64 NATURAL SELECTION FROM 



when most needed.' In some cases the secretion 

 begins at a very early stage of the plant and soon 

 fails, leaving the maturing and adult plants un- 

 protected. The nectaries secrete spasmodically 

 and are often dry; and the nectar of many forms 

 is avoided by ants and other animals. There is 

 no relation between- mutilated flowers, ants, and 

 extra-floral nectaries. Most mutilated flowers 

 produce as many seeds as those that are not; and 

 the honey-seeking ants are not combative and do 

 not attack other insects visiting their host. If 

 these extra-floral nectaries have been developed 

 and perpetuated by Natural Selection, it is an 

 illustration of the selection of harmful structures, 

 for they often attract insects of all kinds, which 

 damage the plant in various ways. The investi- 

 gation showed that individual plants which se- 

 crete little or no nectar are less harmed by insects 

 than are those that produce nectar. 



It is such work that is playing havoc with the 

 " adaptations " of botanical literature, and is 

 forcing botanists to see in these various structures 

 inevitable responses to conditions that have noth- 

 ing to do with adaptation. It would be going 

 too far to say that such results destroy absolutely 

 all faith in the selection and development and fix- 

 ing of adapted structures, but they do tend to 

 weaken faith and to demand that every claimed 

 case shall be subject to rigid experimental inves- 

 tigation. That there must be selection no one 

 pretends to deny, so far as I know, but when the 



