16 THE JOURNAL OP BOTANY 



2. L. multifolius Dunn, sp. n. Arbor magna, ramis exigue 

 tomentosis. Folia 12-15 cm. longa, 7-juga, stipellata ; foliola 

 oblonga, acuminata, basi obliqua, rotunclata, 4-5 cm. longa, char- 

 tacea, supra glabra opaca, subtus prsecipue in costa tomentosa. 

 Racemi axillares ad 25 cm. longi, fere ad basin fioriferi ; floras 

 8-10 mm. longi, nodoso-racemosi, nodis ad 5 mm. longis, pedi- 

 cellis 2-3 mm. longis, ut rachidibus tenuiter tomentosis, apice 

 bracteolis 2 subulatis ; calyx campanulatus, 3-4 mm. longus sub- 

 glaber, dentibus 4, tubi quartam partem aequantibus, ovatis ; petala 

 glabra ; vexillum sine ungui rotundatum, basi breviter cordatum ; 

 stamen vexillare medio connatum ; ovarium tenuiter sericeum 

 multi-ovulatum. 



West Trop. Africa. S. Nigeria, Kitson, March 21, 1909 

 (Brit. Mus. Herb.). 



This species is easily distinguished from L. Grijfoniamis by its 

 more numerous leaflets. 



SPIB^A ULMARIA L. var. DENUDATA Boenn. 

 By a. R. Horwood. 



Mr. Druce does me an honour in singling out my six-year-old 

 note on this plant." / also " hold no brief for (or against) clenn- 

 data," but when one finds that one and the same plant is made up 

 of typical Spircea Ulmaria, the so-called variety denudata, and a 

 leaf of an intermediate type, it is reasonable to say that denudata 

 has no title to rank as a variety, and to assume that it is merely a 

 state, unstable in the same specimen, and of no more value than 

 lupus in man, and like that due to pathological causes. Of what 

 real value, to take a similar case, is the variety Pryorii of Pajjavei- 

 Bhoeas ? In the herbarium the red colour of the hairs of the 

 peduncle, its chief distinction, is lost. Varieties based on whether 

 some portion of the plant is hairy or glabrous throughout are far 

 from satisfactory, e. g. Sisymbrium officinale var. leiocarpum, Acer 

 campestre var. leiocarpon, and the like. In the nomenclature of 

 Lichens far more names rank as " subspecies " and " forms " than 

 as varieties, and this seems much more sensible than adhering to 

 the higher rank, in the last case, of variety. 



For evolution exhibits three modes of progress. We have 

 heredity, which tends to preserve species in their original form, 

 and to perpetuate the type. The active agent of evolution is 

 variation, which tends to multiply its stock by the production of 

 variations. These last, as subdivisions of a former species, 

 heredity tries to maintain without further alteration, whilst the 

 antagonistic factor of variation tends to multiply them afresh. 

 Then Mendel's Law gives us a tendency in some to go back to their 

 original form, while a simultaneous effort is made to produce the 

 same or fresh forms. In this Scylla and Charybdis of conserva- 

 tion and progress a state of instability may be created ; some 



* See .Journ. Bot. 1910, 281. 



