"Radioactive decay is supposed to be the ultimate random process, immutably governed by an element's half life and nothing else.  There is no way to determine when a single radioactive atom will decay, nor any way to speed-up or slow down the process.  This iron clad certainty has always been the best argument of opponents to conventional nuclear fission power generation, as it means that the inevitable nuclear waste will have to be kept isolated from the biosphere for million of years (notwithstanding recent research attempts at stimulated transmutation of some of the longer lasting waste products.) When plotting the activity of a radioactive sample you expect a graph like the following, a smooth decrease with slight, random variations . Detected activity of the 137Cs source. The first two points correspond to the beginning of data taking. Dotted lines represent a 0.1% deviation from the exponential trend. Residuals (lower panel) of the measured activity to the exponential fit. Error bars include statistical uncertainties and fluctuations. (This graph stems from a measurement on the Beta decay of 137CS and was taken deep underground). What you don't expect are variations that follow a discernible pattern in the decay rate of a radioactive element, nor any correlation with outside events. But this is exactly what Jere H. Jenkins et al. found: Plot of measured 36Cl decays taken at the Ohio State University Research Reactor (OSURR). The crosses are the individual data points, and the blue line is an 11-point rolling average. The red curve is the inverse of the square of the Earth–Sun distance. (Error bars are only shown for a limited number of points). And now this surprising result of the sun's influence has been corroborated. The latest research was a collaboration of Stanford and Purdue University with the Geological Survey of Israel, rather reputable research power-houses that make these results difficult to dismiss. Their paper contains the following contour graph for the  measured gamma decay during the day plotted over several years. When comparing this with the same kind of graph of the sun's inclination during the observed date range the correlation is quite obvious:" - Winckel