I have a radium dial luminous watch manufactured in Switzerland back in 1940's. I took a gamma spectrum of the watch with a NaI scintillation probe and found minor but definite discrepancies to Ra-226 spectrum found in the scientific literature. Attachment is the plot of the experimental gamma spectrum I took against Monte Carlo simulation. It's obvious that the areas of the 186 KeV peak (second to left) in both spectra don't agree, with experimental value greatly exceeding theoretical calculation. Let's look at Ra-226 decay chain. Suppose we have a pure Ra-226 source with initial activity of 10^6 Bq, after 70 years (from 1940 to 2011), what remain are: Isotope (e.g. U238):Ra226 Ra226 halflife: 50491081451.52s Initial activity (in Bq):1e6 Time (s/m/h/d/y):70y Isotope:Ra226 Activity:970130.006496587 Halflife:50491081451.52 Isotope:Rn222 Activity:970136.353355819 Halflife:330324.48 Isotope:Po218 Activity:970136.356896194 Halflife:184.26 Isotope:Pb214 Activity:969709.574869973 Halflife:1614.96 Isotope:Bi214 Activity:969709.644640781 Halflife:1188 Isotope:Po214 Activity:969302.409354230 Halflife:0.0001623 Isotope:Pb210 Activity:868672.539092521 Halflife:701510462.917056 Isotope:Bi210 Activity:868610.377548218 Halflife:433036.8 Isotope:Po210 Activity:866860.422794737 Halflife:11955712.32 Isotope:Tl210 Activity:0.04276419539472 Halflife:78 Isotope:At218 Activity:0.04695459967508 Halflife:1.4 Halflives are in seconds. Check radioisotope database and we can get theoretical gamma emission rates of the chain: Isotope (e.g. U238):Ra226 Ra226 halflife: 50491081451.52s Initial activity (in Bq):1e6 Time (s/m/h/d/y):70y Theoretical line: 186.211, probability: 34488.1217309537 Theoretical line: 241.997, probability: 70478.4919015496 Theoretical line: 295.224, probability: 178562.321116557 Theoretical line: 351.932, probability: 345216.60865371 Theoretical line: 609.312, probability: 441120.917347091 Theoretical line: 1120.287, probability: 144583.708015941 Theoretical line: 1764.494, probability: 148462.546594504 ... As you can see, the emission rate of 186.211 KeV line is much less than those of the neighboring peaks, say 241.997, 295.224 and 351.932 KeV. Areas of corresponding peaks in Monte Carlo spectrum reflect the proportionality. Similar experimental and simulation results can be found elsewhere, so I believe my experiment and calculations are reliable. The Monte Carlo simulation, utilizing GEANT4, takes into account detector geometries and lab environment, such that influences of back-scattering peak etc can be excluded. So it must be assumed that Ra-226 is not the sole component of the radioactive material in radium paint. The question is, what is it? Ra-228 was also put into radium paint to extend luminous lifetime, but it does not have lines around 186 KeV. U-235 does have one, but it's unlikely SNM is involved here. I've been confused for a long time and sincerely hope someone can explain this to me. Thanks a lot!