Oecologia (1993) 96:179-185 Oecologia �9 Springer-Verlag 1993

Water relations of native and introduced C4 grasses in a neotropical savanna

Zdravko Baruch, Denny S. Ferndndez*

Departamento de Estudios Ambientales, Universidad Sim6n Bolivar, Apartado 89000, Caracas 1080, Venezuela

Received: February 1993 / Accepted: 15 July 1993

Abstract. Introduced African grasses are invading Neo- tropical savannas and displacing the native herbaceous community. This work, which is part of a program to understand the success of the African grasses, specifically investigates whether introduced and native grasses differ in their water relations. The water relations of the na- tive Trachypogon plumosus and the successful invader Hyparrhenia rufa were studied in the field during two consecutive years in the seasonal savannas of Venezuela. The two C4 grasses differed clearly in their responses to water stress. H. rufa consistently had higher stomatal conductance, transpiration rate, leaf water and osmot- ic potential and osmotic adjustment than the native T. plumosus. Also, leaf senescence occurred much earlier during the dry season in H. rufa. Both grasses showed a combination of water stress evasion and tolerance mech- anisms such as stomatal sensitivity to atmospheric or soil water stress, decreased transpiring area and osmotic ad- justment. Evasion mechanisms are more conspicuous in H. rufa whereas T. plumosus is more drought tolerant and uses water more "conservatively". The evasion mechanisms and oportunistic use of water by H. rufa, characteristic of invading species, contribute to, but only partially explain, the success of this grass in the Neo- tropical savannas where it displaces native plants from sites with better water and nutrient status. Conversely, the higher water stress tolerance of T. plumosus is consis- tent with its capacity to resist invasion by alien grasses on shallow soils and sites with poorer nutrient and water status.

Key words: C4 grasses - Hyparrhenia rufa - Neotropical savanna -Trachypogon plumosus - Water relations

* Present address: Department of Biology, University of Puerto Rico, P.O. Box 23360, Rio Piedras, PR 00931

Correspondence to: Z. Baruch

Neotropical savannas cover extensive areas with low primary productivity attributed to low soil fertility, sea- sonal drought and frequent fires (Frost et al. 1985). In Venezuela, the lowland savannas (Llanos) are located mainly in the center of the country and used extensively for cattle grazing. Because most of the native grasses are fibrous and have low nutritive value, African grasses have been introduced in order to improve livestock production (Parsons 1972). Hyparrhenia tufa (Nees.) Stapf. is one of several successfully introduced forage grasses which has displaced native plants in some hab- itats of the savanna (Parsons 1972; Baruch et al. 1985, 1989; San Jos~ and Farifias 1991). This displacement might have far-reaching consequences in the dynamics of the savanna comparable to those documented in other ecosystems (Vitousek 1988; D'Antonio and Vitousek 1992).

Several probable causes for this displacement have been postulated (Baker 1978) but few have been tested in the field. The superiority of African grasses, as com- pared to South American and Australian native savanna species, has been related to the following characteristics of the former: (a) higher net photosynthetic rates (Chris- tie 1975; Baruch et al. 1985); (b) more efficient use of soil nutrients (Christie and Moorby 1975; Baruch and G6mez, unpublished results); (c) higher proportion of assimilates allocated to leaves (Baruch et al. 1989) and (d) higher tolerance to defoliation (Simoes and Baruch 1991).

Considering the importance of water economy in sea- sonally dry savannas (Frost et al. 1985), we propose that H. rufa in the Llanos is also able to cope better with the seasonal drought than the native grasses. The objective of this study is to reveal further differences, in terms of the water relations, between the dominant native grass Trachypogon plumosus (Humb. & Bonpl.) Nees. and the introduced H. rufa which might help to explain the higher competitive capacity of the latter in some Neo- tropical savannas.

180

Methods

Site and species studied

Field work was performed at the Estacidn Bioldgica de los Llanos (8 ~ 58' N and 67 ~ 25' W), 80 m above sea level, 10 km south of the city of Calabozo (Gu~irico State, Venezuela). This area, protected from fire and grazing, is representative of the well - drained sea- sonal savannas (Ramia 1967) with scattered trees among the her- baceous matrix dominated by C4 bunch grasses (0.~1 m tall). Rainfall is strongly seasonal with 4-6 dry months and averages 1410 mm per year. Mean annual air temperature is 28 ~ C with daily ranges between 10 ~ and 14 ~ C.

