Page:The Zoologist, 4th series, vol 2 (1898).djvu/110

78

much has been written on the relations between insects and flowers, it must be confessed that the information we possess on this fascinating subject is fragmentary indeed in comparison with what might be known; in other words, there is no locality where flowers grow and insects fly in which new and interesting observations may not be made, while there are whole regions from which we have practically no records.

Hermann Müller, in his 'Fertilisation of Flowers,' gives what might at first sight seem a very complete array of facts, but we find him strongly insisting on the incompleteness of his researches. In America the subject has only been seriously attacked by one observer, Mr. Charles Robertson, whose observations are confined to Illinois and Florida.

The subject is more complicated than might at first be imagined. Repeated observation only confirms the validity of the following rules:—

(1.) Observations made in one year should be repeated in other years, as the results of different years may greatly differ.

(2.) Observations made on a plant in one locality should be repeated in other localities throughout the range of the plant, as the insect visitors are often different in different parts of the plant's range.

(3.) Observations made on plants growing in cultivation, away from their natural habitat, prove little regarding the natural visitors of the plants.

(4.) Observations on the Honey-bee prove little regarding the actions of wild bees; each species of bee must be observed separately, its habits cannot be certainly inferred from observations on other species.