Photosynthesis and water use efficiency of the association between Larrea tridentata ( DC ) Cov . and

Recibido/Received 26.III.2006. Aceptado/Accepted 5.V.2008. Abstract. The study was conducted in the Chihuahuan Desert Rangeland Research Center, 32 km north of Las Cruces, NM, USA. Physiological responses of the shrub Larrea tridentata (DC) Cov. and the perennial grass Muhlenbergia porteri Scribn were determined from the summer 1997 to the winter 1998-99 when they grew together. In the summer 1997, Larrea tridentata competed for water with M. porteri, as shrubs growing alone showed higher integrated photosynthesis than shrubs in competition with M. porteri following a heavy rainfall event. Larrea tridentata had lower water use efficiency (WUE) than M. porteri plants. In the summer 1998, M. porteri had greater pre-dawn xylem water potentials (ψL) than L. tridentata when soil water potentials were above -3 MPa. Larrea tridentata showed a higher photosynthetic activity when M. porteri was defoliated than when it remained undefoliated. Muhlenbergia porteri presented a higher WUE. While green in the fall, plants of M. porteri had higher ψL and WUE than those of L. tridentata. Net assimilation was similar between shrub and grass plants. Both species responded to soil moisture availability in this season. Muhlenbergia porteri had higher net assimilation when growing alone than when growing under creosotebush. In the winter season, and while 57o ANIVERSARIO 57th ANNIVERSARY (2008) 77: 297-320 Revista Internacional de BOTANICA EXPERIMENTAL


INTRODUCTION
In the Chihuahuan Desert, a common plant association occurs between the shrub Larrea tridentata (DC) Cov. and the perennial grass Muhlenbergia porteri Scribn.Larrea tridentata is a dominant perennial evergreen shrub in the deserts of North America (Runyon, 1934;Reynolds, 1986).This plant has a stress-tolerant strategy with distinct growth patterns which vary with soil depth.Longevity of its leaves in the Chihuahuan desert is from less than one year to two years.This species is tolerant to extreme temperatures and desiccation: photosynthesis and respiration strongly Castellanos-Pérez E et al., ΦYTON 77 (2008) acclimate to seasonal changes in temperature (Reynolds, 1986).This shrub is metabolically active throughout the year (Oechel et al., 1972;Bamberg et al., 1975).Ecophysiological studies are necessary on this shrub (1) for trying to explain variations in its growth and (2) to understand the relative importance of abiotic factors on its productivity (Reynolds, 1986).Since lack of water has been a major selective force for plant evolution in these ecosystems, it is important to learn about plant strategies which may contribute to cope with water stress (Freitas, 1997).Muhlenbergia porteri is a highly-branched perennial grass which occupies dry plateaus, canyons and rocky deserts in the southwestern United States and northern Mexico (Tidestrom & Kitell, 1941;Welsh & Beck, 1976).This grass species is often growing under the canopy of L. tridentata.Muhlenbergia porteri is one of the grasses that appear to have possibilities for either compete or coexist with shrubs (Pieper, 1994).This shrub-grass association is unusual since creosotebush more often appears to displace grasses from their historical distribution in these warm desert ecosystems.
The C 3 L. tridentata is an evergreen shrub with metabolic activity throughout the year (Giorgetti et al., 2000), and the C 4 M. porteri is a perennial grass dormant during winter.Both species grow successfully under arid conditions.Thereafter, the question is how these plants deal with C fixation and loss of water by transpiration to remain under these conditions.The objective of this study was to observe the effect of large and small plants of L. tridentata on the gas-exchange activity of M. porteri, and vice versa.Several hypotheses related to the success of the L. tridentata-M.porteri association were considered: (a) the resource islands under L. tridentata provide a better microhabitat for M. porteri which will be reflected in its greater photosynthetic activity when compared with M. porteri growing alone, (b) M. porteri and L. tridentata are able to coexist because they stratify periods of greatest physiological activity: creosotebush is physiologically active during dry periods, while M. porteri tends to be more physiologically active during periods of adequate soil water availability, (c) M. porteri has greater photosynthetic rates when growing under small than large plants of L. tridentata, (d) removal of senescent plant material from plants of M. porteri will result in greater photosynthetic rates due to greater light levels available for photosynthesis.

