K. R. Saundkhs 143 



Column 1); or, since only 5 individuals ovontually stood above them in 

 the series, they might be counted as belonging to grades 6 and 7 (the 

 method adopted in Colunni 2). We can also make the comparison in 

 yet another way. A dividing line maybe drawn half- way down in each 

 series and the whole total made up of all the more vigorous individuals 

 (those above the line) compared with the total from the less vigorous 

 grades (those below the line), as shown in Columns 3 and 4. By this 

 last-named method ei-mrs of judgment in grading are reduced to a 

 minimum. 



The results in the F, generation exhibit this same connexion 

 between vigour and preponderance of doubles seen in F„. In this 

 case a considerable number of individuals yet remain to flower and 

 this fact no doubt accounts for too large an excess of doubles in the 

 total as it now stands. When the final totals are available, however, 

 we may expect the usual approximation to the ratio 9 D -.7 S, but no 

 very great levelling up between the excess and deficiency seen in the 

 higher and lower grades respectively, since it is chiefly the less vigorous, 

 and we may suppose single individuals, which remain to be added and 

 which will fall into the lower grades. 



Analysis compiled from those plants of the F^ yeneratioa ivhich have 

 flowered up to date. 



Grades 



The plants here as in F.^ were mixed half-hoary, quarter-hoary and 

 glabrous, being derived horn quarter-hoary parents, but in this case the 

 numbering was made irrespective of surface character'. 



The expenses in connection with the work were defi-ayed in patt by 

 grants fi'om the Royal Society and from the British Association for the 

 Advancement of Science. 



' See note to preceding Table. 



