No. 512] PL' Est. XT l'TOHLEMS IX TLA XT ECOLOGY 475 



have handled roses for a period of years have learned 

 what varieties may he expected to do well in the dry air 

 of the desert, and what ones may he counted out, and so 

 on through a long list of plants which, hy knowledge 

 gained in the costly school of experience, are coming to 

 be depended on, or are being rejected one after another, 

 as they are found to be unsuited to the environment into 

 which they have been brought. Thus, in a purely em- 

 pirical way, it has been found that many plants success- 

 fully cultivated in regions of greater atmospheric hu- 

 midity make an entirely normal growth in the desert, if 

 their roots are well supplied with water, but that others 

 however well cared for in this respect, either fail com- 

 pletely, or come short of making a healthy growth, and 

 that this is especially true in the summer months when 

 desert conditions are most pronounced. 



With the accumulation of such facts the more evident 

 does it become that a very complicated problem is here 

 presented. Why is it that one plant, properly watered, 

 does well in the desert, while another, though treated in 

 the same way, makes a poor growth or fails altogether? 

 At first thought it would seem as though there must be a 

 difference in the capacity of the root systems of the two 

 plants for absorption, and that this may be a sufficient 

 explanation of their different behavior; but it is evident 

 on consideration, that with precisely the same capacity 

 for root absorption, a plant in which transpiration is suc- 

 cessfully regulated may thrive in an atmosphere in winch 



most elaborate experiments and the most exact deter- 

 minations of rate of absorption— assuming that such de- 

 terminations are possible— would be very likely to throw 

 no light on the problem. Comparisons of the transpira- 

 tion rate of the plants in question appear more promising, 

 but the same difficulty arises in an attempt to pursue the 

 investigation along this line, for there is no reason to 

 suppose that two plants of widely different rates of trans- 



