Small pieces of liver from rats subjected to different dietary regimes were fixed by freeze-drying, and postfixed by in vacuo heating and denaturation with alcohol. Specimens were digested with ribo- or deoxyribonuclease, and stained with gallocyanin-chromalum, azure II, the Feulgen procedure or alcoholic platinic tetrabromide. Some specimens were reserved as controls of the effects of enzyme treatment. Stained and unstained specimens were embedded in methacrylate and examined by light and electron microscopy.
Basophilic and Feulgen-positive substances, after contact with watery reagents, were found by electron microscopy to exist as small dense granules embedded in a less dense homogeneous matrix, forming the walls of submicroscopic vacuoles. These granules were absent after digestion with nucleodepolymerases. In specimens (unstained, or stained with platinic tetrabromide) which had not passed through water, the dense (basophile) substances in nuclei and cytoplasm were found to exist, not as granules, but as ill defined submicroscopic concentrates which blended imperceptibly into the homogeneous matrix of the vacuolar walls.
Objections to the use of stains for improving contrast conditions in electron microscopy of tissues are discussed, and it is concluded that the reagents do not necessarily produce the observed increases in contrast by selectively stabilizing certain structures. The concept of microsomes as pre-existing distinct morphological entities in intact (unhomogenized) cells is thought to be inconsistent with the distribution of basophile substances in frozen-dried liver.