126 THE FERTILISATION OF FLOWERS. [PART III. 
of which bears an upright hair. The limb of the corolla forms 
a rosy circle, finely notched at its circumference, indented by five 
deep incisions, and with its white centre surrounded by a purple 
ring, outside which are scattered white spots (Fig. 39,1). As soon 
as the flower has expanded and thus made itself attractive to the 
eminently colour-loving butterflies, the stamens, with their anthers 
coated all round with pollen, elongate and protrude one after 
another in rather quick succession from the narrow tube, so that 
every butterfly which inserts its proboscis must dust at least its 
head with pollen. . The five inner stamens still remain in the tube, 
rendering it too narrow for anything save the butterfly’s proboscis. 
As the outer stamens wither, the five inner ones protrude in succes- 
sion; but the two styles still lie twisted together within the tube 
(Fig. 39,2 and 3). Only after the stamens have withered, and 
have for the most part lost their anthers, do the styles elongate 
and allow the stigmas to emerge from the flower; the stigmas now 
spread apart, still retaining their spiral twist, so that from which- 
ever side a butterfly comes to suck honey it cannot fail to touch 
some part of the stigmatic papilla with its head. 
Pollen-seeking insects, if they only resort to young flowers 
whose anthers still have pollen, cannot act as fertilisers ; but they 
may do so if they alight, as flies often do, upon older flowers with 
mature stigmas, in which case they fly away again disappointed. 
Self-fertilisation is completely prevented by the well-marked 
dichogamy. 
As normal fertilisers of Dianthus deltoides, | -have only observed 
Pieris rape, L., and Satyrus Janira, L., but these I have noted 
repeatedly. Several Syrphide, viz., Rhingia rostrata, L., Meli- 
threptus scriptus, L., M. pictus, Mgn., and Helophilus pendulus, L., 
resort to the flower for the sake of pollen. I once noticed Rhingia 
rostrata make a vain attempt to reach the honey; it plunged its 
proboscis with a great effort into four flowers successively, but 
as its proboscis is only 11 mm. long it never reached the honey; 
in a short time it resigned the attempt, and betook itself to eating 
pollen. 
53. DIANTHUS CARTHUSIANORUM, L., has the same floral mech- 
anism as the preceding species. In Thuringia I have observed 
numerous visitors upon its flowers, consisting almost solely of 
Lepidoptera, 
(a) Rhopalocera: (1) Colias hyale, L., frequent ; (2) Rhodocera rhamni, 
L., very ab.; (3) Polyommatus Phloas, L.; (4) Hesperia, frequent; ()) 
