4 o 4 SCIENCE PROGRESS 



results have been "cooked " or "faked," to use terms familiar to 

 many students and teachers of chemistry. The work is un- 

 doubtedly thoroughly conscientious but either the problem has 

 been incorrectly stated or the conclusions have been uncon- 

 sciously forced to fit the preconceived ideas of the investigator. 

 There is a fascination in arriving at a mathematical "proof" of 

 an hypothesis and especially in obtaining a good "constant" 

 which makes it imperative that chemists should be taught to be 

 most cautious and critical when their results seem to confirm 

 their anticipations. We are apt to overlook the fact that a 

 number of different equations may often fit a given set of facts 

 equally well and that a variety of assumptions may lead to just 

 as good a constant. 



One example taken from the Zeitschrift filr physikalische 

 Chemie may help to show how easily errors are made and over- 

 looked. It is from a paper by A. A. Noyes and J. Seidensticker 

 which confirmed (?) Jakowkin's conclusions that in solutions of 

 iodine and potassium iodide in water the iodine combined with 

 the potassium iodide to form the tri-iodide in the manner 

 indicated by the expression : KI + L = KI 3 . They determined 

 the solubility of iodine in aqueous solutions of potassium iodide 

 of different concentrations and assumed that the concentration 

 of the free iodine was the same as in water alone at the same 

 temperature and that the excess of iodine over that dissolved by 

 water was combined with the potassium iodide in the form of 

 KI 3 . By deducting the solubility of iodine in water, they ob- 

 tained the concentration of the iodine combined as KI 3 for each 

 concentration of potassium iodide — one molecule of iodine, 

 I 2 , forming one of KI 3 . As the concentration of the uncombined 

 iodine was by assumption constant, the mass action equation 

 for equilibrium, viz. : 



(KI,) = k (free KI) (L), 

 simplified to : 



(KI 3 ) = k l (free KI), 



where the brackets represent the molecular proportions of the 

 respective substances. It followed that (KI 3 ) should be propor- 

 tional to (KI 3 ) + (free KI), that is to say, to the original concen- 

 tration of the potassium iodide. They found that this ratio was, 

 in fact, approximately constant, as the following table shows ; 

 hence they concluded that their assumption was confirmed. 



