1916] PRITCHARD—SUGAR-BEET BREEDING 459 
gradual accumulation of small real differences is still an open 
question. Since aside from the progeny test no method of dis- 
tinguishing real differences between beet roots has been in vogue, 
the selection of choice roots by chemical and physical means 
probably has played no part in sugar-beet improvement except 
where an occasional root has mutated and thus given rise to a 
superior physiological species. 
The real differences between sugar-beet families are usually very 
small, as may be noted from the difference in size of corresponding 
SN Row numbers 
SSE SESESSSISERRESERSESERESSERE 
ae a 
32 be Li 
30 Q ‘ 
27. fet A a Mt 
25 | php | x, 
225 ‘ ee at 
2 B 
178 at > Meet ris 
Piel tet et @ ) oe a et 
2 CEE EEC CEEECEEEELEEECEH 
Fic, 5°.—Relationship between actual yield and corrected yield of sugar in 
Consecutive check rows, Madison, 1913; solid line indicates actual yield; broken line, 
corrected yield, 
variations of individual check and progeny rows. In fact, their 
teal differences are greatly exceeded by their fluctuations. e 
Both the best and the poorest families transmit average qualities. 
Hence continuous selection does not seem to be an efficient means 
of improvement. Moreover, it is difficult to conceive how it 
could have played any important part in sugar-beet improvement 
in the past. — 
The isolation of mutations probably offers more promising 
©pportunities for improving beets than continuous selection. Our 
records show no evidence of mutations, but the numbers are too 
small to disprove their occasional occurrence. Moreover, no table 
Presented contains data for more than two consecutive generations, 
