CONSTANCY AND INDEPENDENCE 



14/9 



of immediate effects shown in Figure 14/8/1. What properties 

 must the three variables B have if the systems A and C are to 

 become independent and absolute ? The question has not only 

 theoretical but practical importance. Many experiments require 

 that one system be shielded from effects coming from others. 

 Thus, a system using magnets may have to be shielded from the 

 effects of the earth's magnetism ; or a thermal system may have 

 to be shielded from the effects of changes in the atmospheric tem- 



Figure 14/8/1. 



perature ; or the pressure which drives blood through the kidneys 

 may have to be kept independent of changes in the pulse-rate. 

 A first suggestion might be that the three variables B should 

 be removed. But this conceptual removal corresponds to no 

 physical reality : the earth's magnetic field, the atmospheric 

 temperature, the pulse-rate cannot be ' removed '. In fact the 

 answer is capable of proof (S. 24/15) : that A and C should he 

 independent and absolute it is necessary and sufficient tlmt the 

 variables B should be null-functions. In other words, A and C 

 must be separated by a wall of constancies. 



14/9. Here are some illustrations to show that the theorem 

 accords with common experience. 



(a) If A (of Figure 14/8/1) is a system in which heat-changes 

 are being studied, B the temperatures of the parts of the con- 

 tainer, and C the temperatures of the surroundings, then for A 

 to be isolated from C and absolute, it is necessary and sufficient 

 for the B's to be kept constant, (b) Two electrical systems joined 

 by an insulator are independent, if varying slowly, because 

 electrically the insulator is unvarying, (c) The centres in the 

 spinal cord are often made independent of the activities in the 

 brain by a transection of the cord ; but a break in physical con- 



159 