The soils of the study area are ultisols or oxisols originating from alluvial deposits of the Mesa formation. In some areas, soils are poorly developed because of the presence of a shallow lithoplin- thic stratum of hardened iron sesquioxides called "arrecife" (L6pez et al. 1971) whose depth influences local vegetation patterns (Velfis- quez 1968). In general, the soils are sandy, acidic, poor in nutrients (especially nitrogen and phosphorus), low in organic matter and high in exchangeable aluminum (Medina 1982; Garcia and C/tceres 1990). The graminoid genera Trachypogon, Axonopus, Andropogon, Aristida and Bulbostylis are dominant in these savannas. The in- vasion of H. tufa took place between 1969 and 1977 (San Jos6 and Farifias 1991).

The species studied are perennial C4 bunch grasses: T. plumosus is about 1 m high, reproduces at the middle of the rainy season (August-October) and it is dominant in the Calabozo area as in other savannas of the Venezuelan and Colombian Llanos. Hypar- rhenia rufa grows in large bunches up to 3 m high, it flowers at the end of the rainy season (November-December) and forms distinc- tive stands with very low diversity only in relatively deep soils. The annual growth cycle of both grasses starts with the beginning of the rains in May, peak biomass is attained by August-September and then it decreases to a minimum at the end of the dry season in March-April. The leaves of H. tufa dry out completely during the dry season (February-March) whereas those of T. plumosus senesce gradually and the youngest remain alive throughout the whole dry season (personnal observation).

Measurements

This study encompassed two dry and rainy seasons between April 1985 and June 1986. The climatic variables were monitored at a meteorological station 300 m from the sampling site. Unless other- wise stated, all measurements were performed on individuals grow- ing on deep soils (> 1 m) of the "normal" (without arrecife) savanna (Vellsquez 1968). Plant water relations were assessed on the youn- gest fully expanded leaf of five randomly selected individuals per species. Before dawn, and then every 2 h, leaf water potential (q~) was measured with a pressure chamber (Scholander et al. 1965), water loss was minimized by enclosing the leaf in a plastic bag before cutting and during the measurement. Leaf conductance (Gs), transpiration rate (E) and leaf temperature (T1) were measured every 2 h on the abaxial surface of leaves with a steady state diffusion porometer (LI-COR 1600). Occasionally, dew prevented the first measurement. Three times during the day (pre-dawn, mid- day and afternoon) a segment of the leaves used in water potential measurement was preserved in liquid nitrogen for osmotic potential (17) determinations via dew point hygrometry in C-52 WESCOR chambers. From the values of q~ and II, leaf turgor potential (P) was calculated as P = ~P - 17. Soil water content at a depth of 20 cm was determined gravimetrically in five samples each under 7". plu- mosus and H. rufa. Soil matric potential was calculated using water retention curves determined for similar soils (Cficeres 1983). During each sampling, leaf weight per area (LWA) was measured on 10 leaves per species. Pressure - volume curves (Tyree and Hammel 1972) were determined in three to four leaves per species collected before dawn and rehydrated for 2-3 h. Osmotic potential at maxim-

um leaf turgor (IIloo), leaf modulus of elasticity (e), percentage of apoplastic water (A) and the relation between leaf turgid and dry weight (TW/DW) were determined as in Wilson et al. (1979).

When appropiate, statistical significance of the difference be- tween means was established with the t-test. Statistical significance was assumed whenever P < 0.05.

Results

The results are shown as the mean of the values obtained during each measuring period throughout the year and as hourly means of typical days from the rainy season (19 October 1985), transition (15 December 1985) and dry season (23 February 1986 for H. rufa and 15 March 1986 for T. plumosus). Different sampling dates were needed during the dry season because of the early leaf senescence in H. rufa.

During 1985, rainfall was somewhat higher than the average of the 5 preceding years (1462 vs 1410 ram) (Fig. 1). January, February and March were rainless in 1985 and 1986. There was a small rainfall event (6.5 ram) on 11 April 1986 before the scheduled field measurement with the consequences discussed below. Pan evaporation was approximately twice rainfall (Fig. 1) which, accord- ing to the Thornthwaite water balance, results in only four wet months per year. Maximum air temperature at the nearby weather station was 38 ~ C, but near the grass stratum it was near to 40 ~ C. Air humidity follows the rainfall cycle with minimum relative humidity and vapor pressure saturation deficits of 15 % and 6 kPa respective- ly. The leaf-air water vapor pressure difference also in- creased markedly in both grasses during the dry season up to 5 kPa which represents a high evaporative demand.

