THE MOLECULAR FORCES IN PLANTS. 201 



A thermometer is introduced into the soil, which is then covered 

 with tinfoil, as also is the portion of the stem rising above ground. 

 The pots are now placed in a room kept at a very constant tem- 

 perature, or are introduced into a thermostat. When the soil in 

 the flower pots has assumed the temperature of its surroundings, 

 the observation begins. We note the flow of sap, say, per hour. 

 After several hours, the soil in which the plants are rooted is well 

 watered, and when the temperature has adjusted itself again, the 

 determinations of the flow of sap are continued. Owing to the 

 abundance of water in the soil, the outflow is now found to be 

 much more considerable than before. It must not be forgotten, in 

 this and the following experiments, that the flow of sap in many 

 plants exhibits periodic variations, independent of external condi- 

 tions (see 78). 



The same objects may also be used to prove that conditions of 

 temperature have an important influence on the outflow of sap. 

 The soil in which the plants are rooted is well watered at the 

 commencement of the experiment, and then from time to time the 

 temperature in the thermostat is altered, and in each case, after 

 adjustment of temperature (I.e. when the soil has assumed the 

 temperature of its surroundings) the observations on the flow of 

 sap are resumed. At 16 C. more fluid flows out in the unit time 

 (e.g. in an hour) than at 12 C. At 20 C. the outflow of sap is 

 more considerable than at 16 C. I find that the optimum tem- 

 perature in Cucurbita Melopepo is about 26 C. Still higher 

 temperatures retard the flow of the sap, and at a temperature of 

 43 C. it entirely ceases. 1 



It is frequently necessary as, for example, in the experiments 

 which we have just been considering -to expose objects for a long 

 period to a constant temperature. We will now describe the 

 methods and apparatus to be employed for the purpose. 



First as regards the thermometers, these are to be obtained in a 

 great variety of forms and sizes, and variously graduated, from 

 the Firms named in the Appendix. If we require several thermo- 

 meters for the same experiment, they must be accurately com- 

 pared. In special cases, e.g. in experiments with the auxanometer, 

 it is of advantage to employ a registering thermometer. 



When working with registering apparatus, or with large plants 

 which cannot be introduced into a thermostat, we must take pains 

 to keep the temperature of the laboratory as constant as possible. 

 For small rooms the so-called American stoves, heated with 



