278 THE POPULAR SCIENCE MONTHLY. 



uent parts ; and it is a very simple mechanical deduction that the 

 larger molecules will, as a rule, have a slower rate of vibration than 

 the smaller ones very much in the same way as a short string gives 

 a higher note than a long one. The color of chlorine changes just in 

 the way we should expect if the molecules, instead of going about 

 separately, were hanging together in couj^les ; and tlie same thing is 

 true of a great number of the metals. Mr. Lockyer, in his admirable 

 researches, has shown that several of the metals and metalloids have 

 various spectra, according to the temperature and the pressure to 

 which they ai-e exposed ; and he has made it exceedingly probable 

 that these various spectra, that is, the rates of vibration of the mole- 

 cules, depend upon the molecules being actually of diiferent sizes. 

 Dr. Roscoe has, a few months ago, shown an entirely new spectrum 

 of the metal sodium, whereby it appears that tliis metal exists in a 

 gaseous state in four different degrees of aggregation, as a simple 

 molecule, and as three or four or eight molecules together. Every 

 increase in the complication of the molecules every extra molecule 

 you hang on to the aggregate that goes about together will make a 

 difference in the rate of the vibration of that system, and so will make 

 a difference in the color of the substance. 



So, then, we have an evidence, you see, of an entirely extraneous 

 character, that in a given gas the actual molecules that exist are not 

 all of the same weight. Any experiment which failed to detect this 

 would fail to detect any smaller difference. And here also we can see 

 a reason why, although a difference in the size of the molecules does 

 exist, yet we do not lind that out by sifting. Suppose you take oxygen 

 gas consisting of single molecules and double molecules, and you sift 

 it through a plate ; the single molecules get through first, but, when 

 they get through, some of them join themselves together as double 

 molecules ; and, although more double molecules are left on the other 

 side, yet some of them separate up and make single molecules ; so the 

 process of sifting, which ought to give you single molecules on the one 

 side and double on the other, merely gives you a mixture of single and 

 double on both sides ; because the reasons which originally decided 

 that thei'e should be just those two forms are always at work, and 

 continually setting things right. 



Now let us take the other point in which molecules are very nearly 

 alike; viz., that they have very nearly the same rate of vibration. 

 The metal sodium in the common salt uj)on the earth has two rates of 

 vibration ; it sounds two notes, as it were, which are very near to 

 each other. They form the well-known double line. The two bright 

 yellow lines are very easy to observe. They occur in the spectra of 

 a great number of stars. They occur in the solar spectrum as dark 

 lines, showing that there is sodium in the outer rim of the sun, which 

 is stopping and shutting off the light of the bright parts behind, and 

 all these lines of sodium are just in the same position in the spectrum, 



