ECOLOGICAL PAPERS. 123 



show great irregularity during spring and autumn and a compara- 

 tively high rate during July and August, The general height of 

 the different graphs probably expresses the most instructive and 

 intresting dift'erences in the different habitats. That of the cotton- 

 wood dune is farthest removed from those of the other associa- 

 tions and shows a habitat not only with great evaporating power, 

 but one of great extremes, the difference in rate between two con- 

 secutive weeks being nearly or quite 10 cc. per day during May 

 and the first part of June, and on two occasions amounting to an 

 increase of 100 per cent in one week as compared with the pre- 

 ceding. This occurring early during the growing period would 

 doubtless be very unfavorable for the development of any seed- 

 lings, especially as it was followed by the very high rates of the 

 succeeding months. The high maximum occurring at midsummer 

 would probably prove the excluding factor for all mesophytic 

 plants even if not combined with such other factors as the defi- 

 ciency of soil water at the same time. Such a graph seems to 

 depict rather well a habitat of atmospheric extremes, making large 

 demands upon all available water, and naturally and necessarily 

 resulting in a xerophytic plant association, with a very limited 

 undergrowth and an almost entire absence of herbaceous plants 

 and seedlings. Perhaps nowhere could an association be found 

 so entirely dependent upon vegetative reproduction for its main- 

 tenance, as almost without exception any increase in vegetation 

 is the result of development from subterranean branches. 



The graph for the pine dunes is decidedly lower and more 

 regular in its contour than that of the association w^hich it suc- 

 ceeds. Its four nearly equal maxima would indicate that within 

 its limits there was throughout the summer season a continuous 

 stress rather than a series of violent extremes. On the whole it 

 shows a water demand of little more than half of that occurring 

 in the cottonwood dunes. Its greatest divergence is plainly due 

 to the evergreen character of its vegetation and is seen on its low 

 range in May and the first part of June, and again in October 

 when it falls below that of the oak dunes and is even less than 

 that of the beech-maple forest. This would give good reasons 

 for expecting to find within this association truly mesophytic 

 plants whose activities are limited to the early spring. 



The graph from the oak dune stations shows two surprisingly 

 high points ; one during May that may be partially explained by 

 the absence of foliage ; and the other near the end of June which 



