Frequency of defoliation of native and naturalized species of the Flooding Pampas (Argentina)

Transcription

1 Frequency of defoliation of native and naturalized species of the Flooding Pampas (Argentina) M. G. Agnusdei and A. Mazzanti* Instituto Nacional de TecnologõÂa Agropecuaria (INTA), EstacioÂn Experimental Agropecuaria de Balcarce, Balcarce, Argentina Abstract This study describes the effect of herbage mass of species on defoliation patterns in a multispecies temperate± subhumid natural grassland of the Flooding Pampas at low, medium and high sward herbage mass [761 8á7, and á0 kg green dry matter (DM) ha ±1 respectively] maintained by continuous grazing of cattle at variable stocking rates. Four native and naturalized warm-season species were studied: the grasses Cynodon dactylon, Leersia hexandra and Paspalum dilatatum and the legume Lotus tenuis. Sward state of the four species was described in terms of sward surface height, herbage biomass and the number of individual grass tillers and legume stems per unit area. The defoliation pattern was monitored twice-weekly during an experimental period of 3 weeks on permanently marked individual plants. Proportions of individual plants defoliated and the ranking of species selected differed among sward treatments and species. The results showed that the density of the species was a major factor determining the pattern of species defoliation and suggest that, in continuously grazed mixtures, the pattern of defoliation might be more dependent on the abundance of a species in the plant community than on species preferences. The results show the importance of considering this structural characteristic of swards in understanding the process of species selection by large herbivores in multispecies vegetation. Keywords: defoliation patterns, species selectivity, temperate±subhumid grasslands *Deceased 10 June Correspondence to: Ir M. G. Agnusdei, Instituto Nacional de TecnologõÂa Agropecuaria, Estacio n Experimental Agropecuaria de Balcarce, Ruta 226 km 74.5, CC 276, CP 7620 Balcarce, Pcia. de Buenos Aires, Argentina. magnusdei@balcarce.inta.gov.ar Received 5 May 1998; revised 18 November 1999; resubmitted 24 January 2001 Introduction The Flooding Pampas of Argentina (a triangle with one vertex located at 60 W longitude and 36á5 S latitude and the other two at 57 W longitude, 35±38 S latitude) is a subhumid temperate region of ha where the dominant vegetation is natural grassland. The botanical components of these plant communities are both native and naturalized species, most of which originated from South America and Europe. The mild climate of the region enables the co-existence of temperate and subtropical species (Cabrera, 1970) that confer a year-long variation to the structure and functioning of the natural grasslands. In addition to this temporal variability, natural vegetation is spatially heterogeneous as a result of the differential adaptation of the species to mosaics of alkaline and/or poorly drained soils (Tricart, 1973). The understanding of the interactions between the grazing process and the heterogeneity of vegetation is recognized as critical for optimizing animal production on multispeci c grasslands (Hodgson, 1981; Stuth, 1991). Local research concerning the grazing process on natural grasslands is mainly descriptive and refers to botanical differences associated with grazing management (Sala et al., 1986; Facelli, 1988; Hidalgo and 1 CauheÂpeÂ, 1991), as well as to animal selectivity described by microhistological techniques (CauheÂpe 2 and Fernandez Grecco, 1981; Brizuela et al., 1983). At present, no information on species defoliation patterns of plant species by cattle has been published for these natural grasslands. Several detailed studies on patterns of defoliation in multispecies vegetation have been conducted at other latitudes (Gammon and Roberts, 1978a,b,c; Hodgkinson, 1980; Briske and Stuth, 1982; Heitschmidt et al., 1990; Tallowin et al., 1995). These authors have analysed inter- and intraspeci c variations in the frequency and severity of tiller defoliation, as well as the in uence of several architectural attributes of plants on defoliation patterns under differing grazing systems. 344 Ó 2001 Blackwell Science Ltd. Grass and Forage Science, 56, 344±351