The water retention capacity of the soils is low, with the agronomic wilting percentage ( -1 .5 MPa) corre- sponding to 7.5-10.5% of soil water content (Cficeres 1983). Soil water content, at 20 cm, decreased from Sep- tember 1985 onwards and soil water potential remained below - 1.5 MPa during the dry season (Fig. 2). The low soil water content during April 1985 (2.0%) corresponds to soils in the "arrecife" savanna with very low water

E

z

3 0 0

200

100

ir "~ j

", e ',, r - ' - - , a /

\ i ' , , / _ / i v

F M A M d d A S 0 N D d F M A M J

1985 t986

300 s

<

2 0 0 0_

lal

IOO

Fig. 1. Monthly rainfall (solid circles) and pan evaporation (open circles) throughout the period of field work. Data from Ministerio del Ambiente, Venezuela

1 8 ~ - T H.rufa

16

14 "[plumosus

g; ,a I" ' ~ 8

__a ~ 6

A M a J A S 0 N D J F M A M J

1 9 8 5 t 9 8 6

Fig. 2. Soil water content expressed as percentage of fresh weight under a Hyparrhenia rufa stand (continuous line) and under the native grass Trachypogon plumosus (dashed line). The horizontal line corresponds to soil water content at the agronomic wilting point ( - 1.5 MPa) = 7.5%. Vertical bars indicate 1 SD. Sampling in April 1985 was done on a different soil type (arrecife)

w

z

o

o

1.4

1,2

1.0

0 . 8

0,6

0 . 2 /

1 I I I l I I I I I I F I I I A M J J A S 0 N D J F M A M J

1985 1 9 8 6

Fig. 3. Maximum stomatal conductance during field measurements. H. rufa, solid triangles," T. plumosus, open circles. (During March and April 1986, the leaves of H. rufa were senescent). Vertical bars indicate 1 SD

retention capacity. Only during May and June 1985 did the soil under H. rufa have soil water content higher than that of native savanna.

Maximum leaf temperature was higher during the dry season but generally lower for H. rufa (36.8+0.6 ~ C) than for T. plumosus (41.9 4- 1.9 ~ C). The maximum daily G~ measured throughout the year had a similar pattern in both grasses (Fig. 3). The peak values ocurred during

181

g

8

08I"'~176 * =0

C I I I I I I I 1 l I I

04 ~ T . , , , u . , o . ~ ~ ~----al ~ ~

o I I I i I I i I I I I 7 9 I I t 5 15 17

L O C A L T I M E

Fig. 4. Stomatal conductance during selected field work days. The symbols are: rainy season (R): 19 October 1985 (solid circles); transition (T): 15 December 1985 (open circles); dry season (D): 23 February 1986 for H. rufa; 15 March 1986 for T. plumosus (solid triangles). Vertical bars indicate 1 SD

T / / t ~ T

L I L I I i t I I I I

I i i i I i

L O C A L T I M E

Fig. 5. Transpiration rate during selected field work days. Symbols as in Fig. 4. Vertical bars indicate 1 SD

the middle of the rainy season when the Gs of H. rufa was 30% higher than that of T. plumosus. However, the dry season decrease in the maximum daily Gs was around 80 % for the former but only 50 % for the latter. During the rainy season and the transition to the dry season, the daily maxima of Gs occurred around noon in both grasses (Fig. 4). During the dry season, however, the maximum was somewhat displaced towards mornings in T. plu- mosus whereas H. rufa had lower and relatively constant Gs throughout the day. Maximum E was 20-40% higher in H. rufa during the rainy season and the transition to the dry season but similar to that of T. plumosus during the dry season. In both grasses, the daily course of E had a characteristic maximum around noon which decreased during the dry season (Fig. 5).

