NARDIA. 519 
Subgenus I. EUNARDIA. 
Foliola parva subulata preesentia. Perianthiwm immersum tenerum, 
obscure carinatum, involucro obvelatum eidemque inferne alte adnatum. 
(V. supra).— Hunardie nulla species eequatorialis mihi obvia fuit. 
Subgenus I]. HUCALYX, Lindb. Bot. Notis. (1872). 
Foliola nulla. Bracteew 1-2-jugee, interiores saltem perianthio basi 
adnate. Perianthiwm emersum (nec involucro velatum) sat validum, 
argute 3-8-carinatum, ore constricto interdum tubulari. 
1. N. hyalina. F. suborbiculata undulata. Sr. foliis duplo longiores, 
convolutive, duze interiores perianthio altiuscule adnate. Per. 
ovalia alte 4-6-plicata. 
2. N. callithriz. A priore differt foliis minoribus longioribus, cellulis 
subminoribus, bracteis minus alte adnatis, per. 3-5-plicatis. 
Subgenus III. APOTOMANTHUS, 8. 
Folia late orbiculata, apice rotundata emarginatave. oliola ubique 
presentia, majuscula, lanceolata, ad Harpanthi foliolorum instar torta. 
Bractee foliis subconformes, unijugze, interior saltem perianthio basi ad- 
nata ; br'* foliolis submajor, basi dentata. Perianthiwm subcylindricum, 
a latere compressulum, juvenile apicem versus humiliter 3-plicatum— 
plicis interdum rudimentarie alatis, in etate seepe obsoletis—ore hiante 
breviter 3-9-fidum.—Melius forsan pro genere proprio habendum, a Clas- 
matocolea, nisi pro bracteis perianthio adnatis, haud tam remotum. 
3. N. succulenta. F. latiora quam longa, retuso-emarginata. F!* utrinaue 
unispina, raro bi-triloba. Br. apice 3-4-lobulate. Per. superne 
obscure 2-(4-) plicata, plicis subalatis, ore triloba. 
4. N. tortistipula. F. tam longa quam lata, apice semper fere rotundata 
raro retusa. F'" integra. Per. ore obscurissime plicata, 6—9-loba. 
Obs. The generic character given above applies mainly to Hunardia 
alone, to which the other two subgenera are doubtfully united. For the 
sake of uniformity I place Hucalyxin Nardia, but, as I have stated above, 
I consider it would rank more naturally with the entire-leaved Junger- 
manic, to which in all other respects except the slight adhesion of the 
bracts to the perianth it is very closely related.—The subgenus A poto- 
manthus agrees with Nardia in the presence of underleaves, and the 
bracts are adnate to the perianth (although in a less degree); but the 
perianth, instead of being immersed and at first entire at the constricted 
apex, as in Hunardia, is considerably emersed, and in every stage wide- 
mouthed. This character, and the occasional interpolation of a bifid 
underleaf among the normally simple ones, indicate an approach to Clas- 
matocolea, already described ; although the latter has a very different 
habit, and bracts quite free from the perianth.* 
1. NARDIA HYALINA (Lyell). 
Jungermania hyalina Lyell in Hook. Brit. Jung. tab. 63. 
Hab. In sylva Canelos Andium Quitensium, alt. 400-1000™; necnon in 
m, Pichincha, alt. 3000; in arena saxisque rivulorum inundatis. 
Per Europam totam, et in insulis Canariis inventa est. 
* Fora comparative view of the characters which distinguish Nardia from 
Marsupella see Spruce ‘On Marsupella Stablert” in Husnot’s Révue Bryo- 
logique (1881), p. 92. 