2 Defoliation of Flooding Pampas species under grazing 345 The prediction of species selection from multispecies communities is dif cult to carry out with present knowledge. In this sense, detailed studies on defoliation patterns and sward structure are needed in order to progress understanding of the mechanisms controlling the grazing process in multispecies vegetation. This information will be valuable for the elaboration of ef cient and sustainable grazing systems based on the utilization of temperate natural grasslands. The objectives of the investigation were: (i) to estimate the frequency of defoliation of four native and naturalized species that constituted»0á8 of the herbage mass present in a natural grassland of the Flooding Pampas; and (ii) to relate grazing patterns to the availabilities of species in the grassland. Three sward states were maintained through frequent control of sward by periodic stocking rate adjustments under continuous grazing by cattle. The study was carried out in a natural grassland that, like a great proportion of the natural vegetation in the 3 region, is dominated by grasses (Leo n, 1975) and legumes (Lotus tenuis) (Mazzanti et al., 1988). Although non-forage species (e.g. non-legume dicotyledons) may be co-dominant in certain grasslands in terms of aerial cover (Sala et al., 1986; Facelli, 1988), they generally make a minor contribution to forage production (Fernandez Grecco et al., 1995; Pueyo, 1996). For example, avoided or toxic species (e.g. Ambrosia tenuifolia, Solanum malacoxylon) may be co-dominant in some situations (Leo n, 1975). However, they are scarce components in most grasslands of the region (Vervoorst, 1967). Materials and methods Experimental conditions The study site was in Ayacucho county (37 S latitude and 57 W longitude), Buenos Aires province, Argentina. The climate is temperate±subhumid with average daily temperatures for the coldest and warmest months (July and January, respectively, corresponding to winter and summer in the southern hemisphere) of 8 C and 21 C, respectively, and average monthly precipitation ranging between 30 and 60 mm in winter and 70 and 120 mm in spring, summer and autumn. Hydric de cits are normally registered from November to March (end of spring and summer in the southern hemisphere). The general topography of the area is at, with alkaline humic gley soils as the dominant type. Natural grasslands are dominated by annual and perennial grasses. The species that were studied are major constituents of these grasslands. They were (a) Cynodon dactylon (Cynodon), a naturalized warm-season, rhizomatose grass; (b) Leersia hexandra (Leersia), a native warmseason, caespitose grass; (c) Lotus tenuis (Lotus), a native warm-season legume; and (d) Paspalum dilatatum (Paspalum), a native warm-season, caespitose grass. The experiment was carried out from 4 to 23 March 1994 on a typical natural grassland of the region. The experimental area was subdivided to obtain duplicates of three paddock sizes: 2, 3 and 5 ha (six paddocks in total). Between August 1993 and the start of the experiment, the number of animals was adjusted in order to generate three contrasting levels of herbage mass of around 700 (2-ha paddocks), 1000 (3-ha paddocks) and 1500 (5-ha paddocks) kg green dry matter (DM) ha ±1 [low (L), intermediate (M) and high (H) herbage mass treatments respectively]. Once the desired herbage mass levels had been established, they were maintained at as near a steady state as possible by regular sward surface height controls through variable stocking densities with Aberdeen Angus heifers (see below). Mean stocking densities during the experimental period were 3, 1á5 and 0á8 heifers ha ±1 day ±1 for the L, M and H treatments respectively. Sward measurements Measurements of sward surface height were made twice a week during the experimental period on 50 randomly chosen sites per plot with an HFRO sward stick (Bircham, 1981), individual heights being de ned by the rst touch of a green leaf. At the beginning and end of the experiment, two herbage mass measurements were made. Herbage mass estimates were made on cm turves per paddock that were removed to the laboratory and cut to ground level. Three measurements of sward surface height were made on each turf before it was removed from the eld. Sward surface height values were regressed with herbage mass, and the relationship was used for controlling and maintaining the sward states from eld sward surface height determinations. Before cutting, the turves were divided into two portions comprising one-third and two-thirds of the total surface area. The vegetation on the one-third was separated into the following fractions: Cynodon, Leersia, Lotus, Paspalum, other grasses and graminoids, dicots (dicotyledoneous species) and dead material. Individuals (tillers for grasses and stems for legumes) were counted in order to estimate species density. The fractions were dried at 60 C for 24 h and weighed, and species biomass was expressed on a DM basis [biomass is de ned as the weight of green leaf (lamina and pseudostem) material per unit area from ground level]. The weight of individuals was also estimated. Additionally, the dry weight of the total green biomass remaining on the two-thirds of the turves was determined.

