The sunspot record since 1610 shows cycles of magnetic activity with an irregular distri- bution of amplitudes and with a period around 11 years; they are modulated on longer timescales and were interrupted by the Maunder minimum in the 17th century. Solar magnetic activity is associated with the emergence of magnetic flux at the solar surface, giving rise to sunspots, flares, coronal mass ejections and other magnetic manifestations. Solar activity is also thought to be responsible for changes in solar irradiance and may be connected to climate changes in the past.

In order to understand how the Sun behaved magnetically prior to the first systematic observations (from 1610 onwards) we have to rely on the use of proxy data for solar activity such as the cosmogenic radioisotopes carbon 14 (C14) and beryllium 10 (Be10).

In this talk I will present and comment the results of the analysis of a time series of the solar magnetic activity (reconstructed from the concentration of the isotope Be10 in polar ice-cores) going approximately 10,000 years back into the past. Three important points are [1] whether the solar cycle has existed during all this time, [2] whether its period has always been around 11 years and [3] whether an interruption such as the ‘Maunder minimum’ in the 17th century was an isolated or a recurrent feature of solar magnetic activity.

I will also address the following questions: How cosmic rays produce cosmogenic isotopes in the Earth’s atmosphere, how the Earth’s magnetosphere modulates the production of cosmogenic isotopes, why C14 and Be10 are the best proxies for reconstructing solar magnetic activity and why in spite of Be10 having a much better time resolution than C14 we need both of them.