

44Q 



BOTANICAL GAZETTE 



[JUNE 





by myself on one afternoon, so that I think there can be but little 

 error due to variation of judgment. All plants were taken at random, 

 and none were discarded except because of mutilation which rendered 

 the countings untrustworthy, save three which had an extra lamina 

 at the base of the peduncle. The data for the 400 flowering stalks 

 counted appears in the form of a correlation table as table I. 









TABLE 



I 













■ 



Upper leaf 











4 



5 



6 





7 



8 



Totals 



his 



< 1 A 



2 \ 6 



1 

 4 



4 

 34 



3 



66 





• * 



1 



_ 



• • 

 ft ft 



8 

 105 



/ 7 



1 



. 52 



144 





28 



• ft 



225 



w ] 8 



1(9 



• • 



2 



22 





16 



• ■ 



40 



* • 



- * 



6 





10 



I 



!7 



►3 \io 



• ft 



6 



1 







I 





2 



I 



5 



Totals.. . 



93 



242 



57 



2 



400 



The second series I secured in the woods at Palos, Athens Co., 

 Ohio, in early July, 1908. It was really too late in the season to work 

 to advantage, for many of the plants were so badly dried that they 

 had to be discarded. I see no reason for believing that this discard- 

 ing of individuals too brittle to be counted introduces any element of 

 error into the work, though it did considerably increase the labor. 

 I he countings were made in as nearly the same manner as for the 

 first series as possible. The data are given in table II. 



TABLE II 



Upper leaf 



< 

 •4 



X 



w 

 o 



- 



4 



5 



6 



7 

 8 



9 



Totals 



2 



3 



4 



5 



6 



7 



8 



I 



■ • 



I 



ft • 



* ft 



* • 



• • 



* • 



• • 



13 



24 



• • 



• ■ 



ft • 



♦ • 



2 



12 



80 



4 



1 



I 



I 



• • 



9 



121 



42 



8 



• • 



• * 



• • 



• • 



2 



8 



8 



ft * 



• • 



* • 



ft ♦ 



♦ ■ 



1 



• • 



ft * 



2 



2 



35 



227 



55 



*7 



I 



Totals 



37 

 IOO 



181 



18 



339 



The physical constants for these two lots are laid side by side in 

 table III. To permit of easy comparison the differences between 

 the constants for the two series, and the probable errors of the differ- 

 ences, are given. Comparing the means for both upper and lower 



