2 ARKIV FÖR BOTANIK. BAND 12. NIO 14. 



bias for more stately and conspicuous specimens I had, practi- 

 cally, all specimens gathered at a certain spot of the wood 

 (at the north side of the so called »Persian bay»). In all 

 469 samples were examined. 



From the primary observing list a card catalogue was 

 formed. Such a card catalogue is indispensible as soon as 

 more than one character in a given population is considered. 

 The mode of treatment of the material was that given in 

 my »Grunddragen af den matematiska statistiken».^ 



According to the classification of statistical problems 

 given in this pamphlet, I consider here separately A) Homo- 

 grade statistics, where only the number of samples having a 

 certain character — and not the intensity of the character 

 — is considered and B) Heterograde statistics, where the 

 samples are examined regarding the different degrees of in- 

 tensity of a certain character. 



A) Homograde statistics. 



The characters considered, belonging to this part of sta- 

 tistical problems, were the number of flower-stalks, stamens, 

 petals, sepals and leaves in the rosule (at the top of the 

 stem). 



1. Number of Flower-stalks. All the 469 samples were 

 examined regarding this character. The result of the counts 

 were: 



Number of samples with 1 flower-stalk 340 

 » » » » 2 flower-stalks 117 



» » J »3 » 12 



Summa 469 



Besides one specimen was found having 4 flower-stalks. 

 As, however, this specimen was found, occasionally, at an- 

 other place I have not taken it in consideration in computing 

 the »probabilities» for finding a specimen with a certain 

 number of stalks. (In other words, I suppose the probabi- 

 lity for finding an individual with 4 flower-stalks to be 

 different from 1 : 470.) 



By dividing 469 into 340, 117 and 12 we shall find the 



^ Statistisk Tidskrift (1910). Extrahäfte. Some copies are still to be had. 



