66 INTRODUCTION 



10. Flowers with adaptation B, but in which the pollen also is more or less 

 completely concealed (BB). Zygomorphous. Viola, Anthyllis, Trifolium, Lotus, Vicia, 

 Linaria, Alectorolophus. 



Verhoeff represents the most important relationships between his ten chief 

 stages of adaptation in the following scheme : — 



W 



1 



PoA 



i ' 

 A 



; 



AG ^ AB 

 i 

 ^B 

 ^./^i PoB 



BG ^ BB_— — -^ 



The classification presents no advance upon that of Miiller, for the Zygo- 

 morphous Pollen-flowers (Sarothamnus, Genista, Ulex) are such well-marked 

 Hymenopterid Flowers, that they cannot possibly be separated from the Nectar- 

 flowers of this group (Anthyllis, Trifolium, Lotus, Vicia, &c.). Such separation 

 would be like the artificial division of the Papilionaceae into two separate classes 

 (XVI and XVII) in the Linnean classification, according as they are monadelphous 

 or diadelphous. And it may be added that the term ' nectar-flowers ' is applicable 

 not only to the extremely simple nectar-containing flowers of Salix, but also to 

 Verhoeff's classes A, AG, AB, B, BG, BB. 



Loew (Beob. iiber den Blumenbesuch von Insekten . . . Weit. Beob. liber den 

 Blumenbesuch . . . , Beitr. zur blutenbiolog. Statistik) has arranged the- floral 

 classes of Hermann Miiller (excluding, however, classes D and Kl) in the following 

 three groups : — 



I. Allotropous Flowers : these are adapted to various kinds of insects 

 possessing a short proboscis. To this group belong classes W, Po, A, and AB. 



II. Hemitropous FlOAvers: these are imperfectly adapted to some definite 

 set of insects possessing a proboscis of medium length. To this group belong 

 classes B and B'. 



ni. Eutropous FloTwers : these are more or less exclusively adapted to 

 a definite set of insects possessing a long proboscis. To this group belong 

 Bee Flowers, Humble-bee Flowers, and Lepidopterid Flowers. 



Loew distinguishes three groups of insects, the visitors, respectively, of allotro- 

 pous, hemitropous, and eutropous flowers. These will be discussed later. 



Taking into account the observations first made (1892) by W. Burck, and 

 confirmed (1897) by J. H. Hart, which prove that bats act as poUinating-agents, 

 and collating the groupings of Delpino and Miiller, the follov.'ing classification 

 of plants, according to their flower pollination, may be advanced ' : — 



' Kerner ('Die Scliutzmittel des Pollens,' pp. 45, 46, note) employs the term Kangaroo fiowers in 

 describing the floral arrangements of Dryandra (Proteaceae). The flowers are inserted on the rim 