MATERIALS AND METHODS
The study site is on a northwest-facing piedmont slope of Mt.Summerford in the Chihuahuan Desert Rangeland Research Center (CDRRC), approximately 32 km north of Las Cruces, NM.Elevation is 1327 m.a.s.l.The vegetation classification corresponds to a desert shrub formation, Chihuahuan desert region, creosote series, creosote-bush muhly association (Donart et al., 1978).The vegetation at this site is dominated by L. tridentata with some dispersed individuals of honey mesquite (Prosopis glandulosa Torr.var.glandulosa) and tarbush (Flourensia cernua DC).The principal grass is M. porteri with some fluffgrass (Dasychloa pulchella (Kunth) Steudel) and spike dropseed (Sporobolus contractus A.S. Hitch.).
The soil association corresponds to Haplargids-Torripsamments (Maker et al., 1974).The soil at this site is loamy fine sand about 10 cm thick in the surface layer; the subsoil is sandy loam, and sandy clay loam about 78 cm soil depth.The substratum is sandy clay loam to a depth of 150 cm or deeper (Bulloch & Neher, 1980).Since 1954, people at the site were in the search for grass species which would need to respond to favorable environmental conditions.Finally, M. porteri became established on the site in 1960.In 1965, the site was fenced to prevent grazing by large domestic herbivores (Welsh, 1972).The coolest months are December and January with mean monthly temperatures of 5.5 and 5.3 °C, and minimum temperatures of 3.2 and -3.5 °C, respectively.The warmest months are June, July, and August with mean monthly temperatures of 24.7, 26.4, and 25.3 °C, and maximum temperatures of 34.4, 34.3, and 33.1 °C, respectively (Kunkel et al., 1988).From 1997 to 1999, seasons had an average of 21.1, 14.0, 14.7, and 25.1 MJ per day of solar radiation for summer, fall, winter and spring, respectively.Maximum relative humidity was 71.4, 57.5, 60.1, and 40.8%, and minimum relative humidity was 21.3, 20.5, 23.1, and 13.3% during summer, fall, winter and spring, respectively (1997, 1998, 1999 weather data, http://jornada-www.nmsu.edu/datacat.php?withJS=true).Rainfall data are shown in Fig. 1; these data were obtained using a raingauge located at the Headquarters of the CDRRC, a mile away from the study area.Long-term rainfall data were obtained from Kunkel et al. (1988).
In July 1997, creosotebush plants with or without bush muhly were tagged in the field.On each shrub, the (1) canopy height, (2) largest diameter, and (3) diameter perpendicular to the largest diameter of the canopy were measured.Canopy volume of each shrub was calculated using the prediction equation of Ludwig et al. (1975): where V c is canopy volume, D 1 is the largest canopy diameter, D 2 is the diameter perpendicular to D 1 , and h is height.Shrubs were stratified based on the calculated volume: Small plants = between 25.4 to 122.9 dm 3 ; medium plants = 204.8 to 409.7 dm 3 , and large plants = >491.6dm 3 .A total of Castellanos-Pérez E et al., ΦYTON 77 (2008) 64 plant units were chosen; each unit was a plant of creosotebush and a plant of bush muhly growing together, or each species growing alone.Bush muhly canopy volumes were similar, 593.8 ± 50.2 dm 3 ; this was calculated using the upper-half prolate spheroid shape (Ludwig et al., 1975): where V c is canopy volume, D 1 is the largest canopy diameter, D 2 is the perpendicular diameter to D 1 , and h is height.
