VIII. BIOELECTRIC MEASUREMENTS 255 



coupling. Either oil or mica capacitors .should be used, and only the 

 best grades of each. This is necessary because a very slight leak 

 from the high potential plate of one tube through the coupling capaci- 

 tor can cause a large change in the grid potential of the next tube. 



The size of the capacitor depends on the frequency pass required. 

 If a steady potential is suddenly applied to the input of a capacitance- 

 coupled amplifier, the potential on the grid of the second stage, in- 

 stead of rising suddenly to a new value and staying there, will sud- 

 denly rise but then decay exponentially to its former value. The 

 rate of decay will be a function of the size of the capacitor and the 

 value of the grid resistance. Its time constant, /, is given by the rela- 

 tion: 



t = RC (4) 



where R is the value of the grid resistor in ohms and C the value of 

 the coupling capacitor in farads. The time will then be given in 

 seconds, and represents the time required for the potential to fall 

 to 1/e of its initial value (e = 2.718). The potential on the grid of 

 the third stage will fall correspondingly faster, and so on for suc- 

 ceeding stages. From these facts it should be possible to determine 

 how much error will be committed in any individual problem by using 

 a capacitance-coupled amplifier. 



From this discussion one might infer that it is best to use as large 

 a capacitor as possible. This is true from the standpoint of accu- 

 racy, but is certainly not true for ease of operation. The larger the 

 capacitors the longer one has to wait for equilibrium to be established 

 in the amplifier, and, with 4 yd. capacitors in a four stage amplifier, 

 this can be very annoying. For the recording of most action poten- 

 tials from nerve or muscle, 0.1 /if- capacitors are large enough. 



5. Limitations of Amplifiers 



It might be inferred from the above discussion that there is no 

 lower limit to the potentials that can be measured with electronic 

 amplifiers. But such is certainly not the case. The ultimate limit 

 of sensitivity is imposed by "tube noise." This can be defined as 

 random spurious changes in plate current of a tube, and may be due 

 to a variety of causes. The chief ones are irregularities of electron 

 emission from the cathode and changes in resistance of resistors in the 

 circuit. There are a number of things that can be done to reduce this 

 artifact. The first is to use only wire-wound resistors in the first 



