THE THREE FORMS OF LYTHRUM SALICARIA. 191 
the flowers are visited by insects, which would occasionally bring 
pollen from other flowers of the same or of any adjoining plant, 
as surely as occurs with the short-styled Z. salicaria, of which 
the pistil and corresponding stamens closely resemble those of 
L. hyssopifolia. According to Vaucher and Lecoq*, this species, 
which is an annual, generally grows almost solitarily, whereas the 
three preceding species are social; and this alone would almost 
have convinced me that Z. hyssopifolia cannot be dimorphic, as 
such plants cannot habitually live by themselves any better than 
one sex of a diecious species. 
Nesea verticillata.—1 raised a number of plants from seed sent 
me by Professor Asa Gray,and they presented three forms. These 
differed from each other in the proportional lengths of their organs 
of fructification and in all respects in very nearly the same way as 
the three forms of Lythrum Grefferi. The green pollen-grains from 
the longest stamens, measured along their greater axis and not 
distended with water, were 733, of an inch in length; those from 
the stamens of middle length 254%, and those from the shortest 
stamens 7%, of an inch. 
We have seen that the genus Lythrum affords trimorphie, dimor- 
phic, and monomorphic species. 
The inquiry naturally arises, why do these species differ so 
remarkably in their sexual relations? of what service can reci- 
procal dimorphism or trimorphism be to certain species, whilst 
other species of the same genus present, like the great majority of 
plants, only one form? Ihave elsewhere given too briefly f the 
* Géograph. Bot. de l'Europe, tom. vi. (1857) p. 157. 
+ ‘Origin of Species, 3rd edit., p. 101. Hugo von Mohl has recently (Bot. 
Zeitung, 1863, S. 309, 321), in a most interesting paper, advanced the case of the 
minute, imperfectly developed, closed and self-fertile flowers borne by Viola, 
Ozalis, Impatiens, Campanula, &c., as an argument against my doctrine that no 
species is self-fertilized for perpetuity. I may state that in the spring of 1862 I 
examined some of these flowers, and saw, though less thoroughly, all that 
H. von Mohl has so well described. I can add only one remark, which I 
believe is correct, that in V. canina there is an open channel for the pollen- 
tubes from the extremity of the stigma to the ovarium ; for I gently pressed a 
minute bubble of air repeatedly backwards and forwards from end to end. 
Though the imperfectly developed and the perfect flowers are so different in 
structure, it is a rather curious case of correlation, that in the double purple 
Violet (F. odorata) the minute imperfect flowers are double to the very core, 
80 that a section appears like the head of a cabbage when cut through. There 
can be, as von Mohl asserts, no doubt that these flowers are always self-fertilized ; 
they are moreover specially adapted for this end, as may be seen in the remark- 
able difference in the shape of the pistil in V. canina (and in a less degrec in 
V. hirta and the single V. odorata) as compared with that of the perfect flower ; 
LINN, PROC.—BOTANY, VOL. VIII. Q 
