SciELO - Scientific Electronic Library Online

Home Pagelista alfabética de revistas  

Servicios Personalizados




Links relacionados


Anales del Instituto de la Patagonia

versión On-line ISSN 0718-686X

Anales Instituto Patagonia (Chile) v.36 n.1 Punta Arenas  2008 


Anales Instituto Patagonia (Chile), 2008. 36: 5-24





J. Aidan Kerr1,2 & James H. McAdam3

1 Agricultural Research Centre, Stanley, Falkland Islands, South Atlantic.
2 Department of Agriculture and Rural Development, Countryside Management Branch, Newry, Co. Down, Northern Ireland BT60 1EN, U.K.
3 Applied Plant Science and Biometrics Research Division, Agri Food and Biosciences Institute and Department of Applied Plant Science, The Queen's University of Belfast, Northern Ireland BT9 5PX. U.K.


Patches of short grasses and herbs (Greens) are important grazing resources for sheep in the Falkland Islands. Little was known of their origins or diversity making it difficult to select management options for sustainable grazing.

The floristic, physiographic, biotic and edaphic attributes of 131 Greens were surveyed. They occurred widely, were mainly <5 ha, had variable shapes and physiography and were used frequently by livestock. The most dominant plant species were Agrostis spp, Juncus scheuchzerioides, Gunnera magellanica and Poa pratensis.

Greens were classified into eight communities that differed floristically, physiographically, edaphi-cally and in herbage quality. For management purposes two main types were identified. The first type, dominated by Agrostis spp. and G. mageííanica, occurred on relatively wet and infertile soil and was mainly associated with inland valleys and small paddocks. The second type, dominated by Poa spp., occurred on relatively dry and fertile soil and was mainly associated with coastal sites, ponds and penguin rookeries.

Ordination of sites showed that Greens were floristically similar as no wholly distinct groups were found on any axes combination. Differences were principally in species frequency rather than in diversity. The first ordination axis was positively related to soil Mg and P and dung cover and negatively related to soil moisture. The second axis was positively related to soil Ca and negatively related to K and dung cover. These corresponded generally with the soil features of the classification groups.

Variation between Greens was due mainly to the type and degree of disturbance caused by combinations of livestock, penguins and sheldgeese. As Greens were more similar than dissimilar and inter-Green variation was more continuous than discrete, it is expected that management strategies for sustainable grazing would also be similar. Special management may be needed only for overgrazed Greens.

Key words: grassland management, patches, sheep grazing, sheldgeese, penguins.


Sectores de pastos cortos y especies herbáceas (Greens) son importantes recursos forrajeros para la ganadería ovina de las islas Falkland (Malvinas). Como su origen y diversidad eran poco conocidos, se hacía difícil elegir opciones de manejo para un pastoreo sustentable.

Se muestrearon los atributos florísticos, fisiográficos, bióticos y edáficos de 131 sectores con tales características. Estos se distribuyeron ampliamente, y eran principalmente menores a 5ha, con formas y fisiografía variable y usados frecuentemente por el ganado. Las especies de plantas más dominantes fueron Agrostis spp, Juncus scheuchzerioides, Gunnera magellanica y Poa pratensis.

Estos sectores fueron clasificados en ocho comunidades que diferían florística, fisiográfica y edafi-camente y en calidad herbácea. Se identificaron dos tipos principales para efectos de manejo. El primer tipo, dominado por Agrostis spp. y G. magellanica, estaba en un suelo relativamente húmedo e infértil y principalmente asociado a valles interiores y pequeños potreros . El segundo tipo, dominado por Poa spp., estaba en un suelo relativamente seco y fértil, principalmente asociado a sitios costeros, aguadas y colonias de pingüinos.

El ordenamiento de los sectores mostró que estas areas eran similares florísticamente ya que no la totalidad de los distintos grupos fueron encontrados sobre la combinación de los ejes. Las diferencias eran principalmente en cuanto a frecuencia de especies más que en diversidad. El primer eje de ordenamiento se relacionó positivamente con Mg y P del suelo y cobertura de heces, y negativamente con la humedad del suelo. El Segundo eje se relacionó positivamente con el Ca del suelo y negativamente con K y cobertura de heces. Estos correspondieron generalmente con las características del suelo de los grupos de clasificación. La variación entre sectores se debió principalmente al tipo y grado de perturbación causado por las combinaciones de ganado, pingüinos y gansos silvestres. Dado que estos sectores fueron más similares que distintos y que la variación entre ellos fue más continua que discreta, se espera que las estrategias de manejo para un pastoreo sustentable debieran ser también similares. Un manejo especial podría ser necesario sólo en aquellos sectores sobrepastoreados.

Palabras clave: manejo de praderas, parches, pastoreo por ovejas, gansos silvestres, pingüinos.


The vegetation of the Falkland Islands (51° 53 ° S; 57 ° 62 ° W) has been described as oceanic heath with two associations; those in which the grass Cortaderia pilosa is dominant and those in which dwarf shrubs e.g. Empetrum rubrum, are dominant (Moore 1968). The Cortaderia associations tend to be found on areas where the drainage is poor to moderate whereas the dwarf shrub heath usually occurs on drier ground. The vascular flora of the islands consists of 348 species, 171 of which are believed to be native (Broughton and McAdam 2002).

Agriculture is confined to extensive sheep farming in large enclosures (89% > c. 2000 ha (Davies et al. 1971). Traditionally, pasture improvement through reseeding and fertilising was practiced only on a very small scale. Recently the development of pastures with improved grasses and legumes has received much greater priority than previously.

The nutritional quality of the native pastures is low with C. pilosa, the dominant species, typically having 29 - 73 % oí dead matter in the herbage and consequently low organic matter digestibility (32 - 45 %) (McAdam et al. 1989). Where C. pilosa occurs as separate tussocks intermingled with more productive and better quality grasses, the overall value of the pasture is higher (Summers & McAdam 1993).

