ATOMIC WEIGHTS REFERRED EITHER TO HYDRO GEN OR TO OXYGEN. Being purely relative quan tities, the atomic weights of the elements assume definite values only if referred to some fixed num ber chosen to represent one of them. Different sets of atomic weights may therefore be obtained by choosing different standards, and one such set is obtained by assigning to the atomic weight of hydrogen the value 1. This standard seemed at one time preferable to any other, in the light of Prout's hypothesis. According to this the several chemical elements are not really differ ent from one another in substance, but are con densation products of hydrogen. It was conse quently expected that careful atomic weight determination would yield whole numbers, if the atomic weight of hydrogen were taken as the unit. Experiment of the highest precision has, how ever, shown conclusively that the atomic weights cannot he represented by commensurable num bers; so that the common derivation of the ele ments from one of them is at least extremely improbable, and choosing a 'common unit of measurement' for the atomic weights is as good as entirely useless. Nevertheless, the choice of
a standard need not be altogether arbitrary. The actual determination of atomic weights involves analytical processes in which many of the ele ments are weighed as oxides, the required atomic weights being calculated from the weights of these and from the atomic weight of the oxy gen contained in them. Hence, for the sake of sim plicity in precise calculations, it is desirable to have the atomic weight of oxygen represented by a whole number. plainly for this reason chemists have chosen the atomic weight of oxygen itself as a standard, assigning to it the value 16 (the atomic weight of hydrogen is then about 1.01).