My Account: Log In | Join | Renew
Search
Author
Title
Vol.
Issue
Year
1st Page

Abstract

 

This article in CS

  1. Vol. 32 No. 6, p. 1496-1502
     
    Received: Dec 19, 1991
    Published: Nov, 1992


    * Corresponding author(s):
 View
 Download
 Alerts
 Permissions
 Share

doi:10.2135/cropsci1992.0011183X003200060038x

Combining Vigor Test Results for Prediction of Final Stand of Shrunken-2 Sweet Corn Seed

  1. Dale O. Wilson ,
  2. John C. Alleyne,
  3. Bahman Shafii and
  4. S. Krishna Mohan
  1. P arma Res. & Ext. Cen., Univ. of Idaho, 29603 U of I Lane, Parma, ID 83660
    L a Pastora Propagating Stn., Santa Cruz, Trinidad
    C ollege of Agric. Univ. of Idaho, Moscow, ID 83843

Abstract

Abstract

Supersweet sweet corn (Zea mays L.) based on the shrunken-2 gene suffers from low and erratic field emergence and seedling survival. The purpose of this project was to find seed vigor tests highly correlated with final stand count, and to combine tests to form a multiple regression model for prediction of final stand. Forty-nine commercial seed lots (and one noncommercial lot) obtained in 1988 and 48 lots obtained in 1989 were planted in the field the next year. Final stand (FS) was counted at the six-leaf stage. The standard rolled-towel germination test with counts at 4 (G4) and 7 d (G7) was performed, along with accelerated aging (AA), soil cold test (CT), fat acidity of kernels (FA), kernel weight (WT), embryo weight (EM), ratio of EM to WT, and leachate conductivity (EC). The G7 and G4 counts were not highly correlated with FS (r < 0.40). The test most highly correlated with FS in the 1988 seed was EC (r = −0.62); in the 1989 seed, it was AA (r = 0.87). The EC and AA tests were complementary, and were combined using linear multiple regression into vigor indices. In 1988, the optimum model involved the variables AA, EC, and EM (R2 = 0.69); in 1989, only AA and EC (R2 = 0.78). Models were cross-validated by using the regression equation from one year to predict field emergence in the other year. When the 1988 multiple regression model was cross-validated using 1989 data, it was highly correlated with FS (r = 0.86); the 1989 equation cross-validated using the 1988 data was likewise highly correlated with FS (r = 0.64). Collinearity was detected among the explanatory variables, but was not strong enough to influence the regression results. Since AA and EC proved to be complementary tests, multiple regression was a useful way to combine them and other test results for prediction of final stand.

Contribution no. 91793 of the Idaho Agric. Exp. Stn.

  Please view the pdf by using the Full Text (PDF) link under 'View' to the left.

Copyright © .