Throughout the oceanic heath formations numerous patches oí short, dense hard grazed turf occur near streams, valleys and ponds. Similar ve-getation formations can be found in coastal areas and around settlements in small paddocks. For most of the year the patches appear bright green which contrasts with the straw-brown colour of the surrounding Cortaderia grassland. Consequently, these formations are referred to locally as Greens.

Greens have long been recognised as an important grazing resource (Davies 1939). Davies et al. (1971) estimated that Greens and tussock C. pilosa occupied 15 % of farms on East and West Falkland and up to 25 % of the pastures on the smaller, outer islands and regarded them as 'the most valuable grazings on the islands'. Summers (1985) estimated that Greens composed 9.5% (11,828 ha) of one farm and supported 81.7% (35,381 ha) of the total wild sheldgeese (Chloephaga spp.) population on the farm. Although such data did not exist for sheep, casual aerial observation suggested that a similar pattern of grazing occurred.

The productivity and nutritional value of herbage from Greens was higher than from the surrounding pasture (Kerr & McAdam 1993). Gi-ven the opportunity for free ranging behaviour and grazing selection that the herbivores (sheep, cattle, horses and sheldgeese) can exhibit, utilisation of Greens is much higher than that of the surrounding pasture. Thus the value of a particular large grazing paddock is often assessed on the extent of Greens that it contains.

One of the current thrusts of agronomic re-search in the islands is directed towards improved management and utilisation of the native pastures. Within this broad framework, more detailed knowledge of Greens was essential. Little was known about their underlying features and the inherent differences and similarities between Greens in terms of their origin, floristic composition, physiography and edaphic characteristics.

This paper presents information that describes ordinates and classifies Greens and determines their environmental gradients. Such information will ultimately be of value in;

(a)  Predicting change within and between Green types,
(b) Assessing their contribution to the grazing resources and,
(c)  Prescribing options for their sustainable management.


Sample selection

From an earlier land system analysis (King et al. 1969), a vegetation distribution map (Davies 1939) and examination of aerial photographs, a stratified sampling strategy was adopted based on the main categories of Greens recognised by far-mers viz.; 'Penguin', 'Valley', 'Pond', 'Settlement' and 'Paddock'. Due to the lack of a surfaced road network and difficult terrain as many Greens of each type as possible were sampled on a series of journeys between isolated settlements and into the surrounding countryside between December 1987 and March 1988 i.e. over one austral summer. The routes chosen ensured that Greens were sampled areas that represented the main soil types and geo-graphic regions. In each locality sampling of Greens was random. In all, 35 farms were traversed and a total of 131 Greens were surveyed (Fig. 1).

Data collected for each Green;

(i) Environmental features

Geographic: Latitude and longitude, located on East or West Falkland or the smaller outlying islands.
Physiographic: Area, altitude, proximity to coast, topographic position, slope, aspect, shelter (degree and direction), proximity to freshwater stream or pond.
Biotic: Information from farmers on; the type of livestock carried, the likely stock density in the overall paddock, the type of use of each Green by livestock and the history of fire at each Green. The presence or absence of penguins was recorded.

(ii) Floristic composition

Floristic composition in five 50 cm x 50 cm quadrats randomly located in each Green was es-timated using the Domin scale (Kershaw 1973) of cover/abundance. Where species identification was occasionally difficult e.g. in the absence of flowers, specimens were grouped at the genus level.

(iii) Surrounding vegetation

The surrounding vegetation was classed according to Moore (1968) with slight modification as follows:

Grass heath (C. pilosa dominant);
Dwarf shrub heath (either E. rubrum and or Baccharis magellanica dominant);
Large shrub (Chüiotrichum diffusum dominant);
Cushion plant (e.g. botax gummifera dominant);
Fern (Blechnum penna-marina dominant);
Fen and bog (e.g. Juncus spp. or Rostkovia magellanica dominant);
Other grass types (e.g. Hierochloe redolens dominant).

(iv) Herbage

The herbage in each of the five quadrats was clipped to ground level and bulked to give one sample per Green. This herbage was dried at 80°C for 24 hours and weighed to determine the standing crop. A dried sub-sample was analysed for the propor-tions of N (total), Ca, K, Mg, and P using standard analytical techniques (MAFF 1987).

(v) Soil

One soil core, 2.5 cm in diameter, was sampled to a depth of 15 cm in each of five quadrats and bulked to give one sample per Green. Samples were taken fresh to the laboratory where moisture contení (oven dry %), pH, C:N, nitra-íe-N and ammonium-N, Ca, Mg, P and K were deíermined using slandard analyíical íechniques (MAFF 1987).

(vi) Other

The area of bare ground and the cover/abundance of all faecal maíerial in each quadraí were estimaíed using the Domin scale.

Data analysis

The floristic data were averaged over the five quadrats recorded from each Green.

Four separate data matrices were created each of 131 Greens by the following attributes; 43 species and bare ground cover/abundance, nine environmental (geographic, physiographic and biotic) features, nine soil chemical and dung cover/ abundance, five herbage biomass, species diversity and chemical composition.

Where appropriate the soil chemical, herbage biomass and P composition, dung and bare ground were transformed to their natural logarithms (ln) prior to statistical analysis. Data matrices were edited as recommended by Gauch (1982). "Rare" species (see Appendix I), that occurred in less than about 5% (7) of the Greens sampled, were deleted from the floristic matrix before classification and ordination.


