508 L. H. G R A Y 



number of droplets, as would be expected, since the droplets occur in 

 pairs condensed on the positive and negative ions. The lateral sepa- 

 ration of the ions is out of proportion, having been artificially increased 

 fivefold, by means of an electric field, and will be discussed later. 

 The separation of clusters along the track is about half a micron and 

 we speak of this radiation as producing an average of two ion clusters 

 per micron in tissue. The frequent occurrence of clusters of more than 

 two ions is noteworthy. This is true of all ionizing radiations and is 

 due to the fact that the ionizing particle may expel an electron from 

 an atom with all speeds from zero to that with which it is itself mo^^- 

 ing. We have to remember also that it may transfer energy to an elec- 

 tron without removing it from its parent atom, i.e., it may merely 

 excite the atom through which it passes. These events are, in fact, 

 considerably more frequent than the events leading to ionization, so 

 that between each pair of ions we should, to complete the picture, 

 visualize two or three excited atoms, some of which may or may not 

 contribute to the biological effect. The ratio of the number of atoms 

 excited to the number ionized may well vary considerably from one 

 medium to another, but it is not dependent to any appreciable extent 

 on the nature or speed of the ionizing particle, provided this exceeds 

 that of the orbital electrons of the atom traversed by the particle. 

 Sometimes an electron may be expelled from the parent atom with 

 sufficient energy to produce ions in its turn. Thus, the clusters of 

 ten ions on the extreme left and also third from the right in the figure 

 each represent one pair of ions produced by the swiftly moving /3 par- 

 ticle and four pairs produced by the ejected electrons. On the aver- 

 age a total of three ion pairs are formed for each atom ionized by the 

 primary particle. 



These rules cease to be valid when the maximum energy that may 

 be transferred to the orbital electron is comparable wth its ionization 

 potential. In practice, therefore, some deviation may be expected 

 for electrons, protons, and a particles having energies less than about 

 100 e.v., 200 e.kv. and 1 m.e.v., respectively. The expected devia- 

 tion is in the sense of an increase in the proportion of primary to sec- 

 ondary ionization, and an increase in the ratio of excitation to ioniza- 

 tion. 



In Figure 3 the same /S-particle track as in Figure 2 is shown 

 superimposed on an electron micrograph of Escherichia coli and bac- 

 teriophage taken by Luria and Anderson {27). It will be observed 

 that only rarely are more than two or three clusters formed, within an 



