192 LIFE : OUTLINES OF GENERAL BIOLOGY 



otiicrs and give it the first trial. 'H»e>- will do this repeatedly, even 

 though the meal is not forthcoming. This proves a registration of 

 colour as colour, but it has its limitations. Thus hive-bees seem to 

 mix up colours that are near one another, such as yellow and orange, 

 or blue and violet. Though they can see the ultra-violet rays that 

 arc invisible to us, they seem to be colour-blind to scarlet. Yet they 

 may 1h^ attracted to a brilliant scarlet flower whose surface reflects 

 much light, more than to a dull flower othenvi.se equal. To give 

 llower-visiting insects credit for evolving the colours of flowers is 

 credulous, but given physiological reasons for the presence of floral 

 pigments, and some of these reasons are now becoming apparent, 

 it is justifiable to .say that bees and other colour-discriminating 

 insi»cts may have played a part, throughout the ages, in favouring 

 flowers with certain colours and in eliminating others. It is unneces- 

 sary to use such words as "prefer"; what has been proved is that 

 certain insects build up u.seful associations with particular colours, 

 and more readily with these than with other colours. More experi- 

 ments are necessary before we can say much as to inborn suscepti- 

 bility to certain colours rather than to others. 



To a less extent than odour and colour, the brilliance and the 

 shajx? of flowers may be of imjx)rtance in securing the visits of 

 insects. 



In warm countries, and very markedly in Java, pollination is 

 sometimes effected by birds, such as humming-birds, sun-birds, and 

 honey-suckers, which suck up the abundant and unusually fluid 

 nectar. The list of habitual bird-pollinators, as dependent on the 

 flowers as the flowers are on them, is already long, and it is rapidly 

 growing. Hut whili* insects and birds are both important, other 

 animals are trivial; snails occasionally carry pollen from blossom to 

 blossom in their slow-going way, and Freycinetia is actually pollin- 

 ated by bats. 



Ai>VANT.\c,i:s OF Cross-pollination. — The frequency of cross- 

 IK)llination and of arrangements that prevent self-pollination cannot 

 be overlooked, but it is doubtful whether there is in the latter as 

 such any deteriorative danger, or in the former any positive 

 promotion of vigour or of variability. The results of experiment are 

 strangely discrepant. Probably the most prolonged experiments are 

 those on maize, which is naturally a cross-pollinated plant, but can 

 Ix* artificially self-i>ollinated. This has been studied for twelve or 

 mf)rc successive generations, and the first results confirmed Darwin's 

 general conclusion that self-pollination deteriorates vigour and 

 prcKluctivity. Without there being any actual degeneration, the 

 successive s<'lf-iH>llinated crops of maize plants showed reduction in 

 size and prcKlurtiveness. but only to a certain point, after which 

 there was stability .igain. and a marked uniformity. There was, 

 however, a sifting out into true-breeding sub-varieties marked by 



