Determining the structure of biological macromolecules, such as proteins and nucleic acids, is important for understanding their function. This leads to a mechanistic understanding of biological systems, and suggests pathways for discovery of therapeutic drugs. Structural analysis of biological macromolecules was first demonstrated in 1968 by DeRosier and Klug, with their three-dimensional reconstruction of negatively stained bacteriophage T4 [1]. Since liquid water is not compatible with the vacuum of an electron microscope, yet water retention is essential to maintain high-resolution molecular structure, methods of freezing the water in a vitreous state were eventually developed. This enabled the molecules to be imaged in a near-native state in the electron microscope. Yet, the high-resolution structure of these samples is easily compromised by electron-beam irradiation. As expected, lowering the temperature of the sample in the electron microscope was shown to reduce radiation sensitivity, and the lowest temperatures provided the most protection [2]. Thus, an electron microscope with a cryostage to maintain the low temperature, and a cryo-transfer system to avoid frosting or devitrification during specimen insertion, was called for.In 1983, development of a dedicated cryo-transmission electron microscope (cryo-TEM) was started jointly by Y. Fujiyoshi and JEOL. Membrane proteins were chosen as the first specimens to be imaged, since they play important functions in the cell for trans-membrane transport and signaling, and knowledge of the structure and function of membrane proteins at the molecular level may clarify disease mechanisms. The three-dimensional structure of macromolecules can be obtained by several methods, including electron tomography, single-particle analysis, and electron diffraction. We chose electron diffraction, optimized for cryo-TEM. Either x-ray or electron diffraction can yield atomic or nearatomic results. Previously, x-ray diffraction was favored for obtaining the best resolution, but a threedimensional crystal is required. Electron diffraction is carried out with two-dimensional crystals, which are often easier to produce from biological macromolecules. By embedding two-dimensional crystals in lipid layers, followed by vitreous freezing, we were able to use electron diffraction to study membrane proteins at near-atomic resolution in a near-native state.We used a liquid-helium reservoir centrally located in the TEM column to cool the specimen stage. The helium reservoir was surrounded by a liquid-nitrogen tank to suppress thermal radiation from the surrounding room environment. In order to avoid vibration from bubbling of the liquid helium, we made use of super-fluidity to cool the specimen, by which we could reach a temperature close to 2K. In keeping with best practices at the time, we employed a top-entry stage design for optimum stability. After much trial-and-error, the first-generation cryo-TEM was completed in 1986, which enabled highresolution imaging at 4.2K. Since then, the c...