- Proceedings of the New Zealand Grassland Association 57: 4%53 (1996) 49 Variation and adaptation in limpograss (Hemarthria altissima) P.W. WOODS, J.N. COUCHMAN and H.A. BARLOW AgResearch, PO Box 194. Kqikohe Abstract and anthocyanin content are readily apparent (Oakes 1973). Considerable development for cattle grazing Persistent and productive cultivars of limpograss occurred in Florida, leading to the release of the diploid (Hemarthriu altissima (Poir.) Stapf et C.E. Hubb.) cultivars Redalta (P1299993) and Greenalta (P1299994), are required if the species is to be useful in and the tetraploid cultivars Bigalta (P1299995) and Northland cattle systems. A study was undertaken Floralta (PI364888). Significant variation exists between to investigate plant morphological traits, adaptation ploidy levels, diploids tolerating frost better, and and persistence of 8 accessions. A series of tetraploids having superior in vitro digestibility abandoned research sites was revisited to determine (Quesenberry et al. 1982). long-term persistence and adaptation. Trials were Bigalta was used extensively in comparisons with established at 4 new sites throughout the northern other grasses in Northland (Taylor et al. 1976; Rumball North Island to assess morphological variation and & Lambert 198 1; Davies & Hunt 1983) and in the plant survival. There was considerable variation in Manawatu (Forde unpublished). Under favourable morphological traits assessed and plant survival, conditions limpograss production was comparable or allowing scope for future selection. Bigalta used in better than other grasses and the species was considered earlier research trials had largely failed to persist to have lower weed potential. Paddock scale plantings under normal pasture management practices. Future of Bigalta occurred on four Northland and one South work should use the cultivar Floralta. Auckland dairy farm, during the summers of 1981-84 (Taylor pers. comm.). With good establishment and Keywords: adaptation, flowering, Hemarthria management Bigalta performed well. altissima, leaf, morphological variation, During the late 1970s a subsample of accessions persistence, seed, stolon from the USDA collection representing a range of variation and elite germplasm was imported to New Introduction Zealand. Establishment, productivity, persistence and frost tolerance were evaluated by Davies & Hunt (1989). The potential of limpograss (Hemarthria altissima At least 80% of stolen cuttings from coarse-stemmed (Poir.) Stapf et C.E. Hubb.) as a special-purpose lines established plants compared with only 40-50% of perennial summer forage or componerit of mixed pasture stolon cuttings from fine-stemmed lines. Initially, in Northland dairy and beef cattle systems has been tetraploid cultivars performed better than other lines demonstrated (Taylor et al. 1976; Rumball & Lambert but subsequent yield was directly related to shoot tissue I98 I ; Davies & Hunt 1983). Persistent and productive survival through winter, fine-stemmed lines being less cultivars are now needed if the species is to succeed. damaged by frost than coarser tetraploids. Floralta was Limpograss is most likely to flourish on sites with fertile intermediate in its frost tolerance, yield and persistence. moisture-retentive soils, low frost occurrence (limit about Three tetraploid lines were compared by Rumball (1989) -6”C), supplemented by 200-300 kg N/ha from in a plot trial under grazing, with and without nitrogen November to March, and perhaps irrigation (Taylor et fertiliser. Bigalta and PI349753 were more susceptible al. 1976; Davies & Hunt 1983). Stands require vegetative to frost than Floralta. In vitro digestibility and crude establishment, and remain dormant through winter. protein contents were highest in Bigalta (72.1%, 12.8%) Growth rates from January to March exceed 100 kg with Floralta and PI349753 slightly less. Efficiency of DM/ha/day, with digestibility of 68-72% and protein response to applied nitrogen was greatest with Floralta, content of 12-13% (Rumball 1989). and slightly less with Bigalta. Limpograss is a stoloniferous grass of the tribe Two further accessions from South Africa (I 1 and Andropogoneae and family Poaceae. Many genotypes 12) were imported during 1992, and were included in were collected from southern Africa during 1964 and the work reported here. This paper reports observations 1971 by the USDA (Oakes 1973). Diploid, tetraploid from visits to abandoned research sites, and field trial and hexaploid forms exist (Quesenberry el al. 1982). information on morphological variation and survival Differences in leafiness, growth habit, internode length among limpograss lines available in New Zealand. 