G. M. Thomson. — On the Fertilization of Floioering Plants. 253 



work oil " British Wild Flowers iii relation to Insects," Sir John Lub- 

 bock gives a memorandum of Mr. Darwin's on the fertilization of F. tricolor, 

 the common heartsease or wild pansy, which is applicable, to a great 

 extent, to the whole genus. He remarks how rare it is to see any insect 

 visiting the flowers, and how he watched many times daily in a fortnight 

 before he saw a single bee visiting a certain clump of heartsease. I have 

 myself repeatedly watched patches of om- violets, and have never seen 

 them visited by any insect. 



Species of thrips are to be found in these flowers — as, indeed, in most 

 flowers — and Mr. Bennett considers that the fertilization of F. tricolor is 

 due to this insect. It seems to me, however, that this view must be erro- 

 neous. Thrips is an insect which lives in the flower it frequents, feeding 

 probably on the pollen, and only flying to other flowers to lay its eggs in 

 them. In the course of its running about inside a flower, it probably fre- 

 quently carries pollen from the anthers to the stigma, but this would only 

 bring about self-fertilization, whereas the flowers are specially constructed 

 to avoid this. Even where several flowers are crowded together in the 

 same inflorescence, if the pollen of each was distributed to others, the 

 fertilization — if it took place — would not be that of a true cross. 



I have not verified, experimentally, whether the ordinary flowers of both 

 our violets are self-fertile, but h'om the fact of frequently finding withered 

 flowers in which the ovaries showed no signs of enlargement, I think it 

 improbable that they are. I have never been able to detect any scent in 

 the flowers of either F. filicaidis or F. cunninghamii. 



The genus Melicytus contains four New Zealand species, all of which 

 occur in Otago, but I have only examined two of them. The name of the 

 genus — literally, honey-cavity — suggests the occurrence of abundant nectar. 

 This is secreted by the so-called scale at the back of each anther. This is a 

 thick succulent club-shaped organ, on the apex of which a bead of honey is 

 produced. The flowers are polygamous, ranging from perfect unisexuality 

 to complete hermaphroditism. The male flowers occasionally have im- 

 perfect pistils, though the total want of the pistil is commoner. The female 

 flowers, on the contrary, seem always to be furnished with stamens, though 

 the anthers are fi.-equentiy abortive, and the reason of this may be the 

 adaptation of part of the anther for the secretion of honey. 



M. ramijiorus produces enormous numbers of flowers of a greenish- 

 yellow hue, conspicuous by their mass, fragrant and abundantly supplied 

 with nectar. These cannot be wind-fertilized growing as they do in dense 

 bush where the wind can hardly affect them, and having somewhat waxy 

 pollen. 



M. lanceolatus has relatively larger flowers, and they are individually 

 more conspicuous, being yellowish in then" lower part, and having the 



