[1. Breeding Experiments]
B - Early Corn (Pedigree breeding): Brazilian corn is very slow in growing, producing generally very tall plants with the ears at about 3/5 to 2/3 the height of the plant. Crosses were made between extracts of "Tirol × Early Canadian" (white flint, 40 days from sowing to silking) with Santa Rosa (white dent, 70-80 days to silking) and Cateto (orange flinty 60-70 days to silking.) It was not possible to combine tallness and earliness and it was difficult to suppress completely tillering in the early lines. Reasonably well adapted lines were obtained with the following characteristics; 45-50 days to silking, plant without tassel, 1.3 m., ear height 50 cm., mean ear weight 70 g per ear. Since the plants are completely different from the local varieties, it seems doubtful if these lines will be acceptable to the farmer, especially since earliness is not a necessity in the State of São Paulo.
The experiment was used to study the segregation of quantitative characters and to try out methods of statistical analysis. Some results may be summarized:
The standard error of distribution can be used as a measure of variability only if the means are of more or less the same magnitude. In order to compare P, F1, F2, etc.; a weighted measure has to be used. As can be shown theoretically, and has been proven experimentally, the coefficient of variation (standard error of distribution/mean × 100) should not be used, but instead, a term called the "variance index": (standard error of distribution/square root of mean). Using this term, it can be shown for this index that, as expected:
(P) = (F1) (F2) (F3) ................
The segregation for earliness can be shown only by comparing F3 families. The inevitable phenotypic variation with an error of more in F2.
In studying the relative position (height) of the ear, the ordinary coefficient of linear correlation r is of no use. The correlation for plant and ear height was found in all lines, hybrids or segregates, to be nearly constant and equal to 0.6 (positive and significant). However the index: "ear height"/plant height" should be used and it varies significantly with the following values: imported early lines O.20, Brazilian commercial lines 0.60, some native corn up to 0.7 or 0.8 improved corn 0.5.
F. G. Brieger