a 



S 



153 



M. aculeata the ^>teiii-2jedicels are usually sliort, less often long-; 

 wlien tlie pedicels are long, occasionally iheie are a few simple 

 basal scapes as well. In i/. Prattii one instance— and that 

 recurring- one— of the condition with long pedicels and simple 

 basal scapes in addition to the central stem has been recorded. In 

 M. rudis.on the other hand, the condition with a central stem and 

 short pedicels has not been met with; all the specimens seen have 

 long stem-pedicels, with, in addition, simple scapes round tlie base 

 of the stem. In M. horridvJa we meet, for the first time, a further 

 dcA^elopment. In the state of this species which has to be 

 regarded technically as '' typical,'' because of its being the states 

 to which the original diagnosis of the species alone applies, there 

 is no central stem — or compound scape — at all; the scapes are all 

 simple,^ l"fiowered and radicaL Growing side by side with the 

 plants in which all the scapes are radical we find others agreeic^ 

 with the simple-scaped plants in eYery respect save in having a 

 central stem with long j^edicels in addition to a number of simple 

 .scapes; these agree in habit, though not in foliage, nor as 

 regards their style, with AL rudis. Along with these plants we 

 find growing siill otliers, exactly the same as regards leaf and 

 flower and fruit, but with short-pedicelled flowers on the main- 

 scape, usually, though not always^ all ebracteate, and usuallj-, 

 though not always, without any accompanying simple and 

 solitary-flowered basal scapes. That all three states are condi- 

 tions of the same natural species is, in the writer's opinion, 

 hardly open to discussion. It was felt convenient, however, 

 in 189G to recognise two varieties within the species. The 

 first of these varieties was made to include only those specimens 

 in which all the scapes are simple and have solitary flowers; the 

 second to include all those specimens which have a compound 

 scape, whether accompanied or not by simple basal scapes. The 

 reason for the recognition of the first variety was a two-fold one. 

 On the one hand, this variety corresponds precisely with what 

 was described in 1855 as M. Jtorridvla^ and what had for forty 

 years been accepted as 3f. horridula; on the other hand, the 

 character hj which the variety may be recognised seemed then 

 to be unique within the group Acuteatae^ to which M. horriduJa 

 belongs. The ground for the treatment of the remaining states 

 of M .horridtda as one variety, and not as two or more varieties, 

 was that, although the same degree of variation is manifested 



within the limits of typical M. aculeata ^ in the case of that 

 species this variability had not given rise to a necessity for the 

 segregation of taxonomic varieties. 



In the Pftanzenreich a somewhat different view has been taken. 

 The treatment there accorded to what is technically the typical 

 variety of M. horridula is identical with that proposed in 1896 

 and continued in 1906. The treatment accorded to what in 1896 

 and again in 1906 was recognised as var. racemosa is, however, 

 modified in the Pflanzenreich in two respects. In the first place 

 the state in which the central compound scape is accompanied 

 by a number of simple basal scapes, is separated, to some extent, 

 from the state in which the central compound scape has no 

 accompanying basal scapes; further, only the former of these 



