Makch 13, 1908] 



SCIENCE 



427 



while submarine eruptions took place in the 

 neighborhood of Pantelleria in 1831 and 1891. 

 The small and hitherto undescribed island of 

 Linosa, about one square mile in area, con- 

 tains nine volcanic cones. These are referred 

 to two periods; an earlier one of cones of 

 greenish and yellow tuffs, containing blocks 

 of basalt, and a later one of cinder cones, with 

 lava flows of feldspar-basalt and nephelite- 

 basalt. The Linosa lavas are very uniform 

 in chemical characters. 



Many analyses of the rocks have been made, 

 and they are shown to be closely alike in their 

 chemical features, one of the most notable of 

 these being the uniformly high percentage of 

 titanium. All the volcanoes mentioned are 

 considered, therefore, to be genetically related 

 and to belong to the same eo-magmatic region 

 or petrographic province, which possibly ex- 

 tends into Africa, as far as the Great Eift 

 Valley. This petrographic province is very 

 different in its characters from those of the 

 Italian peninsula and also from that embra- 

 cing the volcanoes of the Grecian Archipelago 

 and Asia Minor, which have also been studied 

 by the speaker. The reputed volcano of 

 Boukournine, near Tunis, was visited and was 

 found to be composed entirely of limestone. 

 A brief account was given of the occurrence 

 of orbicular diorite at Santa Lucia di Tallano 

 in Corsica. The paper was illustrated by 

 lantern-slides. 



Ealph Arnold, 



Secretary 



DISCUSSION AND CORRESPONDENCE 



THE DISTRIBUTION OP CLOSELY ALLIED SPECIES 



The idea that closely related species do not 

 possess identical geographical distributions is 

 a theoretical deduction which should be rigor- 

 ously tested by the examination of actual 

 cases. Recently, R. G. Leavitt has done this 

 with reference to plants (chiefly orchids),' and 

 arrives at the conclusion that this principle, as 

 expressed by D. S. Jordan,' is not well sup- 

 ported. 



' " The Geographical Distribution of Nearly Re- 

 lated Species," American Naturalist, 41, 1907, pp. 

 207-240. 



' " Given any species in any region, the nearest 



Nevertheless, I think that Jordan's sen- 

 tence is fundamentally correct, provided 

 it is changed so as to refer not only ta 

 purely geographical, but also to ecological 

 conditions. 



I have expressed this idea in two sentences :*" 

 (1) " Closely allied species occupy neighboring' 

 areas"; (2) "More or less closely allied 

 species occupying the same or nearly the same 

 territory, generally possess different habits."' 

 Further, I have treated of this subject more 

 especially in a recent publication,'' classifying 

 the forms of segregation of closely allied 

 species, geographically as well as ecologically. 



It will be remarked that, wherever I have 

 talked about this subject, I always have used 

 the word species.'^ Leavitt, in the paper re- 

 ferred to (p. 230), deliberately changes the 

 word species in Jordan's sentence into kind, 

 and investigates the distribution of "pairs of 

 kinds" of plants. This substitution, kind or 

 form in place of species, is also advocated by 

 J. A. Allen.' 



This change, however, is entirely inadmis- 

 sible. I used the above phrases in connection 

 with the bearing of isolation or segregation 

 upon the formation of species (speciation), 

 and wanted to bring out the idea that geo- 

 graphical or ecological segregation is a cri- 

 terion by which species may be recognized. I 

 intended the word " species " in the strict taxo- 

 nomic sense, that is to say, for " forms " which 

 are morphologically separated from the allied 

 forms. My object was to express the opinion 

 that the morphological segregation of true 

 species is connected with and due to some kind 

 of segregation in the physical conditions under 

 which the " species " live, and I pointed out 

 that this might be either purely geographical 

 or ecological. Ecological or geographical 

 segregation is the factor which results in 

 speciation, that is to say, the factor which 



related species is not likely to be found in the 

 same region" (Science, 22, 1905, p. 547). 



'Proceedings American Philosophical Society, 

 44, 1905, pp. 127, 128. 



' American Naturalist, 41, 1907, p. 654. 



' " The Crawfishes of the State of Pennsyl- 

 vania," Memoirs Carnegie Museum, 2, 1906, p. 512. 



" See also Science, 23, 1906, p. 949. 



