84 THE JOURNAL OF BOTANY 



The direction of the leaves is certainly of no value ; no series of 

 these plants can be examined without finding every jDossible position, 

 from perfectly equally imbricated to stronglj^ homomallous, with no 

 correlated characters ; the types of R. loxense and B. coespitosum are 

 neither of them extreme forms, and that of B,. loxense does not 

 show the leaves very sti'ongly homomallous ; while there is a decided 

 tendency to this direction of the leaves on one of the two bi*anches of 



B. coespifosnm at Kew ; and Dr. Andrews, in his notes on the ty})e 

 at Stockholm, writes : " Branches largely erect and not branching 

 further, leaves falcate, secund." He also indicates the variability in 

 the capsule form b}^ describing them as " nearly as wide as long up to 

 twice as long as wide." 



The leaves show some variability as to acumination and as to 

 erectness or otherwise of margin in both plants. The onlj^ point 

 as to which there could be any question of difference, and that a 

 minute one, is that the leaves in B. coespitosum type show frequently, 

 but not constantl}^, a minute close denticulation of the margin near 

 apex. This, however, is not constant, and I have scarcely seen it on 

 any other plant of B. coespitosum that I have examined, at least 

 as at all a constant character. No. 2113 of Mr. E. S. Williams's 

 exsiccata of Bolivian plants, issued as Sematophyllum coespitosum 

 (Sw.), is instructive in this respect. It has many of the leaves finely 

 denticulate as in the type, but at least an equal number on the same 

 stem show no denticulation at all. Most of the West Indian and 

 South American plants under this name show no denticulation what- 

 ever of the leaf -margin. 



There is some slight variation in leaf form, in length of seta, and 

 form of capsule, on the sj>ecimens, and as these variant forms occur 

 on both plants they rather confirm than otherwise their identity. I 

 do not doubt for a moment that an impartial examination of the two 

 types, even without taking into consideration the various forms in 

 which they occur in both their areas, would lead anyone to pronounce 

 them identical, and I am inclined to think that the only cause that 

 they have l^een so long kept apart is due to Hedwig's drawings and 



C. Mueller's description having led to Leskea coespitosa of Hedwig 

 being taken as representing the H. coespitosum of Swartz. 



My conclusions, after studying the types of H. coespitosum Sw. 

 and II. loxense Hook., as well as the original specimens of H. litho- 

 p)Mlum Hornsch. and H. Duisahoanum Mont., can perhaps best be 

 summed up by saying that except for the slight denticulation on the 

 leaves of H. coespitosum, and for some slight difference in the 

 arrangement of the leaves on the branches, I should be exceedingl}'- 

 sorr}^ for an3^one who got the tj'pes mixed up and had to separate 

 them again ! 



Having, as I think, established the identity of the two main groups 

 of the South American plants, and their identity with the African 

 B. Dtdsaboanum, I need scarcel}^ labour the point further, but simply 

 state that the Indian Stereodon tristi cuius of Mitt. Muse. Ind. Or. 

 p. 102 (which it may be noted was called II. spjlicdrotheca by Wilson 

 in sched.), of which I have examined the type, and the Australian 

 B. omde Broth., of which 1 have seen specimens of the original 



