On the 11th of August, 1999, the last total eclipse of the Millennium occured over Europe and parts of the East. Some part of the earth was in the path from about 09:30 to 12:36 UTC. Dick Bierman did a preliminary analysis on the data using 2-minute blocks, and similarly analysed "control data" taken on the 10th of August. The european RNG's showed a continuous trend towards lower variance on both days. There appears to be little difference between eclipse and control days but a quite different trend in the US and the european RNG's on both days. The data analysed using the specifications of the formal prediction don't support an eclipse effect but they might indicate some consistent difference in US and European RNG's as a function of time of day (or local sidereal time?). In order to evaluate the graphs, note that one datapoint is from 18 RNG's in the eclipse data and from 19 RNG's in the control data. Thus the increase in cumulative chi2 at sample 70 in the controldata (sum chi2=60, df = 19) has an extreme low probability (p = 3.87 e-06). Even with correction for multiple analysis, selecting one of 74 points, this remains impressive (p = 0.0003). The divergences of the data sequences over the three hour period are all significant, with probabilities of a few parts in 100 or less. The following figures show comparisons of the data in Europe and the US.

An independent look at the data, using the raw trials and examining the behavior of individual eggs which were actually in the path of the eclipse, shows a strikingly consistent suggestion of positive correlation of deviations in the eclipse data compared with control data from the subsequent day.