ZOOGEOGRAPHY 373 



Endemic 33% 



South American 36% 



Central American 15% 



Antillean 3% 



Cosmopolitan 3% 



Unattached 3% 



Six genera have nearctic species [Thesium, Arthmius, Scalenarthms, 

 Reichenbachia, Decarthron, and Hamotus). None of these genera are pre- 

 ponderantly nearctic; all save Reichenbachia are neotropical in proportion of 

 species, yet the island holds six out of nineteen genera common to the nearctic 

 and neotropical regions. It holds one of the five genera {Reichenbachia) found 

 in the neotropics, nearctic, and also with species outside the Western 

 Hemisphere. 



Since this analysis was possible only after the taxonomic work was finished 

 in the preceding pages, the author feels that the above discussion is without 

 bias, and consequently represents the situation as far as is known. If this is 

 so, the Panamanian Isthmus is represented as just such a cross-roads as theory 

 would expect, with considerable local speciation and a marked South American 

 element the outstanding characteristics. 



Professor Dunn (1940) finds the rich herpetological fauna of Nicaragua, 

 Costa Rica, and Panama to be relatively homogeneous and to have a high 

 percentage of endemic species. From our survey of Barro Colorado we find 

 this high endemism coinciding with the high endemism of Pselaphidae. 



With these data before us we are in a position to estimate the total num- 

 ber of species of neotropical Pselaphidae (PI. XXI). In the first place, the 

 total area of this region is placed at 7,290,000 square miles. If the B.C.I. 

 index of ten species per square mile be used, then the fantastic total of 72,- 

 900,000 faunal units is obtained as a starting point. Let us examine the problem 

 with as much care as possible. Such an assumption would infer that: 



1. There were no subspecies. 



2. There was a maximum taxonomic density for the entire neotropical 

 area. 



3. There was no competition for food. 



4. Within the range of a species there was equal population density. 



5. There was no overlapping in range between species. 



***** 

 1. If there were 72,900,000 taxonomic units, this large number would 

 presuppose a highly developed taxonomy with all ranges mapped. If this were 

 so, these units would not be species but subspecies. The most advanced 

 taxonomy is probably that of the ornithologists. Taking one of the largest 

 families of American birds, the warblers, we find about 50 species and 150 

 subspecies.^ Assuming that the Pselaphidae subspeciate no more than three 



^ These figures on species and subspecies of warblers are taken from Pearson's Birds 

 of America (1936). It should be noted that they do not agree with the 1935 Abridged 

 Checklist of North American Birds of the American Ornithologists' Union. Other sub- 

 speciation studies, within Coleoptera, attest the probability of this general process, for 

 example the work of Valentine (1935, 1936). 



