Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Individual study: The effects of soil disturbance, nutrient addition and seed supplementation on seedling recruitment in an invasive grass (Pennisetum setaceum) and a native grass (Heteropogon contortus), Ka'iwa Ridge, O'ahu, the Hawaiian Islands, USA

Published source details

Goergen E. & Daehler C.C. (2002) Factors affecting seedling recruitment in an invasive grass (Pennisetum setaceum) and a native grass (Heteropogon contortus) in the Hawaiian Islands. Plant Ecology, 161, 147-156


In the Hawaiian Archipelago, many introduced African grasses have invaded disturbed habitats. One of the most aggressive is fountain grass Pennisetum setaceum, which in recent decades has replaced the native pili grass, Heteropogon contortus in many arid habitats on O'ahu and Hawai'i. Both species are perennial bunchgrasses that rely on recruitment by seeds. Therefore, in addition to faster vegetative growth of P.setaceum reported in earlier studies, differences in seed bank dynamics, and in conditions favoring seed germination and establishment, may partly explain the invasion by P.setaceum into former H.contortus-dominated grasslands. To help elucidate factors promoting the spread of feather grass and decline of pili grass, the effects of soil disturbance, nutrient addition and seed supplementation on seedling recruitment were investigated.

Study site: The study was undertaken at Ka'iwa Ridge on the island of O'ahu(21º23.262'N, 157º43.095'W). Average annual rainfall is of 900 mm, falling mostly between December and March. In 1968, most of the ridge was H.contortus grassland, P.setaceum was absent. At the time of study 30 years later, the ridge was dominated by P.setaceum with only scattered remnant patches of H.contortus.

Seed bank estimates: Soil samples were collected in October 1999 from three grassland types:

i) P.setaceum-dominated (>80% cover of P.setaceum, no H.contortus);
ii) H.contortus-dominated (>80% cover of H.contortus, no P.setaceum),
iii) mix of both species (approximately 40% cover of each species).

Each area had approximately 20% bare ground or rock cover, and no other species had more than 5% cover. From each, 15 soil samples (0.005 m²; 5 cm deep) were collected at 1 to 3 m intervals. The soil was spread over peat moss flats in a greenhouse and kept moist under natural light; seedling emergence was recorded over 4 months.

Effect of disturbance, seed addition, and nutrients: Plots were established near the start of the rainy season (January 1999), at the edge of a remnant H.contortus grassland patch (c.15 m diameter) surrounded by P.setaceum. The treatments were two levels of disturbance (undisturbed or vegetation cleared), two of nutrients (low or high), and two of seeds (seeds of both species added or no seeds added), in a factorial design. Forty 0.25 m² plots were divided into 5 blocks with similar slope and aspect, each block with the eight possible treatment combinations. The plot size simulated patches of disturbance caused by human (i.e. by hikers) or natural disturbance (small-scale slope erosion). All treatments were watered only by natural rainfall.

Disturbance: At the end of January 1999 'disturbance plots' were cleared of above and most belowground, biomass by hand pulling, and the plant material was removed.

Seed addition: In late February, 300 P.setaceum and 50 H.contortus seeds were sprinkled over the soil surface (or over a thin litter layer in uncleared plots) in the relevant plots. The number of seeds of each species sown per plot was based on estimates of the average reproductive output of a single plant of each species

Fertilizer application: In mid-April, mid-September 1999 and January 2000, dry fertilizer was sprinkled on to the high nutrient plots (1 tablespoon/plot/treatment date, Scott'sMiracle Grow 20-20-20). The amount of fertilizer was determined after identifying the amount of nutrient addition that promoted an obvious growth response in greenhouse plants grow in soil from the field.

Plant monitoring: Plots were monitored about once a month 12 months; the number and type of seedlings present were recorded. They were resurveyed in the following, wetter, year.

1999: In the first year (April-December 1999), there was a drought and seedling recruitment was greater for H.contortus (typically >< 2.5 – 8.2 in disturbed, and 0.5 – 4.9 in undisturbed plots) than for P.setaceum under all treatments; most P.setaceum seedlings died between sampling periods (averaging <1 seedling/plot in all months). Greenhouse comparisons of seedling survival (see: revealed that H.contortus seedlings tolerate drought better than P.setaceum seedlings.

Recruitment was not increased by seed addition for either grass despite only 49 and 4 seeds/m² in the seed bank for H.contortus and P.setaceum, respectively.

2000: Due to the drought in the 1999, the plots were resurveyed in 2000, a wetter year. The pattern of seedling recruitment was reversed, more P.setaceum seedlings (e.g. in April 2000, 2-9 seedlings/plot) being present than H.contortus (2-5 seedlings/plot) in all but one treatment (i.e. disturbed + nutrient addition: H.contortus - c.5 seedlings/plot; P.setaceum - c.2 seedlings/plot).

Conclusions: Seedling recruitment of these two grasses on O’ahu appears to be primarily dependent on water availability, with the non-native P.setaceum doing better than H.contortus in wetter years. Once seedlings of the long-lived P.setaceum are established, it then appears able to maintain dominance over H.contortus, even during periods of drought.

Note: If using or referring to this published study, please read and quote the original paper, this can be viewed at: