172 THE JOURNAL OF BOTANY 
tinction, one ceed to the distribution of the subordinate 
forms so far as to define clearly the limits of the respective 
cardinal species, it is not long before one is n un- 
e a 
expected obstacle. The lamentation of Linneus over a similar 
critical genus—‘‘ Species Rosarum difficillime limitibus circum- 
scribuntur et forte naturé vix eos posuit’’—would here be equally 
applicable. 
or instance, there are found side by side with typical speci- 
mens of Mentha aquatica and M. arvensis ambi s and varying 
forms, in which are combined the distinguishing characteristics of 
these two species. They are linked each to each by a continuous 
chain of intermediate individuals, among which one vainly seeks to 
determine where the first species ends and the next begins. It 
seems to be a case in which one might adopt the formula of Kunth, 
Prodromus, cutting the ribbon pretty nearly in the middle, have 
united ei o M. arvensis or to M. aquatica the portions nearest 
their own analytical principles, have drawn from, without exhaust- 
served for the original illustratio 
wo reasonable sol 
arvensis extreme variations of a single species, or hybrid products 
hem ? 
hin lered or interrupted, has been carried on for twenty years. 
Desiring, however, to complete it in several details before de- 
scribing its phases and results, we shall confine ourselves here to 
poe out some of the most striking facts that have been actually 
ained. 
our observations have been frequently directed 
towards the Mints of other countries, yet, as pre on not in 
* Malinvaud, Trois Genres Critiques, 1890, 
t Relatively and parallel to Eumentha, Mentha iendi cum. and 
cervina constitute each a subgenus or distinct genus, Sane. De 