On 18 July 1997, the first measurement of gas exchange was carried out in half of the plots; however, more than 1.5 hour was spent measuring them.Due to the data variation we encounter investing more than one hour in measurements, we decided to discard the medium-size creosotebush plants from the study.Plants selected for measurements are in Table 1.In order to measure the treatments units in less than one hour, half of these units were measured on one date, and the other half on the day after.Each plant unit was around 1 m in area, and contained one shrub and/or grass clump.Defoliation of 50% of M. porteri biomass was done according to the equations of Almeida (1981), where the dependent variable was biomass and the independent variable was height.M. porteri was defoliated on 15 July 1997.P Pl la an nt t a an nd d S So oi il l W Wa at te er r P Po ot te en nt ti ia al ls s. .Pre-dawn water potentials (ψ L ) were measured on plants using a 3000 series Plant Water Status Console (Soilmoisture Equipment Corp., Santa Barbara, CA) (Scholander et al., 1965).For creosotebush, plant water potential and gas-exchange were taken on the same days.Muhlenbergia porteri xylem water potential was determined during the summer 1998.Prior to this date, we did not have the appropriate rubber gasket to hold the bush muhly in the pressure bomb.
Soil water potential (ψ s ) was measured with soil psychrometers, calibrated in the laboratory and installed at 0.2 m depth in November 1997 under the following plants units (treatments): 2 in SL, 2 in M, 2 in LL, 4 in SLMD, 6 in LLM, and 6 in SLM (see Table 1 for plant unit references).The first measurements of soil water potential were made in December 1997.These measurements were carried out either on the same day the gasexchange measurements were conducted, or within 24 hours of these measurements.G Ga as s E Ex xc ch ha an ng ge e. .One marked apical stem segment with several leaves (L.tridentata) and one marked leaf segment (M. porteri) on each experimental plant were used for repeated measurements of gas exchange, usually five sets of measurements per day.When M. porteri was in a quiescent or senescent stage, two apical stem segments of L. tridentata were used per plant to Table .1. Plant units (treatments); one half of all plants (n/2) were sampled on each study date.
Tabla. 1. Unidades de plantas; la mitad del total de las plantas (n/2) fue muestreada en cada fecha de estudio.-Pérez E et al., ΦYTON 77 (2008) measure gas exchange; a mean of these two measurements was obtained to avoid pseudoreplication.Each set of gas exchange measurements took approximately one hour: 8:30-9:30, 10:00-11:00, 11:30-12:30, 13:00-14:00, and 15:00-16:00 h on relatively cloudless days, using a LI-6200 Portable Photosynthesis System (LI-Cor, Inc., Lincoln, NE).Data were integrated (area under the curve) by plant for each sampling day prior to statistical analysis to minimize the effects of environmental variability on photosynthetic activity, e.g. ( 1) vapor pressure deficit (Shirke & Pathre, 2004), (2) reversible photoinhibition (Gray et al., 1996), or (3) high temperature which affects water use efficiency (WUE) (Polley, 2002).For each twig of L. tridentata or leaf of M. porteri, daily integrated values of photosynthesis and transpiration were obtained between 8:30 to 16:00 h.Water use efficiency was calculated by dividing integrated photosynthesis by integrated transpiration.At the end of each day gas-exchange was measured, leaves of M. porteri used for gas exchange measurements were excised, and leaf area was determined in the laboratory using a LI 3000 Leaf Area Meter (LI-Cor, Inc., Lincoln, NE).The single-surface area on creosotebush leaves used for gasexchange was determined in situ using the method of Franco et al. (1994).

Castellanos
S St ta at ti is st ti ic ca al l A An na al ly ys si is s. .For each date, data were analyzed in a completely randomized design, using the Duncan's multiple-range test.Also, the data were grouped by season.For each season, data were arranged factorially in a completely randomized design where one factor was the plants and the other was the sampling dates.Data were analyzed with analysis of variance procedures using indicator variables (Litell et al., 1991). .For each season, the comparison of each response variable of L. tridentata was carried out between dates.Within a season, dates when M. porteri was green were compared for the respective variable.Dates when M. porteri was senescent or quiescent were excluded.Mean comparison for sampling dates was conducted using Duncan's multiple-range test.