To characterise the vegetation and to provi-de a classification that could be used during other studies on Greens, the floristic and bare ground cover/abundance data were divided into groups using a divisive hierarchical classification, the two-way indicator-species analysis "TWINSPAN" (Hill et al. 1975). The default values were used, except as follows:

a. The number of pseudospecies cut off levels was set to four. These were 0.1, 4.0, 6.0, 8.0 on the Domin scale and represented the typical species cover/abundance categories of "present, a little, a lot, and more or less dominant" as suggested by Hill (1979a),
b. The maximum level of divisions was set to three. This produced eight groups, beyond which clear ecological interpretation was not readily obtained.

Analyses of variance compared species, soil and herbage attributes between the communities of Greens identified by the classification and dominance between species within each community. Chi-square tests, with Yates' correction, were conducted to test for associations between the environmental features and the groups of Greens. Tables were constructed in which the environmental features, either binary or categorical, were the columns and the community groups were the rows. Tests were conducted only for groups formed at the first and second level of división. Beyond this level table enfries were too small (<1) for valid tests to be conducted.


To examine inter-site and inter-species relations the floristic data were ordinated using the detrended correspondence analysis program "DECORANA" (Hill 1979b). To facilítate interpretation of possible environmental gradients, the communities formed in the classification were superimposed on the ordination co-ordinates. Additionally, arbitran; boundary lines were drawn in ordination space to highlight discrete communities.

Step down predictive regression

To examine relations between the vegetation and the soil and dung factors step down predictive regression equations were calculated for the first four ordination axes. Ordination positions of the sites were regressed against the soil chemical variables.



Greens occurred widely on West Falkland (62), East Falkland (53) and on the three outlying islands (16) that were surveyed (Table 1a).


The size of individual Greens ranged from about 0.01 ha to 136 ha with a mean of 10.7 ha. 64%, 29% and 7% were <5 ha, 5-20 ha and >20 ha, respectively. They varied greatly in shape, although on valley bottoms, the most common position (46%) (Table Ib), they often followed the shape of the valley. Greens on open plains (33%), mid-slopes (17%) and hill crests (4%) generally had irregular shapes. However, where Greens were as-sociated with fence lines, paddock corners, or small paddocks (26%) their shapes often followed those of the fences. Greens found near penguin rookeries often followed the shape of the coast and sometimes extended up ío several hundred metres inland. 69% of the Greens were well shelíered (i.e. from ai least two direction s) and most commonly from the west(Table 1b). 57% were close to sttreams and 15% ío ponds, where the green ofíen surrounded the pond (Table 1b). All Greens were low lying (<83 m.a.s.l.) and 55% of siíes were sloping.

Management and Utilisation

Accuraíe histories of fire in each paddock were difficulí ío obtain as many farms had changed ownership since sub-division of large farms began in 1980. However 18% of Greens had been burned within five years and ai least anothe r 13% in the previous five years. 72% of Greens were in large paddocks in which the annual stocking rafe was low (<0.4 sheep/ha) (Table 1c). Sheep alone were present ai 56% of Greens, while ai the rest caítle and/or horses were also present, particularly ai Greens on the outlying islands (Table 1c). 74% of Greens had been used frequently by livestock (Table 1e). Well-worn tracks and abundaní deposiís of dung of varying age were present . These Greens were probably eithe r bedding areas, or areas where livestock regularly congregated, were herded, or confined íemporarily e.g. before and afíer sheep shearing.


Of the 43 species or groups of species of higher plant s identified, 23 were dicotyledons (for-bs), 16 were monocotyledons (graminoids) and 4 píeridophyíes (Appendix 1). The species were mainly herbaceous as only three species of shrubs were found. There were more forbs species than graminoids, but the graminoids occurred more frequently and had greaíer cover than the forbs. The average number of species per green was 11. The species or groups comprised c. 12% of the vascular flora of the Falkland Islands (Broughton & McAdam 1999) and c. 29% was alien ío the Islands (Moore 1968, Broughton & McAdam 2002) (Appendix 1). The most dominaní species were Agrostis spp, Juncus scheuchzerioides, Gunnera magellanica, and Poa pratensis. Ai least one of these species occurred in all Greens. Some Agrostis spp and P. pratensis are alien species. Lycopodium spp. was the most common píeridophyíe. Greens were surrounded most frequently by grass and dwarf-shrub heath.


Soil type ranged from highly organic, acidic peat on wetter sites to less acidic mineral clay loam on drier sites and occasionally sandy loam at coastal sites. The ranges and the means of soil pH, moisture contení, exchangeable Ca, K, Mg, P, NH4-N, and N03-N, C: N and dung cover are shown in Appendix II.


The range and means for the amounts of standing crop, bare ground cover, species diversity and the levels of N, K, Mg, Ca, and P in the dried herbage are shown in Appendix II.


Penguins were present, often in large populations, at 22 sites. Magellanic penguins (Spheniscus magellanicus) and Gentoo penguins (Pygoscelis papua) were the most common species. King penguins (Aptenodytes patagonicus) were present at fewer sites. Native sheldgeese (Chloephaga rubidiceps, C. picta and C. hybrida) were frequently observed at sites near to ponds and in valleys but population densities were not recorded.


Eight plant communities (I -VIII) were recog-nised (Fig. 2). They differed significantly (P<0.05) in standing crop, the levels of Mg and P in the dried herbage (Table 2), the number of species and the cover/abundance of 16 of the main species (Table 3), the cover of dung and the levels of K, Mg, P, NH4-N in the soil (Table 4). The main environmental features of each community were summarised in Table 5.


The eigenvalues for the four axes of the ordination of sites were 0.268, 0.236, 0.167 and 0.139 respectively. The best separation occurred along axis one and was between communities I-IV to the left and VI and VIII to the right (Fig. 3). Also communities I-IV were almost distinct from VIII. The separation of the communities was not improved when the Greens were ordinated on the other axes combinations. Separation between the groups of communities also occurred at the first level of división in the classification. However, the separation was not as clear as that between the communities formed at the third level. Communities I-IV had relatively lower values on axis one than communities V-VIII (Fig 3). About 30% of all the sites were closely ordinated in an intermedíate zone near the middle of axis 1.

