because of their affinities that are too close’. It was not, 
however, until 1923 that he considered definitely the in- 
frageneric divisions of Hevea: ** While still awaiting fuller ma- 
terial of certain species, I have already been able to affirm that 
the sections Euhevea and Bisiphonia ... are not so well de- 
fined as one has thought: I have found, amongst the many 
specimens of H. guianensis, some that have the anther-whorl 
slightly irregular due to the insertion of one of the anthers a 
little too low. In this same species, in trees of one single 
locality, the staminate buds vary from wholly obtuse (almost 
globose) to rather distinctly acuminate’. This point of view he 
reiterated in 1935. Ducke’s silence on this matter in later publi- 
cations may be taken as an indication of abandonment of the 
whole system of grouping the species into subgeneric affinities. 
I know this to be true, for when we discussed this point in 
depth, he stated that he had no further use for the proposed 
infrageneric classifications that had been published. And our 
refusal to recognize these classifications was crystallized 
when, in 1945, we jointly reduced Hevea lutea (up to that time a 
typical member of Bisiphonia) to varietal status under H. 
guianensis (the only species of Euhevea). 
Ducke spent more than half a century studying wild Hevea in 
the Amazon, and he was, undoubtedly, the taxonomist most 
thoroughly acquainted with Hevea over most of its natural 
range. Ducke’s taxonomic outlook in Hevea underwent three 
distinct periods. In his earlier years, still under the influence of 
his teacher, Huber, Ducke often described minor variants as 
species (H. gracilis, H. Huberiana, H. humilior, H. mar- 
ginata). In what we may consider his intermediate stage, he 
reduced some of these *‘species”’ to varieties and forms and 
described a large number of additional infraspecific variants. 
Towards the end of his life, he recognized a limited number of 
species and fewer varieties and forms, reducing many of the 
concepts that he himself had previously described. 
In his papers on Hevea, Baldwin failed to discuss in- 
frageneric classification. That he did not consider the available 
treatments as natural, however, may be inferred from several 
of his statements. An example is the following opinion: 
**. Ducke found a tree which he considered to be inter- 
247 
