It’s the first half of the twentieth century. Mendeleev has just come out with a basic table of elements that looks like it might, one day, order all the possible known elements known to science. You want to improve it. How do you do that?
If you’re econometrician Anton van den Broek, you come up with two ideas. One is an expanded table of elements. It incorporates newly discovered elements, and leaves space for not-yet discovered ones. It’s a great step forward, and looked at as a contribution to science. You also come up with the concept of the alphon. An alphon is half a helium nucleus, and has an atomic weight of two units. Looking back, with more complete knowledge, you can’t help but think, “Oh. So close.” Van den Broek believed that all elements would increase by two atomic mass units, and that the alphon was the basic building block of the world. He hadn’t separated the half-helium nucleus into protons and neutrons. That would come later, when people learned more about mass and charge inside the nucleus.
Alphons are a footnote of science history. Even van den Broek didn’t mention them in more than one paper. They are a reminder, however, of the way science actually works. No concept is a fait accompli. The available data can be interpreted many different ways, and many ideas that seem so promising are actually wrong turns. Even science wreckage can be interesting. Edited from io9