In the ordination of species on axes 1 and 2 (Fig. 4) four species (H. lanatus, L. scariosa, Sagina procumbens and B. magellanica) had re-latively extreme values compared to the rest and were regarded as outliers. Bare ground cover was ordinated centrally. The other species ordinated into three groups (labelled X, Y and Z, Fig 4). Although the species groups formed in the ordination did not exactly match those formed by the classification the order of species was similar in both methods.

Multiple regression

The soil and dung factors explained low pro-portions (20-32%) of the variation in the first four axes of the ordination (Table 5). Explanatory gradients that included the soil and dung factors were slight and difficult to define from these factors alone. However, axis 1 represented positive gradients of Mg, P and dung cover and a negative gradient with soil moisture. Axis 2 represented a positive gradient with Ca and negative gradients with K and dung cover. Axis 3 represented positive gradients with NH4-N and a negative gradient with pH. Axis 4 represented a positive gradient with P and a negative gradient with Ca.



Several short-growing graminoids and for-bs, some of which were alien species, dominated Greens. Generally Greens classified as Agrostis spp.- G. magellanica - J. scheuchzerioides - Poa spp. grassland. Although many of the species were common to the Cortaderia and Empetrum associa-tions (Davies 1939, Moore 1968) the shortness of the swards and the paucity of both tussock forming and dwarf shrub plants were distinctive floristic features of Greens.

Greens had many species in common with other grassland communities at similar latitudes. Greens were most similar to the Mesic grassland "meadows" of the Patagonian Steppe in Tierra del Fuego; dominated by J. scheuchzerioides and G. mageííanica and which occurred in valleys or alongside watercourses (Moore 1983). Greens were similar to the Fuegian Alpine "meadows" which had poor drainage and supported grassy vegetation, often dominated by Agrostis mageííanica (Moore 1983). The co-dominance of P. pratensis also occurred on grazing lawns in the steppe of Tierra del Fuego (Posse et ai. 2000). In the Northern Hemisphere, Greens were similar to the Agrostis/ Festuca vegetation types in Scottish (Frame et al. 1985) and Welsh hills (Hughes 1958).

Floristically two main types of Greens could be distinguished viz. a. those with high cover of the forb G. magellanica in association with the grass Agrostis and b. those dominated by the grasses P. annua and P. pratensis. When the communities classified at level three were superimposed on the ordination (Fig. 3) the separation between communities was generally restricted to those at the extremes of the classification and dominated by these four species. Despite this simple distinction, the high constancy of a few species, especially J. scheuchzerioides, across most communities (Table 3) suggested that floristically Greens were more alike than they were different. It also suggested that the variation between Greens was more continuo-us than it was discrete. The floristic composition of the two largest communities, IV and V, which together contained 47% of all sites, supported this conclusión in that they both contained all the main species found in the other communities. Our data were typical of the "fuzzy" data sets which Equihua (1990) and other authors (reviewed in Belbin & Mc-Donald 1993) regarded as suitable descriptions of the continuous (non-discrete) distribution of species along environmental gradients.

The floristic similarity of Greens and their distinctness from the dominant native grasslands suggested that they were specialised communities. A few dominant species have become adapted to a specific set of environmental factors. Greens occurred in a variety of topographic positions, most of which had relatively high levels of livestock occupancy in comparison to the other heath communities. The consequent disturbance, mainly through the grazing, treading and manuring by sheep was common to most Greens and was probably the primary cause of vegetative change from the Cortaderia/Empetrum dominant vegetation.


The soils of all Greens were acidic. While most had pH values in the expected range (4.1 5.0) for the land systems in the Falkland Islands (King et al. 1969) 23% of the soils had pH ranges which were more favourable for plant growth. These soils were from valley, pond or penguin Greens.

Most land facets had adequate levels of K and Mg in their soils and all had very low values for total P and exchangeable Ca, except soils from marine terraces and valley bottoms, which had higher than average values (King et al. 1969). Our results con-firmed that soils from Greens on these land facets had high values for all these minerals.


The range in standing crop on Greens (Appendix Ilb) was lower than for Cortaderia communities (2 - 22 t/DM/ha) (Davies et al. 1990). This was probably due more to differences in utilisation by herbivores than to differences in the herbage productivity between the two communities. Annual productivity of Greens was 4 -141 DM/ha of which 65 - 94% was utilised (Kerr & McAdam 1993). As well as the differences in utilisation between different plant communities, they also found that utilisation was greater on coastal Greens, occupied by penguins, than on inland valleys, which were not.

The chemical composition of Greens herbage (Appendix II) may have encouraged their selective grazing by livestock in preference to other communities. N levels in herbage from Greens (1.3-3.8%) were higher than those found in a range of herbaceous species (0.7 - 2.5%) selected from other pastures in the Falkland Islands (Davies 1988). The levels of Ca (0.1-0.5%), Mg (0.1 - 0.6%), and K (0.9 - 2.8%) found by Davies (1988) were similar to those found in Greens, (0.2 -1.0%, 0.2 - 0.5% and 0.6 - 2.5%, respectively) in our survey. He also found that the forb Gunnera magellanica, an abundant species in Greens, had relatively higher levels of N, Ca and Mg than the other species sampled, which were not abundant in Greens. This species may have been the source of the higher levels of Mg found in herbage from community II.

The chemical composition of herbage from Greens also compared favourably with that from similar communities in other cool températe grasslands. The average N level (2.4%) of Greens was higher than that from grasslands on Soay (2.2%) (Milner & Gwynne 1974), an extensively grazed Scottish island, and from upland grasslands in Wales (2.3%) (Hughes et al. 1964). The latter herbage had slightly higher levels of K (1.7%) and similar levels of Ca (0.4%), P (0.2%) and Mg (0.25%) to those from Greens.

