GENERAL SUMMARY AND CONCLUSIONS. 73 



many of which are halophytes, with the perennial grass, Siipa ienacissima, 

 and Artemisia herba-alba away from salt spots. Such low forms are present 

 in sufficient numbers as to give character to the landscape and to conceal 

 the surface of the ground fairly well. 



South of the Saharan Atlas a marked change occurs. Here, with a rain- 

 fall of 200 mm. and less, the trees are confined to the dayas, a narrow belt, 

 the vicinity of oueds, and the oases, exclusively. The shrubs of the hamada 

 also decrease in numbers as one goes south, and where the annual precipita- 

 tion is least, as on the Gantara between Ghardaia and Ouargla, large barren 

 areas extend. At no place on the hamada of the M'Zabite region are the 

 shrubs present in sufficient numbers or size to conceal the surface of the 

 ground or to give character to the landscape. 



Aside from the effects following a lessened annual precipitation there is 

 also to be taken into account the increasing uncertainty of rain, or itsirregu- 

 larity, which is also a marked characteristic of the Saharan climate. In 

 the desert, also, storms are likely to be of the torrential type. I did not 

 observe vegetation characteristics which appeared directly traceable to 

 the irregularity in rainfall, but Ha3rv\'ard reports an interesting condition 

 observed by him in the southern Sahara, near Kidal, where large tracts of 

 Mimosa had died from an unusually long period (five years) without rain.* 

 It is not at all improbable that to the cause named much is directly trace- 

 able which is generally attributed to insufficient rainfall taken in the usual 

 sense. It is probable that the vegetation of the desert — the amount as 

 well as kind — ^is due to the capacity of desert forms to meet successfully 

 the occasional, even rare, conditions, of whatever sort. 



THE SOIL RELATION IN SOUTHERN ALGERIA. 



A very important environmental factor, although one which can not at 

 this time be adequately presented, is the soil relation. Nowhere is the 

 edaphacic factor more important than in the desert, where quantity and 

 quality are always important and occasionally even determining factors. 

 In this connection I do not refer particularly to dunes or to chotts, but to 

 country soils, that is, the sort most commonly to be found, which in south- 

 ern Algeria is a clay with sand present in greater or less amount. So far 

 as the relation of plant to soil refers to the presence or the absence of the 

 plant, the problem can be briefly stated thus: Given similar lands of soil 

 and an equal precipitation, areas where, within limits, there is greatest 

 depth of soil will have the largest number of plants, and areas with light soil 

 covering will have few or no plants. Also, having given sufficient soil, the 

 kind of plants present, together with certain root-types, will depend on the 

 soil depth. It should be understood that these generalizations are supposed 

 to apply to southern Algeria and not the deserts in general, or at least not 



* Through Timbuctu and across the Great Sahara, 1912, p. 266. 



