WINTER KILLING AND HARDINESS 253 



longer sustain itself when too many molecules of water are removed 

 from its support. In the great majority of plants this point lies so high 

 in the water content that it is passed very soon after the inception of ice 

 formation, hence the death of many plants at this period. Others may 

 be able to exist with so little water that a very low temperature is 

 necessary before a sufficient quantity is abstracted to Qause death. 

 From some plants enough water cannot be abstracted by cold to kill 

 them." Several investigators have shown that certain tissues, cooled 

 to a temperature which is fatal if ice formation occurs, will withstand 

 that same temperature if ice formation does not occur. 



After investigating several possible ways in which bud scales and wool 

 packing might serve to protect the embryo flowers and shoots during the winter, 

 Wiegand concluded that their main function is not to shut out cold or even to 

 retard temperature changes (about 10 minutes seemed the limit for the greater 

 part of any change), but rather to retard the loss of moisture and to prevent me- 

 chanical injury especially when the buds are frozen. Lilac buds lost in 3 days, 

 at temperatures between — 18°C. and — 7°C., 2.8 per cent, of water when bud 

 scales were left on ; with bud scales removed the loss of water was 39 per cent. 

 Heat absorption due to the color of bud scales in horse-chestnut buds amounted 

 to 15°F. Chandler^* also found that "scales of peach buds do not serve to 

 protect them from low temperature. Buds frozen in the laboratory with the 

 scales removed were shghtly more resistant to low temperature than were buds 

 with the scales not removed." 



Freezmg and the Deciduous Habit. — The view that death from low tempera- 

 tures is due to a withdrawal of water is supported by the consideration that the 

 deciduous habit is in most cases essentially a protection against water loss during 

 the winter and that the leaves of evergreen plants are particularly adapted to 

 reduce the rate of transpiration to a minimum. 



An interesting and suggestive parallelism exists between the autumnal be- 

 havior of trees in temperate regions and the changes in trees in regions subjected 

 to prolonged dry but warm weather. In both cases they assume a distinctly 

 xerophytic character. The most obvious phenomenon accompanying this 

 transition is leaf-fall. Of this Coulter, Barnes and Cowles^^ say: "The leaf 

 behavior of deciduous trees and of tropical evergreens obviously is related to 

 external factors, in the former being associated with climatic periodicity (either 

 of moisture, as in the monsoon forests of India, or of temperature, as in the north- 

 ern deciduous forests), while in the latter it is associated with uniform moisture 

 and tem.perature. That the deciduous and the evergreen habits are related to 

 external conditions may be inferred from many trees and shrubs (e.g., poison 

 i^y, Virginia creeper, various oaks) which shed their leaves in regions of cold 

 winters, but retain them in warmer climates; furthermore, various plants (as 

 the grape and the peach) become evergreen in uniform tropical climates, and 

 even those species that remain deciduous (as the persimmon and the mulberry) 

 have much longer periods of leafage. 



"The exact factors involved in leaf-fall, that is, in the development of the 

 absciss layer, are imperfectly known. In the monsoon forest and in other regions 



