548 HINRICHS— TRUE ATOMIC WEIGHT OF BROMINE. [April 4. 



made, what the outcome thereof will be," or perhaps more strikingly 

 still, we might say that "we get (within sufficiently large limits to 

 count) whatever value r we would like to get." 



But the value of r directly determines the value of the atomic 

 weight itself, as we shall show in detail ; hence the fact just stated 

 for the ratio applies with equal force to the atomic weight itself. 



I\". Systematic Errors and Chemical Perturbations. 



The existence of such systematic errors in the most refined 

 laboratory work of renowned chemists, from Stas to the present, is 

 not a new discovery, for I have proved the existence thereof twenty 

 years ago. See the note presented by Berthelot at the Seance of 

 the twelftli of December, 1892.^ 



I here insert (Plate XXXVI.) a reduction to half the original 

 scale of the diagram (no. 215) published in the note just mentioned, 

 together with the diagram (no. 216) of the next note (February 

 2y, 1893). The new cut (no. 752) is a like reduction of Plate 

 I. of my "True Atomic Weights" of 1894 and represents the 

 systematic errors of Stas in his famous syntheses of silver nitrate 

 (no. 251) and of lead nitrate (no. 252). See plate XXXV. 



Indeed it is even forty years since I first pointed out the existence 

 of definite perturbations (or disturbances) in the chemical work of 

 Stas, namely at the Salem meeting of the American Association for 

 the Advancement of Science in 1869,® of which the part here in 

 question is reprinted in my "True Atomic Weights," 1894, pp. 

 65-69, under the regretfully appropriate heading: vox clamantis in 

 deserto. 



We shall, however, in the future restrict the term " chemical 

 perturbation " to such systematic errors as are expressible by a 

 definite function and therefore representable by a definite curve. 

 Such are the systematic errors in the recent work of Mr. Weber, 

 being represented analytically by an equation of the first degree (i) 

 and geometrically by the straight line A-B (cut no. 7 so. Plate 

 XXXIV.). 



' Comptes Rendus, T. 115, p. 1074. 

 ^Proceedings, pp. 1 12-124. 



