Degree Type


Date of Award


Degree Name

Master of Science



First Advisor

Michael Blanco

Second Advisor

Thomas Lubberstedt


The allelic diversity (AD) project of the Germplasm Enhancement of Maize program utilized the double haploid (DH) breeding method to expedite release of lines from 300 exotic maize races. Differential effects on haploid induction rates (HIR) and chromosome doubling rates (DR) associated with the recurrent parents PHB47 and PHZ51, the elevation that the race is traditionally grown at, and the race itself were examined. Races from the AD project were grouped by elevation (high, middle, or low altitude) of their country of origin. Six races per elevation were randomly selected and backcrossed using both recurrent parents to generate 36 populations. Ten replications were randomized in a complete randomized design for two growing seasons (2011, 2012) to estimate HIR, and one growing season (2013) to estimate DR. In the HIR study, populations were pollinated with a haploid inducer line. Seed progeny were screened the following winter and the number of haploid, hybrid, and contaminate seeds were recorded and analyzed. The recurrent parent effect was significant with PHB47 having a higher HIR than PHZ51. Elevation was significant with higher HIR associated with low elevation origin, and race also proved to be significant. In the DR study, the populations were treated with colchicine as seedlings to promote chromosome doubling. Haploid plants that successfully doubled and had functional ear and tassel structures were self-pollinated. Recurrent parent, elevation, and race effects were all non-significant for DR and kernel set per ear.

The doubled haploid breeding method can produce maize inbred lines more rapidly than traditional methods, but there are challenges. Sorting haploid from hybrid seed based on visual color markers is time consuming, and can be difficult due to color inhibitors that obscure pigmentation needed to distinguish between haploid, hybrid, and outcrossed seed. In this study, weight was evaluated as a method to sort haploid from hybrid seed. A preliminary experiment utilized two families for analysis. Eleven haploid and eleven hybrid kernels from each family were weighed for a total of 44 experimental units. A second experiment used six families and the same format as the previous, for 132 experimental units. Hybrid seed weighed significantly more than haploid seed in both experiments. However, the interaction between line and kernel type was significant in the second experiment. In conclusion, efficacy of sorting haploid from hybrid kernels based on weight depends on the genotypes involved.


Copyright Owner

Andrew Smelser



File Format


File Size

63 pages