3 346 M. G. Agnusdei and A. Mazzanti Measurements of defoliation pattern At the beginning of the experimental period, individual grass tillers and legume stems were labelled in each of the six paddocks at soil level with coloured wire rings. A total of 60 individuals per species and per treatment were placed along 12 randomly allocated 1-m-long transects. Twice a week during the 21 days of the experimental period of detailed measurements (Hodgson and Ollerenshaw, 1969), defoliations on each leaf were recorded. Grazed leaves were marked at the tips for the identi cation of further defoliations. With this information, the proportion of individuals defoliated (PID) was estimated daily for each species. In accordance with other authors (Gammon and Roberts, 1978a; Heitschmidt et al., 1990), PID for each species was considered as a measure of the degree of grazing of that species. Statistical analysis Rankings of species were calculated for each paddock. The species with the highest PID was ranked 1, and the values for the remaining species were calculated as the ratio between their PID and that corresponding to the most defoliated species (Gammon and Roberts, 1978c). In order to explore the global relationships between the variables density, green leaf biomass and weight of individuals with PID, the ratio between the values estimated for each species and the corresponding totals within each paddock were calculated. The relative contribution of each species to the total number of individuals defoliated (CD) was the ratio between the number of individuals defoliated from each species and the total number of individuals defoliated in each paddock. For each species, the selectivity index (SELd) was de ned as PID/proportion of grass tillers or legume stems of the species in the grassland. The experimental design was completely at random within blocks with three treatments (L, M and H) and two spatial replicates. Analysis of variance was performed on sward and pattern of defoliation variables. Main factors, sward state and species, and their interaction were tested using the residual mean square. Duncan's range test was used for comparisons of means (P <0á05). Results Sward characteristics Table 1 summarizes the structural characteristics of the swards during the experiment. There was a range of herbage mass as desired, but only the L and H heights were signi cantly (P <0á05) different. The density of grass tillers and legume stems decreased signi cantly (P <0á05) between L, M and H herbage masses. Data on species density and biomass (Table 2) showed high variability, with no signi cant differences between treatments for either variable. However, herbage biomass tended to be consistently greater on the H than on the L and M treatments. Cynodon and Paspalum tended to be the most abundant species in all treatments. On the other hand, Leersia was always the least abundant one, whereas Lotus showed variable contributions, being a low-abundant component at L and M and one of the dominants at H. The spatial distribution of individuals varied among species. Leersia was absent in 0á60 of the quadrats, whereas the other three species were present in more than 0á65 of the cases (Figure 1). When only the quadrats in which a species was present were considered, Cynodon, Lotus and Paspalum tended to have a more symmetric distribution than Leersia. The data indicated the existence of two contrasting spatial arrangements: (i) low-abundant, spatially localized Herbage mass treatments L M H Table 1 Structural characteristics of the swards. Mean s.e. Mean s.e. Mean s.e. Height (cm) 4á9 1á3 12á4 5á6 24á5 3á2 Green dry matter (DM, kg á á á6 green leaf DM ha )1 ) Tiller and stem density (no. m )2 ) á á á1 L, M and H denote sward states maintained at low, medium and high herbage mass respectively. Mean values were calculated from measurements made at the beginning and end of the experiment (4 and 23 March 1994 respectively).

