Latent periodicity in the interannual fluctuations of temperature was analyzed using, as an example, the data on annually mean temperature in Tomsk for the period since 1881 until 1998. Wavelet and Fourier transforms were used for the analysis. The power spectrum |Wk(s)|2 of observations of the surface air temperature revealed periodic structures of the scales of 2-3, 7-10, 12-25, and 30-50 years. The smallest scale of 2-3 year was assigned to the quasibiannual cycle. Wavelet analysis of the Wolf number for the period of 1700-1998 and critical frequencies of the F2 ionospheric layer for the period of 1936-1998 was carried out. The correlation analysis showed the presence of significant correlation only between the 22-year cycle of the Wolf number and temperature fluctuations on the scales of 12-25 years. The quasi-periodic structure on the scale of 30-50 years was noticed; in the 20th century, this structure tends to a smaller scale. The time interval between 1940 and 1960, during which the solar activity increased markedly, was revealed. It was assumed that this increase could cause the change in the atmospheric circulation, which was completed by the late 70s. This change, in its turn, caused changes in the temperature conditions.