Following the first electron micrographs of cotton in 1940, the development of transmission electron microscopy applied to native cellulose has been evolving in a series of successive advances. At first, faced with the weak contrast of the early images, the operators had to use specific electron dense contrasting agents to reveal the ultrastructure of their samples. It was thus found that all native celluloses consisted of microfibrils, with some size variations depending on the sample origin. Following this, a major advance was achieved when the electron microscopes could be adjusted with low electron doses, allowing the recording of diffraction diagrams from the electron beam-sensitive cellulose samples. Under these conditions, one could obtain information of cellulose itself and not, as before, of the contrasting agent. This important development applied to microdiffraction conditions revealed that some large cellulose microfibrils could yield spot diagrams typical of single crystals. Their recording led to a decisive progress for resolving the molecular and crystal structure of the two cellulose allomorphs, cellulose I and I. Using various combinations of diffracted beams to create the images, the so called "diffraction contrast images" could then be developed. These micrographs showed many aspects of the crystalline core of cellulose, including spectacular high-resolution images showing the molecular planes of cellulose in their crystalline environment. Today, electron diffraction, diffraction contrast imaging and low-dose electron microscopy have become major tools to follow the effect of various physical, chemical and biochemical processes at the cellulose crystalline level.