John D. Anderson has done it again! It seems every few years he discovers a fascinating anomaly, and his papers are great for data-driven theorists like me because he publishes while freely admitting not to know the cause: old-style empiricism.

In 2005 I read the mammoth papers of Anderson et al. (1998, 2002) which first presented the Pioneer anomaly to the world (although most now consider this explained with a complex thermal model, I do not agree). In 2008 I read his (to me) 'mind-bomb' observational paper on the spacecraft flyby anomalies in which he pointed out an intriguing symmetry of these anomalies about the Earth's spin axis, which at the time sent me into a roller coaster of emotions, made me obsessed with applying MiHsC to it, and stimulated me into realising that MiHsC could explain it if I used 'mutual' accelerations (discussion). A crucial step.

Now he's back with a paper in the journal EuroPhysics Letters (Anderson et al., 2015, see refs) looking at the values of Newton's gravitational constant (big G) that have been puzzling alert physicists for years. G can be measured by suspending a dumbbell arrangement of two masses from a wire and then bringing a known mass up to a distance r away from one of them, and measuring the twist in the wire to work out the force (F) of gravitational attraction. Newton's F=GMm/r^2 then gives you G because M, m, r and F are known. Mysteriously, the 13 recent values of G found using this method vary by more than the likely error in the experiments, which suggests the real uncertainty is bigger, because something unexpected is going on.

The new result is that Anderson et al. (2015) have found that most of the 13 values of G vary with a pattern: a sine curve with a period of 5.899+/-0.062 years. Intriguingly they point out that Holme and de Viron (2013) showed that the length of our day varies with the same period. Anderson et al. (2015) argue that the changes in G cannot be causing the length of day changes, since a greater G would shrink the planet, spinning it up and reducing the length of day: the opposite to the correlation seen. It is possible that something else is influencing both G and the length of day, but what?

I've been fascinated by these unexplained variations in G for a while, and I attended a Royal Society workshop about it last year. I have found a 'weak' correlation in values of G with latitude, and I have been trying to investigate links with MiHsC, with no success so far. The new periodic pattern in G shown up by Anderson's paper is a potential clue, and at the limit of science any clue is priceless.

**References**

Anderson, J.D., G. Schubert, V. Trimble, M.R. Feldman, 2015. Measurements of Newton's gravitational constant and the length of day. EPL, 110, 10002. Paper

Holme, R. and O. de Viron, 2013. Nature, 499, 202. Paper

*End of The Magnificent Amberson's by Orson Welles: Old-style George Amberson wanders around a polluted city, confused and disoriented by the industrial society that has developed around him*.