Relationships with animals

Disturbance by animal grazing, treading and defecation influenced the ecology of Greens to a great extent. Communities I-IV had probably been disturbed to a lesser extent than V-VIII since the latter had a greater cover of dung, more bare ground and generally less standing crop. Communities V-VIII also contained a greater proportion of species typically found in disturbed environments. Rumex acetosella (Davies & McAdam 1989) and Poa pratensis (Davies 1939) are known colonisers of disturbed grasslands. Davies (1939) regarded J. scheuchzerioides as "highly resistant to grazing". P. annua was a characteristic species of overgra-zed tussock grassland in South Georgia (Walton & Smith 1973) and P. pratensis had the same status in the steppe of Tierra del Fuego (Posse et al. 2000) and in the Scottish uplands (King & Nicholson 1964). Colobanthus quitensis, a common species in community VIII, was associated with maritime tussock grassland in Tierra del Fuego, which was often disturbed by penguins, albatrosses and seáis (Moore 1983). Exotic plant species were associated with highly disturbed grasslands in températe areas of Australia (Mclntyre & Lavorel 1994).

Inputs of nutrients from biotic sources probably had the greatest influence on the floristic and the chemical composition of Greens herbage and soils. Small deposits of calcareous shells exist and their potential to enhance Ca levels for agricultural purpo-ses has been investigated by Miller (2000) and more recently by Radie & McAdam (2008). Additionally, the traditional remoteness of the Falkland Islands from markets has, for economic reasons, resulted in little use of imported lime and fertilisers. Inputs of N via fixation in sown introduced legumes have been provisionally estimated as 66-69kg N/ha (Radie & McAdam 2008). Only two greens contained a legume, viz. Trifolium repens. There are no native legumes. Furthermore it is only since 2000 that significant areas have been sown with introduced species (Miller et al. 2000).


The main input of nutrients to coastal Greens was from penguins. Their rookeries were exclusive to communities V-VIII and were associated strongly with communities VII-VIII. Treading and manuring by penguins may have contributed to the better soil fertility of communities V-VIII compared to that of I-IV Community VIII had the highest level of soil and herbage R In an associated study, one of the penguin Greens in community VIII produced about 14 t dm/ha in one year (Kerr & McAdam 1993). This is one of the highest herbage production rates reported for any pasture in the Falkland Islands. The Green was grazed almost continuously and heavily by sheep and sheldgeese and was dominated by J. scheuchzerioides, P. annua and L. scariosa.

There were similarities in soil chemical status between Greens and some of the sub-Antarctic islands. Smith & French (1988) found 11 - 93 ppm of available P on Marion Island. One sample from a community heavily influenced by seal and penguin excreta measured 179 ppm R This was close to the máximum for Greens in our survey, which was sam-pled from a pond Green near to a penguin colony. Smith & French (1988) also found higher levels of Ca, (120 - 460 ppm), K (117 - 627 ppm) and Mg (106 - 3086 ppm) than those found in our study.

Superior plant productivity related to enhan-ced soil fertility associated with penguin rookeries has been reported for other south Atlantic islands. On Marión Island, penguin excreta were composed of 80 % uric acid, which was rapidly converted to ammonia (Lindeboom 1984). The soils, enriched in NH4-N, N03-N (Smith 1978) and inorganic-P (Smith 1976), sustained the improved vitality of many plants, such as Agrostis magellanica. The plants also had high levels of N and P in their leaves.


Communities I-IV had similar floristic compositions associated with high levels of treading, manuring and grazing by sheep. Agrostis spp., G. magellanica and J. scheuchzerioides were common co-dominants of these communities. Their association with Greens where sheep habitually congregated or were herded suggests that the activities of sheep also had a major influence on Green ecology. Floristic differences between these communities were minor and confined to the relatively high abundance/cover of Leptinella scariosa in community III. This may have been related to flock management and the higher level of disturbance associated with the heavily stocked small paddocks in which these Greens occurred. Their soils had high levels of NH4-N and NO3-N compared to those of the other communities in this group.

Occupancy of Greens by sheep may also have been related to site exposure. The availability of shelter f rom the seasonally severe chilling winds (McAdam 1985) may have influenced the densities of sheep between exposed and sheltered sites, particularly in the large, low-lying and extensively managed paddocks. Sheltered Greens, such as those in communities V-VI, may have been frequented more often by livestock and consequently disturbed to a greater extent than more exposed Greens. A comparable mechanism also controlled the distribution of sheep on Soay. There, the weather, plant growth and animal heat loss in winter interacted to control sheep population densities on similar communities (Milner & Gwynne 1974).

Floristic composition may also have been related to site drainage. The soils of the communities V-VIII were drier than those of communities I-IV and this may have favoured the higher abundance/cover of Poa pratensis and Aira praecox. The dryness of these sites may also have encouraged sheep to lie down more often there. In turn this may have concentrated the return of nutrients via excreta to these more fertile sites. Similarly, the composition of Festuca-Agrostis grasslands on Scottish hills was closely related to soil base status and moisture regime (King 1962). There the effects could not be separated from those of grazing pressure and other environmental factors that influenced the selective grazing of species.

Cattle and horses

Other livestock may also have influenced the floristic composition of Greens communities through their selective preference for particular species and the subsequent dispersal of the seeds in excreta. The establishment of P. annua, P. pratensis, Agrostis spp, Juncus spp. and Rumex acetosella graminoids in cattle and sheep dung on Scottish moorland was favoured over that of dicotyledons (Welch 1985). He proposed that cattle, through their defecation, enhanced the spread of grasses that attracted them. Thus cattle had a homogenising influence on the vegetation within their range. We did not find any difference in floristic composition between Greens occupied by mixed livestock and sheep alone. However the homogenising effect of cattle may have been true for all Greens in that, they were generally species-poor, dominated by short-growing graminoids (Table 3) and had abundant deposits of faeces (Appendix IIa) from cattle, horses and sheep.


