262 THE VARIATION OF ANIMALS IN NATURE 



However, it would appear that in many instances the species 

 of models are extraordinarily variable. Thus in Acraea, 70 

 out of 133 species have at least two distinct colour-forms (a 

 good number of the species with no known variety are still 

 very rare in collections) ; 46 species have three or more named 

 forms. Sometimes as many as half a dozen forms of a species 

 occur sporadically throughout the range, while in other cases 

 there is sexual polymorphism or marked geographical variation. 

 Something of the same sort would appear to be usual in 

 Heliconius also. In view of this variability it is difficult to 

 maintain that the broad features of colour-pattern are essential 

 specific differences. Of course there are a certain number 

 of species with a distinct colour-pattern unlike any other, 

 but the more general position would appear to be that the 

 main lines of colour-pattern are non-specific, and that specific 

 characters are found more in the male and female genitalia 

 and in the finer details of the pattern, such as the exact shape 

 of bands or the exact number and position of spots. In this 

 connection we may note the example given by Jordan (1896, 

 pp. 449-50). In Malaya, Papilio caumis is a striking mimic 

 ofEuploea rhadamanthus. Races of P. caumis, inhabiting Malacca 

 and Sumatra, Borneo and Java, may be separated by slight 

 differences in the size of the white markings. These subspecific 

 characters do not affect the general resemblance to the model, 

 which is unmodified throughout the area. 



When we turn to the mimics we find the same extreme varia- 

 bility. The association between polymorphism and mimicry 

 has long been emphasised, and in many cases, as in the well- 

 known Papilio dardanus (Eltringham, 1910, p. 91), several forms 

 of one species may all occur in one place. In these highly 

 polymorphic mimetic species colour-pattern by itself is almost 

 of no value in specific diagnosis, and we appear justified in 

 maintaining that, in a number of cases, mimetic differences 

 are not of the same nature as specific differences. In simple 

 Batesian mimicry colour-pattern is much more closely asso- 

 ciated with specific difference, but this sort of mimicry does 

 not appear to form a very large proportion of the known 

 examples in butterflies. 



It will be objected that, even if the colour differences 

 between these species involved in mimicry are not actually 

 specific, they are still analogous to the differences observed 



