210 THE JOURNAL OF liOTANY 



less developed small central " boss " as well as larger lateral elevations, 

 which vary from nearly smooth to much wrinkled. 



E. viridi^ora still seems to be mixed up with E. latifolia^ but 

 there is no need for this, because, although it varies very much in 

 other ways, the reproductive organs always afford a good distinction. 

 They are arranged for self-fertilisation ; the friable pollen falls on the 

 receding stigma, and there is little or no rostellum. In E. latifolia, 

 atropurimrea, and imrpnraia tlie pollen is less friable, the rostellum 

 is well developed, and the stigma is pushed so far forward that the 

 pollen cannot fall on to it. As far as the rest of the plant is con- 

 cerned, it sometimes resembles slender greenish specimens of E. lati- 

 folia quite closely, but usually is more delicate if growing in woods, 

 more slender and wiry and yellower if growing in sand-dunes. It 

 flowers at the very least a fortnight earlier. The leaves are often 

 small, the flowers small to medium, the colour green or whitish green, 

 with not much brown tinge. Lip a triangle with a long point in 

 woodland forms, the point curled under in dune-forms ; usually not 

 much development of roughened elevations on the epichile. Ovary 

 smoother than in the other species, may be almost glabrous. 



E. latifolia in our scheme is a name which includes E. onedia Fr. 

 and E. atroviridis Linton as synonyms. We hope to amplify our 

 results on this species in a later paper, but may give the conclusion 

 here as follows : — 



i. E. media. — Col. Godfery in this Journal (1919, 80) has already 

 shown that E. media Fries is a nomeii nudum, and that E. media 

 Bab. is E. viridijlora, so that the name " media " is completely wiped 

 out. We have independently concluded that no distinction is possible 

 between E. latifolia proper and E. ''media'' as it has been under- 

 stood before Col. Godfery's papers. The supposed differences between 

 these two were, that in E. latifolia the leaves are broad and there is 

 a sudden transition from the fairly large top leaf to the small lowest 

 bract, and that the lip has two smooth bosses or elevations on the 

 epichile ; whereas in E. " media " the leaves are narrower and the 

 transition to bracts is gradual, the epichile having two *' plicate- 

 ruu'ose" bosses. We have rarely found plants with quite smooth 

 bosses on the epichile, and these usually have the " media " type of 

 leaf ! There is every sort of gradation from smooth to veiy rough 

 and pitted elevations, and one linds every sort of elevation combined 

 with every possible sort of leaf-scheme ; for there are all grades of 

 leaves too, from broad to narrow, and one gets narrow leaves with 

 abrupt transition to bracts, and broad leaves with gradual transition, 

 as well as the reverse, and all sorts of other variations as well. In 

 other words, one finds ''media'' flowers with " latifolia" lea^YCS and 

 vice versa, and the majority of plants are not exactly one or the other, 

 every grade of variation being present which affects these characters. 



ii. E. atroviridis Linton. — The chief supposed diagnostic cha- 

 racter of this is that there are three roughish bosses on the epichile. 

 We find that the original description of E. atroviridis would fit a 

 specimen of E. latifolia here and there, but that it is as common to 

 find 3 bosses as 2, and that the two conditions grade into each other — 

 the middle one may be lai-ge, small, minute and distinct, or confluent 



