204 MERRILL. 
cosmopolitan tropical weeds have originated in tropical America, 
and at the same time that region seems to lead in the number of 
plants of economic value, especially those cultivated for food. 
These facts may indicate a very ancient civilization in some parts 
of tropical America, by which the various weeds have become i 
adapted to the environment of cultivatioi^ and by which many * 
of the economic plants have become domesticated. It is more 
probable, however, that at least the weed-flora has been devel- 
oped in a region with thin forests, or with very scattered trees, 
or perhaps in nonforested areas. Such types of vegetation as 
the above have probably existed in some parts of tropical Amer- 
ica, especially in the drier regions, from very ancient times, 
but in the eastern tropics with their high and continuous 
humidity, such as is found in the entire Malay Archipelago, and 
in a large part of tropical Asia, extensive areas of open country 
could hardly exist under primeval conditions. 
In tropical countries like the Philippines in respect to anthesis 
there is a constant succession during all months of the year, 
and hence, in order thoroughly to investigate the flora of any 
region, explorations and observations must be continuous 
throughout the year. Very many species flower only during 
the dry season, some at its beginning, others as it progresses, 
and still others toward its end, while very many others flower 
only during some part of the rainy season. While most of the 
species found in and about Manila, and in other parts of the 
Archipelago as well, have well defined and restricted periods of 
anthesis, a comparatively large number flower throughout the 
year, some continuously, and some at intervals of greater or less 
duration. 
With the exception of those species that flower continuously, 
the time of anthesis is found to be closely correlated with the 
distribution of the rainfall. In the Philippines where the time 
and duration of the rainy season is very largely modified by 
local conditions, the influence of the monsoons, geological and 
geographical features, etc., it is a well-known fact that the 
period of greatest precipitation in one part of the Archipelago 
may be and often is in the period of greatest drought in other 
parts. Generally speaking the western part of the Archipelago 
is dry while the regions bordering on the Pacific coast are wet, 
and vice versa, but the local distribution of mountain masses 
frequently modifies this general rule, and in the vicinity of some 
mountains the precipitation is often fairly continuous through- 
out the year. Manifestly then, observations as to the periods 
