A Regression and Path Analysis of Yield Related Traits in Two Populations of Grain Amaranths.

Abstract

The grain amaranths, which originated in Incan Andes and Guatamala (Sauer, 1950), belong to the Amaranthaceae family which has at least 60 known species. They are ubiquitously distributed across the tropical world either as a Ip'ain ill Central South America and Suthern. Asiatic countries, or as a vegetable crop in Central and South America, Southern Asia and Africa (Hauptli et aI., 1979). Among the presently cultivated grain types are Amaranthus caudaIUs, A. cruenlUs and A. hypochondriacus. The grain amaranths are classified as outcrossing (Simmonds, 1979). A recent study by Jain et al. (1982) indicated that grain amaranths have an average outcrossing rate of 30%. 1971; Few studies have been reported in grain amaranths which relate seed yield to other traits except those of Hauptli and Jain (1980, 1985). The objective of this study was to identify potentially useful quantitative predictors of grain amaranth yield. Plant height and seed weight measurements, harvest index threshing percentage, seed—yield: height ratio, seed weight and days to flowering were assessed for their influence on seed yield per plant using simple correlations, multiple regression and path-coefficient analysis. MATERIALS AND METHODS Materials for this study were developed from two heterogeneous grain amaranth populations, namely UC87 (Amaranthus cruentus) and U. C99 (A. hypochondriacus). The two populations were originally obtained from Guatemala and Mexico respectively. The study was conducted in the experimental farms of the University of California, Davis. USA Each population was planted in a 34-metre by 30-metre plot consisting of 34 rows during the summer of 1982. Between-and-within row spacing were one metre and 15 cm respectively. Excluding the outer row on either side, the plot was subdivided into 24 subplots, each four rows by nine metres, to compensate for microenvironmental fluctuations. Data were taken from 7 subplots by sampling 20 plants per subplot, giving a total of 140 plants for each populations. Mature individual plants were scored for the plant size traits (plant height, plant dry weight, head length and head dry weight) and seed production efficiency parameters (threshing percentage, seed yield: height ratio, harvest index and seed yield per plant) and 500 seed weight. The top 21 plants, three per subplot, were selected for harvest index. Equal amounts of seed from each of the 21 plants were bulked to form the selected accessions in each population.