American Species of Marchantia. 237 



chambers, as Schiffner has pointed out; in var. alpestris the 

 thallus is uniformly green on the upper surface, and air-chambers 

 are everywhere present. In var. aquatica the margin of the 

 thallus is entire or nearly so, and the upper surface completely 

 lacks epidermal papillae or bears them very rarely; in var. 

 alpestris the margin of the thallus is more or less denticulate from 

 projecting cells, and epidermal papillae occur in greater or less 

 abundance in the marginal portions. In var. aquatica the 

 appendages of the ventral scales are entire or nearly so, and 

 the spreading rhizoids are smooth; in var. alpestris the appen- 

 dages are distinctly denticulate, and the spreading rhizoids 

 tuberculate. 



About fifteen years ago another so-called variety was distin- 

 guished by Hagen under the name var. mamillata. It was based 

 on a supply of specimens collected by its author at Opdal in 

 Norway and distributed by Schiffner in Hep. Europ. Exsic. 15. 

 Apparently Hagen himself did not publish his variety. Schiff- 

 ner^^ did so, however, and quoted Hagen's original diagnosis, 

 as follows : "Cellulae epidermicae et f rondis dorsalis et carpo- 

 cephali acute mamillosae." In commenting on this diagnosis 

 Schiffner showed that the mamillose appearance, so strongly 

 emphasized, was due to epidermal papillae and that these were 

 restricted to the marginal portions of the thallus. He showed 

 further that the female receptacles in all forms of M. poly- 

 morpha were mamillose in Hagen's sense. Var. mamillata, there- 

 fore, is based on exceedingly vague characters and has little or 

 nothing to distinguish it from var, alpestris. 



Although var. aquatica and var. alpestris are at first sight so 

 distinct from each other they are connected by intermediate 

 forms, and their differences seem to be associated with definite 

 dift'erences in environmental conditions. They represent, tliere- 

 fore, modifications rather than varieties in the taxonomic sense. 

 Probably the most logical disposition to make of them is to regard 

 them as forms, as Miiller^- has done, and to cite them as forma 

 aquatica (Nees) K. Miill. and forma alpestris (Nees) K. MiilL, 

 respectively. Other forms, less distinct than these, might like- 

 wise be distinguished, but it would hardly be a profitable task 

 to designate them by names. 



" Lotos 49 : 93. 1901. 



^' Rabenhorst's Kryptogamen-FIora 6:306. 1907. 



