KUNGL. SV. VET. AKADEMIENS HANDLINGAR. BAND 63. NIO 8- 21 



and among them Phillips deserves special attention. In Ann. of Bot. XII (1898) p. 

 177 he pointed out that the cells of the central axis in D. sanguinea a. o. never divide by 

 transverse intercalary walls, contrary to D. sinuosa. Further, he pays attention to the 

 position of the procarps: on the costa in D. sanguinea, ruscijolia, alata, etc, spread över 

 the frond, without primary relation to the costa or nerves in D. sinuosa, just as in Ni- 

 tophyllum; he is inclined to restore Kutzing's genus Phycodrys for sinuosa, and says that 

 it is more related to Nitophyllum than to Delesseria (p. 191, 195). Nienburg does not 

 quote Phillips, and several of the Nithopylla examined by the former are compared with 

 the growth-type of sanguinea, not of sinuosa. But all genera and species with the apex 

 of sinuosa ha ve the anatomical structure of the frond as in Nitophyllum, without the inner 

 rhizoids or hyphae characteristic of D. sanguinea and its relatives (see my paper on the 

 Californian »D. quercifolia»). Among the species with the apical growth of sanguinea 

 I have only met with two exceptions from this rule: Apoglossum ruscifolium and Mon- 

 tagneanum. J. Ag., where no internal hyphae are developed; still, their costa is not 

 built after the regular fashion of Phycodrys or Nitophyllum. 



These differences in growth, position of procarps and anatomy are important 

 enough to justify the establishment of two subfamilies or tribes, Delesserieae 

 and Nitophylleae. From what I have seen of this subfamily I should like to 

 distinguish two assemblages, the Phycodrys group and the N itophyllum group, as a 

 provisional arrangement only. 



Delesserieae. 

 Delesseria Lam. 1 



D. salicifolia Reinsch. — Kylin & Skottsb. p. 45, Taf. 1, f. 3. — Fig. 4. 



South Georgia: Cumberland Bay, Boiler Harbour, sublit. 5 m (St. 48, 

 20. 4. 09). 



1 Howe, Mar. Alg. of Peru p. 136, retains this name with much reservation, remarking that it was put on 

 the list of nomina conservanda by the Brussels Congress, and adds: »inasmuch as Delesseria was originally based 

 upon thirty seven species, representing several genera and several families of modern writers, it offers a good il- 

 lustration of the futility of »conserving» a generic name unless it is specified for what it shall be »conserved», or 

 in other words, unless it is fastened to a certain species as »type». Delesseria sanguinea, the first species of La- 

 mouroux's first section of the genus, might fairly be considered as the type of Delesseria, but this is also the evident 

 type of Hydrolapatha Stackh., which (altered to Hydrolapathum) has been maintained as generically distinct from 

 other »Delesserias » by Kuprecht, Le Jolis, J. Agardh and others». This criticism hits a weak point in the list of 

 nomina conservanda, and Howe is of opinion that Delesseria is an untenable name and that it is undesirable to con- 

 serve it. But in this I do not agree with him. If we want to use Hydrolapathum, we have to fix this name to a 

 species as type, and this type will be H. sanguineum. Then, nothing prevents ns, from agreement, to adopt Deles- 

 seria instead of Hydrolapathum in spite of the younger date of the former. The combination D. sanguinea is uni- 

 versally known, the plant is mentioned under this name in innumerable floristic papers, and its cytological etc. 

 features were described under the same name. I find it more desirable to conserve Delesseria than to use Hydro- 

 lapathum. But of course no list of nomina conservanda ought to be proposed by any committee unless it associates 

 itsclf with specialists able to fästen a type to every genus. If not, uncertainty and confusion are sure to follow. 



