THE ANGIOSPERMAE 1307 



need for food or shelter, which bring insects to flowers, but it has consider- 

 able secondary value as a means whereby flowers may be distinguished from 

 one another and in this sense it is an attraction to an insect like a bee which 

 is engaged in working one particular type of flower. The colour perceptions 

 of insects appear, however, to be limited and in some cases difl'erences of 

 colour may mean no more than variations in brightness to the eye of the 

 insect. It is important to remember that the eye of an insect does not 

 recognize colour as we do and that some colourings which to us look 

 similar may to the insect visitor seem sharply contrasting, while in other 

 cases colours which to us appear quite distinct may to the perception of the 

 insect appear almost alike. (See also p. 1261.) 



Flower colour is frequently not restricted merely to the petals. In many 

 cases, both petals and sepals become coloured. Very often they assume the 

 same basic colour, though the tint may vary, but in other species the two 

 whorls of floral parts may develop contrasting colours. In certain cases the 

 filaments of the stamens become coloured and may contrast with the colour 

 assumed by the anthers or the perianth. 



Miiller originally suggested, what has subsequently been largely proved 

 to be true, that insect visitors show a preference for certain colours. Bees 

 in general seem to prefer blue flowers; butterflies, red; hover flies, yellow; 

 and carrion flies, brown or purple. In early times considerably greater 

 importance was attached to these matters than is done today. Indeed it has 

 been questioned whether insects can distinguish colour at all. Experiments 

 by von Frisch with Hive Bees have, however, proved that they have the 

 ability to recognize difl^erent colours and there is therefore no reason to sup- 

 pose that a similar faculty is not shared in some degree by other insects. 



Some plants, like the Foxglove, produce one-sided inflorescences even 

 when growing in entirely open situations. It has been suggested that such 

 inflorescences ensure that more flowers are pollinated, because insects tend 

 to travel upwards from the bottom to the top of a spike rather than to work 

 round it. Radially arranged inflorescences are not so uniformly pollinated 

 and there is a marked difference in the number of fruits developed on the 

 symmetrical inflorescence of a Lupin compared with the one-sided spike 

 of a Foxglove. 



Sce7it 



To what extent the scent or odour of flowers assists in pollination is a 

 matter of some dispute. The question as to the ability of insects to distin- 

 guish different odours has not been fully settled. There seems to be little 

 doubt, however, that flies are attracted by the nauseous odours emitted by 

 some flowers. Many such flowers combine purple and dull red colours and 

 are much sought after by carrion flies. Other yellowish-green or white 

 flowers emit malodorous scents to which dung flies and some beetles are 

 attracted. Both smell and colour seem necessary to attraction. On the other 

 hand many flies are equally attracted to strongly scented, but not malodorous 

 flowers. It is obvious, however, that smells may produce entirely different 



