230 CHEMISTRY 



symbols of uncertain elements. That some of them were tri- and 

 others quadrivalent was well established; but their number was 

 undetermined, and the places which they should occupy were even 

 more doubtful. Some of the uncertainties still remain, and some 

 have been cleared away; but the main problem is as yet unanswered, 

 and therefore the metals of the rare earths are still actively studied. 

 Supposedly definite earths have proved to be mixtures; others, like 

 cerium, lanthanum, yttrium, and scandium, seem to be definitely 

 placed; but what shall we say of the rest? Didymium was thought 

 to be a distinct element, and yet it has been split in two; samarium, 

 gadolinium, erbium, and ytterbium are probably definite; but sev- 

 eral other metals are claiming recognition; and so, notwithstand- 

 ing the progress which has been made, a large part of the field is still 

 obscure. Through the study of the rare earths, one side of our pro- 

 blem, the nature of the elements, is open to attack; but only the 

 outworks have been carried so far. 



According to modern ideas, the integrity of an element is deter- 

 mined by two conditions; it must have a distinct spectrum and a 

 definite atomic weight. In the study of the rare earths these criteria 

 have been systematically applied, and to great advantage; but what 

 has been done elsewhere? To answer this question we must go back 

 more than forty years in time and make a new beginning. 



It was near the middle of the century that August Comte, seeking 

 to find some limits to positive knowledge, argued that it would be 

 impossible for us ever to determine the nature of the heavenly 

 bodies. Are they composed of matter like that which forms the earth, 

 or are they different in kind? on that theme we might speculate, 

 but we could never know. The prophecy was futile; Kirchhoff and 

 Bunsen, with the spectroscope, swept the limitations away, and all 

 the universe, as far as eye could reach, was found to contain familiar 

 elements, but under conditions not always like our own. Astronomy, 

 physics, and chemistry had gained a new weapon, and discovery 

 followed discovery along widely different lines. 



In the chemical laboratories the value of the new instrument was 

 immediately proved. Two metals, caesium and rubidium, were pre- 

 sently discovered by its aid, thallium and indium were found a little 

 later, and their analogies to other elements made them comparatively 

 easy to classify. The periodic system, which was developed later still, 

 gave them their proper positions among the metals, and they in turn 

 made the classification more complete, and therefore easier to esta- 

 blish. In each case the double criterion was applicable, and definite 

 spectrum was connected with definite atomic weight. I speak now of 

 emission spectra; but they are not the only kind. Certain solutions 

 give absorption spectra, and they have been of great assistance in the 

 study of the rare earths. In the identification of the elements, then, 