Also, contrasts were used for mean comparisons of plants in each season.Contrasts were: (a) L. tridentata vs M. porteri; (b) large vs small L. tridentata; (c) L. tridentata growing alone vs L. tridentata growing with M. porteri; (d) large L. tridentata growing alone vs large L. tridentata growing with M. porteri; (e) small L. tridentata growing alone vs small L. tridentata growing with M. porteri; (f) small L. tridentata with M. porteri undefoliated vs small L. tridentata with M. porteri defoliated; (g) M. porteri growing alone vs M. porteri growing underneath L. tridentata; (h) undefoliated M. porteri growing underneath a small L. tridentata vs M. porteri defoliated growing underneath a small L. tridentata; (i) M. porteri growing underneath a large L. tridentata canopy vs M. porteri growing underneath a small L. tridentata canopy.

RESULTS AND DISCUSSION
In the summer season, and although measurements were carried out every week from 31 July 1997 to 28 September 1997, and from 06 July to 29 September 1998, only four dates will be shown per year in this season.However, when results are shown without a specific date, it implies that all measurement dates were taken into consideration.In the summer 1997, M. porteri was green during five out of nine sampling dates.In summer 1998, M. porteri was green during seven out of 12 sampling dates.P Pl la an nt t a an nd d S So oi il l W Wa at te er r R Re el la at ti io on ns s S Su um mm me er r s se ea as so on n. .During summer 1997 (Fig. 2A), pre-dawn ψ L was greatly influenced by rainfall.However, there were no significant effects of either L. tridentata plant size or presence/absence of a M. porteri understory underneath L. tridentata.The only exception was on 30 August, where large plants (LL) had greater ψ L (-3.25 MPa) than small plants (SL; -4.75 MPa).
During the summer of 1998, ψ L measurements on July 22 (Fig. 3A), coincided with increased soil water potentials (-2.6 MPa; Fig. 4); however, there were no ψ L differences among plants.On 28 July, ψ s and ψ L increased, and measurements were also possible for M. porteri.On this date, ψ L were greater in the grass than in L. tridentata; means ranged from -0.62 to -1.05 MPa in the grass, and -1.73 to -2.3 MPa in L. tridentata.By 25 August, ψ s decreased to almost -4.0MPa, and MDSL had greater ψ L than MSL, SLM, SL, M, and LLM; at the same time, SLMD had higher ψ L than SL, M, and LLM.Finally, LL and MLL had higher ψ L than M. On 29 September, ψ L were significantly greater in LL (-4.50 MPa) than in SLM .Muhlenbergia porteri had greater ψ L than L. tridentata when soil water potentials increased above -3.0MPa.In 1998, ψ L was not affected either for L. tridentata plant size or for the presence of M. porteri.These results agree with those found in a nearby area (Franco et al., 1994).
F Fa al ll l S Se ea as so on n. .The results of soil water potential and the predawn ψ L readings for the fall season are shown in Fig. 4 and Fig. 5A.On 5 October 1997, there were no significant differences (p>0.05) between plants (from -3.45 in SL to -3.6 MPa in LL).On 9 November 1997, this variable was so low that we could not measure it.This was the result of the absence of precipitation.
On 5 October 1998, ψ L readings for L. tridentata were similar to those in September 1998, and soil water potentials were below -8.0 MPa.On 28 October, soil water potentials increased due to rainfall, and ψ L in L. tridentata increased as a result.These were the greatest ψ L observed in the study plants, although they were similar to those on 3 November (p>0.05).-Pérez E et al., ΦYTON 77 (2008) However, M. porteri was still quiescent, and measurements were then not possible.On 3 November, M. porteri was green most likely due to the high soil water potentials coming from the October precipitations.It showed a ψ L range from -0.8 MPa in M to -1.07 MPa in MSL.These values were significantly higher (p<0.001)than those in L. tridentata, which showed values from -1.95 MPa in LL to -2.4 MPa in LLM.While there were significant differences (p<0.05) between M. porteri and L. tridentata plants, there were no differences within either M. porteri or L. tridentata plants.On the remaining dates, there were no significant differences (p>0.05) in ψ L between plants.In general, M. porteri showed a greater ψ L than L. tridentata when it was green during its growing season.W Wi in nt te er r S Se ea as so on n.Low temperatures during winter prevented us from taking valid ψ L measurements.Minimum and mean temperatures during 24 January were -3.9 °C and 6.7 °C respectively, and -1.7 °C and 8.9 °C, respectively, during 25 January.On 19 February, minimum and mean temperatures were -1.1 °C and 8.9 °C, respectively, and on 20 February values were 2.7 °C and 9.4 °C, respectively.Despite very high soil water potentials were observed during winter 1997-98, predawn ψ L in L. tridentata could only be measured after March 1998 (Fig. 6A).During January and February, however, this variable was too low to be measured (below -7.0 MPa).This was very likely due to the low winter temperatures.In March 1998, predawn ψ L increased from -3.44 MPa on 7 March to -2.82 MPa on 22 March.The minimum and mean temperatures were 10.0 °C and 14.4 °C, respectively, on 6 March, and 2.2 °C and 7.2 °C, respectively, on 7 March.On 21 March, minimum and mean temperatures were 5.0 °C and 15 °C, respectively, and they were 6.7 °C and 18.3 °C, respectively, on 22 March.

