42 THE AMEBIC AN NATURALIST [Vol. LII 



organisms, like Daphnids and Kotatoria, different forms 

 appear in different regions. The ornithologists describe 

 different races for every river basin of the affluents of the 

 Amazonas; the mammalogists do exactly the same thing 

 for every area which was thoroughly covered. Where 

 breeding experiments have been carried on it has been 

 shown that the geographic races may be perfectly fertile 

 with each other and may produce fertile offspring. In 

 some cases, however, the transitional stages toward steril- 

 ity are found. Thus the production of intersexual moths 

 in crossing geographic races can be regarded as a step 

 toward increasing incompatibility, which in one of the 

 crosses attempted by us was an absolute one. In other 

 cases only a small percentage of the offspring of the hy- 

 brids could be reared, as in the crosses of the North and 

 South European Callimorpha dominula. We, therefore, 

 with many evolutionists, feel convinced that the geo- 

 graphic races are the most important visible steps in 

 species-formation in nature. 



If we now look into the characters distinguishing geo- 

 graphic races, we very often find certain qualitative dif- 

 ferences most conspicuous, for example, exchange of red 

 and yellow color in the moths. A close study of definite 

 examples, however, reveals that these differences are 

 often more conspicuous than important. This is shown 

 by the only group of information in the animal kingdom 

 which we have both by ecological and genetic work— the 

 geographic variation of land snails. The facts about the 

 extreme variability of Helix, Achatinella, Partula, etc., 

 are well known, as well as the irregularities in the con- 

 finement of definite types to definite localities. We have 

 been so fortunate as to gain some insight into these facts 

 through a very interesting collection which Dr. Haniel 

 made in Timor and studied under our direction (not yet 

 published). It was evident here, as in the other cases, 

 that a series of unit factors for number, color, form of 

 bands and ground color, which recombined freely, was 

 involved. And practically all the combinations could be 



