THE TAXONOMY OF THE AGARICACEAE 
based on the study of American rather than European material. We 
want no "pounds, shilhngs, and pence" in the form of cumbersome 
subgenera, subsections, and subspecies. The taxonomist must know 
but not recognize varieties, which are essential to the gardener, the 
physiologist, and the plant breeder. Let the grouping be as natural 
as possible, but artificial when conducive to clearness. The absence 
of sex in the gill-fungi gives one considerable liberty and our knowledge 
is still far from complete. A system of classification representing the 
genetic relationships of the higher fungi is hardly yet in sight. If the 
species could be collected and grown together under cultivation, the 
glad day might be hastened, but they cannot. Every house has a 
garret; so has every family a genus or two which catch everything not 
wanted elsewhere. Do not be too particular with the misfits, but do 
not throw them out of the window; they may fit in when you move to 
the next house. 
In seeking suitable characters for classification, one must use what 
comes to hand, and the same characters may not be available for 
different groups. The best and most constant primary character for 
the gill-fungi seems to be the color of the spores. Earle tried to use 
the "partial veil" but failed on account of its evanescent and variable 
character. The form and surface markings of the spores may be 
quite characteristic, as in Entoloma and Inocybe, but other good 
characters should be associated with them. I believe that Patouillard 
goes too far with microscopic characters, and, moreover, that these 
should be used in keys as little as possible, in order to save time and 
trouble. A key character need not necessarily be the most important 
generic character, but only the most convenient. 
Recent researches on color in the flowering plants have shown 
this character to hang on a very slender thread sometimes in that 
group, but what would we do in the fleshy fungi without the recog- 
nition of color? Poisonous properties alone would hardly seem to be a 
sufficient basis for the separation of species, and still there might be 
a good practical reason why they should be recognized in certain 
instances. I have in mind the variations in the poisonous properties 
of Venenarius muscarius, V. pantherinus, and Chlorophyllum molybdites, 
and the separation of Clitocybe sudorifica, a poisonous species or 
variety, from Clitocybe dealbata, generally considered harmless but 
not differing morphologically from C. sudorifica. Ordinarily, physi- 
ological properties would seem to have no taxonomic standing, but 
