2004-10-212013-01-041996-01-012013-01-041996-011996-01-01January 19http://hdl.handle.net/10388/etd-10212004-000152Brassica rapa cultivars occupy about 44% of Canada's five million hectares of canola. However, B. rapa cultivars yield 15 to 20% less seed than those of B. napus. In order to increase the competitiveness of B. rapa , significant increases in seed yield must be achieved. The development of hybrid cultivars of B. rapa could provide the basis for high yield. The objective of this research was to evaluate the performance of B. rapa doubled haploid (DH) lines and their potential use as parents in hybrid cultivars. A total of 162 DH lines, derived from five B. rapa breeding populations were evaluated in field tests at Saskatoon. Bud pollination was used to obtain selfed seed for evaluation of the DH lines. Sixteen top cross and 27 polycross progenies and 45 single cross hybrids were evaluated in the field to measure combining ability of DH lines. Many B. rapa DH lines were chlorophyll deficient as a result of expression of recessive alleles, a classical inbreeding phenomenon. Average seed and biological yield and number of seeds/pod of DH lines were only 24, 48, 46% of their donor populations, indicating severe inbreeding depression. Inbreeding greatly extended days to flowering. However, seed weight, pod length and days to mature were less severely affected than other traits measured. Several DH lines equaled their donor population in plant weight and height at specific stages of growth, however, on average the overall growth and development of the DH lines was slower than their respective donor populations. One DH line (BC-3015) equaled the seed and biomass yields of its donor population, suggesting that dominance deviation not overdominance was the genetic basis of high yield in B. rapa . It is suggested that chlorophyll deficient, late flowering DH plants could be discarded on the basis of greenhouse performance. Top cross and polycross procedures were equally effective in ranking DH lines for general combining ability (GCA). The top cross method of predicting GCA is the preferred method since it will allow the use of a weak, recessive tester which will not mask dominant alleles present in DH lines. Hybridity of top cross seed was high as measured by the erucic acid marker. The single cross procedure identified heterotic combinations which were different from those identified in the top cross and polycross methods. It was concluded that this difference was caused by the differential effects of male parents used to determine GCA. One single cross hybrid yielded significantly more seed than the check cultivar Tobin (130%). It is concluded that DH lines of B. rapa will be useful in developing inbred parents for hybrid development and procedures for combining ability testing and maintenance of SI DH lines for the production of hybrids is proposed.en-UStext