— 73 — 



Found growing on bark of deciduous trees (beech, oak, chestnut), bank of 

 Swannanoa River, at Swannanoa, Buncombe Co., July 9; North Fork, some 

 5 miles above its confluence with Swannanoa River, July 10; Grandmother 

 Gap, Avery Co., Aug. 13; all in North Carolina, 1919. 



Apparently specifically identical and to be provisionally referred to this 

 species are specimens from Mexico: Etzatlan, Jalisco, Oct. 6, 1908 (Pringle 

 10618, in part); Cuernavaca, Morelos, Oct., 1908 {Pringle I'i.^oS, in part). 



Several species of Tortula growing on the bark of trees and reproducing 

 partly or almost entirely by means of propagula are already known, the classic 

 type being T. papulosa Wils., widely spread in Europe and also known from the 

 eastern coastal region of North America, as in fact from various parts of the 

 world. From this the species described above differs in a great number of 

 characters, the older species having for example leaves without recurved borders, 

 with excurrent costa, papillose only on dorsal surface, where the number of 

 papillae does not exceed i to the cell, the papilla larger, measuring about 10 [l 

 from end to end of the crescent. The propagula originate mostly from the 

 ventral surface of the costa or the cells immediately adjoining it, they are small, tend 

 to be spherical or slightly oval in shape and are divided in part by oblique planes. 

 The points mentioned are enough to show that there is no very close relation 

 between the two species. T. latifolia Bruch is a distinct species sometimes pro- 

 ducing propagula similar to those of T. papulosa, though apparently mostly 

 from the \entral surface of the leaf blade. It has also a dififerently shaped leaf, 

 different papillae, and in fact shows no close relationship to our species. The 

 propagula of T. laevipUa (Brid.) De Not. and the closely related T. pagorum 

 (Milde) De Not. are modified leaves having their origin in the stem. Lim- 

 pricht^ also assigns cylindrical propagula exceptionally to the leaves of T. laevi- 

 pUa and T. pulvinata (Jur.) Limpr., which would suggest something more nearly 

 resembling our species. Exception is however taken to this statement by Cor- 

 rens,'^ who was unable to demonstrate such propagula as actually belonging to 

 these species. The statements of Limpricht are however so definite that it seems 

 almost incredible that they are not based upon actual observations. The same 

 statements are also given byAmann and Meylan,^ whether from original observa- 

 tions or on the authority of Limpricht is not clear. Both species show a long 

 €xcurrent leaf-costa and are also otherwise entirely distinct from the North 

 Carolina moss. Its relations are in fact not boreal or European at all, but south- 

 ern. In a specimen of Pringle's Plantae Mexicanae, No. io6ih, labeled Fa- 

 bronia octoblepharis var. americana Card., collected from the bark of Ipomoea 

 trees in the first Mexican locality noted above I found a bit of Tortula which is 

 clearly related to our species. It shows the same type of propagula, no longer 

 attached, but loose among the plants, while its leaves clearly show the cells 

 from which they had originated. Mr. R. S. Williams identifies this with T. 



sRabenhorst, Kryptogamenflora, IV, I, 680, 683. 1888. 



Vermehrung der Laubmoose, 78, 84. 1899. 



^Flore des Mousses de la Suisse, I, 82, 83; II, 120. 1918. 