The native sheldgeese (Chloephaga spp.) may also exert a considerable effect on Greens. The high constancy of J. scheuchzerioides in Greens may have been due to its preferential selection by the sheldgeese (Summers & Grieve 1982) and its ability to reproduce clonally (Moore 1983). The sheldgeese may have encouraged the adventive spread of J. scheuchzerioides, particularly on pond and coastal Greens that they grazed frequently (Summers 1985). Similarly, on tundra vegetation in Canadá, the foraging activities of native sheldgeese (Anser caeruíescens) enhanced the establishment of stoloniferous graminoids from vegetative propagules (Chou et al. 1992).

Overall, sheldgeese may benefit the grazing value of Greens. Sheldgeese selectively graze Greens in preference to C. pilosa communities, but they consume (15% of the herbage with minimal adverse effects on pasture production (Summers 1985). Additionally, the high productivity of Greens (Kerr & McAdam 1993) may be attributed to close grazing of the sward by sheldgeese. Jefferies et al. (1986) found that grazing by sheldgeese increased the net aboveground primary production of stands of Puccinellia phryganodes in Canadian tundra. This was attributed to the maintenance of high relative growth rates of axillary shoots of P. phryganodes in grazed areas. We found that the more productive Greens contained species that were preferentially grazed by sheldgeese, such as P. annua and P. pratensis (Summers & Grieve 1982).

Sheldgeese may also have influenced Greens vegetation through their defecation. Summers & Grieve (1982) estimated that the daily output of faeces was 115 - 253 g dm/bird with N levels of 3.5 -5.3 % OM. Given that geese population densities on Greens averaged 0.1 - 7.8 birds/ha, the output of N was calculated to be 4 - 718 kg ha/yr. As about 66% of N in the faeces of Canadian geese was estimated to be ammonia (Jefferies et al. 1986) it was probable that sheldgeese may have also positively affected N cycling on Greens.

The effects of the activities of each of sheep, cattle, horses, penguins and sheldgeese were difficult to sepárate. The effects of sheldgeese may only have been important around pond and coastal Greens where they gathered in large numbers. Defecation by penguins may have improved soil fertility and in turn enhanced herbage nutrient levels and greater herbivore activity. This agrees with the conclusión of Summers & Grieve (1982) that the main mechanisms of Poa Green development were intensive grazing, trampling and manuring.


Greens were floristically more similar than they were dissimilar and their management may now be easier to plan than if discrete floristic groups had been found. The species composition was a continuum in which the ubiquitous J. scheuchzerioides was associated with G. magellanica and Agrostis spp. on relatively wetter, infertile and undisturbed sites and with Poa spp. on drier, fertile and disturbed sites. Also, the variation in species composition was not associated strongly with the traditional distinctions between Greens based on landscape position or use by livestock. Although communities I-IV were most frequently located in valleys and holding paddocks and communities V-VIII were most frequently coastal and found near ponds, Greens at these four locations occurred in all community groups.

The high productivity and nutritive value of Poa Greens has long been recognised but as they composed only about three per cent of the land, mainly along coasts and around settlements, their contribution to sheep diets was limited due to severe uncontrolled grazing (Davies 1939). Kerr & McAdam (1993) confirmed their high productivity and utilisation by sheep and sheldgeese. This study showed that communities V and VIII, which were dominated by Poa spp were barer than communities I-IV (Table 3) and had fewer species than most other communities, except community I. Additionally, community VI had the highest cover/abundance of club mosses (Lycopodium spp.). These have little grazing value for livestock. Thus the productive potential of these Greens was degraded. Many of these Greens occurred on slopes and crests where the shallow soils had been eroded by frequent trea-ding by livestock.

To minimise further degradation of Greens the heterogeneous patterns of grazing and utilisation that exist between Greens and the surrounding pastures need to be evened out. In the extensive hill pastures in the south island of New Zealand, Thorrold et al. (1985) found that increased stocking rates produ-ced more even distributions of sheep grazing and dung than sub-division alone, although sub-division may have enhanced the effect of stocking rate. In the Falkland Islands land sub-division since 1980 has resulted in higher stocking rates than before (Summers et al. 1993). Additionally, greater utilisation by sheep of C. pilosa pastures was achieved in summer by adjusting stocking rates in small (3 ha) paddocks according to changes in leaf length (Davies et al. 1989).

Recent research to improve the sustainability of sheep grazing in the Islands has demonstrated the potential of rotational grazing compared to continuous grazing, to improve sheep productivity on native pastures (Kerr & McAdam 2000, Kerr 2002) and to use hay made from Cortaderia pilosa and fortified with supplements (Miller et al. 2005). Under rotational grazing, areas containing Greens and the associated tussock C. pilosa were rested during summer and reserved for grazing during winter. Over three successive winters rotationally grazed wethers continued to gain weight in contrast to continuously grazed sheep, which lost weight (Kerr 20021). Consequently, annual production from rotational grazed sheep was also superior. The superior productivity of the rotational sheep was due mainly to improved availability and utilisation of Greens forage by the sheep during winter. This improvement was attributed to the four-month rest from grazing these pasture types received during late summer.