4 Defoliation of Flooding Pampas species under grazing 347 Table 2 The effect of herbage mass treatments on species density (number of grass tillers and legume stems m )2 ) and green leaf biomass (kg DM ha )1 ). Herbage mass treatments L M H Species Mean s.e. Mean s.e. Mean s.e. No. of grass tillers and legume stems Cynodon dactylon á á á9 Leersia hexandra á á á2 Lotus tenuis á á á7 Paspalum dilatatum 579 5á á á6 Green leaf biomass Cynodon dactylon á á á1 Leersia hexandra 30 39á á á1 Lotus tenuis 70 80á á á2 Paspalum dilatatum 250 0á á á3 L, M and H denote sward states maintained at low, medium and high herbage mass respectively. Mean values were calculated from measurements done at the beginning and end of the experiment (4 and 23 March 1994 respectively). Figure 1 Frequency distribution of individuals (grass tillers and legume stems) in the grassland. h, Cynodon dactylon;, Leersia hexandra;, Lotus tenuis;. Paspalum dilatatum. species ˆ Leersia; (ii) mid- to high-abundant, spatially conspicuous species ˆ Cynodon, Lotus and Paspalum. Frequency of defoliation The PID was affected by herbage mass treatments and species (Figure 2), and there was a signi cant (P <0á05) species sward state interaction. PID was higher on the L treatment than on H for all species. On the M treatment, Leersia and Lotus had intermediate values between L and H, whereas no differences were detected between this treatment and L for Paspalum and Cynodon. PID was highest for these two species on the L and M treatments, respectively, whereas on the H treatment, Paspalum, Cynodon and Lotus showed the highest PID values. On the H treatment, Leersia had the lowest frequency of defoliation. The selection ranking of species differed among sward treatments and species. Differences between the highest and the lowest ranking increased from the L to the H treatments (0á40, 0á55 and 0á66 respectively). With reference to species, although Cynodon and Paspalum had similar rankings among sward states (between 0á80 and 1á00), Leersia and Lotus showed large variations in their ranking. Values for Leersia and Lotus were 0á68 and 0á60 on the L treatment, 0á45 and 0á50 on the M treatment and 0á34 and 0á83 on the H treatment respectively. The weight of the individuals and their density are the two structural components of the sward that determine herbage biomass. Although the weight of individuals showed no relationship with PID, density explained around 0á70 (P <0á05) of variation in PID in the L and H sward treatments. Species selectivity A general function of the type y ˆ ax b was used to describe the relationship between the selectivity index

5 348 M. G. Agnusdei and A. Mazzanti Proportion of individuals defoliated Figure 2 Effect of sward state on the proportion of individuals (grass tillers and legume stems) defoliated. L, M and H denote sward states maintained at low, medium and high herbage mass respectively. h, Cynodon dactylon;, Leersia hexandra;, Lotus tenuis;, Paspalum dilatatum. Figures denote ranking of species selection. [ The species with the highest proportion of individuals defoliated (PID) was ranked 1. The values for the remaining species were calculated as the ratio between their PID and that of the most defoliated species.] Large and small letters denote within- and among-sward state differences (P < 0á05, Duncan test), respectively, between mean values. Species selectivity index Figure 3 Relationship between the proportion of green leaf biomass and the species selectivity index [proportion of individuals of the species defoliated (PID) daily /proportion of grass tillers or legume stems of the species]. h, Cynodon dactylon; s, Leersia hexandra; n, Lotus tenuis; ), Paspalum dilatatum. (SELd) and the contribution of the species to the total green leaf biomass (Figure 3). SELd increased strongly at low levels of green leaf biomass but not above values of 0á25. Within the range of 0á25±0á50 contribution to the total green leaf biomass, a mean SELd of 1á12 0á14 (no to low selection) was estimated.

