664 eeport— 1884. 



indigo by Baeyer, and lastly the preparation by Fischer, of kairine, a febrifuge as 

 potent as quinine, are some of the well-known recent triumphs of modern syntheti- 

 cal chemistry. And these triumphs, let us remember, have not been obtained by 

 any such ' random haphazarding ' as yielded results in Priestley's time. In the 

 virgin soil of a century ago, the ground only required to be scratched and the seed 

 thrown in to yield a fruitful crop ; now the surface soil has long been exhausted, 

 and the successful cultivator can only obtain results by a deep and thorough 

 preparation, and by a systematic and scientific treatment of his material. 



In no department of our science has the progress made been more important 

 than in that concerned with the accurate determination of the numerical, physical, 

 and chemical constants upon the exactitude of which every quantitative chemical 

 operation depends. For the foundation of an accurate knowledge of the first of 

 these constants, viz., the atomic weights of the elements, science is indebted to the 

 indefatigable labours of Berzelius. But ' humanum est errare,' and even Berzelius' 

 accurate hand and delicate conscientiousness did not preserve him from mistakes, 

 since corrected by other workers. In such determinations it is difficult if not im- 

 possible always to ascertain the limits of error attaching to the number. The 

 errors may be due in the first place to manipulative faults, in the second to inaccuracy 

 of the methods, or lastly to mistaken views as to the composition of the material 

 operated upon ; and hence the uniformity of any series of similar determinations 

 gives no guarantee of their truth, the only safe guide being the agreement of 

 determinations made by altogether different methods. The work commenced by 

 Berzelius has been worthily continued by many chemists. Stas and Marignac, 

 bringing work of an almost astronomical accuracy into our science, have ascertained 

 the atomic weights of silver and iodine to within one hundred-thousandth of their 

 value, whilst the numbers for chlorine, bromine, potassium, sodium, nitrogen, sul- 

 phur, and oxygen may now be considered correct to within a unit in the fourth 

 figure. Few of the elements, however, boast numbers approaching this degree of 

 accuracy, and many may even still be erroneous, from half to a whole unit of 

 hydrogen. And (as Lothar Meyer says) until the greater number of the atomic 

 weights are determined to within one or two tenths of the unit, we cannot expect 

 to be able to ascertain the laws which certainly govern these numbers, or to 

 recognise the relations which undoubted!}' exist between them and the general 

 chemical and physical properties of the elements. Amongst the most interesting 

 recent additions to our knowledge made in this department we may r note the 

 classical experiments, in 1880, of J. W. Mallet on aluminium, and in the same 

 year of J. P. Cooke on antimony, and those, in the present year, of Thorpe on 

 titanium. 



Since the date of Berzelius' death to the present day no discovery in our 

 science has been so far-reaching, or led to such unforeseen and remarkable con- 

 clusions, as the foundation of Spectrum Analysis by Bunsen and Kirchhoff in 1860. 



Independently altogether of the knowledge which has been gained concerning 

 the distribution of the elementary bodies in terrestrial matter, and of the dis- 

 covery of half-a-dozen new elements by its means, and putting aside for a moment 

 the revelation of a chemisti'y not bounded by this world, but limitless as the 

 heavens, we find that over and above all these results spectrum analysis offers the 

 means, not otherwise open to us, of obtaining knowledge concerning the atomic and 

 molecular condition of matter. 



Let me recall some of the more remarkable conclusions to which the researches 

 of Lockyer, Schuster, Liveing and Dewar, Wiillner and others in this direction 

 have led. In the first place it is well to bear in mind that a difference of a very 

 marked kind, first distinctly pointed out by Alex. Mitscherlich, is to be observed 

 between the spectrum of an element and that of its compounds, the latter only 

 being seen in cases in which the compound is not dissociated at temperatures neces- 

 sary to give rise to a glowing gas. Secondly, that these compound spectra — 

 ns, for instance, those of the halogen compounds of the alkaline-earth metals — 

 exhibit a certain family likeness, and show signs of systematic variation in the 

 position of the lines, corresponding to changes in the molecular weight of the 

 vibrating system. Still this important subject of the relation of the spectra of 



