Il8 Kleinere Mitteilungen. 



Further researches in the statistics of Cojf'ea. 

 (Second communication.) 



By P. C. van der Wölk (Buitenzorg). 

 Eingegangen: ic. Mai 1913. 



§ I. 



My further investigations respecting the statistics of the varieties of 

 Coffea have led in the first instance to results which may up to a certain 

 point be indeed called a direct proof of the interpretation given in my 

 former treatise 1) dealing with increased or diminished variation in the 

 number of leaves in consecutive branch-pairs of one and the same tree 

 per one meter branch length of Coffea Uganda and C. Robusta respec- 

 tively. I was then enabled, on the basis of curves and tables, to demon- 

 strate a striking parallelism between the greater variability designated 

 in hganda (as contrasted with Robusta) and the fact that in the former 

 plant the greater number of leaf-length-properties are irregularly distri- 

 buted over the plant, while on the other hand with C. Robusta, in which 

 in every part of the tree the greater number of properties of leaf-length 

 occur in the same way, in the same proportions, in the same constellation, 

 the above designated variation in the number of leaves per meter of branch 

 length is practically nil. 



In the same manner as with the properties of leaf-length I sub- 

 sequently investigated the lengths of the internodes of the side branches. 

 Since with each branch-pair the length of the internodes of both branches 

 practically coincide in their variations and dimensions we may consequently 

 take it as sufficient in future to record only one of the branches of each 

 branch-pair. A certain periodicity in the length of internodes naturally 

 manifests itself in every such branch. We see however by this time that 

 this periodicity in the successive branch-pairs of one and the same tree 

 appears as those which I have been able to demonstrate in the previous 

 publication on leaf-length properties; viz: at different heights of the tree 

 a definite curve is repeated, definite curve-tops persistently make their 

 appearance again and again-). We must regard this phenomenon as the 

 expression of the fact that a greater number of different properties are to 

 be found in the tree which bring about the lengths of the internodes, but 

 that these qualities are not continually actuated in the same way; so that 

 at different heights of the tree different of these qualities predominate: 

 I need not however go further into this question, since it is amply set 

 forth in my former treatise. A similar kind of presentment of a frequency 

 curve I have named the analysis of frequency curve. 



1) See vol. 10 of this periodal. 



-) See, for instance, curve III of the first communication in vol 9. 



