We have already briefly mentioned the discovery of scandium in the chapter devoted to REEs (see p. 130). Although many of scandium’s properties are similar to those of rare earths, D. I. Mendeleev predicted that the element would be a boron analogue in the third group of the periodic system. His prediction proved to be accurate enough. Scandium was discovered by the Swedish chemist L. Nilson; on March 12, 1879, his article “On Scandium, a New Rare Metal” was published and on March 24 it was discussed at a session of the Paris Academy of Sciences.

Nilson’s results, however, were in many respects erroneous. He considered scandium to be tetravalent and gave, therefore, the formula of its oxide as ScO2. He did not measure the atomic mass and gave only its probable range (160-180). And, finally, Nilson suggested that scandium should be placed in the periodic table between tin and thorium, which ran counter to Mendeleev’s prediction.

The discovery of scandium excited the scientific community and Nilson’s compatriot P. Cleve set out to study the newly discovered element. He studied it thoroughly for almost five months and came to the conclusion that many results obtained by Nilson were erroneous. Cleve reported to the Paris Academy of Science on August 18, and the academicians learnt much new about scandium. It turned out to be trivalent; its oxide’s formula was Sc2O; its properties differed somewhat from those determined by Nilson. According to Cleve (and this was especially important) scandium was the eka-boron predicted by Mendeleev; Cleve showed a table in the left-hand column of which eka-boron properties were given and in the right-hand one those of scandium the following day Cleve sent a letter to Mendeleev in which he wrote: “I have the honour to inform you that your element eka–boron, has been obtained”. It is scandium discovered by L. Nilson this spring.

And, finally, on September 10 Cleve published a long article about scandium from which it is clear that he had a much better understanding of the new element than Nilson. Therefore, those historians are who consider Cleve and Nilson as co–discoverer of scandium right.

For a long time Nilson was working under an illusion about some of scandium’s properties and refused to recognize its identity with eka-boron. Cleve’s investigations, however, impressed Nilson very much; in the long run he was forced to admit that he was wrong, thus doing justice to the prediction power of the periodic system.

All of Mendeleev’s predictions were confirmed in the long run. The last to be confirmed was the prediction of the density of metallic scandium; only in 1937 did the German chemistry W. Fischer succeed in preparing 98 per cent scandium. Its density was 3.0 g/cm3, that is exactly the figure predicted by Mendeleev.