74 BOTANICAL FEATURES OF THE ALGERIAN SAHARA. 



to semi-deserts where in certain regards a very different condition obtains. 

 The plant distribution thus dependent on soil depth, and the root character 

 also having relation thereto, will be presented below ; the soil conditions, so 

 far as they are known to me, can be given briefly in this place. 



From a few excavations on the hamada near Laghouat, on an analogous 

 area by Tilrempt, on the hamada at Ghardaia, between Ghardaia and 

 Ouargla (Gantara), and by Biskra, in each instance with analogous topo- 

 graphical conditions, it was learned, in short, that on formations of this 

 sort the soil is usually less than 50 cm. in depth, although where there are 

 rocks or in drainage depressions the depth may be greater. In the oueds, 

 where a different type of soil occurs, greater soil depth is naturall}'- found; 

 also on the reg, or alluvial plain, frequently at least the flood-plain of oueds, 

 the soil conditions are peculiar and the soil is deeper than on the higher 

 hamada. The special significance lies not so much in the differences in the 

 soil per se, but in the differences in the water relation occasioned b}^ varia- 

 tion in depth. Owing to want of data in regard to penetration of water and 

 its retention on the Algerian soils, the soil-moisture relation can only be 

 gathered from inference. As regards rains, it is probable that light showers, 

 those so slight as not to penetrate over i.o cm., have little or no direct influ- 

 ence on the perennial vegetation, but greater penetration directly benefits 

 such plants. The first conclusion is drawn from the observation that fila- 

 mentous rootlets (seemingly like the deciduous rootlets of the perennials of 

 the Tucson region) on Haloxylon at Biskra were not found nearer the surface 

 than 8 cm. Should the more superficial soil layer be moist for any con- 

 siderable time, there would apparently be no reason why such temporary 

 rootlets should not be formed nearer the surface. However, should there 

 be sufficient moisture in the soil to permit absorption by roots, provided a 

 slow rate was adequate to replace the transpiration loss, which rate was 

 made lower by a more moist air, a sHght rain would be of great significance, 

 even if it did not penetrate to any appreciable depth. On the intensely 

 arid desert such slight modifications of the water relations as the lowering 

 of the temperature as winter approaches, causing decreased evaporation 

 or rains, although actually small in amount, may be of large moment to 

 plants. Such a condition was noted at Ghardaia,* where there had been 



* Soil samples were taken from the plain about 3 kilometers north of Ghardaia from 

 an area where the vegetation is relatively good (see p; 40). The surface of the soil at 

 the place is slightly depressed. Soil samples from the depressed area and samples 

 from portions of the plain adjacent to it were settled under water, with the result that 

 the proportion of fine soil was found to be less in the depressed area. Samples of the 

 soil from the lower area were placed in air-tight cans and the moisture content deter- 

 mined subsequently. The soil was found to contain 0.8 per cent water. Through the 

 kindness of the Bureau of Soils, U. S. Department of Agriculture, the critical moisture- 

 content of the same soil was determined, which was 5 per cent. The critical moisture- 

 content of mesa soil, taken from the creosote-bush slope at the foot of Tumamoc Hill, 

 Desert Laboratory, as determined by the Bureau of Soils, is 10.5 per cent, which forms 

 an interesting comparison of nearly similar situations. 



