A correlation analysis shows that the sunspot numbers at the peaks of the last eight solar cycles are well-correlated with the sunspot numbers in heliolatitudes 20o-40 ~ (specially in the southern hemisphere) occurring in the solar minimum years immediately preceding the solar maximum years.
Among the various methods used for attempting to predict solar activity, some are based upon identifying some parameters which could serve as precursores of sunspot number. For example, Ohl (1968) showed that the level of geomagnetic activity (as indicated by kp index), in the descending branch of a solar cycle, was well correlated with the height of the following cycle. Recently Sargent (1977) and Kane (1978) showed that the 'aa' index (Mayaud, 1973) at the minimum of a solar cycle, could serve a similar purpose. However, the total annual sunspot numbers themselves at sunspot minimum do not correlate well with sunspot numbers at the following maximum. The correlation coefficient is only +0.27 (Kane, 1978). Recently, Schatten et al. (1978) suggested on physical grounds that the Sun's polar fieldstrength near a solar minimum, may be closely related to the following cycle's solar activity. They tested this hypothesis by plotting the sunspot number at maximum versus four different measures of polar magnetic field strength viz. (a) Ludendorf index indicating coronal flattening, (b) bending of high heliolatitude polar plumes, (c) flattening of the current sheets in interplanetary space, obtained from the towardsthe-Sun polarity of interplanetary magnetic field and (d) the number of faculae at the solar poles. The plots, in their Figure 1, are indicative of a positive relationship, but the scatter is very large, indicating poor correlations. On the other hand, the physical grounds, on which their analysis is based, appear quite reasonable. It strikes us rather strange, that no worthwhile telltale effect is left on the Sun, which could give some indication of the future activity. Though the gross sunspot number at solar minimum has provided a poor precursor, Brown (1976) has found a good correlation between sunspot minimum and the following sunspot maximum, by using the secularly smoothed data (moving averages over successive cycles, Gleissberg, 1967). Also, it is known that sunspot activity has strong dependence on heliolatitudes. Could it be, then, that the tell-tale evidence is depicted not in the polar field strength, as such, but in sunspot numbers in some heliolatitude belts?