In both species, the daily maximum and minimum values of leaf �9 decreased gradually as drought pro- gressed (Fig. 6). The decline of predawn �9 was slower than that of midday �9 and the native grass had lower leaf

during the dry season. The low �9 measured in T. plu- mosus during April 1985, as compared to April 1986, (Fig. 6) could have been caused by: (i) the rainfall event

182

-0 .5

U,l

~" - 2 5 _u

,,r -1 5

- 2 5

5

. . . . ~ ~p

r im

H . r u f o

I I [ I I I I I I I I I I i I

i i i i I I

1 9 8 5 1 9 8 6

Fig. 6. Predawn (xCp) (solid circles), and midday (~m) (open circles) leaf water potential and psychrometric midday osmotic potential ( ~ ) throughout the field work. Vertical bars indicate 1 SD. During April 1985 measurements were performed on plants growing on a different soil type (arrecife)

-1 .O - - - ' ( 2 T D

o - 2 . 0

- 3 ,0 i

,~ ~ R

-Z.O

'-~ T. p l u m o = u $

I I I i I i J i I i I I - 3 ,0 7 9 I I 13 15 I T

LOCAL TIME

Fig. 7. Leaf water potential during selected field work days. Symbols as in Fig. 4. Vertical bars indicate 1 SD

E o

z

1,2 "-' i i i i i

O T. plumo~s t.O \ \ O �9 �9 H. rufa

0 . 8 �9 I ~ \ \ O ~

0.6 ~

o 0.4

0.2

~ o~ | I [ ~I I

0"%~0 -0.5 -1.0 - 1.5 -2.0 -2.5 -3.0

WATER POTENTIAL (MPo)

Fig. 8. Relationship between stomatal conductance and leaf water potential. Measurements taken between 10.00 a.m. and 2.00 p.m. Each point represents the average of five measurements. H. rufa, full circles and broken line; T. plumosus, open circles and continuous line. Quadratic linear regressions were fitted to the data: T. plumosus, y=O.96-0 .53x +0.07x 2 (R=0.73); H. rufa, y = 1 .19-0 .38x-0 .11x z (R=0.84)

just before measurements in 1986 and (ii) the former were performed on individuals growing on soils of the arrecife savanna which had a lower water content (Fig. 2). The daily courses of leaf ~ typically decreased towards mid- day and stabilized or slightly increased in the afternoon (Fig. 7). The monthly and hourly averages of leaf q? were always lower in T. plumosus. Hyparrhenia rufa showed higher stomatal sensitivity to leaf �9 as the lowest G~ measured occurred near -2 .0 MPa in contrast to the native grass when the lowest G~ ocurred between -2 .5 and -3 .0 MPa (Fig. 8). Also, the decrease of Gs with increased qJ was steeper in H. rufa (Fig. 8).

The decrease of midday 17 started in December 1985 (Fig. 6) indicating osmotic adjustment (as confirmed by the pressure-volume technique). Although the differences

Table 1. Comparison of the means of the parameters obtained from the analysis of the pressure-vokune curves and teaf weight per area (LWA) between the rainy season (May and July 1985) and the dry

season (March and April 1986 for T. plumosus," January and Feb- ruary 1986 for H. tufa)

H. rufa

Parameter May July 1985 Jan-Feb 1986 Difference

II~o o (MPa) (n= 10) -0.97_+0.13 A (%) (n=4) 16.40___4.1 TW/DW (n= 10) 3.18 +0.46

(MPa) (n=3) 3.70_+0.52 LWA (rag- cm -z) (n= 15) 4.32_+0.90

(n=6) --1.79_+0.19 0.82"** (n=2) 15.50_+4.0 0.90 ns (n=6) 2.18_+0.20 1.00 *** (n=5) 8.14+2.09 4.44 ** (n=20) 13.25+1.94 8.93 ***

7". plumosus

Parameter May-July 1985 Mar-Apr 1986 Difference

Fifo o (MPa) (n = 6) -- 1.55 +_0.06 (n = 8) -- 1.83 +_0.18 0.28 ** A (%) (n=6) 32.11-t-_9.45 (n=8) 30.00_-t-6.93 2.11 ns TW/DW (n=6) 2.25_+0.21 (n--8) 2.17_+0.06 0.08 ns

(MPa) (n=4) 5.77_+3.71 (n=6) 10.93__+3.72 5.16 * L W A ( m g . c m -2) (n=15) 6.50_+1.95 (n -19) 11.21_+2.83 4.71 **

(ns=not significant; * 0 .05>P>0.01 ; ** 0.01 >P>0.001 ; *** P<0.001)

