72 THE JOURNAL OF BOTANY 



divisions of tlie subfamily Monandrce — JiasitoncB a.u([ Acrofonce — 

 for cross-fertilization can and does take place without caudicles, but 

 not without viscid glands. All the other differences between Cephal- 

 anthera and Epipactis, such as the presence or absence of a peduncle, 

 the position of the llower, the extent to which it opens, the shape of 

 the lip, &c. are of but secondary importance, and not differences 

 of the first rank, like the absence of a rostellum. 



The author's argument that the rostellum is smaller in E. pah/s- 

 tris, and therefore on the way to disappearance, seems to be mere 

 special pleading — so long as a rostellum is effective, its size is of no 

 importance. It will be noticed that the author does not refer to 

 lieichenbach's characters — *' gymnostemium semi-teres, gracile " for 

 Cephal anther a ^ and " gymn. breve " for Epipactis. Yet this is a 

 distinction of considerable importance, the relatively long cylindrical 

 column of the former being in sharp contrast to the short squat 

 column of the latter. 



The author having thus minimized or explained away the differ- 

 entiating characters of previous writers, not unnaturally considers 

 that the time has come to re-unite the two genera. He gives a 

 beautifully drawn series of lips to show how they gradually merge 

 into each other, and form a homogeneous whole. 



He further points out that the epichile in CepJi. cucidJata has a 

 distinct though short spur, and that in the very similar lip of Limo- 

 dorum abortivi/m the spur is merely more developed. He therefore 

 considers that the genus Livwdoru)?i does not essentially differ from 

 Epipactis, and proposes that it should be included in the latter. 



Having never seen Ceph. cuciillata, 1 was anxious to learn what 

 I could about a plant which bulked so largely in the author's argu- 

 ment. Turning to the figure of it in lleichenbach's Icones (Tab. 

 120) I was much struck by its remarkable aspect. It suggests 

 abnormalit^^ The three lower leaves are represented by loose funnel- 

 shaped sheaths enclosing the stem, the 3 or 4 upper are bract-like, 

 erect, embracing the stem. The flowers resemble those of Cep)h. 

 (/7'andiilora, but have a very short spur. Altogether it is a most 

 remai'kable plant. Turning next to the figure of Limodorum ahor- 

 tivum in Tab. 129, I was astonished to find how extraordinarily 

 similar its leaves are to those of C. cucullata — the same three lower 

 perfoliate funnel -sha])ed sheaths, the same upper bract-like amplexicaul 

 leaves. In view of the unique character of the leaves of Limodorum 

 this resemblance can hardly be accidental. I think that anyone 

 expert in recognizing the characters of the parents in orchid-hj'brids, 

 on comparing these figures, will admit that there is a strong pre- 

 sumption that C. cucullata is a hybrid between Limodorum and some 

 species of Cephalanthera. The leaves are essentially those of the 

 former, just so much modified as might be expected from the influence 

 of Ceplialanthera, the column is that of the latter; whilst the short 

 spur (unknown in any other species of Cephalanthera) is just what 

 might occur in a cross between a moderately long-spuiTed and a non- 

 spurred flower, lleichenbach states (Z. c. 133) that he had very 

 recently been informed that a specimen had been found with a long 

 spur, /. e. nearer to Limodorum than the one figured. As is well 