Castellanos
Soil water potentials were also high during winter 1998-99.It was then possible to measure ψ L , which were higher during this second winter season (1998-99).On 19 December 1998, for example, mean value of -1.7 MPa (p<0.05) was obtained.The minimum and mean temperatures on 18 December were 8.3 °C and 12.2 °C, respectively, and they were 11.2 °C and 11.7 °C, respectively, on 19 December.Mean xylem water potentials on L. tridentata plants were -3.82 MPa in January and -3.38 MPa in February 1999.Within each date, there was no difference among plant units (treatments), except on 19 December 1998 where SLMD (-1.05 MPa) had greater ψ L than SL (-2.7 MPa; p<0.05).
S Sp pr ri in ng g S Se ea as so on n.High winter rainfall in 1998 led to high soil water potentials: -0.86 and -1.03 MPa on 10 April and 19 April, respectively (Fig. 4).Leaf water potentials measured on 10 April were lower (p<0.05)than those determined and 18 April (means were from -1.97 to -2.3 MPa).We were unable to measure ψ L from 18 May throughout 23 June.This was because of the extremely low ψ L registered at this time.In turn, this coincided with low ψ s (ranging from -7.6 to -8.1 MPa) which probably resulted from the low precipitation at this time.On 10 April, SL (-2.35 MPa) and SLMD (-2.38 MPa) had greater ψ L (p<0.05) than LL (-2.9 MPa).On 18 April, there were no significant differences (p>0.05) in ψ L .-Pérez E et al., ΦYTON 77 (2008) P Ph ho ot to os sy yn nt th he es si is s S Su um mm me er r s se ea as so on n. .On 7 August 1997 (Fig. 2B), M. porteri was in a quiescent state, L. tridentata increased its photosynthetic activity in relation to the first sampling data, and LL showed a higher photosynthetic activity than SLM plants (p<0.05).On 14 August, L. tridentata showed the highest photosynthetic activity of the whole season, partially as a result of the rainfall on 13 August (40.1 mm).Photosynthetic activity was higher (p<0.05) on SL, LLM, and SLMD than in the other plants.Muhlenbergia porteri exhibited less (p>0.05)photosynthetic activity than L. tridentata.Photosynthetic rates were similar (p>0.05) on all M. porteri plants.

Castellanos
Photosynthetic activity decreased in both species on 30 August in part due to the absence of rainfall.At this date, the standard error was so high that there were no statistical differences among plant units (treatments).On 28 September, both species showed activity because of a previous rainfall.However, no differences (p>0.05) were detected between species.The exception was L. tridentata which showed the highest activity (p<0.05) in the LL treatment.Also, activity was lower (p<0.05)at the M than at the LLM, SLM and SLMD treatments.
Larrea tridentata exhibited an opportunistic role due to its increased photosynthetic activity after the high rainfall.Our results suggest that L. tridentata competed with M. porteri for water resources.This is because L. tridentata plants growing alone had higher net assimilation after a large rainfall event [i.e., in summer 1997, where rainfall (132.1 mm) was above the long-term average (116.3mm)] (Kunkel et al. 1988).These results differ from those of Dodd et al. (1998): they found that the woody Atriplex canescens and grass species used water from different soil horizons.
