44 



days with a solution of potassium chloride ranging from 4.02 per 

 cent normal at first to 16.08 per cent normal toward the last showed 

 no increase in either sap density or hardiness over that of plants not 

 so treated. It is only in sand cultures that we have been ahle to 

 increase the sap density of plants by watering them with solutions. 



If freezing to death results from too complete withdrawal of 

 water from the protoplasm, then, even with winter buds and wood, if 

 the sap density could be increased with some material that stays in 

 solution at temperatures lower than that at which the tissue kills, 

 it would hold more water in solution and thus lower the temperature 

 at which a killing degree of desiccation is reached. Muller-Thurgau's 

 measurements showed that a very large percentage of the water in 

 plant tissue is in the form of ice. In case of winter resting tissue 

 very little water can be left unfrozen at a killing temperature. If 

 freezing to death pf plant tissue results from desiccation, then very 

 small quantities of water must be sufficient to protect the protoplasm 

 in case of the more hardy plants. Some writers hold that since some 

 plants continue to live after nearly all the water is frozen out, desic- 

 cation of the protoplasm can not be the explanation of death from 

 low temperatures. Others seem to hold the opinion that the pro- 

 tective action of the sap solute ends at the freezing point of the sap. 

 It should be remembered that so long as the temperature is above 

 the eutectic point of any of the sap solute, there will be some water pre- 

 sent in the liquid state. The number of degrees which the killing tem- 

 perature is lowered by an increase in sap density should be greater 

 with plants that kill at the lower temperatures. Thus in case of a 

 solution containing a gram molecule of a non-electrolyte in a liter of 

 water, the freezing point is -1.86° C. In case of a solution with 

 one-half of a gram molecule to a liter of water, the freezing point 

 would be -0.93° C. The following figures would evidently be approxi- 

 mately true: 



