The feasibility of using two different methods of assaying for DNA polymer phisms has been assessed. They were Restriction fragment length polymorphisms (RFLPs) as revealed by a range of characterised Brassica cDNA sequences and Random amplified polymorphic DNA (RAPD). These techniques have been shown to reveal DNA polymorphisms between varieties of Brassica napus L. Further more, the sequence and organisation of a Hind III family of highly repetitive DNA sequences were also studied on Brassica napus L. RFLPs associated with rape extensin, ext A, and Brassica oleraceae self- incompatibility genes were observed when DNA samples from 19 commercial varieties of B. napus were analysed using the cDNA probes pRR566 (coding for root-specific extensin) and pBOS2 (coding for S(_5) self-incompatibility allele in B. oleraceae). Both cDNA clones were able to reveal RFLP patterns with varying degrees of polymorphism depending on the restriction enzymes used in the digestion of genomic DNAs. Although both probes could generate complex RFLP band patterns, those revealed by pB0S2 were generally easier to analyse and more suitable for DNA fingerprinting while those revealed by pRR566 were less distinct as a result of extensive background hybridisations. The probe pRR566, with certain restriction enzymes generated simpler RFLP band patterns that were more suitable for segregation analyses. Segregation analysis of F(_1) individuals revealed additive RFLP band patterns of both parental varieties, while that of F(_2) individuals revealed RFLP band patterns of each parental varieties as well as the additive pattern. When analysed for possible association with varietal glucosinolate content, none of the RFLP band patterns showed such linkage. A cDNA library was constructed from pod material of a high glucosinolate variety in an attempt to obtain clones which could reveal RFLP patterns associated with glucosinolate content. Differential screening using total cDNAs from pod materials of high and low glucosinolate varieties failed to isolate any cDNA clones useful as RFLP markers. Another DNA polymorphism assay studied, RAPD, was able to detect inter- and intraspecific variation in Brassica sp. Analysis of six phylogenetically-related but distinct Brassica sp. revealed extensive variation in the RAPD band patternsof amplification products; with some amphidiploid species sharing conserved band patterns with their ancestral species. RAPD analysis on 17 varieties of rape revealed polymorphic as well as highly conserved RAPD band patterns depending on the primer used. One of the primers was able to amplify a polymorphic band which could be associated with low glucosinolate varieties i.e. present almost exclusively in low glucosinolate varieties. Species-specific as well as variety-specific band patterns were also observed during the RAPD analysis. Finally, sequence and organisation of a Hind III family of repetitive sequences was studied. The monomeric and polymeric forms (trimer and tetramer) of the repetitive sequences were successfully cloned into pUCl8. Sequence analysis of the two clones containing the polymeric forms revealed that the monomers were arranged in tandem array and that all internal Hind III recognition sites were lost due to point mutation(s) which occurred within the six basepair recognition site. A consensus monomeric sequence was deduced from sequence comparison of the 8 copies of the monomeric sequences present in the 3 clones and the deviation from the consensus sequence of each of the eight monomers was less than 3%. No two monomeric sequences were identical. It was estimated that the number of copies of the monomeric sequences in a haploid genome was approximately 0.3 million copies. Estimates of the proportional representation of each of the polymeric sequences based on the number of copies of the monomers in each polymer were also calculated.