Partnerships between researchers and restoration practitioners can improve restoration outcomes, which is especially important for restoration in challenging settings. Here, we describe one such partnership in the Great Basin, United States, which used trait-based methods and practitioner knowledge to identify the most promising seed sources for restoration. Managers in this region can either use widely available commercial seeds, which are often sourced from far outside the seeding area, or take extra steps to use locally collected seeds. We asked whether local, wild-collected seeds of two native plants, Elymus elymoides and Poa secunda, had traits more conducive to seedling establishment in degraded sites, relative to commercial sources. Seeds were collected from four remnant native populations within lands managed by the Winnemucca Bureau of Land Management. Collections were screened for seed and seedling characteristics previously identified as associated with increased seedling performance in degraded Great Basin systems, and we provide a detailed methodology for these measurements. Relative to commercial seeds, wild-collected seeds had more characteristics identified as beneficial for seedling establishment, including phenology, root allocation, root form, and overall size characteristics that suggest locally sourced populations would be likely to establish better than commercial sources. Using phenotypic traits as criteria, the most promising wild populations were selected for agronomic production to increase the quantities of seeds available for restoration, and field trials are ongoing using these field-increased seeds. These results provide support for collaborative efforts to identify, collect, screen, and increase the availability of local seed sources to improve restoration success.

Grasslands Kara cocksfoot, Grasslands Nui ryegrass, Grasslands Roa tall fescue, Grasslands Gala grazing brome and Grasslands Maru phalaris were sown as single grass species with either white or Caucasian clover during December 1993, on a Lismore stony silt loam in Canterbury. Plots were irrigated and rotationally grazed by sheep. Total herbage produced during the first measurement period (September 1994 to July 1995) averaged 13.6 t DM/ha. All white clover-based swards produced similar yields, but with Caucasian clover, phalaris-based swards produced the most and ryegrass and grazing brome the least. Production tended to be lower from pastures sown with Caucasian than white clover (12.9 cf. 14.2,kg DM/ha). During the second measurement period (July 1995 to May 1996) total production was similar with both legumes. When evaluated over both measurement periods, phalaris and tall fescue gave the highest and cocksfoot and ryegrass the lowest production when sown with white clover, while cocksfoot and grazing brome were the lowestproducing swards with Caucasian clover. Ten --months-after-sow.ing,-white-cIover-contributed-in excess of40% of the DM in all swards except with ryegrass and grazing brome where it contributed 20%. The proportion of white clover then declined in all pastures over the following 20 months, ryegrass-based swards having the highest (10%) and cocksfoot the lowest (1%). When sown with Caucasian clover the legume content of swards increased during the 12 months from spring 1994 in swards containing phalaris, tall fescue and grazing brome. The proportion of legume in -cocksfoot swardsmdeclined-continuously from 22 to 4% over the 20 months while it increased continuously from 4 to 28% with ryegrass. The legume content in lo-month-old pastures was higher when sown with white than with Caucasian clover but this ratio was reversed by 28 months. This occurred with all pasture types but was most pronounced with ryegrass and cocksfoot pastures. The present experiment shows Caucasian clover can make a major contribution in irrigated pastures in Canterbury. 'Wattie Bush, RD 22, Peel Forest, Geruldine Keywords: Bromus stamineus, Dactylis glomerata, Festuca arundinacea, grazed pastures, herbage production, Lolium perenne, Phalaris aquatica, Trifolium ambiguum, Trifolium repens

The results of a three-year study in New Zealand undertaken to compare the effect of alternating sheep and cattle grazing and of different pasture species on internal parasitism in lambs is reported. Despite the commonly held belief that the use of cattle in sheep production systems improves sheep health by reducing nematode burdens, this did not occur. Cattle substantially reduced pasture larval numbers. However, this failed to reduce parasite burdens in lambs as those managed in the absence of cattle apparently developed increased immunity to parasites in response to the greater larval availability. Varying the number of successive grazings by sheep or cattle from 1 to 4 did not change this result. Lamb carcass weights were improved by the presence of cattle, primarily because of improvements in pasture quality. Substituting 30% of the ryegrass area with lucerne or replacing ryegrass with a multi-species mix A 96082 Received 3 December 1996; accepted 24 June 1998 consisting predominantly of bromes, tall fescue, phalaris, timothy, and red and white clover, had no effect on lamb faecal egg counts or worm burdens. Lamb liveweight was not affected by herbage species. Both lucerne and the multi-species mix increased the requirement for conserved feed during winter compared with ryegrass-based systems.

Increasing resistance of gastro-intestinal nematode parasites to anthelmintics and consumer resistance to the possibility of residues in animal products have prompted research on the effect of pasture species on nematodes and animal performance. Lambs (either infected with high rates of gastrointestinal nematodes or maintained nematode-free) were grazed on pure swards of chicory, high- or low-endophyte ryegrass, cocksfoot, tall fescue, lucerne, lotus, white clover or plantain. Infected lambs that grazed chicory had lower faecal egg counts and adult nematode populations, and higher carcass weights, than lambs grazed on plantain or the grass species; lambs that grazed legumes generally had intermediate counts, populations and weights. When kept parasite-free, carcass weights were up to 48% greater than in the nematodeinfected treatments. On farmlets run over 3 years, substituting 30% of the ryegrass area with lucerne or replacing the ryegrass with a multi-species mix consisting predominantly of bromes, tall fescue, phalaris, timothy and red and white clover, had no effect on gastrointestinal nematode larvae, lamb faecal worm egg or adult nematode numbers. It is concluded that a diet of pure chicory affects internal parasite populations but the small proportion included in the farmlet studies had no effect. Keywords: Cichorium intybus, Dactylis glomerota, Festuca arundinacea, gastro-intestinal nematodes, lambs, Lolium perenne, Lotus corniculatus, Medicago sativa, pasture species, Plantago lanceolata, Trifolium repens

A desirable option for increasing milk yield per farm is to increase milk production in summer without compromising peak-season production or the duration of lactation. The dairy industry has a goal to achieve a 4% per month post-peak decline of milk production. The effect of five pasture types and two summer water regimes on predicted farm milk yield, in Northland, Waikato, Manawatu and Canterbury was measured over 1 year. Two pasture treatments, resident pasture and Grasslands Nui ryegrass, were common to all sites. A third pasture treatment varied between sites: triple mix (Grasslands Advance tall fescue, Grasslands Kara cocksfoot, and Grasslands Maru phalaris) at the Manawatu site; the same triple mix but with Grasslands Raki paspalum at the Northland site; and low-endophyte ryegrass and Grasslands Kahu timothy (LER) at the Canterbury site; all sown treatments included red and white clover. All five pasture types were included at the Waikato site. On average for all pasture types, water deficit reduced summer herbage accumulation (HA) to 79, 68, 43 and 18% of irrigated controls, and annual HA to 88, 80, 73 and 63% of irrigated controls in Northland, Waikato, Manawatu and Canterbury, respectively. Since farm milk yield predicted by UDDER was highly correlated with annual HA (R2 = 83%), the effects of pasture type and water deficit on milk yield were similar to effects on HA. Seasonal and annual HA differed among the five pasture types at each of the four sites. The most important factor affecting predicted farm milk yield was annual HA, with relatively little effect from the seasonal pattern of pasture growth. Differences in shapes of the predicted milk supply curves for various treatments occurred, particularly when drought forced early drying-off. The best predicted summer milk yield was for the Canterbury irrigated LER pasture treatment, a 6.8% per month postpeak decline of milk production. Keywords: farm milk yield, farm system modelling, forage supply, pasture species, water deficit