When Dmitri Mendeleev laid out his periodic table he left gaps in it, and for some of them he described the element that ought to be there. One, below aluminium, he called eka-aluminium and gave it a weight, a low melting point and a density of about six times that of water. In 1875 Paul-Émile Lecoq de Boisbaudran found a new metal in a zinc ore by its spectrum and named it gallium. It fitted the description well, except that he measured its density at 4.7. Mendeleev, who had never held a grain of the stuff, said the number was wrong. Lecoq de Boisbaudran, who had the metal in front of him, measured again and got about 5.9. It is an odd position to be in, holding a new element and being corrected about it by a man who has never seen it. But Mendeleev had not been looking at the metal. He had been looking at the hole it had to fit.