In the summer of 1998 (Fig. 3B), photosynthetic activity was lower than 50 mmol/m 2 /h on 22 July, and there were no differences (p>0.05)among plants (treatments).On 28 July, M. porteri was green and its photosynthetic activity was greater in the MSL and M than in the SL, MLL, LL and SLM treatments (p<0.1).Photosynthetic activities were higher than 50 mmol/m 2 /h.On 25 August, SLMD had greater (p<0.05)photosynthetic activity than LL and LLM plants.On 29 September, M. porteri was senescent, and L. tridentata has mean photosynthetic activities lower than 15 mmol/m 2 /h.There were no differences among plants within each date (p>0.05).Small L. tridentata plants showed a greater (p<0.0005)photosynthetic activity when M. porteri was defoliated (40.9 mmol/m 2 /h) than when it remained undefoliated (29.8 mmol/m 2 /h).This indicates that L. tridentata plants were more active when M. porteri was defoliated than when it remained undefoliated.Even though 1998 was marginally drier than 1997, it was still above the long-term average with 120.1 mm of rainfall.However, only 4.6 mm were recorded in September whereas the long-term monthly average is 31.8mm (Kunkel et al., 1988).There were no clear patterns of L. tridentata -M.porteri interactions.This was likely due to limited water availability in the soil profile where roots of L. tridentata and M. porteri occur together.F Fa al ll l S Se ea as so on n. .On 5 October 1997, photosynthetic activity was similar (p>0.05) in both species and plants (treatments), having values no greater than 70 mmol/m 2 /h (Fig 5B).On this date, L. tridentata had the highest (p<0.05)photosynthetic activity during the fall season in 1997.This value was also greater than those found in the 1998 fall season.On 9 November 1997, only L. tridentata was physiologically active, and photosynthesis was lower (p<0.05)than that observed on 5 October, when LLM had greater (p<0.05)photosynthetic activity than all others.
During fall 1998, we could not measure photosynthesis in M. porteri plants on 5 October and 28 October.The lowest photosynthetic activity for L. tridentata was recorded on 5 October 1998.Photosynthetic activity was similar (p>0.05)during 5 and 28 October.On 3 November, M. porteri plants showed photosynthetic activity because of the high soil water potentials.The greatest (p<0.1)net assimilation was found in M plants.In turn, MDSL and SL treatments had greater photosynthetic activity than LL (p<0.1).
We made a statistical comparison between 5 October 1997 and 3 November 1998, times when M. porteri was green: photosynthetic activity was lower (p<0.05) on 3 November 1998, than on 5 October 1997.Large plants of L. tridentata growing alone had lower (p<0.05)photosynthetic activity than LLM plants.Muhlenbergia porteri growing alone had greater (p<0.005)photosynthetic activity than M. porteri growing underneath the shrub canopy.Both species responded to soil moisture availability, in agreement with desert plants which have the ability to respond to moisture during late summer or early fall (Bamberg et al., 1975).No differences in the ψ L or water use efficiency (see below) in M. porteri plants suggest that the lower photosynthetic activity of these plants when growing underneath the shrub canopies could be due to the shadow effect on this warm season grass (Ballaré et al., 1990).
W Wi in nt te er r S Se ea as so on n. .On 25 January 1998, there was no photosynthetic activity for any of the study plants because of the low temperatures.At the other measurement dates during the winter season, only L. tridentata showed photosynthetic activity, with values lower than 50 mmol/m 2 /h (Fig. 6B).On 28 January 1999, 7 March and 22 March 22, 1998, L. tridentata plants had higher photosynthetic activity than on 8 February 1999, and 19 December and 20 February 1998. On 20 February, 7 March and 22 March Castellanos-Pérez E et al., ΦYTON 77 (2008) 1998, there were no differences (p>0.05) in the assimilation rates between L. tridentata plants.On 19 December, SL had greater (p<0.05)photosynthetic activity than LL and SLM.On 28 January 1999, there were no significant differences (p>0.05)among plants.On 8 February 1999, photosynthetic activity was greater (p<0.05) in SL than in the other plants.In L. tridentata plants, SL also had higher (p<0.005)photosynthetic activity than SLM and SLMD plants.Also, LL and SL plants had higher (p<0.01)photosynthetic activity than LLM, SLM and SLMD plants.
