ABSTRACT


Raúl Barnier B.1, Alejandro Violic M.2, Surendra P. Sinha3 e Ismael Parker V.4
 

A sample consisting of 380 ears of corn was taken and its weinght, moisture content of the grain at harvest time, and grain, weight, were determined. Moisture content was calculated by direct method drying a sample of grain in an oven at 80ºC until constant weight was obtained.
The grain weight of each ear was expressed on a constant moisture content basis of 15%. this value was expressed through a regression equation as a function of total weight of the ear and moisture content of the grain determined at harvest time.
Regression was calculated by a IBM 360-40 computer using an uncorrected model. Several predicting equations were tested, obtained by increasing the degree of the polynomial as power's of the moisture content variable.
A better fit was obtained by increasing the degree of the equations, as expressed by the multiple correlation coefficient. However, no imporvement was observed when the large range in moisture content of the original data.
The results obtained with the different regression models were efficent, therefore its use in experimental work in corn is advantageous.

Keywords:
1Ing Agr. Proyecto Estadísticas, Estación Experimental La Platina, Instituto de Investigaciones Agropecuarias, Casilla 5427, Santiago, Chile. Profesor de Estadísticas, Facultad de Agronomía, Universidad Católica de Valparaíso. Profesor Auxiliar de Estadística, Facultad de Agronomía, Universidad de Chile.
2Ing Agr PhD, Proyecto Mejoramiento de Maíz, Estación Experimental La Platina, Instituto de Investigaciones Agropecuarias. Profesor de Genética Vegetal Aplicada, Facultad de Agronomía, Universidad de Concepción.
3Ph D, Profesor de Estadística, Facultad