The increasing number of laboratory experiments related to the solar wind and the magnetosphere calls for a critical review of the conditions under which a laboratory experiment can provide information about space phenomena. Naturally, the starting point is a discussion of the problem of physical similarity. For a unified presentation of the several sets of scaling laws that have been used, it is convenient to use the similarity laws of the Vlasov theory as a reference system. Clearly, the biggest obstacle for a strict reproduction of the solar‐wind‐magnetosphere interaction in the laboratory is collisional interaction. In the solar wind at the earth's orbit, strict simulation is possible only for phenomena with a dimension of the order of a few hundred kilometers or smaller, such as the bow shock structure. A discussion of the existing terrella experiments shows that most of them obtain similarity by scaling the interplanetary field and the geomagnetic field differently. This determines the processes for which quantitative agreement can be expected, e.g., the shape of the frontal part of the cavity boundary and the approximate shape and position of the bow shock. Strict similarity can be achieved in simulating the structure of the bow shock wave. The over‐all ranges of the shock parameters of seven collision‐free shock experiments discussed cover substantial parts of the bow shock parameter ranges. A comparative discussion of the magnetic‐field profiles of the estimated shock width and of the appearance of oscillatory structures shows strong resemblances between space and laboratory observations. Several other experiments related to the solar wind and the magnetosphere are al