Value of families in 1914 for. 
448 BOTANICAL GAZETTE [DECEMBER 
sugar in rows 435-449 has not been accounted for. These rows 
belong to strip 2, as indicated in the figure, and not to strip 3, as 
might be inferred from their behavior. 
A variation of 2 per cent sugar occurs between different parts 
of the same field, as shown in 
M4 Average peat of sugar figs. 27 and 28. ‘These local 
= © » %& -& 2% %& & differences are not limited to 
se GN eS Ss 
SNe tS of <<. S the progeny. rows, but appear 
f /265 also in the check rows, and 
& 325 / _} hence cannot be attributed to 
: ae ae accidental groupings of good or 
& y poor families. The yields ex- 
S445 hibited in fig. 30 are fully as 
we cece ye variable as the percentages and 
show striking progressions with 
frequent changes in direction. 
TRANSMISSION OF QUALITIES EXHIBITED BY SELECTED FAMILIES. 
—Average root weight and average percentage of sugar are the 
chief factors which determine the relative merits of sugar-beet 
Fic. 31.—To accompany table XXVII 
Value of families in sore for oe ie ig of sugar 
expressed in percentage of their checks 
in percentage of their checks 
&> 
te 
> 
percentage of pe expressed 
/ 
Ss (Mom we Se a ke a 
Sse kee Se 8 eS 
95.5 == 
97.5 | 
me Y 
Fic. 32.—-To accompany table XXVIII 
families grown under uniform spacing, as they represent both 
yield and quality; but since these characters are very easily 
modified by the environment, it is doubtful whether real family 
differences are ordinarily distinguishable. In fact, fluctuations 
appearing under field conditions probably exceed the real differ- 