It is also suggested that the forage contení of Poa Greens, the preferred grazing areas in the drier Empetrum heaths in the Islands, could be enhanced through controlled mixed grazing and nutrient input, particularly P. In Scotland cattle grazing and defecation depressed the growth of the Calluna vulgaris in favour of palatable graminoids e.g. Agrostis, Festuca and Juncus (Welch 1985). P fertilisation to improve grass composition over that of the shrubs has been advocaíed for the Empetrum heaths in Tierra del Fuego (Mendoza et al. 1995) and in the Falkland Islands (Broughton 1999).

In conclusión, the characteristics of the Greens studied reflected the combined effecís of defoliation , defecation and treading of the sheep, caííle, horses, sheldgeese and penguins thaí frequeníed the m. Their value for sheep grazing would be improved and sustained using management strategies thaí utilise and integrate the beneficial effecís of the livestock and birds.


We acknowledge the financial support of the UK Overseas Development Administration and the cooperation and hospitality of the landowners and farm managers in the Falkland Islands. We also thank; Messrs. F Cheong and D. Legg of the Agricultural Research Centre, Falkland Islands for the chemical analyses of the herbage samples and the preparation of the soil samples; Dr. J. stevens and Mr. G. McCollum of the Department of Agriculture and Rural Development (DARD) for Northe rn Ireland, for the chemical analyses of the soil samples and data processing respectively; Dr. Sally Watson of AFBI for preliminary analyses of the data; Dr. Rodrigo Olave for help with manuscriptpreparation ; Dr. N. Gremmen and Mr. G. Hoppeand the provisión of helpful comments with the latter; Mr. R. Le Feuvre of CSIRO, Division of Tropical Crops and Pastures, Australia for assistance with the full data analyses; Dr. R. J. Jones of the CSIRO and Dr. D.W.H. Walton of the British Antarctic Survey for the ir valuable comments on early drafts.



Belbin, L. & C. McDonald 1993. Comparing three classification strategies for use in ecology. J. Veg. Sci. 4: 341-348.         [ Links ]

Broughton, D.A. & J.H. McAdam 1999. Taxonomic changes in the Falkland Islands vascular flora. Falkland Islands Journal 7: 131-139.         [ Links ]

Broughton, D.A. & J.H. McAdam 2002. The non-native vascular flora of the Falkland Islands. Botanical Journal of Scottland 54:153-190.         [ Links ]

Chou, R., C. Vardy & R.L. Jeffries 1992. Establishment from leaves and othe r plant fragments produced by the foraging activities of Geese 1992. Functional Ecology 6: 297-301.         [ Links ]

Davies, A.B., J.Riley, R. Poultney & A.O. Carter 1989. The impact of utilisation on a Cortaderia pilosa Whitegrass pasture in terms of production , botanical composition and tiller dynamics. Proceedings of the 16th International Grassland Congress, Nice, France, pp 1475-1476.         [ Links ]

Davies, A.B., J. Riley & D.W.H. Walton 1990. Plant form, tiller dynamics and above-ground standing crops of the range of Cortaderia pilosa communities in the Falkland Islands. J. Appl. Ecology, 27: 298-307        [ Links ]

Davies, T.H., LA. Dickson, C.T. McCrea, H. Mead & W.W. Williams 1971. The sheep and cattle industries of the Falkland Islands. Overseas Development Administration, London.         [ Links ]

Davies T.H. & J.H. McAdam 1989. Wild flowers of the Falkland Islands. Bluntisham Books for the Falkland Islands Trust. Bluntisham, UK.         [ Links ]

Davies, W. 1939. The grasslands of the Falkland Islands. Government Printer , Port Stanley.         [ Links ]

Davies, A.B., J. Riley & D.W.H. Walton 1990. Plant form, tiller dynamics and above-ground above ground crops of the range of Cortaderia pilosa communities in the Falkland Islands. J. Appl. Ecol. 27: 298-307.         [ Links ]

Equihua, M. 1990. Fuzzy clustering of ecological data. J. Ecol. 78: 519-534.         [ Links ]

Frame, J., R Newbould & J.M.M. Munro 1985. Herbage production from the hills and uplands. In: Hill and Upland Livestock Production . Occasional publication of the British Society of animal production , 10: Edinburgh, pp 9-37.         [ Links ]

Gauch, H.G. 1982. Multivariate analysis in community ecology. Cambridge studies in ecology. Cambridge University Press.         [ Links ]

Hill, M.O. 1979a. TWINSPAN - A FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and the atribuities. Section of Ecology and Systematics, Cornell University , Ithaca, New York.         [ Links ]

Hill, M.O. 1979b. DECORANA: A FORTRAN program for detrended correspondence analysis and reciprocal averaging. Section of Ecology and Systematics, Cornell University , Ithaca, New York.         [ Links ]

Hill, M.O., R.G.H. Bunce & M.W. Shaw 1975. Indicator species analysis, a divisive polythetic method of classification, and its application to a survey of native pine woods in Scotland. J. Ecol. 63: 597-613.         [ Links ]

Hughes, R.E. 1958. Sheep population and environment in Snowdonia North Wales. J. Ecol. 46: 169-189.         [ Links ]

Hughes, R.E., C. Milner & J. Dale 1964. Selectivity in grazing. In Proceedings of the 3rd Symposium of the British Ecological Society. Grazing terrestrial and marine environments. Ed. Crisp, D.J. pp 189-202. Blackwell Scientific Publications, Oxford.         [ Links ]

Jefferies, R.J., D.R. Bazel & S.M. Cargill 1986 Effects of grazing on Tundra vegetation - a positive feedback model. In Proceedings of the 2nd International Rangeland Congress, Adelaide 1985. Rangelands: A resource under siege. Ed's: Joss P.J., Lynch RW. & Williams O.B. p50.         [ Links ]

Kerr, J.A. & J.H. McAdam 1993. The productivity and nutritive value of coastal and valley-bottom plant communities in the Falkland Islands. Proceedings of the 17th International Grassland Congress, New Zealand & Australia, pp 1611-1612.         [ Links ]