183

were not statistically significant, the seasonal decrease of H was larger in T. plumosus. Leaf turgor remained pos- itive throughout the year except in T. plumosus during April 1985 and midday leaf turgor decreased gradually as the drought advanced in both species (Fig. 6). Osmotic potential at full turgor (II,00) , as obtained from the pres- sure-volume curve analysis, was significantly lower in T. plumosus during the rainy season (Table 1). During the dry season, II,oo of both grasses decreased significantly but much more in H. rufa than in T. ptumosus (Table 1). The largest change in IIloo is the maximum osmotic adjustment (AII~oo) and it was higher in H. rufa whereas the value for the native grass was only 34% of this (Ta- ble 1). It is possible to quantify the role of leaf structural changes (increase in dry weight and decrease in leaf water; changes in TW/DW) and that of the active solute accumulation to total AH~oo (Ludlow et al. 1983). Be- cause the variation in water content per unit of dry weight (TW/DW) was significantly higher in H. rufa than in T. plumosus (Table 1), the contribution of structural changes to AI-Iloo was much higher in the former (70% of total AIIloo vs 19% in T. plumosus). The remaining AHloo reflects active solute accumulation which was then higher in T. plumosus (81% of total AT~oo vs 30% in H. tufa).

In both grasses LWA increased significantly as the dry season progressed (Table 1). During the rainy season H. rufa showed the lowest LWA but its seasonal increase was also proportionately larger. Apoplastic water con- tent (A) in T. plumosus was double that in H. rufa but there were no seasonal differences in either species (Ta- ble 1). The leaf bulk modulus of elasticity (e) had values characteristic of herbaceous plants, with a significant increase during the dry season in both grasses (Table 1).

Discussion

The two grasses differ clearly in their water relations throughout the year. Hyparrhenia rufa consistently showed higher Gs, E, ~ , II and osmotic adjustment than T. plumosus. Also, the rapid leaf senescence early in the dry season in H. rufa contrasted markedly with the per- manence of living leaves in the native grass.

The G~ reported here was in the range of non-crop grasses (Pavlik 1984; Knapp 1985) and the higher Gs of H. rufa confirms previous laboratory results (Baruch et at. 1985; Simoes and Baruch 1991). In addition, stomata of this grass appeared to be more sensitive to water stress because they closed at higher leaf ~ than those of T. plumosus. Low G~ in T. plumosus, even during the rainy season, acts as a water conservation mechanism and contributes to its relatively low carbon assimilation rate (Baruch et al. 1985; Simoes and Baruch 1991). Also, the presence of silica bodies around the stomatal pore in the native grass (Fonseca 1983) might contribute to its low Gs through decreased stomatal aperture. It is likely that the stomatal closure under positive leaf turgor ob- served in both grasses was promoted by direct response to either low air humidity or low soil matric potential (Baruch et al. 1985; Gollan et al. 1986; Passioura 1986; Schulze 1986; Turner 1986).

The decrease in E during the dry season was more prominent in H. rufa than in T. plumosus. When water supply was not limiting, E increased in phase with air evaporative demands. As the dry season advanced, there was a gradual decrease of E that ended with leaf senes- cence in H. rufa and with low E in the few leaves remain- ing alive in T. plumosus. However, the gradual decrease of E was not found in a previous study (San Jos6 and Medina 1975) probably due to the low sensitivity of the method employed. The higher water demands of H. rufa, even at the beginning of the dry season, are probably sustained by its profound root system (Daubenmire 1972; Baruch, unpublished results) and might explain its presence only on deep soils which have more water avail- able.

The minimum leaf ~s measured throughout this study were higher than those reported for grasses from tem- perate grasslands (Redmann 1976; Sala et al. 1981; Knapp 1985) and from subtropical savannas (Peake et al. 1975; Wilson et al. 1980). Leaf ~t' remained high in H. rufa until the soils dried at the beginning of the dry season when its leaves wilted. In T. ptumosus relatively high ~gs could be attributed to low G~ and E and the gradual decrease of functional leaf area as the dry season advanced. At the peak of the dry season, only the youn- gest leaves of the plant remained alive, reducing drasti- cally total plant E, which might help to maintain some leaf turgor and carbon gain in the native grass during the dry season.

