Seedset Rating as a Measure of Fertility in Grasses 1

Contribution from the Department of Plant Sciences, University of Idaho, Moscow, Idaho. Research Paper No. 660 of the Idaho Agricultural Experiment Station. This investigation was part of a Western Regional Project (W-40) and was supported in part by re ional funds. Received July 1, 1965. 'Research Fellow and Assistant Agronomist, University of Idaho. also be impaired as sterile florets tend to remain attached to normal florets. Therefore, any improvement in the general fertility level of a species may well result in increased yield and quality of seed. Fertility has been expressed in many ways by different workers. Generally, fertility evaluation falls into 1 of 3 categories: (a) number or weight of seeds produced per inflorescence, (b) weight of seed produced per inflorescence expressed as a percent of the gross wheight of inflorescence, or (c) percent of florets setting seed. Slinkard,3 working with intermediate wheatgrass (Agropyron intermedium), expressed fertility as percent fertile primary and secondary (basal) florets from 5 spikelets near the center of each spike, using 10 spikes per plant. Although this proved to be an accurate and precise estimate of fertility, the determination was time-consuming and tedious. If a less time-consuming method could be substituted to provide essentially the same information, many man-hours would be available to study larger populations in more detail. Seedset rating, as used by Ross and Bullis,4 was compared with the percent fertile basal florets as an alternative estimate of fertility in intermediate wheatgrass. Seedset rating was determined by dividing the unthreshed weight of 10 randomly chosen spikes per plant into the weight of clean seed threshed from these 10 spikes and was expressed as a percentage. Included in this preliminary trial were 70 spaced plants of 'Greenar' and 72 spaced plants of 'Oahe' intermediate wheatgrass. Simple pheno typ ic correlations between percent fertile basal florets and seedset rating of ,898 and ,890 were obtained for Greenar and Oahe, respectively. In 1964 percent fertile basal florets and seedset rating were compared in a more comprehensive study of characters affecting seed yield in intermediate wheatgrass. A six clone diallel cross nursery was used. Thirty single crosses were entered in 2-plant plots in 8 replications resulting in 480 single cross plants. A nested analysis of covariance was computed on the data from these single crosses. Estimates of components of variance and covariance among single crosses, replications within single crosses, and plots within replications were obtained. Genotypic correlations were calculated among the characters affecting seed yield using the "among single crosses" variance and covariance components. Phenotypic correlations were calculated using "total" variance and covariance components. As shown in Table 1, the genotypic and phenotypic correlations between percent fertile basal florets and seedset rating were .936 and 314, respectively. Correlations of percent fertile basal florets and seedset rating with other