Kleinere Mitteilungen. 



New researches into some statistics of Coffeo. 



(Third communication.) 

 By P. C. van der Wölk (Buitenzorg-Java). 



§ I. 



The further statistical working of my coffee material, based upon the 

 analysis of frequentation curves, has made me acquainted with a new and 

 astonishing case, through which the results of both my previous communi- 

 cations (i) concerning this matter are exhibited in a new light. 



The question which arises in this third communication is exemplified 

 in the curve following here below. This curve again has reference to the 

 leaf-lenghts of Coffea bganda. It is the frequency curve of the leaf- 

 lenghts of one and the same tree. The method of curve analysis is again the 

 same as that in both of my former communications. The measurements 

 began at the base of the tree, and extended gradually of the top. At 6 

 different heights of the tree the results already obtained were each 

 represented by a curve. Finally the 7''' curve, represented by a thick line 

 is the definitive end-curve of all the leaf-lengths. In both of the former 

 communications we saw how that generally the different curves, ivithin the 

 definite end curve, are present in a greater or smaller number of removings; 

 the tops of the different curves remove in all directions, whereby the 

 crucial point is still that the place of those tops is not so arbitrary. But 

 that definite tops during the development of the tree return again and again. 

 The tops i. e. the different curves are representative of different properties. 

 Thus the tree possesses a greater number of qualities of leaf-length and 

 those different qualities are not regularly and not in like capacity of 

 expression distributed over the tree. We have seen in the second communi- 

 cation how that one or more of these properties suddenly becomes latent, 

 whereby in some cases these qualities which had become latent can appear 

 later on, while an illustration also was given of how that after the 

 disappearance of the properties they never manifested themselves again. 

 In the first communication indeed a case was given of Robusta, in which 

 the different curves by means of which the tree was analysed were all 



(1) See vols. 10 and 11 of this periodical. 