Low temperatures reduce assimilation rates because Rubisco activity and electron transport capacity are reduced (Farquhar & Sharkey, 1982).Our results are similar to those of Reynolds et al. (1999).These authors found a low net assimilation rate in the winter season, even when soil moisture was high.The high temperature acclimation of L. tridentata (Mooney et al., 1978;Reynolds et al., 1999), may allow this species to have net assimilation during winter.Increasing temperatures during the day may allow carbon assimilation in spite of the low temperatures.Avoidance of freeze-induced cavitation, caused by air bubbles formed in situ during freezing, is associated with the small vessel size in L. tridentata (Pockman & Sperry, 1997).Another associated factor is leaf age.New leaves grow in spring and summer; therefore, winter leaves are, in general, older than those of previous seasons (Reynolds et al., 1999).
During the winter season, M. porteri, although dry, appears to compete with L. tridentata when they grow close to each other.This is because L. tridentata plants growing alone showed higher photosynthetic activity.This may have been the result of higher levels of photosynthetically active radiation reaching these plants without interference by the grass (Welsh & Beck, 1976).S Sp pr ri in ng g S Se ea as so on n.In April, M. porteri was green because of the high soil moisture and the fact that this grass has photoperiodic control for flowering in long-days (Carr-Smith et al., 1989).On April 10, both species showed almost equal photosynthetic activity; the exception was SLMD which had higher photosynthesis than MLL plants (Fig 7B).Both species still showed active by 18 April, although photosynthetic activity was lower (p<0.05)than at the previous sampling dates, and differences were not significant (p>0.05).On 18 May, 25 May, 8 June and 23 June, only L. tridentata showed photosynthetic activity with values lower than 40 mmol/m 2 /h for all plants; means at each date were lower than 34 mmol/m 2 /h.The lowest photosynthetic activity during spring (22.6 mmol/m 2 /h) was reached on 23 June; this reading was similar (p>0.05) to that on 18 May and 8 June.On 18 May, photosynthetic activity was greater (p<0.05) on SLMD than LL plants.There were no differences (p>0.05) in photosynthetic activity between plants from 25 May to 23 June.Low values in photosynthesis were probably a result of low ψ s .On average, photosynthetic activity was greater (p<0.01) on small than on large L. tridentata plants.Photosynthetic activity of L. tridentata was lower (p<0.05)when growing alone than when it grew with M. porteri.Photosynthetic activity of undefoliated Muhlenbergia porteri was greater (p<0.005) when it grew under small than large L. tridentata plants.
Greater net assimilation values for small than large L. tridentata plants are consistent with data of de Soyza et al. (1996).They concluded that these results allow L. tridentata to compete with grasses.The continued net assimilation rates of L. tridentata at soil water potentials below -8.0 MPa show its ability to survive under these conditions (Strain & Chase, 1966).Leaf water potentials could not be measured at these low soil water potentials.However, Oechel et al. (1972) found that L. tridentata have a continuous net assimilation either at temperatures of 50 °C or ψ L of -7.3 MPa.W Wa at te er r U Us se e E Ef ff fi ic ci ie en nc cy y S Su um mm me er r S Se ea as so on n.Water use efficiency (WUE) increased from 31 July to 7 August 1997, with a WUE of 2.41 [mmol C/mol H 2 O] / m 2 /h for LL (Fig 2C).This value was statistically similar to that on LLM and SLM, which had the highest (p<0.05)WUE ratio (Fig. 2C).The lowest (p<0.05)WUE was on plants of SLMD, SLM, SL and LLM.On 14 August, WUE of M. porteri was similar in MSL (mean = 2.08) and MLL (mean = 2.15) plants.However, these WUE were higher (p<0.05)than those in L. tridentata which showed WUE ranging from 1.11 (LL plants) to 1.40 (LLM plants).On 30 August, MDSL on L. tridentata plants had a high value (10.35), similar (p>0.05) to that in MLL and MSL M. porteri plants and SLM plants of L. tridentata.Except on MDSL plants, lowest WUE were found in SL, although values were similar (p>0.05) in all other treatments.On 28 September, when M. porteri was green again, MDSL plants showed a high WUE (mean = 6.65).This value was greater (p<0.05)than that on L. tridentata and M. porteri plants in all other treatments.Except on MDSL plants in M. porteri, all other plants in this species showed similar (p>0.05)WUE values, ranging from 3.14 on MSL plants to 4.01 on MLL plants.L. tridentata plants had lower (p<0.05)WUE than M. porteri plants.There were no significant differences (p>0.05) in WUE between plants of different size in L. tridentata; values ranged from 1.08 in LL plants to 1.74 in LLM plants.
