250 



SCIENCE. 



[N. S. Vol. IV. No. 87. 



It'ought to be added that the evidence 

 against Front's hypothesis seemed to many 

 to be decisive, even without the knowledge 

 of the atomic weight of oxygen which has 

 recently been acquired. But the evidence 

 can now be stated in a much more direct 

 and simple manner ; and it has gained in 

 force, for to the seven fit instances at hand 

 before there is added an eighth, which 

 happens to be the most weighty of the 

 whole. 



In order to present the evidence against 

 Front's hypothesis when we lack an accu- 

 rate knowledge of the atomic weight of 

 oxygen, we have first to assume this value. 

 We may, for one trial, assume that this 

 value is the whole number 16.00, which is 

 required by Front's hypothesis, and see 

 whether, on this assumption, the other 

 seven atomic weights in question are very 

 nearly such as the hypothesis requires.* 

 But the average deviation from the numbers 

 required by the hypothesis is one part in 

 five hundred, and one deviation amounts 

 to more than one part in three hundred. 

 We may make another trial by assuming 

 for oxygen, not the whole number 16.00, 

 but that value which shall make the sum 

 of all the deviations the least possible ; and 

 we may also take one quarter of the atomic 

 weight of hydrogen as our divisor.f But 

 the average deviations from the numbers 

 required by the theory is, even in this case, 

 one part in six hundred and the atomic 

 weight of that element for which the deter- 

 minations of friends of the hypothesis agree 

 with those of its opponents to one part in 

 thirty-five hundred, is supposed, after all, 

 to be in error by one part in five hundred. 

 The atomic weight of oxygen, computed 



* The values on this assumption are as follows : 

 Li=7.02, C=i2.00, N=14.04, O==16.00 (assumed), 

 Na=23.07, S=32.04, Cl=35.46, K=39.14. 



t The values are as follows: Li=7.00, C=11.96 

 N=13.99, 0=15.94, Na=-22.96, S=31.96, 01=35.33^ 

 K=39.00. 



expressly to give every possible advantage 

 to the hypothesis, differs from the whole 

 number required by the theory by one part 

 in two hundred fifty. 



We read in our school books of the bed 

 of Frocrustes, to which the tyrant fitted his 

 compulsory lodgers; if they were too short 

 he stretched them on the rack; if they were 

 too long he lopped off the superfluous length. 

 This fable was really a prophetic vision ; 

 the bed is Front's hypothesis ; our friends 

 who admire it want to stretch the most un- 

 yielding quantities, and to lop off numbers 

 which have been determined with the great- 

 est precision. Either the experiments are 

 in error by an amount which seems incredi- 

 ble, or the hypothesis is an illusion. If the 

 supporters of the hypothesis would avoid 

 the conclusion they must supply better de- 

 terminations, or they must detect real and 

 tangible sources of error in those already 

 made. 



The hypothesis was most interesting and 

 attractive ; it promised, if sustained by ex- 

 perimental evidence, to give the means of 

 such insight into the nature of the matter 

 and into the intimate structure of the atoms 

 that it was well worth all the attention 

 which has been given to it. That it should 

 fail of support, that its promises could not 

 be kept, is a matter of regret; but it is time 

 to recognize that our hopes are quite cut off. 

 That other elements are composed of the 

 same substance as hydrogen may or may 

 not be true, but we have now no hope of 

 proving it by determinations of atomic 

 weight. It would not be diflacult, perhaps, 

 to modify Front's hypothesis again and 

 again, so as to to bring it into some accord 

 with the facts. We may imagine, if we 

 will, that the observed numbers, if deter- 

 mined without error, would all be divisible 

 by the eighth part of the atomic weight of 

 hydrogen, or the ninth, or the tenth, or by 

 some smaller fraction . But such a hypothe- 

 sis is of no interest and of no utility, be- 



