IX WARNING COLORATION AND MIMICRY 253 



fact, that insect-eating birds only learn by experience to 

 distinguish the edible from the inedible butterflies, and in 

 doing so necessarily sacrifice a certain number of the latter. 

 The quantity of insectivorous birds in tropical America is 

 enormous ; and the number of young birds which every year 

 have to learn -wasdom by experience, as regards the si)ecies of 

 butterflies to be caught or to be avoided, is so great that the 

 sacrifice of life of the inedible species must be considerable, 

 and, to a comparatively Aveak or scarce species, of vital im- 

 portance. The number thus sacrificed will be fixed by the 

 quantity of young birds, and by the number of experiences 

 requisite to cause them to avoid the inedible species for the 

 futui^e, and not at all by the numbers of individuals of which 

 each species consists. Hence, if tAvo species are so much 

 alike as to be mistaken for one another, the fixed immber 

 annually sacrificed by inexperienced birds will be divided be- 

 tween them, and both will benefit. But if the two species are 

 very unequal in numbers, the benefit "wall be comparatively 

 slight for the more abundant species, but very great for the rare 

 one. To the latter it may make all the diff"erence between 

 safety and destruction. 



To give a rough numerical example. Let us suppose that 

 in a given limited district there are two species of Heliconidge, 

 one consisting of only 1000, the other of 100,000 individuals, 

 and that the quota required annually in the same district for 

 the instruction of young insectivorous birds is 500. By the 

 larger species this loss will be haixlly felt ; to the smaller it 

 Avill mean the most dreadful persecution resulting in a 

 loss of half the total population. But, let the tAvo species 

 become superficially alike, so that the birds see no diff"erence 

 betAveen them. The quota of 500 Avill noAv be taken from a 

 combined population of 101,000 butterflies, and if joropor- 

 tionate numbers of each suff'er, then the Aveak species A\ill 

 only lose five individuals instead of 500 as it did before. 

 NoAV Ave knoAv that the diff"erent species of Heliconid^e are 

 not equally abundant, some being quite rare ; so that the 

 benefit to be derived in these latter cases Avould be very im- 

 portant. A slight inferiority in rapidity of flight or in poAvers 

 of eluding attack might also be a cause of danger to an in- 

 edible species of scanty numbers, and in this case too the being 



