STOPPING POWER OF A POLYATOMIC GAS 151 



drogen atoms is 15.00 ev, and this must be very accurate. The differ- 

 ence between these two values of I is equivalent to a difference in the 

 stopping powers for alpha particles of 3 per cent at 10 Mev, 4 per cent at 

 5 Mev, and 5 per cent at 1 Mev. 



In lieu of theory, interpretation of the stopping phenomena of poly- 

 atomic gases has always been based on the so-called "Bragg rule," which 

 supposes each type of atom to have a definite stopping power for fast 

 particles which is a function of particle charge and velocity, and the 

 stopping powers of atoms bound together in a molecule to be additive. 

 In practice this rule works surprisingly well. Thus, for example, stop- 

 ping powers of H, N, and have been obtained by taking one-half of 

 the empirical stopping power of the corresponding diatomic gas, and 

 that for C by difference from the stopping power of any of a number of 

 gases (or solids) containing carbon. In this way a basic table for H, C, 

 N, and 0, and for other atoms as well, is compiled, and is tested with 

 data from as many different compounds as possible, or is used to pre- 

 dict the stopping power of a compound not yet studied experimentally. 

 The Bragg rule can be expressed equivalently — and more conveniently — 

 in terms of the relative stopping powers of the elements. However, 

 ranges can in general be measured much more accurately than can 

 stopping powers (although even for range determinations there has been 

 an abundance of overoptimistic assessment of experimental error), and 

 there have been many more measurements of ranges than of stopping 

 powers. For this reason the Bragg rule has often been tested by the 

 additivity of values of Sr, rather than of s. (The distinction between s 

 and Sr was explained above.) But additivity of Sr is not the same thing 

 as additivity of s, because of the variable dependence of these quantities 

 on the particle velocity. The statement of the rule which we have 

 adopted, that is, additivity of values of s, is the more fundamental, and 

 is equivalent to additivity of Sr if, and (barring highly unlikely ac- 

 cidental compensations) only if, the ratios of all s values for atoms in the 

 molecule, at all velocities less than the initial particle velocity, are 

 velocity-independent. This tacit assumption is never exactly valid, and 

 is more in error, the greater the disparity in atomic numbers of the 

 atoms in the molecule, and the smaller the initial particle velocity. 



However, it should always be borne in mind that the Bragg rule does 

 not rest on any quantitative theoretical basis. As data on stopping 

 power improve in accuracy we must anticipate ever more numerous 

 and more consequential departures from the rule. At the present time, 

 because of inaccuracies in stopping-power data, the observed departures 

 are not certain enough to be considered as established, far less to have 

 revealed any pronounced regularities. The smallness of the departures 