6 Defoliation of Flooding Pampas species under grazing 349 Discussion The PID tended to decrease as sward biomass and height increased (Figure 2). These results agree with those reported for simpler swards (Hodgson, 1966; Mazzanti and Lemaire, 1994) and con rm the high dependence of frequency of defoliation on stocking density (Wade 4 and Baker, 1979). In addition, the species selectivity rankings and the rank order of species also differed among sward states, particularly for Leersia and Lotus (Figure 2). Such effects of sward treatment on species rankings contrast with the results obtained in a rangeland by Gammon and Roberts (1978c), who reported no consistent differences in the ranking of species selection under different grazing managements. Their results suggested that species themselves may be strong determinants of the pattern of defoliation by the animals and are in line with classi cations of species into selectivity (Heady, 1975; Stuth, 1991) and functional nutritional categories (Huston and Pinchak, 1991). Interspeci c variations observed in PID, as well as in the rankings of species selection, may be interpreted as a consequence of between-treatment differences in the pattern of defoliation by the heifers. However, the contribution of the species to the total green leaf biomass of the grassland (B) varied among treatments (Table 3). Such heterogeneity may be a main factor controlling species selection in mixed vegetation (Senft, 1987) and, hence, must be taken into account in understanding the mechanisms that determine diet selection from multispecies pastures (Parsons et al., 1994). PID tended to increase with the increase in a species' contribution to total green leaf biomass (Figure 3). The fact that dominants were more frequently defoliated than less abundant species agreed with data obtained for some other multispecies grasslands (Senft et al., 1985; Heitschmidt et al., 1990). A decrease in the rate of encounter associated with the localized or sparse distribution of low-abundant species (Figure 1) and consequent constraint on intake rates (Newman et al., 1995) may partly explain the trend observed in the data reported here. The data for relative density and weight of an individual species can be interpreted as associated with the respective relative horizontal and vertical availability of those species in the grassland. Although individual availability showed no association with PID, the horizontal component appeared as a main factor determining the pattern of species defoliation in this multispecies grassland. Although within a species population, longer-leaved or taller tillers would be consistently the most frequently defoliated (Hodgson and Ollerenshaw, 1969; Gammon and Roberts, 1978c), our ndings suggest that, in this type of grassland, maintained under relatively stable sward state conditions, the vertical availability would be less important than the horizontal distribution of the species in in uencing the pattern of interspecies defoliation. The species selectivity index increased with relative herbage biomass but at a decreasing rate at higher relative herbage biomass (Figure 3). Although a pattern of low defoliation was evident for less abundant species (biomass values» <0á25), dominants tended to be defoliated in proportion to their density (SELd around a value of 1) or with high defoliation frequency (SELd up to 1á25). According to the predictions of the mechanistic model proposed by Parsons et al. (1994), the costs of grazing selectively have little effect on the diet selected until the relative abundance of the preferred species becomes less than about 0á20 of a two-species Table 3 Contribution of each species to the total number of individuals defoliated (CD) and species contribution to the green leaf biomass (B) in the grassland. Herbage mass treatments L M H Species Mean s.e. Mean s.e. Mean s.e. Cynodon dactylon CD 0á49 0á083 0á49 0á058 0á38 0á087 B 0á44 0á060 0á35 0á031 0á30 0á020 Leersia hexandra CD 0á11 0á047 0á16 0á205 0á02 0á014 B 0á07 0á038 0á09 0á072 0á05 0á011 Lotus tenuis CD 0á10 0á078 0á14 0á060 0á28 0á107 B 0á04 0á024 0á20 0á065 0á23 0á054 Paspalum dilatatum CD 0á30 0á050 0á21 0á093 0á31 0á034 B 0á45 0á046 0á37 0á106 0á43 0á045 L, M and H denote sward states maintained at low, medium and high herbage mass respectively. CD, number of individuals of the species defoliated/total number of individuals defoliated.

