ECOLOGY OF ISLE ROYALE. 187 



criterion can be illustrated by reference to the-Ajax Butterfly {I. ajax) . 

 Tbe sole food plant of the Ajax larva is the Pawpaw, a shrub clearly 

 of tropical origin. The allies of Ajax are also tropical; thus the asso- 

 ciated biogeographic (plant and animal) aflSnities clearly point to the 

 tropics. It is this combination of certain ecological relations or associa- 

 tions which show biogeographic affinities. Thus food and other habits 

 and instincts become of special value. Here also belongs a large class 

 of ecological relations, particularly those related to the succession of 

 insect associations. The great dependence of insects, as a class, upon 

 vegetation necessitates a close relation between the succession of plant 

 associations or societies and certain species of beetles. If certain mem- 

 bers of a biotic (plant and animals) association or society have certain 

 geographic affinities, others associated with them are likely to have 

 similar affinities (cf. Horn '72, ji. 384). This phase is not identical with 

 the idea ol faunal or floral affinities, it includes them and the relation of 

 Motic af<sociation, pai-ticularly as members of a climax association or 

 formation, when geographic affinities are to be determined. 



This criterion is of very extensive application. It is really a group 

 of criteria and not a single one, because associations include not only 

 organisms in close proximity, but also commensals, symbiots, parasites, 

 etc. Seasonal phenomena might well be included within this class. 



10. Least dependence upon a restricted habitat. From the stand- 

 point of animal associations this is a criterion which may be ex- 

 pected to have a rather extensive application. Its most conspicious 

 application is to that of dispersal. Out-lying colonies tend to have 

 a limited or restricted range. At the same time such colonies are 

 particularly liable to become extinct, as they are usually near the 

 limit of favorable conditions. Often beetles in such a location are 

 dependent upon a single food plant, etc. This is true of the "boreal 

 islands" in swamps within the glaciated portion of the continent. For 

 example, members of the tamarack bog association, toward their southern 

 limit, have very restricted or local range; but to the north, the bog 

 forest conditions, as it were, spread from the bogs proper and become 

 of extensive geographic range, as the water beetles invade the damp 

 mosses (Wickham, '97, p. 126). The outlying tropical "islands" border- 

 ing the Rio Grande, as described by Schwarz ('01) and Wickham ('97a), 

 apparently illustrate the same phenomena. ' These restricted, attenuated, 

 or isolated colonies, dependent upon special conditions, are clearly 

 indicative that they are pioneers or relicts, which point toward 

 the region where their range is spread out and becomes of geographic 

 extent. But it does not follow that every isolated habitat has such 

 a meaning. In general, a study of succession in the region will deter- 

 mine to which class the colony belongs, pioneer or relict. 



There is an exception to this criterion in the case of semi-aquatic 

 or aquatic animals in an arid region. In such regions the springs, 

 streams, and water basins are so limited in extent that their isolation 

 is conspicuous; and yet these conditions may be very favorable to the 

 formation, or at least preservation, of new variations and species. 

 Thus an arid region may be particularly favorable, in a sense, to the 

 formation of varieties and species, although individuals may- not be 

 numerous. In such cases the amount and kind of differentiation with- 