Leaf H was significantly lower in T. plumosus than in H. rufa and other grasses (Morgan 1984). Low leaf I-I is important in seasonally dry environments because it fa- vors water absorption and allows the maintenance of turgot at low leaf ~ . The seasonal decrease of II~oo was evident in both species and constitutes the osmotic ad- justment which is important as a mechanism of tolerance to water stress (Ludlow 1980; Ludlow et al. 1983; Santa- maria et al. 1990; Munns 1988). The AIIloo obtained here (0.3-0.8 MPa) are in the range of values reported for tropical grasses (Wilson et al. 1980; Baruch, Ludlow and Wilson, unpublished data), for temperate grasses (Pavlik 1984; Knapp 1985) and for numerous crops (Hsiao et al. 1976; Henson 1982; Morgan 1984; Turner et al. 1986).

The large osmotic adjustment in H. rufa was a conse- quence of leaf structural changes probably caused by the increase of the proportion of cell wall material or the production of smaller cells in the leaves produced under water stress at the beginning of the dry season (Cutler et al. 1977; Wilson et al. 1980). In T. plumosus TW/DW remained relatively low throughout the year, which is characteristic of plants from dry habitats (Wilson et al. 1980) and AIIloo was mostly accounted for by active solute accumulation, probably a mixture of organic ac- ids, aminoacids, simple sugars and potassium ions as in other tropical grasses (Ford and Wilson 1981).

As the velocity of leaf u/ decrease was higher than AIIloo increase, it is doubtful that osmotic adjustment alone was responsible for leaf turgot maintenance and survival in T. plumosus during the drought. In H. rufa AII,oo is probably not very important because its leaves die early in the dry season.

184

Bulk leaf cell elasticity (~) and the amount of apoplas- tic water (A) were in the range reported for other grasses (Wilson et al. 1980; Pavlik 1984) and both decreased during the dry season. These changes, as well as those in LWA and TW/DW were probably consequence of changes in cell wall thickness due to lignin and pectin deposition (Baruch et al. 1985).

Both grasses showed a combination of evasion and tolerance mechanisms to deal with water stress (Kramer 1980; Ludlow 1980; Turner 1986). Stomatal sensitivity to atmospheric and soil water stress and decreased tran- spiring leaf area were found in both grasses, but these responses were more prominent in H. rufa. Deep rooting in H. rufa constitutes another evasive response. Both grasses also had certain degree of drought tolerance evidenced by osmotic adjustment and changes in ~. The most sensitive species to drought was H. rufa which relies mainly on evasion mechanisms for coping with it. The native grass is more drought-tolerant and its low Gs and E even during the rainy season indicate "conservative" water use. Also, high Gs and E in H. rufa, even at the beginning o f the dry season, and a pronounced leaf senescence as water stress increased indicate rather "op- portunistic" water use which also characterizes other invading species (Bazzaz 1986). These results and other on carbon assimilation (Baruch et al. 1985; Simoes and Baruch 1991) indicate that H. rufa should have a higher water use efficiency during the rainy season whereas the native grasses should be more efficient during the dry season.

These contrasting responses to drought permit us to postulate different survival strategies. In the drought- tolerant T. plumosus, the remaining live leaves would allow rapid regrowth at the begining of the rains. How- ever, the significance of this strategy would probably depend on the severity of the dry season. During 1985 rainfall was somewhat higher than average, which might explain the relatively high values of �9 measured during the next dry season. In contrast to the native grass, the leaves of H. rufa wilted completely early in the dry sea- son, showing high sensitivity to even moderate water stress and an effective drought evading mechanism. Car- bon assimilation in H. rufa also stops at significantly higher tp than in T. plumosus (Baruch et al. 1985). The evasion strategy of H. rufa seems to be more effective than the tolerance strategy of the native grasses because leaf loss in the former is compensated by higher carbon assimilation and growth rates under non-limiting water conditions of the beginning of the rainy season (Baruch et al. 1985; Simoes and Baruch 1991). However, the high water demand of the invader grass restricts its distribu- tion to relatively deep soils.

The success of H. rufa cannot be attributed exclusively to drought evasion mechanisms and its opportunistic use of water. It shows other characteristics of successful invading species (Bazzaz 1986) such as high photosyn- thetic rates under optimum conditions (Baruch et al. 1985), a prevalent allocation of carbon to leaf produc- tion, (Simoes and Baruch 1991), high seed germinability (Baruch, unpublished data) and compensatory growth after defoliation (Simoes and Baruch 1991) all of which

contribute to the success of H. rufa in the Neotropical savannas.

