It is apparent, on examining the evidence thus far at hand, that the fact of permanent, progressive changes in climate during historical times has not yet been definitely established.
Periodic Oscillations of Climate: Sunspot Period.—The discovery of a distinct eleven-year periodicity in the magnetic phenomena of the earth, naturally led to investigations of similar periods in meteorology. Numerous and varied studies along this line, extending back even into the seventeenth century, but beginning actively about 1870, have been and are still being prosecuted by a considerable number of persons, and the literature on the subject has assumed large proportions. The results, however, have not been satisfactory. The problem is difficult and obscure. It is natural to expect a relation of this sort, and some relation certainly exists. But the results have not come up to expectations. Fluctuations in temperature and rainfall, occurring in an eleven-year period, have been made out for certain stations, but the variations are slight, and it is not yet clear that they are sufficiently marked, uniform and persistent over large areas to make practical application of the periodicity in forecasting possible. In some cases, the relation to sunspot periodicity is open to debate; in others, the results are contradictory.
Köppen has brought forward evidence of a sunspot period in the mean annual temperature, especially in the tropics, the maximum temperatures coming in the years of sunspot minima. The whole amplitude of the variation in the mean annual temperatures, from sunspot minimum to sunspot maximum, is, however, only 1.3° in the tropics and a little less than 1° in the extra-tropics. There are, however, long periods during which there appears to be no influence, or at least, an obscure one, and the relation before 1816 seems to have been opposite to that since then. More recently Nordmann (for the years 1870-1900) has continued Köppen's investigation, using the mean annual temperatures of certain tropical stations, and finds that the mean temperatures run parallel with the sunspot curve, but that the minimum temperatures occur with the sunspot maxima (amplitude 0.7°). This seems to contradict Köppen's conclusion, and also the fact that the sun is hotter at a time of maximum sunspots. The latter difficulty has been explained on the ground that the rainfall and cloudiness, both of which are at a maximum with the sunspot curve, lower the temperature, especially in the tropics. It is obvious that the situation in this matter is rather confusing just at the present time, and that the relation of sunspots and terrestrial temperatures is not wholly clear. The sunspots themselves are probably not the immediate or sole control. "There seems little doubt," says Sir Norman Lockyer, 'that we must look to the study of the solar prominences not only as the primary factors in the magnetic and atmospheric changes in our
