THE HEVEA RUBBER TREE IN" THE AMAZON VALLEY 7 
But such extensive collections are not likely to be made in the near 
future. The genus is widely spread over a sparsely settled country 
which contains few people with sufficient scientific interest to induce 
them to make botanical collections. Spruce (4-0) long ago pointed 
out the difficulty encountered in collecting specimens from Hevea 
trees 100 to 125 feet high, which no native can be induced to climb 
and which cost much time and much money to fell. The writer can 
testify that Spruce did not exaggerate the obstacles. 
As things stand at present, it is by no means impossible that certain 
species differing little from other related species may represent merely 
fluctuating variations and may not be constant at all. Material 
from a single collection can tell nothing conclusive about such things. 
The writer's experience (17) with Hevea trees on plantations in the 
Orient, especially in Sumatra, shows that the trees there display a 
great range of variation in their characters. In this case we are deal- 
ing with Hevea brasiliensis, though we must admit the possibility 
that the oriental race may represent some admixture with other 
species, such as E. collina. Huber (10) is inclined to doubt the hybridi- 
zation, as does the writer. Also it may be found that the cultivated 
species is more variable than it was in nature. The common notion 
that cultivation induces variation has been attacked in recent years, 
although its possibility is admitted here. But if other species in 
Brazil are anything like as variable as E. brasiliensis in the East, we 
need to be very sure of the stability of the characters used for species 
delimitation. 
Where differences are known to be constant, their extent is not of 
such great importance. As in other groups of plants, the determina- 
tion of a difference as specific or only varietal will largely depend 
upon the measuring stick used by any given student of the group. 
For example, it does not much matter whether we consider E. ran- 
diana Huber a separate species or with Pax reduce it to E. brasiliensis 
var. randiana (Huber) Pax, but it does matter greatly whether or 
not we know that its deviations from E. brasiliensis are constant and 
heritable. 
GEOGRAPHICAL DISTRIBUTION 
The results of this expedition add very little to our knowledge of 
the geographic distribution of the species of Hevea. Ule (1+2) dis- 
cussed the distribution of the different species and published a map 
showing the entire range of the genus. He stated that the area of 
distribution is divided by the Equator into a northern region through 
which rivers of black waters flow and a southern region drained by 
rivers of white water. The northern region he reported as containing 
E. guianensis, E. collina, E. benthamiana, E. duckei, E. rigidifolia, 
E. discolor, E. minor, E. pauciflora, E. membranacea, E. microphylla, 
and E. lutea, 11 species in all. 
The southern area, he said, was inhabited by E. guianensis, E. 
nigra, E. brasiliensis, E. nitida, E. paludosa, E. spruceana, and E. 
similis, seven species. 
Huber (11) points out that the Equator does not represent the 
division between the black-water rivers and those with white water; 
even the Rio Negro lies south of the Equator. The area north of 
the Equator contains, according to Huber, E. guianensis, E. ben- 
tJiamiana, E. rigidifolia, E. microphylla, H. minor, E. pauciflora, E. 
