352 HANS AND OLOF TEDIN 



reached about their full length, the length of six consecutive inter- 

 nodes was measured. We have thought it needless to divide the 

 numbers thus obtained by six, and in the following is always given 

 the length of six internodes as the standard of internode length. 

 When analysing the material mathematically it proved convenient to 

 group it in classes of 5 cm. each. 



The stem thickness was measured on the part of the plant where 

 the stem has gained its full thickness. This measuring was made with 

 an accuracy of 0,5 mm. When working through the material it 

 proved, however, that the number of plants in the V2 mm. classes 

 (viz. 2,5; 3,5; 4,5 mm. etc.) was very often smaller than the number 

 of plants in the neighbouring whole mm. classes (viz. 2; 3; 4 mm. 

 etc.). Certainly this is due to incorrect observation and of no real 

 significance. When taking the classes together two and two it is, 

 however, impossible to decide how many of the plants in an ^/o mm. 

 class are to be calculated in each of the neighbouring whole mm. 

 classes. Hence the original data are given in the following. 



In 1919 the total length of undamaged plants was determined. 

 Already in 1919 several plants were damaged and in 1920 there were 

 too many plants damaged to allow any analysis of the full length. 



The place of the first flower is indicated by the number of sterile 

 nodes below the first fertile one. According to Keeble & Pellew 

 (1910, in the following only cited as K. & P.) there is »a considerable 

 range of variation in this character». It is, however, of great im- 

 portance that every sterile node is calculated, even the first, which 

 bears no developed leaf and is concealed in the soil. Further, basal 

 branches may on no account be calculated, as they mostly have four 

 or five sterile nodes less than the corresponding main axis. This 

 error is easily committed in suth cases, where the main axis has died 

 off in an early stage. If necessary precautions are taken, distinct 

 differences between different pure lines may easily be seen (cfr. 

 Tedin 1897), and these differences are quite constant, even if the 

 number of sterile nodes may vary a little in different years. 



DESCRIPTION OF THE LINES USED FOR CROSSING. 



Ragiinda-i . A pea from northern Sweden. Comparatively long 

 internodes, thin stem, number of sterile nodes 9 — 13. (For more 

 detailed account, see the tables in the following). 



