THE CENTRAL AMERICAN SPECIES OF QUERCUS 3 



constant, and hence more significant, characters. There are some 

 species in north and central Mexico (notably Q. chrysophylla Hum- 

 boldt and Bonpland) that are difficult to refer conclusively to one or 

 the other of the groups. 



Erythrobalanus is distinguished from Lepidobalanus by having its 

 leaves aristate-tipped (when toothed), its anthers usually only six, the 

 stigmas elongate and gradually dilated, cup scales thin, acorn shell 

 tomentose within, and abortive ovules apical. This imposing array 

 of differences would certainly deserve generic distinction if they were 

 constant. However, certain species of Lepidobalanus (e. g., Q. turbi- 

 nella Greene) and the American species of the subgenus Protobalanus 

 have teeth on their leaves clearly intermediate between the aristae of 

 Erythrobalanus and the merely mucronate tips of Lepidobalanus. The 

 gradually dilated and abruptly dilated style types occasionally merge 

 into one another as do also the stamen numbers. The cup scales of 

 some species of both groups are slightly thickened and not clearly 

 representative of either. The acorn shell of Protobalanus is as tomen- 

 tose as that of Erythrobalanus. The abortive ovules of many species 

 of Erythrobalanus are deeply lateral or even basal on occasion. 



These differences usually hold true in a great preponderance of 

 species, but they may scarcely be depended upon. Apparently only 

 the tomentose acorn shell of Erythrobalanus (which it shares with 

 Protobalanus) is constant. The other characters are being revealed 

 to be less and less dependable as more fruiting material becomes 

 available. It seems, therefore, that the sum total of differences may 

 best be regarded as a rather poor subgeneric distinction. In no sense 

 is the difference to be regarded as generic. However, it must not be 

 concluded that there exists too little difference between the groups to 

 warrant subgeneric rank. For a more detailed discussion of this 

 question see Muller (3). 



Williams (8) studied the wood anatomy of several species of Ameri- 

 can oaks and reaffirmed a distinction between the subgenera described 

 by several earlier workers. He found eight species that had at one 

 time or another been considered members of Lepidobalanus but that 

 had Erythrobalanus wood. He proposed to delimit the subgenera on 

 this line. Tillson and Muller 4 found that the so-called Erythrobalanus 

 wood type ranges widely in Lepidobalanus and is scarcely correlated 

 with any other character. It is to be concluded from their evidence 

 that the wood type may serve as an indication of phylogenetic rela- 

 tionship of the species within Lepidobalanus rather than as the basis 

 of a realignment of Lepidobalanus and Erythrobalanus. 



The series of species brought together in named groups by Trelease 

 are for the most part too finely divided for practicability. In a few 

 instances all the species of a single series have been reduced to syn- 

 onymy under one of their number. His numerous series containing 

 single species seem to serve no purpose, and these have been grouped 

 together whenever that proved possible. A great deal of grouping 

 may still be done with profit, but this must await a study of the central 

 Mexican species. These series are partly natural and partly artificial. 

 Most of those in which he placed several species, e. g., the Virentes, 

 the Acutifoliae, and the Reticulatae, are quite natural. The ones to 



4 Tillson, Albert H., and Muller, C. H. anatomical and taxonomic approaches to subgeneric 

 segregation in American quercus. Unpublished manuscript in the files of the Division of Plant Ex- 

 ploration and Introduction. 