50 Materials and methods Kaitaia; and Pekerau Rd, Taipa. During the period of uncontrolled management, Bigalta had virtually disap- To assess persistence, abandoned trial sites where peared from three of the four sites. A strong population limpograss had been previously established in Northland was present at Kohukohu. The paddock was in a frost- were visited during September 1992. Historical reports free area on a Takahiwai clay, and was grazed by cattle of naturalised populations were investigated. A set of and used for making hay each spring. field trials was established during December 1993 to During 1982 and 1983, Davies & Hunt (1989) gain further information on the available range of established plots of 5 cultivars on a dairy farm near accessions in New Zealand. Kauri. PI364891 and PI364884 were the most prevalent. Floralta was present but at relatively low density. Field trials PI349753 and Bigalta had failed to persist. With Trials comparing 8 lines of limpograss (Table 1) were favourable management the three surviving lines quickly established at 4 sites in the northern North Island dominated. (Kaitaia, Dargaville, Hamilton and Te Puke). Two weeks Naturalised limpograss was reported near Kaitaia prior to planting, sites were sprayed with Roundup at 6 (Edgar & Shand 1987), but the population could not be l/ha. Plants were propagated from single node cuttings relocated. A further wild population of a Hemarthria of each accession established under glasshouse species of unknown origin was found at Haruru Falls, conditions at Kaikohe during early spring 1993. Plants Paihia. Plants occupied a dry roadside site and had were transplanted on a 1 .O m x 1 .O m grid, 10 plants per compact rhizome systems with erect culms, similar to plot, in a randomised complete block design with 3 diploid types of Hemarthria altissima reported else- replicates, repeated at each of the four sites. At planting where. Flowering occurred during January and February. stolons were less than 100 mm in length. Young plants Stolons were not present. Oakes (1973) noted that were watered for 2 weeks after transplanting. To allow descriptions failed to include the term stolons or full expression of morphological traits, plots were left stoloniferous, but most accessions are strongly stolon- ungrazed from December to April, but were grazed by iferous. Other local material may be found in the future, dairy cows for the remainder of the year. Basal fertiliser and this may provide useful comparative genetic material was applied annually during December and nitrogen as for selection programmes. urea applied to each plant at every site visit. Assessments were made during late autumn 1994 and again in late Yield and survival in field trials spring 1994. A single whole plant typical of the plot Yield and persistence will affect the success of was removed for dissection, dry weight and morpho- limpograss under farming conditions. P1349753, SAl, logical assessment. Basal stolons at the plant crown and Bigalta and Floralta had the highest yields of both leaf aerial stolon branches were counted and measured. Ten and stolon material (Table 1). Stolons formed the largest samples from each plant were measured for mature stolon component of harvested plant material. Small differences width, and mature leaf width and length. in the percentages of leaf and stolon material occurred among accessions. Percentage of dead material harvested Statistical analysis was relatively minor. The number, width and length of Field trial data across environments were pooled, and stolons differed significantly among accessions (Table measurements for all variates rank transformed (Conover 2). PI364884 had many relatively fine short stolons, & Iman 1981) for analysis. Pooled site analyses taking compared with accession 11 with relatively few coarse into account all sources of variation, were performed long stolons. SAl, Bigalta and Floralta had moderate using the balanced analysis of variance procedure of numbers of coarse relatively short stolons which were Genstat. Standard errors of differences (SED) were reflected in the dry weight per m of stolon. Some fine- calculated from rank transformed data. Pooled means stemmed accessions had poor dry weight per m of stolen across environments were also calculated from raw data. which conforms with poor vegetative establishment recorded by Davies & Hunt (1989). Total length of Results and discussion stolon per plant was least in the improved cultivars. Under grazing conditions it is unlikely stolons would Abandoned trial site visits and naturalised achieve the lengths and sizes recorded in this trial. populations Nevertheless stolons will form a large and very Observations were made where large scale limpograss significant component of plant material available for plantings of Bigalta had occurred lo-13