Acknowledgements. We wish to thank Decanato de Investigaciones of Universidad Sim6n Bolivar for providing funds for this work. Sociedad Venezolana de Ciencias Naturales gave permission to work in the Estaci6n Bioldgica of Calabozo and offered living and research-facilities. We thank Dr. Antonio Vivas for the use of a four wheel drive vehicle. Jos6 G6mez Milton Simoes and Teobaldo P6rez competently helped during the field work.

Thanks are also due to two anonymous reviewers whose com- ments improved the original manuscript.

References

Baker HG (1978) Invasion and replacement in Californian and Neotropical grasslands. In: Wilson JR (ed) Plant relations in pastures. CSIRO, Melbourne, pp 368 384

Baruch Z, Ludlow MM, Davis R (1985) Photosynthetic responses of native and introduced C4 grasses from Venezuelan savannas. Oecologia 67:288-293

Baruch Z, Hernfindez A B, Montilla MG (1989) DinAmica del creci- miento, fenologia y repartici6n de biomasa en gramineas nativas e introducidas de una sabana Neotropical. Ecotropicos 2:1-13

Bazzaz FA (1986) Life history of colonizing plants: some demo- graphic, genetic and physiological features. In: Mooney HA, Drake JA (eds) Ecology of biological invasions of North Ameri- ca and Hawaii (Ecological Studies 58). Springer, New York, pp 96-110

Cficeres A (1983) Caracteristicas fisicas y quimicas del suelo y contenido de nutrientes en la vegetaci6n en un transecto mata- pastizal de una sabana de Trachypogon. Thesis, Universidad Central de Venezuela, Caracas

Christie EK (1975) Physiological responses of semi-arid grasses. IV. Photosynthetic rates of Thyridolepis mitcheliana and Cenchrus ciliaris leaves. Aust J Agric Res 26:459-466

Christie EK, Moorby J (1975) Physiological responses of semi-arid grasses. I. - The influence of phosphorus supply on growth and phosphorus absorption. Aust J Agric Res 26:423-436

Cutler JM, Rains DW, Loomis RS (1977) The importance of cell size in the water relations of pIants. PhysioI Plant 40:255-260

D'Antonio CM, Vitousek PM (1992) Biological invasions by exotic grasses, the grass/fire cycle, and global change. Annu Rev Ecol Syst 23 : 63-87

Daubenmire R (1972) Ecology ofHyparrhenia rufa Nees. in derived savannas in northeastern Costa Rica. J Appl Ecol 9 : 11-23

Fonseca AS (1983) Identiflcaci6n de los cuerpos silicosos de las gramineas de la sabana de Trachypogon de los Llanos Altos Centrales. Bol Soc Venez Cienc Nat 141 : 17-83

Ford CW, Wilson JR (1981) Changes in levels of solutes during osmotic adjustment to water stress in leaves of four tropical pasture species. Aust J Plant Physiol 8:77 91

Frost F, Medina E, Menaut JC, Solbrig O, Swift M, Walker B (1985) Responses of savannas to stress and disturbance. Interna- tional Union of Biological Sciences, Special Issue I0. Paris

Garcla M J, Cficeres A (1990) Soil chemistry changes in a forest- grassland vegetation gradient within a fire and grazing protected savanna from the Orinoco Llanos, Venezuela. Acta Oecol 11 : 775-781

Gollan T, Passioura JB, Munns R (1986) Soil water status affects the stomatal conductance of fully turgid wheat and sunflower leaves. Aust J Plant Physiol 13: 459-464

Henson IE (1982) Osmotic adjustment to water stress in pearl millet (Pennisetum americanum (L.) Leeke) in a controlled environ- ment. J Exp Bot 33:78-87

Hsiao TC, Acevedo E, Ferreres E, Henderson DW (1976) Water stress, growth, and osmotic adjustment. Philos Trans R Soc Ser B 273 : 479-500

185

Knapp AK (1985) Water relations and growth of three grasses during wet and drought years in a tallgrass prairie. Oecologia 65: 35-43

Kramer PJ (1980) Drought, stress and the origin of adaptations. In: Turner NC, Kramer FJ (eds) Adaptation of plants to water and high temperature stress. Wiley-Interscience, New York, pp 7-20

Ldpez D, Roa F, Ramirez I (1971) Estudios en un sedimento ferruginoso llamado localmente "ripio". Bol Soc Venez Cienc Nat 119/120:27-49

