492 Agnes A rber . — On the Structure of the Androecium in 
I have also to thank Mr. George Goode, Miss Ida Roper, F.L.S. and 
Dr. Vigurs, to whom I am indebted for supplies of Hypericum Elodes. 
In examining the anatomy of flowers, in cases where herbarium material 
alone is available, I have obtained good results by boiling, treating with 
medium chromacetic acid for 48 hours, and then washing and dehydrating 
in the usual way. When such material is to be microtomed, it seems 
important to make the passage from xylol into paraffin a very gradual one. 
When whole flowers are used, it is an advantage to keep them immersed in 
paraffin for an unusually long period — even as much as three weeks— in 
order to ensure good penetration. Material treated in this way stains well 
with the Bismarck brown, gentian violet, and orange combination. 
This investigation, which was originally begun at University College, 
London, has been completed at the Balfour Laboratory, Cambridge. I am 
indebted to the Balfour Laboratory Committee for giving me all facilities 
for my work. 
II. The Anatomy of the Androecium in Parnassia. 
(i) The Anatomical Relation of the Androecium to the other Whorls of the 
Flower in PARNASSIA PALUSTRIS, L. 
The structure of the androecium of Parnassia is more easily followed 
when the general vascular symmetry of the flower is understood, and it may 
hence be well to begin with an outline of the anatomy as a whole. So far 
as I am aware, the only contribution to this subject, up to the present, is 
Eichinger’s 1 observation that in P. palustris the staminode bundles are 
connected with the petal bundles. I can confirm this statement, as will be 
seen from the following account. 
In the four species which I have examined (P. palustris, L.; P.fim- 
briata, Banks ; P. montanensis y Fern, et Rydb.; and P. parviflora, DC.) there 
are generally three arcs of vascular tissue in the pedicel below the base of 
the flower (Text-fig. 1, A), although, especially in P. fimbriata, it is not 
unusual to find a closed ring instead of three detached arcs. The mode of 
origin of the bundles destined for the different whorls is naturally somewhat 
complicated, since none of these whorls are trimerous, and yet they are all 
supplied from the three pedicel strands. The process, which I have followed 
in detail in the case of three flowers of Parnassia palustris, is represented 
diagrammatically in Text-fig. 1, in which the strands destined for the 
gynaeceum are omitted. Slight variations occur, due, apparently, to the 
fact that one of the three pedicel strands may be smaller than the others, 
and hence may take less part in providing for the flower, but in all essentials 
the scheme proved to be identical in each of the cases examined. The 
1 Eichinger, A. : Beitrag zur Kenntnis und systematischen Stellung der Gattung Parnassia. 
Beikefte zum Bot. Centralblatt, Bd. xxiii, Abth. ii, 1908, p. 303. 
