A major problem in the determination of the magnitude of a possible solar effect on climate is that no physical parameter of solar energy output exists that has been observed long enough to be used for long-term analyses. Therefore, a number of indirect parameters have been proposed, with the sunspot number as the most commonly used parameter. Recently it has been suggested that climatic effects may be more directly associated with the length of the solar cycle. Whereas the magnitude of the sunspot number is only believed to be reliable back to 1750, determination of solar activity minima may be based on other types of data. A recent reconstructed series of solar cycle lengths back to 1500 gives new information about solar activity in particular before and during the Maunder Minimum. A comparison with reconstructed temperature records has revealed that the good agreement between the solar cycle length and the global temperature found for the modern instrumental temperature record is also characteristic for the total series of reconstructed temperature data. A further result is that the response of the temperature during the pre-instrumental era is the same as for the modern temperature record. This finding confirms the close association beween terrestrial temperature and solar activity measured in terms of the solar cycle length.