306 МВ. Н. N, DIXON ОХ А 
CALYMPERES Beccari, Hampe. Tree in jungle, Tawao, c.fr. (No. 133). 
I have compared this with Hampe’s specimen leg. Beccari, with which it 
agrees quite well. The characters of C. serratum and С. Beccarit as given by 
Bescherelle, Brotherus, and Fleischer do not altogether agree (as to serrature 
of margin, leaf-apex, etc.). Nor am I able to find the alleged difference in 
the form of the cancellina as given by Bescherelle, followed by Brotherus. 
I am not able, in fact, to arrive at any clear distinction between the two 
species from the descriptions, beyond the point of size and habit ; and am 
inclined to doubt whether C. Beccarii be more than a tall form of С. serratum. 
The fruit appears to agree with that of С. serratum as described by Fleischer, 
except that the ripe capsules are cylindric rather than “ liinglich ovoidisch,” 
not narrowed below the mouth, and the lid is rather longly subulate (about 
1 length of capsule), not “kurz geschnübelt." These may be characters 
of importance. 
The cancellina of the leaf-base appears in this and C. serratum to be often 
abnormally developed, with the cells small and passing gradually into those 
of the lamina (or even with the large hyaline cells scarcely developed at all) ; 
this is the form figured in the Bryologia javanica, and reproduced in Brotherus 
(Musei, i. p. 379) ; but it is certainly not the normal form, in which the 
cancellina is well developed. The same abnormal condition obtains very 
frequently in C. Бессати. I have not been able to make out any difference 
in the outline of the cancellina in the normal leaves of the two species, as 
given by Brotherus, op. et loc. cit. 
C. SERRATUM, А. Br. Boulder in stream in shade, Sapong, near Tenom 
(No. 158). 
C. SALAKENSE, Besch. (РІ. 26. fip. 8.) Coconut palm, Sandakan (No. 294). 
This agrees exactly with Bescherelle’s type at the British Museum. Fleischer 
is no doubt correct in reducing C. scalare, Desch. toa synonym of C. salakense. 
Bescherelle’s principal characters are the cancellina rows highly sealariform 
above, and its cells principally quadrate in C. scalare, in C. salakense the 
rows very shortly scalariform and the cells for the most part rectangular 
in shape. Now M. Thériot writes that in Mr. Binstead’s plant he has 
observed the cancellina very shortly, scarcely at all scalariform, with all the 
upper cells quadrate, in the same leaf. 
С. sUBSALAKENSE, Thér. & Dixon, sp. nov. (Pl. 26. fig. 7.) 
C salakensi Besch. affine, sed folii margine in parte basilari multo fortius 
serrato, cellulisque laminæ multo majoribus 10-15 p latis. 
Hab. Near Tenom (No. 164). 
In C. salakense (cf. Pl. 26. fig. 8) the cells are very small, 3 р wide as 
described by Fleischer—1 find them up to 5 or 7 p in Bescherelle's type ;— 
