METHODS FOR THE ELECTRON MICROSCOPE 111 



the other hand, with electron microscopes, at least as at present 

 conceived, perception is based entirely upon the scattering of 

 electrons. Scattering is a function of the atomic number of the 

 atoms in the material upon which the electron beam impinges. 

 In biological objects, the great bulk of the scattering material is 

 carbon, oxygen, and nitrogen, which differ relatively little in 

 scattering power. Consequently, the image obtained in an elec- 

 tron microscope depends almost entirely upon the mass of mate- 

 rial present in different parts of the specimen, and there is little 

 prospect of perceiving chemical differences between different 

 parts of a specimen. Furthermore, the application of reagents 

 to the specimen will facilitate the perception of chemical differ- 

 ences only so far as the reagent adds selectively to the effective 

 electron-scattering power of individual chemical components. 



Thus the sensitivity of methods for use in the electron micro- 

 scope is limited by the sensitivity of the instrument to differences 

 in scattering power. With the techniques available at the mo- 

 ment, it appears that, to obtain reproducible results, the scatter- 

 ing power of a component must be altered by at least 10 percent. 

 To increase the scattering power by this amount is difficult, 

 compared with the small changes in chemical composition which 

 will give a significant change in the transmission of visible light. 

 Consequently, one cannot expect, without marked improvement 

 in the instruments available, to obtain methods of the same de- 

 gree of sensitivity as those available for the light microscope. 

 On the other hand, the great advantage in resolving power pos- 

 sessed by the electron microscope justifies the development of 

 even crude cytochemical methods. In this direction we have 

 had some success. The work has been carried out in the main 

 by Mr. Stuart-Webb and Mr. Bell of my department, and Dr. 

 Bovet and Dr. Lamb of Professor Randall's department. 



After some search for a suitable test object, it was found that 

 thin gelatin films could be obtained which were sufficiently uni- 

 form in thickness for preliminary investigations. With these thin 

 films the following procedure was worked out, in which the scat- 

 tering power of components which will react with dinitrofluoro- 

 benzene is markedly increased, partly by combination with a 

 large organic group, and partly by subsequent linkage with a 

 heavy metal. 



