUT BUSH, 
Another modification of the milpa system that avoids or defers 
the danger of grass invasion is applicable to districts that have a 
long dry season. It was observed in northwestern Guatemala, in the 
district of Nenton, Department of Huehuetenango, in May, 1906, and 
