48 
BULLETIN 1489, U. S. DEPARTMENT OF AGEICULTUEE 
tainecl that Shull recommended selection within selfed lines as a 
method of corn breeding. 
The data reported by Shull are chiefly of historical interest from 
the standpoint of the present discussion. In a similar category are 
the experiments with crosses at the Connecticut station begun by 
East and reported in detail by East and Hayes (19). The latter 
have the added interest that some of the lines started by East and 
continued by Hayes and finally by Jones are the parents of the 
double cross known to all corn breeders as Burr-Learning. Burr- 
Learning is a cross of two selfed lines of the Burr variety crossed 
with a cross between two selfed lines of the Learning variety. The 
productiveness of this double cross under many of the conditions 
under which it has been tested has done more, perhaps, than any 
other single factor to arouse interest in selection within selfed lines 
as a basis for corn improvement. The comparative yields of Burr- 
Learning and five of the highest yielding commercial varieties grown 
at Mount Carmel, Conn., during the 5-year period 1918 to 1922, as 
reported by Jones and Mangelsdorf (4#), are given in Table 6. 
Table 6. — Comparative acre yields of shelled com of the double cross, Burr- 
Learning, and of five high-yielding varieties at Mount Carmel, Conn., 1918- 
1922 
[Data from Jones and Mangelsdorf (43, p. 161). The highest yielding of the commercial varieties in each 
year is marked with an asterisk (*) ] 
Designation of variety 
Acre yields of shelled corn 
(bushels) 
1918 
1919 
1920 
1921 
1922 
Average 
Burr-Learning (double cross) 
116 
88 
55 
95 
63 
83.4 
Beardsley's Learning 
*96 
54 
*79 
62 
75 
51 
51 
38 
*57 
32 
55 
85 
81 
73 
*87 
77 
48 
50 
49 
66.8 
Luce's Favorite .. - . 
62.0 
Webber - . . ... 
81 
84 
68 
64.4 
Northern White . - __ .-. 
69.5 
Century . _ ■__ 
63.6 
182.3 
64.2 
46.6 
80.6 
153.5 
65.4 
i Average of four varieties only. 
The data on the varieties are typical of varietal experiments in 
that no one variety was most productive in any two years. It is 
this condition which makes still more striking the fact that Burr- 
Learning yielded more than the most productive of the five varieties 
in 3 of the 5 years. In the entire period Burr-Learning yielded 18 
bushels (27.5 per cent) more than the average of the varieties. 
The results of later experiments at the Connecticut station also 
are of interest. The yields of combinations between lines of Century 
and Learning after four generations of selfing and selection, as 
reported by Jones and Mangelsdorf (£3) are shown in Table 7. The 
yields of the parent varieties, which are among the higher yielding 
varieties under the conditions of the experiment, are shown for 
comparison. 
In these experiments, although the yields of the single crosses 
exceeded those of their respective parent varieties but slightly, the 
double cross yielded 22.3 bushels (47 per cent) more than Century, 
the higher yielding variety. 