Ludlow MM (1980) Stress physiology of tropical pasture plants. Trop Grassl 14:136-145

Ludlow MM, Chu ACP, Clements RJ, Kerslake RG (1983) Ad- aptation of species of Centrosema to water stress. Aust J Plant Physiol 10:119-130

Medina E (1982) Physiological ecology of neotropical savanna plants. In: Huntley B J, Walker BH (eds) Ecology of tropical savannas (Ecological Studies 42). Springer-Verlag, Berlin, pp 303-335

Morgan JM (1984) Osmoregulation and water stress in higher plants. Annu Rev Plant Physiol 35:299-348

Munns R (1988) Why measure osmotic adjustment? Aust J Plant Physiol 15:717-726

Parsons JJ (1972) Spread of African grasses to the American tropics. J Range Manage 25:12-17

Passioura JB (1986) Resistance to drought and salinity: avenues for improvement. Aust J Plant Physiol 13:191-201

Pavlik BM (1984) Seasonal changes of osmotic pressure, symplastic water content and tissue elasticity in the blades of dune grasses growing in situ along the coast of Oregon. Plant Cell Environ 7: 531-539

Peake DCI, Stirk GD, Henzell ET (1975) Leaf water potentials of pasture plants in a semi-arid subtropical environment. Aust J Exp Agric 15 : 645-654

Ramia M (1967) Tipos de sabanas en los Llanos de Venezuela. Bol Soc Venez Cienc Nat 112:264-288

Redmann RE (1976) Plant water relationships in a mixed grassland. Oecologia 23 : 283-295

Sala OE, Lauenroth WK, Parton WJ, Trlica MJ (1981) Water status of soil and vegetation in a shortgrass steppe. Oecologia 48 : 327-331

San Jos6 JJ, Farifias MR (1991) Temporal changes in the structure of a Trachypogon savanna protected for 25 years. Acta Oecol 12:237 247

San Jos6 JJ, Medina E (1975) Effect of fire on organic matter production and water balance in a tropical savanna. In: Golley FB, Medina E (eds) Tropical ecological systems (Ecological Studies 11). Springer-Verlag, New York, pp 251-264

Santamaria JM, Ludlow MM, Fukai S (1990) Contribution of osmotic adjustment to grain yield in Sorghum bicolor (L.) Moench under water limited conditions. I. Water stress before anthesis. Aust J Agric Res 41 : 51-65

Scholander PFL, Hammel HI, Broadstreet ED, Hemmingsen EA (1965) Sap pressure in vascular plants. Science 148:339-346

Schulze ED (1986) Whole plant responses to drought. Aust J Plant Physiol 13: 127-142

Simoes M, Baruch Z (1991) Responses to simulated herbivory and water stress in two tropical C4 grasses. Oecologia 88:173-180

Turner NC (1986) Adaptation to water deficits: a changing perspec- tive. Aust J Plant Physiol 13:175-190

Turner NC, O'Toole JC, Cruz RT, Yambao EB, Ahmad S, Namu- co OS, Dingkuhn M (1986) Respons e of seven diverse rice cultivars to water deficit. II. Osmotic adjustment, leaf elasticity, leaf extention, leaf death, stomatal conductance, and photosyn- thesis. Fields Crop Res 13: 273-286

Tyree MT, Hammel HT (1972) The measurement of the turgor pressure and water relations of plants by the pressure bomb technique. J Exp Bot 23:267-282

Vel~squez J (1968) Estudio fitosociol6gico acerca de los pastizales de las sabanas de Calabozo, Edo. Gu/tfico. Bol Soc Venez Cienc Nat 109:59-101

Vitousek FM (1986) Biological invasions and ecosystem properties: can species make a difference ? In: Mooney HA, Drake JA (eds) Ecology of biological invasions of North America and Ha- waii (Ecological Studies 58). Springer-Verlag, New York, pp 163-178

Wilson JR, Fisher M J, Schulze ED, Dolby GR, Ludlow MM (1979) Comparison between pressure-volume dew point hygrometry for determining the water relations characteristics of grass and legume leaves. Oecologia 41 : 77-88

Wilson JR, Ludlow MM, Fisher MJ, Schulze ED (1980) Adapta- tion to water stress of the leaf water relations of four tropical forage species. Aust J Plant Physiol 7:207-220