290 ME. G. BE^THAH OF OKCHIDE^!. 



turbed by the process of pressing in drying, that its true form and 

 relation to the rostellum is difficult to ascertain. It is very likely, 

 therefore, that some of the statements here made may require 

 considerable modification from further observation. 



In the Epidendreae generally, before the anthers open, the 

 pollen-masses are either quite distinct or are more or less con- 

 nected in each cell, on the side next to the opening of the cell, 

 by a substance formed of pollen-grains loosely connected by 

 a tissue of highly elastic threads. This substance is exceedingly 

 variable in appearance and amount in different genera. In 

 those forming several of the first subtribes it is usually very small, 

 or more frequently disappears altogether; in the latter tribes it 

 is more often abundant and definite, and has been inappropriately 

 termed caudicle, and taken as the distinctive character of Epi- 

 dendreae as separated from MalaxideaB. In several genera of the 

 Erica? it is so variable in different species that the genera have 

 been placed by some in Malaxidea?, by others in Epidendreae. In 

 some Bletiea? it is very abundant, and almost envelops the waxy 

 masses, or these are distinctly formed only at so late a stage that 

 they have been overlooked, and the genera placed in Arethuse© 

 (now merged in Neottiea?). "When the flower expands, the anther 

 will sometimes fall away entire, with its pollen-masses ; but, 

 generally speaking, as soon as the anther-cells open, which often 

 takes place in the bud, the exposed part of the pollen-masses (their 

 points when they are pear-shaped or more or less acuminate or pro- 

 duced into short caudicles) becomes endued with a transparent, 

 very viscid, almost liquid substance or viscum, by means of which 

 the masses are connected together and adhere to any insect or 

 other extraneous object with which they come in contact. This 

 viscum has been shown by Darwin, Hooker, and others to exude 

 in many cases, and perhaps in all, from the rostellum. It is 

 sometimes so scanty as only to be detected by the fact of the 

 pollen adhering to extraneous objects, and even to have failed 

 entirely where the pollen-masses are seen loose about the flower 

 in or out of the anther-case ; in others so abundant as to com- 

 pletely envelop the rostellum and anther ; sometimes, on opening 

 a bud just ready to expand, I have found the whole inside a mass 

 of viscum, from which it was difficult to extract the pollen. 

 After the flower expands, if the pollen is not immediately carried 

 away, this viscum will in some instances dry into a highly elastic 

 thread, which may remain attached at one end to the rostellum 



