Nine subterranean clover cultivars and 11 white clover lines were evaluated for persistence and production, under a combination of rotational grazing or cutting, for 4 years on rolling pumice hill country. The most successful subterranean clover cultivarwas 'Tallarook' , whichfor 4 yearsproduced the most dry matter (DM) (average 700 kg DM/ha per year) and accumulated the largest soil seed reserve (1500 kg seed/ha), Soil moisture stress over the summer severely limited persistence and hence production of all white clovers (average 400 kg DM/ha per year). Persistence was not assured by larger roots associated with large-leaved lines. Hill country types (e.g., 'Whatawhata Early Flowering') with a high stolon densityand/or free seedingability were the best adapted to this environment

Subterranean clover and white clover plant material was collected from 85 sites located in summerdry northern North Island hill country. Plants of both species originating from this collection were grown and typed at Whatwhata Hill Country Research Station. Using several cultivars and selections as standards, measurements were made of morphological, flowering and reseeding characteristics. The subterranean clover population consisted of 44% Mt Barker type, 28% Tallarook type and the remainder an undetermined intermediate type. Mt Barker types in the collection tended to flower later and for longer than their Mt Barker standard, while Tallarook types tended to flower earlier and for longer than StandardTallarook. White clover plants were smaller and more densely leaved than 'Grasslands Huia', the population mean being similar to 'Grasslands Tahora'. A vast majority (7035%) of plants flowered earlier and yielded more seed than Huia and Tahora. An apparent shift in plant type has occurred within clover populations in summerdry hill country, the most notable being the development of a more flexible reseeding pattern. This type can be a future guide to breeding clovers suitable to these environment . Keywords: white clover, subterranean clover, flowering, seeding, morphology, type hill country

Nine cultivars of subterranean clover

Nine subterranean clover cultivars and 10 white clover varieties, differing in characters such as morphology and flowering date (sub clover), or growth habit and seeding ability (white clover), where evaluated for persistence and production at 8 summer-dry hill country sites. Results for the first 3-4 years suggest the sub clovers on the New Zealand Acceptable Herbage Cultivars List (Mt Barker, Tallarook, Woogenellup, Glare) should be revised. Regeneration of Woogenellup and Glare was consistently poor, while Tallarook performed well at most sites. The current unavailability of Tallarook seed means Mt Barker is the only effechve option for grasslands in most of New Zealand. Cultivars of the sub clover subspecies yanmnicum (Larisa, Trikkala) showed promise for winter~wet, summer dry environments, as did Nangeela in winter-cold environments. A late-flowering, prostrate, low oestrogenic sub clover similar to Tallarook would be well suited to large areas of summer-dry hill country. There is a clear need for a white clover cultivar adapted to summer-dry hill country as none of the varieties tested survived severe moisture stress at 2 sites (Hawke's Bay, North Canterbury), and none performed consistently well at the other sites. New Zealand vaneties (Hula, Pitau, G18 and a hill country selection) showed best persistence and production. The overseas cultivars Haifa, Tamar, Louisiana and Clarence Valley, which are all adapted to dry conditions in their country of origin, were consistently poor. Firm selection criteria for dry hill country were not identified, though stolen density and seeding ability should be incorporated and the strong influence of management and soil fertility on genotype performance must be recognised. Keywords: subterranean clover (Trifolium subterraneum L.), white clover (Trifolium repens L.), hill country, dryland, cultivars, persistence, genotype-environmental interaction.

Eleven white clover lines ofcontrasting plant habit and/or flowering pattern were evaluated for