136 



case the observation does not furnish the complète fre- 

 quency curve. We hâve not the number of houses of 

 value between £ and 1 £; between 1 £ and 2 £ etc. 

 We know only that below 10 £ the total number is 

 3 175 000, which is more than half of ail the houses 

 together. The conséquence is that really the reaction curve 

 can be assigned with certainty only from x = 15 £ unwards. 



As, from this point on, this curve is evidently well 

 represented by a straight line, w^ are, however, naturally 

 led to assume that it will still be represented by this line 

 for lower values of x. That therefore also the z curve 

 will be logarithmic throughout. In this supposition we find 

 (and for this interpolation it seemes somewhat better to 

 rely on the simple mathematical computation given Ist paper 

 p. 42) that the reaction curve cuts the x-axis at x = £ 2.2. 



Even this resuit, uncertain as it still is, is in accordance 

 with what we should expect. For it is évident that there 

 is a lower limit différent from zéro to the value of a 

 house. Our resuit places this limit at 2.2 £ which seems 

 reasonable enough, though of course we lay no stress 

 whatever on the accuracy of this détermination. As soon 

 as more detailed data for the very low values shall be 

 available, the last remaining uncertainty will be removed. 



The frequency curve as shown in our figure has been 

 drawn for the very small values of x, in accordance with 

 the above supposition, that is to say in the supposition 

 that for thèse small values too the log. {x — 2.2) are nor- 

 mally distributed. 



Examplc IV. Diameter of Spores of Mucor Mucedo, 

 measured by Mr. G. Postma in the botanical laboratory 

 at Groningen (unpublished) (tab. 7, fig. 9). 



What draws the attention in the frequency curve is the 

 enormous accumulation of individuals near x = 20.5. 

 Corresponding therewith we find a very strong minimum 

 in the reaction curve. We thus get an indication that at 



