572 PEOP. E. B. POULTON: KATUEAL SELECTIOlf 



(5) Resemblances even within the Limits of an Order are 

 entirely independent of Affinity. 



The entire independence of affinity is specially well seenia the 

 groups of convergent moths and butterflies which are found in 

 different localities in South America. Although the resemblance 

 is clear enough in all the members of a large group, it is far closer 

 in certain species than in others. When these are examined they 

 are found not to be more nearly related than other members of 

 the group, but frequently the reverse. Thus it is very common 

 for a species of Heliconius to resemble with the most remarkable 

 precision a species of MeUncea or some other Ithomiine genus in 

 its locality. Such resemblance is in these cases far closer thari 

 that of the former to the species of the other genus (^ueides) in 

 its subfamily, and than that of the latter to any species in the 

 numerous related Ithomiine genera. Thus, to illustrate this 

 conclusion from some examples in the Hope Collection, in Hon^ 

 duras by far the strongest resemblance is to be found between a 

 Heliconius and a MeUncea ; and this is also the case (both 

 species being different) in Surinam, In Trinidad the resemblance 

 is closest between a Heliconius and a Tithorea. 



Under the theory of natural selection this association is readily 

 explicable. The pairs which thus form the centres of local 

 groups are probably the dominant forms, relying more com- 

 pletely than the other members upon the defence afforded 

 by their warning colours and the associated unpalatability. 

 As a matter of fact tbere is some evidence for their exceptional 

 abundance as compared with tbe otber members of their groups. 

 Ao'ain, they are usually more nearly of the same size than the 

 other members, so that the perfection of the resemblance in colour 

 and pattern is aided by resemblance in another quality. 



The theory of external causes entirely fails to account for 

 these facts. Uniform local conditions, if they can produce 

 any effect at all, must be expected to produce the closest likeness 

 where there is the closest constitutional similarity — in other 

 words in the more nearly related forms, in preference to the less 

 nearly related, in each locality. With the theory of internal 

 causes we should also expect the facts to be the reverse of those 

 which exist. At the best it is unable to account for the observed 

 phenomena. 



Any theory of selection (natural, artificial, or sexual) affords a 



