Mr. J. Miers on the genus Catliecb-a. 455 



tion of an anther advancing to maturity, where nine such cells 

 radiate upon its inner face, and in fig. 4 he gives a section of one 

 of these distinct cells filled with pollen-grains. Eudlicher repeats 

 the same view of the structure of the anthers in Viscum, as being 

 " multicellulosse, poris plurimis deliiscentes." Prof. Lindley con- 

 firms the same statement (Veg. Kingd. p. 790), where he says, 

 " in the genus Viscum the anther forms its pollen in a number 

 of distinct cells, as in JEgiceras, this being beautifully illustrated 

 by DeCaisne " (Mem. Acad. Brux. loc. cit.). Not having any 

 opportunity of examining the structure of the anthers in Viscum 

 album, I will not attempt to deny the facts vouched for upon 

 such authority, but I can speak with confidence to what I have 

 unquestionably observed in regard to their structure in dried 

 specimens in all the Brazilian species of Viscum that I have ex- 

 amined. Here they are constantly free and sessile upon the 

 base of the lobes of the perigonium, and on the margin of an 

 adnate cupuliform disk ; they are bilocular, the cells being par- 

 allel and of the whole length of the anther ; these cells are filled 

 with numerous pollen-grains, and are formed of very thick cry- 

 stalline walls, consisting of large cellules, radiating around each 

 central pollen-cell, as in Cathedra ; and these cellules are closed 

 at both ends by the external and internal facings of the walls, 

 which are thin and mai-ked with close-set parallel interrupted 

 lines, as are also the lateral divisions between these cellules. 

 On a future occasion I shall be able to show that other dif- 

 ferences exist in the structure of the seed, in the tropical 

 species of Viscum, which are not less striking than the fact 

 here recorded respecting the conformation of the stamens. 

 In my examination, I have not been able to perceive in any in- 

 stance, an indication of the bursting of these two cells for the 

 escape of the pollen, but at the apex the spaces over the cells 

 become depressed, and may be mistaken for two open pores, 

 which are also covered by a network screen, as before described, 

 in the anthers of Cathedra. I have met with several other cases 

 of somewhat similar structure in many genera of the Sanialacea, 

 and in some instances also in Olacacece, where however a form of 

 anther prevails, which, though somewhat analogous, must not be 

 confounded with the structure above described. Here the an- 

 thers are distinctly 4-lobed, 4-celled and 4-valved, often more or 

 less crustaceous in texture, the cells opening alternately right 

 and left by the evolution of the valves, which separate from the 

 connective longitudinally by one of their margins, so that when 

 these are rolled back the pollen is discharged, and they appear 

 as if the anthers had been 2-celled and 4-valved, according to 

 the usual mode of construction. The same development often 

 occurs in Loranthacea proper, in Rhamnacece, Celastracece, and 



