166 INTRODUCTION 
Class H, and also show the same liking for dark-coloured blossoms. Anthophora 
pilipes /. (=A. acervorum Z.), which has a proboscis 19-21 mm. long, visits almost 
exclusively the flowers of Class Hh, with nectar concealed as deeply as possible, 
and agrees precisely in this respect with Bombus hortorum Z., which has a 
proboscis of the same length. Owing to its earlier appearance, however, it is 
only found on spring-flowers, such as Corydalis cava, Pulmonaria officinalis, 
Lamium purpureum, and species of Primula. Later on in the year its visits to the 
longest-tubed humble-bee flowers are replaced by those of Bombus hortorum. 
Eucera longicornis Z., in accordance with the length of its proboscis (10-12 mm.) 
visits bee flowers by preference, especially papilionaceous ones. This bee also shows 
a special liking for dark-coloured flowers, and the same is true for species of Mega- 
chile, Osmia, and most other long-tongued Apidae (except Heriades truncorum Z.). 
Loew (‘ Blumenbesuch,’ I, p. 54) shows that this colour-preference is not dependent 
upon the length of the proboscis, but is a specific peculiarity. 
The long-tongued bees exhibit very great constancy in their choice of flowers. 
I saw (‘ Bliitenbesucher,’ I, p. 4), for example, the honey-bee, for the space of several 
minutes sink its proboscis almost every second into one flower after another of 
Trifolium repens, without seeking out any other species. Sometimes it spent a few 
seconds sucking from one flower, but on an average it made thirty to forty visits 
in a minute. 
I noticed exactly the same thing when I observed the visits made to flowers 
by another long-tongued bee. On the 2nd of May, 1897, I had the opportunity 
of counting, watch in hand, such visits in the case of Anthophora pilipes /. 9 
(=Podalirius acervorum Z.). The bee visited Lamium purpureum, and had already 
collected considerable masses of orange-yellow pollen: it now flew from flower to 
flower of the species named, doing nothing but suck nectar. Only once, in passing, 
it sucked from Viola tricolor var. arvensis. In two and a half minutes it made 
seventy-two visits, so that on an average a visit lasted about two seconds. The 
duration of the act of sucking was as long as five seconds in some flowers which 
apparently afforded a richer booty than most others. As a rule, however, the 
bee stayed for a second, sinking the proboscis rapidly into the base of the flower, 
withdrawing it as rapidly, and then buzzing away to another blossom. During 
the acts of sucking the humming noise emitted during flight was of course inter- 
rupted, so that the number of separate visits could be determined from these 
interruptions. As I had heard the Anthophora from one to two minutes’ before 
I saw it, and during this time the humming was interrupted every second or two, 
it follows that the bee exerted its activity among the flowers for at least four minutes, 
and in this time pollinated far more than roo flowers. 
E. Loew (‘ Blumenbesuch,’ I, p. 93) has proved for almost all genera of bees, 
that besides length of proboscis, some other adaptational factor helps to determine 
the selection of flowers, and may have such an influence that, e.g., a long-tongued 
1 I did not measure this period with the help of my watch, but by counting the number of my 
respirations. As I breathed nineteen times a minute, this gave about three seconds for each 
respiration, and afforded me an excellent way of estimating small periods of time with great 
accuracy. 
