158 THE COTTON PLANT IN EGYPT chap. 



The next character " S," or distance from petiole to 

 sinus, is inherited in much the same way as L. We can 

 eliminate the diversities in absolute size by expressing S 

 in terms of L, i.e., plotting for the ratio S/L. We then 

 find that an Upland leaf, in which S is relatively long 

 when crossed with an Egyptian in which ^ is relatively 

 short — or, in other words, slightly and deeply dissected 

 leaves — give rise in Fj to a leaf very like the Egyptian, 

 but with a slightly longer S. In Fo we obtain the same 

 modal curve, stretching to the parental extremes. Cor- 

 relation diagrams for L and S in Fa show modal grouping 

 and no general correlation. In other words, the length of 

 the sinus is inherited independently of the length of the 

 mid-rib. Now this sinus-measurement is determined by 

 the plant at a very early stage in the primordia, since we 

 find freshly expanded leaves showing the same form as 

 adults ; hence a factorial determination is highly probable. 

 It is necessary to express S in terms of L, on account of 

 the diversities of L. Thus, if we plot S simply, we find, 

 e.g., in the King x Charara cross, that the parents are 

 identical, while the Fi is 60 per cent, longer. It is, of 

 course, quite possible that our use of L as the expression 

 for mere size may be fallacious, but this would not affect 

 the general argument for factorial inheritance of form. 



The last component of the form of the central segment 

 is L s. We have formerly seen that the position of the 

 sinus fluctuates along the angle-lines, so that instead of 

 using S and ^ s we might employ ^ s and l t, the latter 

 being the angle made by the sinus with the mid-rib from 

 the tip of the latter. 



On crossing the narrow angle of Upland with the 

 wide angle of Egyptian we obtain a wide angle in Fi. In 

 F2 we obtain the same wide and modal curve on plotting 

 the angle values, with some indication of simple segregation 

 of narrow angle from wide, which is probably fallacious. 



The next point to consider is the inter-relationship of 