When there was high soil water potentials due to winter rainfall, L. tridentata showed high WUE.This may contribute to increase the survival ability of this shrub to drought periods.The C 4 pathway in M. porteri alows a greater C fixation with lower transpirational water loss than a C 3 plant, contributing to its greater WUE.

CONCLUSIONS
The hypothesis that the resource islands under L. tridentata provide water and a better microhabitat for M. porteri needs to be rejected in this study.This is because grass plants growing under L. tridentata and those growing alone showed no differences in the study variables.In addition, M. porteri growing alone had a greater photosynthetic activity than M. porteri growing under the canopy of L. tridentata in the fall.
The hypothesis that M. porteri and L. tridentata coexist because they stratify periods of greatest physiological activity was partially true.Predawn ψ L and photosynthetic activity responded to soil water availability during spring, summer, and fall in both C 3 and C 4 plants.
Soil moisture availability in spring, summer, and fall, and winter temperatures contribute to dryness of M. porteri.Muhlenbergia porteri showed a good and rapid response to the episodic soil moisture availability in spring, summer and fall.We demonstrated that M. porteri, like several C 4 grasses, is an opportunistic plant (i.e., see Bamberg et al., 1975).As such, it quickly responded to greater soil moisture and warm temperature conditions by increasing its C assimilation and growth.These results partially agree with those of Nobel (1980) in the C 4 plant Hilaria rigida; this grass was more influenced by water availability than by low temperatures.
Larrea tridentata showed net C assimilation during the entire study period, except on 25 January 1998, demonstrating its resistance to extreme both soil moisture and temperature conditions.These results agree with findings of Strain & Chase (1966) and Bamberg et al. (1975).Larrea tridentata was always metabolically active (Oechel et al., 1972).It was not dormant like M. porteri under moisture or temperature stress.Larrea tridentata behaved as an opportunistic species because of its increased photosynthetic activity during periods of high soil water potential.A general pattern of water competition with M. porteri was shown when L. tridentata growing alone showed a greater net assimilation after a high rainfall event.However, we were unable to detect if M. porteri was affected by the presence of L. tridentata.Dodd et al. (1998) found that the woody plant Atriplex canescens used water from different soil horizons than the grasses growing under its canopy.Castellanos-Pérez E et al., ΦYTON 77 (2008) The hypothesis that M. porteri has greater photosynthetic rates when growing under small than large L. tridentata had to be rejected.This was because this grass did not have greater photosynthetic rates when growing under small shrubs.However, differences were found with plant size in L. tridentata and when this shrub either was or was not in competition with M. porteri.During winter, SL had higher photosynthetic activity than SLM and SLMD.At the same time, LL and SL had higher photosynthetic activity than LLM, SLM and SLMD.
The hypothesis that removal of senescent plant material in M. porteri would result in greater photosynthetic rates due to greater light availability for photosynthesis was rejected.This was because defoliation of 50% biomass of M. porteri did not result in greater photosynthetic rates compared to undefoliated plants.During summer 1998, photosynthetic activity of small L. tridentata plants was greater when M. porteri growing under its canopy was defoliated than when it remained undefoliated.We conclude therefore that L. tridentata shows the major effects when it grows associated with M. porteri.These effects are shown by a reduced (1) ability to gather resources (i.e., water), and (2) capacity for photosynthetic C assimilation and subsequent growth.
porteri plant growing underneath a small L. tridentata 8