176 BOTANICAL GAZETTE [march 



that there may be a similar change in the two cases. The change 

 occurs rapidly at the high temperature, but slowly at the low 

 temperature. 



Some investigators have used the time required to kill all 

 seeds as the end point. In this work we have selected the time 

 required to kill 75 % per cent of the seeds as the end point. This is 

 more desirable because there seems to be considerable discrepancy 

 in the resistance of a few stronger seeds. This end point for 12 per 

 cent moisture and various temperatures was obtained as shown in 

 table II. While there are some irregularities, there is a definite 

 relation between temperature and time of exposure necessary for 

 killing 75 per cent of the seeds. 



The time-temperature formula suggested by Lepeschkin (22) 

 has been used here to calculate the life duration of the seeds. By 

 determining the time required to kill seeds at any two definite 

 temperatures, the time for killing seeds at any other temperature 

 can be calculated. The formula (referred to as formula 3) is: 



T=a—b log Z 



in which T is the temperature in degrees Centigrade, Z is the time 

 in minutes, and a and b are constants. If the loss of viability 

 of seeds during storage is a matter of coagulation of cell proteins of 

 the embryo, this time-temperature formula for the coagulation of 

 proteins should be applicable as a temperature-life duration formula 

 for seeds. In experiment the life duration determined must be at 

 relatively high temperatures, ranging from 50 to ioo° C. for 

 air-dried seeds. 



In formula 3 constants a and b may be calculated by substitut- 

 ing the time and temperature of any two trials and solving for a 

 and b in the two equations. This is the 

 by Lepeschkin (22). In order to weigh all determinations equally, 

 the constants in this paper are calculated by the method of least 

 squares. The values of the constants and the life duration found 

 in each experiment were substituted in the equation and the theo- 

 retical temperatures were calculated. The values are shown in 

 table II- A comparison of these found and calculated temperatures 

 shows that a comparatively close agreement exists. The dis- 

 crepancies are within the limits of experimental error. 



me 



t 



