How Thin Should a Section Be?

In considering the electron microscopic study of sections of tissue as a practical approach to the study of cell structure, the cytologist is faced with a number of possible decisions with respect to the way in which he will prepare his material. Not the least of these is the problem of how thin the section of tissue should be to serve his purpose. A useful section of a sample must satisfy the following demands: (a) it must provide the observer with a preparation which is sufficiently transparent (one which has a sufficiently low electron-scattering "power") to be viewed by transmitted electrons; (b) it must provide a sample which is sufficiently limited in thickness to permit structural analysis in the other two dimensions without the confusion of structural overlap; and (c) it must be thin enough to permit the optical system to resolve the structure to be studied. Because cutting involves bond rupture and plastic flow, the surface layers of a section are badly deformed. The thinner the section, the larger the fraction of its volume which has been permanently distorted. A more accurate picture of the distribution of structures within a tissue can therefore be expected from thick rather than thin sections, provided the three requirements mentioned above are ,satisfied. I t almost goes without saying that a thicker section with a given percentage of deformation is easier to cut and handle than a thinner one. What, then, are the considerations which set the upper limit of "usable" section thickness? A section with a density of approximately 1 gin. per cc. and 0.2 micron thickness will transmit from 1 to 10 per cent of the incident electrons in the voltage range of 50 to 100 kv. with the usual range of objective numerical apertures. (See Gettner and Ornstein, 1.) This usually constitutes sufficient "transparency" for study. The problem of overlap will, of course, vary from structure to structure, but often it will be found that the overlap in a section of 0.2 micron thickness will not lead to confusion in analysis. Let us then examine the dependence of resolution on section thickness. Electrons can lose energy in passing through a specimen. A loss in energy involves a change in the associated de Broglie wave length of the electron. Since electron microscopes have monochromatic lenses, i. e., they have no correc53