134 Journal of Agricultural Research voi. xxvi, no. 3 
The results here shown give a correlation coefficient of 0.462 ±0.0188 
between the percentage of sugar in the beet at harvest and the decrease 
in percentage in individual beets during storage. 
These tables also indicate that the beets of high percentage at harvest 
tend to decrease more in percentage of sugar during storage than do the 
low-testing beets; as a consequence, beets of high test at harvest tend 
to be low in the spring, while beets of low test at harvest tend to be 
relatively high in the spring. The inclination on the part of the breeder 
is, therefore, to select the low fall-testing beets, since these show a rela¬ 
tively higher percentage of sugar at planting time. A breeder, for ex¬ 
ample, guided only by the spring-testing data (see Table I) would select 
for seed production the beets tagged 74, 103, no, and 155 as the most 
desirable among the first 100 listed. In making his selections he would 
choose beets testing the highest according to his data—that is, beets 
showing 17 or more per cent of sugar. Now, on examining the fall tests 
one finds that there were really 11 beets containing 20 or more per cent 
of sugar. Their tag numbers are 39, 79, 101, no, 119, 134, 140, 152, 154, 
155, and 156. So our breeder would select two beets containing 20 per 
cent of sugar and would neglect nine other beets containing 20 to 21 
per cent. With these two 20 per cent beets he would select two others 
of less than 20 per cent. Therefore, after a few years his lines should 
show a decrease in the average percentage of sugar. 
Friedl did not give the weight of each beet along with the percentage 
of sugar which it contained, and, consequently, we do not know just how 
much of the decrease in sugar was due to the water intake of each beet 
during storage. He made dry weight determinations on other beets 
and found an increase of 6 per cent of water during storage which would 
correspond to about 1 per cent decrease in the percentage of sugar. 
Therefore, his results show that there was a too great difference between 
the fall and spring tests to be accounted for by decrease in the percentage 
of sugar (due to intake of water) and it is evident that there was also a 
marked loss of the actual sugar in the beet. 
In the present work the order of weighing and sampling beets was 
arranged so that comparable weights and percentage of sugar were 
obtained for each beet. The order was as follows: First, sample (at 
harvest time), weight of sampled beet, storage for 98 days, weight of 
sampled beet; and second, sample (98 days after first sample was taken). 
Table IV gives for each beet these weights before and after storage, the 
sugar content before and after storage, and the loss of sugar in grams 
during storage. The same tag numbers in this table and Table I denote 
identical beets. The sugar content of each beet before and after storage 
is the product of the weight and its corresponding percentage of sugar 
(from Table I). For example, beet tagged 1 before storage tested 19.0, 
weighed 1,100 gm. and contained 209 gm. of sugar; after storage this 
beet weighed 1,110 gm., tested 13.2, and contained 147 gm. of sugar. 