Kerr, J.A. & J.H. McAdam 2000. Rotational and continuously grazed Whitegrass pasture in the Falkland Islands. 6th B.G.S. Research Conference, British Grassland Society, Sept. 2000. pp 93-94.         [ Links ]

Kershaw, K.A. 1973. Quantitative and dynamic plant ecology. Edward Arnold, London.         [ Links ]

King, J. 1962. The Festuca-Agrostis grassland complex in south-east Scotland. J. Ecol. 50: 321-355.         [ Links ]

King, J. & LA. Nicholson 1964. Grassland of the forest and sub-alpine zones. In: Burnett, J.H. Ed. The vegetation of Scotland, pp. 168-215. Oliver & Boyd, Edinburgh and London.         [ Links ]

King, R.B., D.M. Lang & A. Blair-Rains 1969. Land systems analysis of the Falkland Islands, with notes on the soils and grasslands. Miscellaneous Report of the Land Resources División, No. 72, Reprinted Directorate of the Overseas survey, London. 24pp.         [ Links ]

Lindeboom, H.J. 1984. The nitrogen pathway in a penguin rookery. Ecology, 65: 269-277.         [ Links ]

M.A.F.F 1987. The analysis of agricultural materials. Technical Bulletin, M.A.F.F, HMSO, London.         [ Links ]

McAdam, J.H. 1985. The effect of climate on plant growth and agriculture in the Falkland Islands. Progress in Biometeorology. 2: 155-176.         [ Links ]

McAdam J.H., A.B. Davies, B.M.R. Harvey, G.M. Hoppe & D.W.H. Walton 1989. The agronomy and ecophysiology of Whitegrass Cortaderia pilosa. Proceedings of the 3rd International Grassland Ecology Conference Czechoslovakia, 1989. Czechoslovak Scientific and Technical Society, Banksa Bystrica, pp. 117-124.         [ Links ]

Mclntyre, S. & S. Lavorel 1994. How environmental and disturbance factors influence species composition in temperate Australian grasslands. J. Veg. Sci. 5: 373-384.         [ Links ]

Mendoza, R.E., M.B. Collantes, J. Anchorena & S. Cainzos 1995 Effects of liming and fertilisation on forage yield and vegetation in dry heath soils from Tierra del Fuego. J. Pl. Nutr. 18: 401-420.         [ Links ]

Miller, S., D.Parson, B.D. Reid & R. Thompson 2000. The pasture improvement Programme. Establishing improved pastures in the Falkland Islands. Stanley. Department of Agriculture, Falkland Islands Government 47 pp.         [ Links ]

Miller, S., G. Lennie & D. Clelland 2005. Fortifying native pasture hay with molasses: urea mixtures improves its digestibility and nutrient intake by weaner sheep. Animal Feed Science & Technology 119:259-220.         [ Links ]

Milner, C. & D. Gwynne 1974. The Soay sheep and their food supply. In: Island survivors -The ecology of the Soay Sheep of St. Kilda, eds R A. Jewell, C. Milner & J. Morton-Boyd, pp. 273-325. Athlone.         [ Links ]

Moore, D.M. 1968. The vascular flora of the Falkland Islands. British Antarctic Survey Scientific Reports No. 60. NERC London.         [ Links ]

Moore, D.M. 1983. The flora of Tierra del Fuego. Anthony Nelson, Oswestry, Wales.         [ Links ]

Posse, G., J. Anchorena & M.B. Collantes 2000. Spatial micro-patterns in the steppe of Tierra del Fuego induced by sheep grazing. J. Veg. Sci. 11: 43-50.         [ Links ]

Radie, S & J.H. McAdam 2008. Establishment and growth of legumes in acid soils in the Falkland Islands. Proceedings of XXI International Grassland and Rangeland Congress, Hohhot, China. Vol 1, p342.         [ Links ]

Smith, V.R. 1976. Above ground crop and nutrient status of Marion Island Sub-Antarctic vegetation. S. Afr. J. Bot. 42: 231-263.         [ Links ]

Smith, V.R. 1978, Animal-plant-soil nutrient relationships on Marion Island Sub-Antarctic. Oecologia 32: 239-253.         [ Links ]

Smith, V.R. & D.D. French 1988. Patterns of variation in the climates, soils and vegetation of some subantarctic and Antarctic islands. S. Afr. J. Bot. 54: 35-46.         [ Links ]

Summers, O., R. Haydock & J.A. Kerr 1993. Land sub-division in the Falkland Islands. Proceedings of the 17th International Grassland Congress, New Zealand & Australia, pp. 812-814.         [ Links ]

Summers, R.W. 1985. The size and composition of sheldgeese populations and their food consumption on different vegetation types in the Falkland Islands. J. Appl. Ecol. 22: 1-17.         [ Links ]

Summers, R.W. & A. Grieve 1982. Diet, feeding behaviour and food intake of the Upland Goose Chloephaga picta and the Ruddy-headed Goose C. rubidiceps in the Falkland Islands. J. Appl. Ecol. 19: 783-804.         [ Links ]

Thorrold, B.S., K.F. O'Connor & J.G.H. White 1985. Management influences on sheep behaviour, dung distribution and soil phosphate. Proceedings of the New Zealand Grassland Association 46: 127-134.         [ Links ]

Walton, D.W.H. & R.I.L. Smith 1973. The status of the alien vascular flora of South Georgia. British Antarctic Survey Bulletin, No. 36: 79-97.         [ Links ]

Welch, D. 1985. Studies in the grazing of heather moorland in north-east Scotland. IV Seed dispersal and plant establishment in dung. J. Appl.Ecol. 22: 461-472.         [ Links ]


Received: Nov., 09, 2007

Accepted: Jul., 11,2008


Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons