Parental choice and selection in the early generations of a potato breeding programme.

A potato breeding programme can be divided in four phases, choice of parents, seedling generation, visual selection in the first and second clonal generation and selection from the third clonal generation based on individually assessed characters. The first three phases were subject of this thesis.<u>Choice of parents.</u> The choice of the most promising parental combinations is crucial for the success in the subsequent phases of the breeding programme. The identification of a method which predicts the value of a hybrid progeny properly will therefore increase the efficiency of a breeding programme. The value of five predictors was investigated by relating the predicted progeny performance to the actual value of 20 hybrid progenies from an incomplete 6x6 diallel. The five predictors were the 'mid-parent value', glasshouse grown seedling populations, diploid offspring obtained after pollination with Solanum phureja , selfed populations, and test-cross progenies using diploid 2n-FDR pollen producing clones as male parent. In addition, for the prediction of tuber yield, the use of a crop physiological approach has been investigated.Of the five predictors, the mid-parent value gave acceptable predictions of all examined characters at time of seed potato harvest, and for most characters at ware potato harvest. However, this predictor failed for the important character ware potato tuber yield. Better predictors of ware potato tuber yield were obtained using the crop physiological approach and the test-crosses.For the crop physiological approach, potential parental clones are harvested at intervals during the growing season. Time of tuber initiation and the relative rate of tuber bulking are important parameters in the prediction of tuber yield of the hybrid progenies. The harvests at intervals, however, askes for high labour inputs, but no loss in time occurs for the production and growing additional hybrid progenies. The test-cross method can be incorporated in a breeding scheme, as presented, with relatively low labour inputs. Compared to the traditional breeding scheme this alternative scheme strongly reduces the chance that inferior hybrid progenies will enter the time consuming and expensive third and fourth phase, but also adds an extra year to the breeding programme.<u>Seedling generation.</u> With the growth of seedlings in the glasshouse botanical seeds are converted to seed tubers. In this phase only a mildly selection in practised and there is need for additional selection traits. Therefore, for a number of characters the prospects of early-selection were examined.For tuber yield, number of tubers and the average tuber weight, no relation was observed between the performance in the glasshouse and in the field. Selection for these characters using glasshouse grown seedlings is not to be advised. The glucose content, as a parameter for the chip and French fry quality, of seedling progenies grown in the glasshouse was satisfactorily correlated to the contents of field grown progenies. The glucose content can be determined simply and cheaply with glucose test strips and early selection of progenies in the seedling generation is recommended. For the dry matter content prospects for early selection are even better and a mild direct genotype selection seems to be feasible.<u>First and second field generation.</u> During the first two field generations, clones are selected for visual characters. A selected clone should meet a number of minimum requirements to be admitted to the last phase of the breeding programme. In the last phase clones are examined accurately for a great number of traits and clones superior to the existing varieties are identified.Selection in the early field generations is generally severe, while based on only a low number of plants per genotype. Such a procedure is not very efficient, and three methods to increase the efficiency were compared.Enlarging the number of plants per genotype in the first clonal generation from one to two plants did increase the reliability of selection, but was found to be very expensive.Analysis of the components of the selection trait 'plant appearance' did not reveal sub-traits with a poor heritability and a large influence on the plant appearance score. Therefore altering the relative weight of the individual components of 'plant appearance' is not likely to improve to selection in the first two field generations.The best prospect for increasing the efficiency seemed to be offered by manipulating the selection intensity in the first field generation. By decreasing the selection intensity from 10% to about 30% in the first field generation, not only the reliability of the selection decisions improved, also the scheme was found to be more economic, at least for the Dutch situation. In the second field generation, a selection intensity of 20% at ware potato harvest is recommended. The number of first year clones should of course be reduced to one third of the initial number to accomodate the changes.A detailed scheme of a potato breeding programme incorporated some of the suggested improvements is presented