This review traces early research on the Earth's magnetic environment, covering the period when only ground‐based observations were possible. Observations of magnetic storms (1724) and of perturbations associated with the aurora (1741) suggested that those phenomena originated outside the Earth; correlation of the solar cycle (1851) with magnetic activity (1852) pointed to the Sun's involvement. The discovery of solar flares (1859) and growing evidence for their association with large storms led Birkeland (1900) to propose solar electron streams as the cause. Though laboratory experiments provided some support, the idea ran into theoretical difficulties and was replaced by Chapman and Ferraro's notion of solar plasma clouds (1930). Magnetic storms were first attributed (1911) to a “ring current” of high‐energy particles circling the Earth, but later work (1957) recognized that low‐energy particles undergoing guiding center drifts could have the same effect. To produce the ring current and aurora, plasma cloud particles required some way of penetrating the “Chapman‐Ferraro cavity”: Alfvén (1939) invoked an electric field, but his ideas met resistance. The picture grew more complicated with observations of comets (1943, 1951) which suggested a fast “solar wind” emanating from the Sun's corona at all times. This flow was explained by Parker's theory (1958), and the permanent cavity which it produced around the Earth was later named the “magnetosphere” (1959). As early as 1905, Birkeland had proposed that the large magnetic perturbations of the polar aurora reflected a “polar” type of magnetic storm whose electric currents descended into the upper atmosphere; that idea, however, was resisted for more than 50 years. By the time of the International Geophysical Year (1957–1958), when the first artificial satellites were launched, most of the important features of the magnetosphere had been glimpsed, but detailed understanding had to wait for in situ observations.