610 



ECOLOGY AND EVOLUTION 



tion rate and low vagility, with the result 

 that distinctive populations may be recog- 

 nized in valley woodlands separated by 

 treeless ridges. Topographic and phyloge- 

 netic sequences are readily recognized 

 without significant ecologic differences. 

 Welch (19c>8, 1942) has given us a de- 

 tailed study of the subspecies and races of 

 two species of snails {Achatinella mustelina 

 and A. apexfulva) in Hawaii. Their distri- 

 butions show the effect of partial isolation, 

 and in general confirm the conclusions of 

 other students of speciation in these ani- 

 mals. Although many of the data can be 

 explained only by the action of heritable 

 variations and topographic isolation, Welch 

 does show a correlation of certain color 

 types with altitude (white colors of ind#^- 

 pendently evolved subspecies of A. apex- 

 fulva in highlands and dark patterns in low- 

 lands). Some selection causing convergent 

 evolution is thus indicated, even among 

 these animals that have provided a classi- 

 cal example of divergence based upon iso- 

 lation. 



Gall wasps (Cynipidae) show many 

 linear series indicating topographic isolation 

 (Kinsey, 1936). Species inhabiting differ- 

 ent isolated "islands" of oak woods are 

 more sharply distinct from their relatives 

 and have a more homogeneous population 

 than the "continental" species, which tend 

 to form clines of gradation without sharp 

 topographic demarcation. 



Topographic isolation is, in all proba- 

 biUty, the most important factor initiating 

 the separation of speciating populations. It 

 seems to be much more common on the 

 whole than habitat or other ecologic isola- 

 tion, and usually precedes physiologic, psy- 

 chologic, and genetic isolation, as well as 

 hybrid incapacitation. However, it is by no 

 means the exclusive mechanism by which 

 populations are divided. In our opinion, the 

 role of geographic isolation has been some- 

 what overemphasized by some authors 

 (e.g., Mayr, 1942; 1947). Without dimin- 

 ishing the importance of geographic separa- 

 tions in evolution, consideration of other 

 isolating mechanisms in the following 

 pages should give a broader perspective 

 on various types of reproductive isolation. 



HABITAT ISOLATION 



Instances of habitat isolation. Like those 

 of topographic isolation, that do not in- 



volve other types of separation are difficult 

 to discover. Ecologic differences emphasize 

 the role of selection in relation to isolation 

 even when combined with topographic or 

 spatial isolation. Closely related species hv- 

 ing together in the same geographic region 

 {sijmpatric species), but occupying differ- 

 ent habitats, should be carefully studied 

 and analyzed. Ancient topographic or ge- 

 netic isolation may be primary, and ecologic 

 isolation secondary in many sympatric spe- 

 cies, although this is not necessarily univer- 

 sal. The discussion in the following para- 

 graphs illustrates the importance of habitat 

 isolation without wholly excluding other 

 factors. 



Mayr (1942, p. 215) concludes that 

 ecologic specializations help to preserve 

 discontinuities between species, but are not 

 responsible for their origin (also see Lack, 

 1944), The present isolation of two sub- 

 species of mice, Peromysciis maniculatus 

 bairdii and P. m. gracilis, affords an ex- 

 ample of secondary habitat isolation, the 

 former living on lake beaches and the latter 

 in woods, both occurring together in Wash- 

 tenaw County, Michigan (Dice, 1931). In 

 this instance we are sure that geographic 

 isolation preceded habitat isolation because 

 of the circle of subspecies connected by 

 intermediate forms that occur to the west 

 {gracilis, borealis, osgoodi, nebrascensis, 

 bairdii). 



Blair (1947) reports different frequency 

 of color genes in deer mice (Peromysciis 

 maniculatus blandus) occupying different 

 colored soils 18 miles apart in New 

 Mexico, but no difference in the gene fre- 

 quencies of mice on different colored soils 

 only 4 miles apart. These differences in 

 color of the more separated populations 

 correspond to the shade of the background 

 soils. Studies of the pelage color of the 

 cactus mouse (P. eremicus) show com- 

 parable relations to the background color 

 in somewhat separated habitats (Blair, 

 1947a). Blair says: "The existence of the 

 local cactus-mouse populations distin- 

 guished by shade of pelage is attributed to 

 selection pressure that exceeds the homo- 

 genizing effect of interbreeding between 

 the several adjacent local populations" (see 

 pp. 627, 650, 668). 



Under special circumstances it would 

 seem possible that initial isolation may be 

 ecologic (Thorpe, 1945; Crombie, 1947), 



