POSITION REGRESSION— MAIN STEM. 



83 



In almost every case one finds that the whorls which were originally 

 present in this region of the plant have been in whole or in part destroyed, 

 so that it becomes quite impossible to tell how many leaves they bore. 

 Only scars and short stumps of leaves remain to indicate the original 

 presence of whorls. It is probable that the destruction of the leaves in 

 this region is due in the main to the action of plant-feeding aquatic 

 animals. In any event it is impossible to get a determination of the 

 regression of leaf -number on position in the first-formed whorls of the 

 main stem. The extent to which this destruction of leaves has gone 

 varies very much in different plants. In some cases only a few whorls 

 are found to be mutilated, while in other plants a considerable number 

 are completely destroyed. 



Table 42. — Correlation between leaf-number and position in main-stem ivhorls. 

 Series I, II, and III combined. 



r = 0.238±0.033. 



In order to show the nature of our material, so far as it goes, in this 

 respect, table 42 has been prepared. This gives for Series I, II, and III 

 combined the actually observed frequencies of main-stem whorls with 

 various leaf-numbers, in different positions. Position is indicated in 

 the same way here as in the case of the branches; that is, the most 

 proximal whorl on the stem is numbered 1, the whorl next it 2, and so 

 on in regular succession. In the cases where the main stem has 

 dichotomized the whorls on each subdivision are numbered in order, both 

 beginning at the same point. For example, if the main stem divided 

 after the formation of the 25th whorl, the first whorl on each of the 

 subdivisions would be numbered 26, since both occupy the same position 

 relative to the main axis. Since the whole number of entries for main- 

 stem whorls is so small, I have not entered each whorl separately in the 

 table here shown, but have combined them in groups of 10. This, of 

 course, would not be legitimate if we expected to get very fine results, 

 but the material is too scanty to give close results in any event, and by 

 arranging it in this way we are able to see better the general trend. 



