434 
bending outwards and backwards ; the filaments are red- 
coloured towards their end ; the style overtops the longest 
stamens by some millimetres. A species of bee, MZac- 
vopis labiata Pz, frequently visits these flowers for 
pollen. It comes first into contact with the stigma, and 
supplies it with pollen from previously visited flowers, 
thus regularly effecting cross-fertilisation. But if we pre- 
vent the visits of insects by covering over the stems bya 
Fic. 9.—Luphrasia officinalis. Lateral view of a flower of the largest o-; 
just opened. 
Fic. 10 —Position of the stigma (s?), and of the anthers (a*, a*) of the same 
flower in a more advanced state. 
Fic. 11.—Two anthers, seen from the inner side, showing the slits fringed 
with hairs, 
net, self-fertilisation scarcely takes place, in consequence 
of the style overtopping all the stamens, 
Specimens of the same species with less conspicuous 
flowers are found in shady ditches. The petals of these 
plants are lighter yellow, uniform in colour, without any 
red at the base, on an average 10 mm. long, and 5 mm. 
wide ; they only open slightly, remaining nearly upright, 
Fic. r2.—Lateral view of a flower of the smallest form, just opening. 
Fic. 13.—Position of the stigma (s¢) and the anthers (a*, a?) in this flower. 
Fic. 14.—Front view of a flower of the same form, in a more advanced state. 
(All the figures are magnified in the proportion ct 7:1. The hairs of the 
calyx and the coloured spots and lines of the corolla are omitted.) 
but diverging obliquely ; the filaments are greenish yel- 
low, without any red towards their end ; the style hardly 
equals the two lowest and longest stamens. The 
stigma comes without any external agency into con- 
tact with the pollen of the same flower thus regularly 
experienciyg self-fertilisation. This manner of producing 
seeds is an indispensable condition for preservation of 
NATURE 
| Sepz. 25, 1873 
this variety of Lysimachia vulgaris. For in consequence 
of its shady habitat, and of its lower degree of attractive- 
ness for insects, its flowers are but very rarely visited, and 
it would be exposed to extinction without the possibility 
of propagation by self-fertilisation. I but once observed 
a little fly of the family of Syrphide, Syrit/a pipiens i 
eating the pollen of this shady form of flowers. Although 
this fly might possibly transport pollen from one flower to 
the stigma of. another, cross-fertilisation was nevertheless 
by no means more probable than self-fertilisation. 
The two forms here described of the flowers of Lysi- 
machia vulgaris graduate into each other by connecting 
forms, which are met with in intermediate localities, for 
instance on the sunny edges of ditches. 
Another example of the same sort of dimorphism, 
even more striking than that just mentioned, is pre- 
sented by Luphrasia officinalis. Of this species flowers 
are found in different localities of a very different size. 
But the more the attractiveness for insects is increased 
by the size of the corolla, the more is cross-fertilisation 
secured in case insects visit the flowers, self-fertilisation 
at the same time being prevented ; while on the contrary, 
the smallest flowers regularly fertilise themselves, even 
without the visits of insects.’ I will attempt to explain 
these peculiarities by drawings of the largest and of the 
smallest form of flowers I have hitherto been able to 
find. 
In the flower just opened of the largest form (as shown, 
in Fig. 9), the stigma, already in a mature condition, 
greatly overtops the anthers. Therefore an insect,* in- 
serting its proboscis into the tubular corolla in order to 
gain the nectar contained at the bottom of its tube, first 
grazes the stigma charging it with pollen-grains from — 
flowers previously visited, and then pushes against the 
two hairs (sf) which project from the two lower anthers : 
(2°) into the middle of the entrance to the corolla. This 
shaking of the hairs is transmitted to all the four anthers, 
which lie close together and are soldered together by 
their upper margins, and a small quantity of the smooth 
powdery pollen-grains falls out of all the pollen sacs. 
The slits in the pollen sacs being fringed with hairs 
directed downwards (as shown in Fig 11) a lateral dis- 
persion of the pollen-grains is prevented ; all the pollen- 
grains shaken out fall directly downwards upon the pro- 
boscis, enabling it to fertilise the next flower visited by 
the insect. 
In the state just described the corolla has not yet at- 
tained its full size. Growing farther, it at length equals 
the stigma by which it was at first so much overtopped, and 
now the mutual position of the stigma and the anthers is 
that shown in Fig. to, When occupying this position, 
the stigma is always already shrivelled and brownish 
coloured, and is no longer capable of being fertilised, 
Self-fertilisation is therefore quite impossible, ; 
The probability of cross-fertilisation and of self-fertilisa- 
tion is directly opposite in the flowers of the smallest 
form, presented by Fig. 12—14. Whilst in the flowers of 
the largest form, as just described, the anthers remain 
soldered together, and do not scatter their pollen unless 
the hairs are shaken, in the flowers of the smallest form the 
anthers separate from each other, and scatter nearly all 
their pollen long before the corolla has fully opened. The 
end of the style, moreover, bends inwards so much as to 
bring the stigma (as Fig. 13 shows) close beneath the 
upper anthers. Therefore, on examining a flower hardly 
half-opened (Fig. 12), we always find the stigma already 
largely charged with pollen-grains of the same flower. 
When fully opened, the flowers of the smallest form show 
the stigma in a shrivelled and brownish coloured condi- 
tion, lying between the separated and emptied pollen- 
sacs (as shown in Fig. 14). Hence cross-fertilisation — 
could scarcely be effected, even if insects (which I never 
* I observed four species of bees and three species of Diptera visiting the 
flowers of Euphrasia officinalis for honey. 
