No. 1810. DESCRIPTIONS OF NEW MILLIPEDS— COOK. 149 



been ascribed, of course, to a deficiency of rainfall, though there are no 

 records to establish the fact, and other factors may be responsible. 



The extension of grazing and the limitations of forest fires has 

 allowed woody vegetation to spread in recent years over large tracts 

 of land that were formerly covered with heavy growths of grass. It 

 is known that little water, relatively speaking, enters the naked sur- 

 faces of the soil when the land is occupied by the usual sparse growth 

 of mesquite, prickly pear, and other desert vegetation. The formerly 

 more heavy and general growth of grass may have afforded more 

 favorable conditions for holding water on the surface and allowing it 

 to soak gradually into the soil. With ordinary gentle rains the effect 

 of grass is, of course, to keep the soil dry, but with the very heavy 

 rains that come occasionally in south Texas, nearly level grass-cov- 

 ered areas might be expected to hold the water longer than denuded 

 surfaces, and give it a better opportunity to penetrate to the lower 

 layers of the soil. 



The recent open growth of mesquite probably represents the most 

 unfavorable condition for the retention of water in the soil. With 

 sufficiently old and abundant forest growth to accumulate fallen 

 timber and other debris, obstacles to drainage would be formed and the 

 earth would be shaded from the sun. Conditions would again be- 

 come more favorable to the retention of permanent moisture, and 

 many areas would eventually change from naked deserts into humid 

 forests or swamps. Tendencies in this direction begin to appear 

 in the region of Brownsville. The finding of another new genus of 

 tropical millipeds near Brownsville serves as an additional indica- 

 tion that present tendencies are only a return toward a former con- 

 dition of much more abundant forest growth in south Texas. 



The rapidity with which these prairies are now being covered with 

 woody vegetation is a further indication that the previous treeless 

 condition did not represent the natural state of the country, but 

 might be ascribed to the same agencies that have exterminated for- 

 ests in many parts of Mexico and Central America. Lands that have 

 been cleared by the natives for agricultural purposes eventually 

 become occupied by perennial grasses that put an end to agriculture, 

 as practiced by primitive Indian methods. The burning of accumu- 

 lations of dead grass kills young trees and thus hinders or prevents 

 the extension of forests. Large areas that are naturally well adapted 

 to forests have been gradually denuded into artificial deserts or 

 grassy fire-swept tracts, now quite unoccupied by human inhabitants, 

 and more or less completely devoid of millipeds and other humus- 

 inhabiting types of arthropoda. 



On the other hand, many areas now covered with heavy growth do 

 not represent truly virgin or primeval forests. The absence or scarce 

 representation of the humus-inhabiting animals shows that many 



