Some aspects of seed production and the evaluation of herbicides for tuber production of the Hammett "Figaro" series semi-dwarf dahlia : a thesis presented in partial fulfillment of the requirements for the degree of Master of Horticultural Science in Seed Technology at Massey University, Palmerston North, New Zealand

Seed yield and quality information was collected from 14 clones of a double flowering semi-dwarf (0.75-1m) bedding Dahlia series, bred by Dr Keith Hammett of Auckland, in the 1994-95 season at Palmerston North. Half-sib progeny were grown out to assess flower colour and doubleness in the 1995-96 season. Seed yield (0.03-6.66g/plant) varied widely among the clones. Seed yield was affected more by the fertility of disc florets than by their number and so the highest seed production potential was likely to be maintained in clones of high bloom quality (degree of doubleness). Clones with yellow, orange or red flowers had greater fertility than clones with purple - magenta, white, or pale colours. This may well reflect a fertility- colour link related to the original hybridization of the garden Dahlia from two wild species If such a link exists then careful manipulation of clonal ratios may be required to maintain a good overall balance of colours. Maintaining seed quality required drying seed without delay, especially when seed was harvested under cooler conditions. Very low levels of primary dormancy were detected, but some clones produced seed which when germinated at a later date varied both in time to 50% germination (over six days) and spread of germination (over four days). This could have significant implications during plant establishment under nursery conditions, and dry storage or possibly a longer period of pre-chilling is suggested to reduce this variation. This requires further evaluation. Oxyfluorfen, oxyfluorfen plus oryzalin, oxadiazon, and oxadiazon plus simazine herbicide treatments did not affect tuber yields and subsequent forced re-sprouting of seedling material under glasshouse conditions. Trifluralin and oryzalin reduced tuber yields, similar to the unweeded control. This was due to competition from inadequately controlled weeds, rather than any obvious toxic effects from these two herbicides. No visual phytotoxicity was observed in any of the treatments in either the initial growing season or the subsequent forced resprouting although the herbicides were not applied directly over the top of the plants. [From Introduction