7 350 M. G. Agnusdei and A. Mazzanti sward. The decrease in SELd below a relative biomass of around 0á25 observed in this study may be associated with the greater overall bene ts of grazing the dominant forage matrix of the sward instead of moving to marginal or localized areas. The fact that, beyond a critical relative availability, the selectivity did not increase markedly suggests a lack of preference categories between the component species of the grassland, whether it was a grass or a legume. However, the restricted proportion of the legume species compared with the grasses might have limited the expression of species preferences by the animals. Species with a green leaf biomass below 0á20 contributed in minor proportions to the total number of individuals defoliated daily, whereas higher availabilities determined the major contributions (CD between 0á20 and 0á16 in the rst case and 0á28±0á49 in the second; Table 3). The lack of difference between preferences for the species implies that the species were similar in nutritive value and hence the most abundant species increase; this generally affords the animal with the opportunity to maximize herbage consumption (Stuth, 1991). Conclusions The results showed that the grazing process had a general pattern that tended to be independent of species themselves and was basically determined by the availability of the species in the grassland. This suggests that at least some plant±animal interactions may be general across contrasting structural conditions of this grassland. The variability observed in the selection index above a relative species biomass of around 0á25 suggests that grazing was basically non-selective. Further, there was no strong avoidance (selectivity ratios close to 0) at any relative availability or for any particular species evident from the data. However, extreme selection for species is commonly reported in the literature (e.g. Hodgson, 1981; CauheÂpe and Fernandez Grecco, 1981; Broom and Arnold, 1986), this being considered the most prominent pattern when cattle select species 5 within multispecies communities (Senft, 1987). Discrepancies may be a consequence of the fact that the study only included four forage species and that the grazing horizon under the applied management mainly comprised young high-quality leaves (Mazzanti and Lemaire, 1994; Pueyo, 1996). Under such conditions, trends in defoliation patterns by cattle grazing in multispecies communities would not differ markedly from those observed for less complex temperate swards (Hodgson and Ollerenshaw, 1969; Hodgson, 1981). References BIRCHAM J.S. (1981) Herbage Growth and Utilization Under Continuous Stocking Management. Ph.D. thesis, University of Edinburgh, 300 pp. BRISKE D.D. and STUTH J.W. (1982) Tiller defoliation in a moderate and heavy grazing regime. Journal of Range Management, 35, 511±514. BROOM D.M. and ARNOLD G.W. (1986) Selection by grazing sheep on pasture plants at low herbage availability and responses of the plants to grazing. Australian Journal of Agricultural Research, 37, 527±528. BRIZUELA M.A., CAUHEPE AUHE PE M.A., CID M.S., VIVIANI ROSSI R.C., FERNANDEZ GRECCO R.C. and YAGUEDDU C. (1983) Botanical composition of the diet of cattle in a natural grassland. Revista Argentina de ProduccioÂn Animal, 10, 395±403. CABRERA A.L. (1970) Vegetation of the Buenos Aires Province. ColeccioÂn Cientõ ca Instituto Nacional de TecnologõÂa 6 Agropecuaria (INTA), Buenos Aires, II, 624pp. CAUHEPE AUHE PE M.A. and FERNANDEZ GRECCO R.C. (1981) Cattle diet under grazing in a natural grassland of the Flooding Pampas. Revista Argentina de ProduccioÂn Animal, 8, 85±95. FACELLI J.M. (1988) Response to grazing after nine years of cattle exclusion in a Flooding Pampa grassland, Argentina. Vegetatio, 78, 21±25. FERNANDEZ GRECCO R.C., MAZZANTI A. and ECHEVERRIA H. (1995) Effect of nitrogen fertilization on herbage growth in a natural grassland of the Flooding Pampas (Argentina). Revista Argentina de ProduccioÂn Animal, 15, 173±176. GAMMON D.M. and ROBERTS B.R. (1978a) Patterns of defoliation during continuous and rotational grazing of the Matopos Sandveld of Rhodesia. 1. Selectivity of grazing. Rhodesian Journal of Agricultural Research, 16, 117±131. GAMMON D.M. and ROBERTS B.R. (1978b) Patterns of defoliation during continuous and rotational grazing of the Matopos Sandveld of Rhodesia. 2. Severity of defoliation. Rhodesian Journal of Agricultural Research, 16, 133±145. GAMMON D.M. and ROBERTS B.R. (1978c) Patterns of defoliation during continuous and rotational grazing of the Matopos Sandveld of Rhodesia. 3. Frequency of defoliation. Rhodesian Journal of Agricultural Research, 16, 147±164. HEADY H.F. (1975) Rangeland Management. New York: McGraw-Hill, 460 pp. HEITSCHMIDT R.K., BRISKE D.D. and PRICE D.L. (1990) Pattern of interspeci c tiller defoliation in a mixed-grass prairie grazed by cattle. Grass and Forage Science, 45, 215±222. HIDALGO L.G. and CAUHEPE AUHE PE M.A. (1991) Effect of seasonal rest on the aboveground biomass for a native grassland of the ooding Pampa, Argentina. Journal of Range Management, 44, 471±475. HODGSON J. (1966) The frequency of defoliation of individual tillers in a set-stocked swards. Journal of the British Grassland Society, 21, 258±263.

