248 



SCIENCE. 



[N. S. Vol. XXIII. No. 581. 



later he says : ' ' The contention is not that 

 species are occasionally associated with 

 physical barriers, which determine their 

 range, and which have been factors in their 

 formation. It may be claimed that such 

 conditions are virtually universal ' ' ; and 

 again: "Given any species in any region, 

 the nearest related species is not likely to 

 be found in the same region nor in a re- 

 mote region, but' in a neighboring district 

 separated from the first by a barrier of 

 some sort." 



J. B. Steere, in a paper on the distribu- 

 tion of birds in the Philippines, published 

 in 1894, formulated the same idea in the 

 following words: "No two species near 

 enough alike structurally to be adapted to 

 the same conditions will occupy the same 

 area," and added that the facts "show 

 isolation to be the first and the necessary 

 step in the formation of species."" 



I fully admit the potency of isolation in 

 the production of species, but can not for 

 a moment admit that complete isolation is 

 a necessary factor in their evolution, mere 

 divarication from a common center being 

 in. many eases sufficient. Neither can I 

 admit that a barrier must be absolute, or 

 that it must be interpovsed ietiueen species 

 in order to be effective. This is proved in 

 the case of climatic barriers, which, while 

 not keeping contiguous species apart, never- 

 theless restrain each from trespassing far 

 on the territory of its neighbor. 



It is quite possible that Dr. Jordan 's use 

 of the word barrier, which occurs over and 

 over again in his recent paper' is suffi- 

 ciently elastic to cover most of the apparent 

 discrepancies, so that the exceptions to his 

 rule are really few. Still, there are in na- 

 ture many groups of closely related species 

 whose ranges foUow one another in geo- 

 graphic series, the one beginning where the 

 other stops, with no barrier of any kind 



'Ank, XL, 239, 1894. 



' Science, N. S., XXII., November .3, 190.5. 



between, as will be .shown directly. In 

 such cases it is often difficult to say whether 

 the adjacent species have been established 

 by differentiation from one another under 

 existing geographic conditions, or have been 

 developed from preexisting species along 

 geographically divergent lines, and after- 

 ward have met by extensions of range. 

 These points may be made clear by actual 

 examples : 



Among the kangaroo rats of the genus 

 Dipodoniys are two large and closely re- 

 lated species inhabiting our .southern des- 

 erts. One of these, D. deserti, ranges from 

 the Colorado and Mohave deserts in Cali- 

 fornia easterly to a little east of Phoenix, 

 Ariz. ; the other, D. spectabilis, from a little 

 east of Phoenix to western Texas. The 

 geographic division between the two is an 

 invisible line a little southeast of Phcenix 

 in a desert area devoid of bari-iers of any 

 kind, even climatic. The case, thei-efore, 

 appears to be an exception to the law of 

 isolation. 



Turning now to another groiip, illustra- 

 tions of at least two kinds of variation 

 among closely related species are afforded 

 by the ground squirrels of the genus Am- 

 mospermophilus, of which three distinct 

 species and three sub.species are recognized. 

 These animals resemble one another in form 

 and markings and in the habit of carrying 

 the tail closely aj^pressed against the back 

 so that its under side is uppermost and is 

 presented to the view of the observer. In 

 one species the under side of the tail is 

 dark; in the others white or nearly M^hite. 

 The three species and their geographic 

 ranges are (see map. Fig. 1) : 



Ammospermophihis nelsoni, restricted 

 to the hot southern end of the San Joaquin 

 Valley in California, where it occupies a 

 detached area and is isolated by the en- 

 circling mountains, which by interposing a 

 climatic barrier cut its range off from that 

 of the parent species, leucurus. It has ac- 