8 Defoliation of Flooding Pampas species under grazing 351 HODGSON J. (1981) In uence of sward characteristics on diet selection and herbage intake by the grazing animal. In: Hacker J.B. (ed.) Nutritional Limits to Animal Production from Pastures, pp. 153±166. Farnham Royal: Commonwealth Agricultural Bureaux. HODGSON J. and OLLERENSHAW J.H. (1969) The frequency and severity of defoliation of individual tillers in set-stocked swards. Journal of the British Grassland Society, 24, 226±234. HODGKINSON K.C. (1980) Frequency and extent of defoliation of herbaceous plants by sheep in a foothill range community in northern Utah. Journal of Range Management, 33, 164±169. HUSTON J.E. and PINCHAK E. (1991) Range animal nutrition. In: Heitschmidt R.K. and Stuth J.W. (eds) Grazing Management, pp. 27±63. Portland, OR: Timber Press. LEON EO R.J.C. (1975) Plant Communities of the Castelli-Pila Region. MonografõÂa no. 5, pp. 75±107. La Plata, Buenos Aires (Argentina): Comisio n de Investigaciones Cientõ cas. MAZZANTI A. and LEMAIRE G. (1994) The effect of nitrogen fertilization upon the herbage production of tall fescue swards continuously grazed with sheep. 2. Consumption and ef ciency of utilization. Grass and Forage Science, 49, 353±359. MAZZANTI A., MONTES L., MINON INÄ D., SARLANGUE H. and CHEPPI C. (1988) Use of Lotus tenuis in livestock ranches of the Flooding Pampas. Revista Argentina de ProduccioÂn Animal, 8, 301±306. NEWMAN J.A., PARSONS A.J., THORNLEY J.H.M., PENNING P.D. and KREBS J.R. (1995) Optimal diet selection by a generalist grazing herbivore. Functional Ecology, 9, 255±268. PARSONS A.J., THORNLEY J.H.M., NEWMAN J. and PEN- NING P.D. (1994) A mechanistic model of some physical determinants of intake rate and diet selection in a twospecies temperate grassland sward. Functional Ecology, 8, 187±204. PUEYO D.J. (1996) Growth and Utilization Dynamics in a natural Grassland of the Flooding Pampas. M.Sc. thesis, Facultad de Ciencias Agrarias de la Universidad Nacional de Mar del Plata, Argentina. SALA O.E., OESTERHELD M., LEON R.J.C. and SORIANO A. (1986) Grazing effects upon plant community structure in subhumid grasslands of Argentina. Vegetatio, 67, 27±32. SENFT R.L. (1987) Domestic herbivore foraging tactics and landscape pattern. In: Proceedings of the Symposium on Plant±Herbivore Interactions. Snowbird, Utah, pp. 137±140. Ogden, UT: Intermountain Research Station, Forest Service, US Department of Agriculture. SENFT R.L., RITTENHOUSE L.R. and WOODMANSEE R.G. (1985) Factors in uencing patterns of cattle grazing behaviour on shortgrass steppe. Journal of Range Management, 38, 82±87. STUTH J.W. (1991) Foraging behaviour. In: Heitschmidt R.K. and Stuth J.W. (eds) Grazing Management, pp. 65±83. Portland, OR: Timber Press. TALLOWIN J.R.B., BROOKMAN S.K.E. and SANTO S.G.L. (1995) Leaf growth and utilization of four grass species under steady state continuous grazing. Journal of Agricultural Science, Cambridge, 124, 403±417. TRICART J.L.F. (1973) Geomorphology of the Flooding Pampas. ColeccioÂn Cientõ ca Instituto Nacional de TecnologõÂa Agropecuaria (INTA), Buenos Aires, XII, 202 pp. VERVOORST F.B. (1967) The vegetation of Argentina. VII. The communities of the Flooding Pampas (Buenos Aires Province). Serie Fitogeogra ca no pp: Instituto Nacional de TecnologõÂa Agrozecuaria (INTA) Buenos Aires. WADE M. and BAKER R.D. (1979) Defoliation in setstocking grazing systems. Grass and Forage Science, 34, 73±74.