Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Annals of Botany 80 : 731–739, 1997 Frequencies, Microclimate and Root Properties for Three Codominant Perennials in the Northwestern Sonoran Desert on North- vs. South-facing Slopes P A R K S. N O B E L* and M A T T H E W J. L I N T O N Department of Biology, Uniersity of California, Los Angeles, CA 90095-1606, USA Received : 1 April 1997 Accepted : 25 July 1997 At a site in the northwestern Sonoran Desert the percent ground cover for the C subshrub Encelia farinosa was eight$ times higher on more arid 20° south-facing slopes than on 20° north-facing slopes at 820 m elevation, and was sixtimes higher on north-facing slopes at a 300-m-lower elevation, also the more arid condition. The ground cover of the C bunchgrass Pleuraphis rigida decreased over 50 % from 20° north-facing slopes to the more arid conditions of % a 36° north-facing slope, a 20° south-facing slope and a 20° north-facing slope at a 300-m-lower elevation. The CAM leaf succulent Agae deserti also had greater ground cover for the 20° north-facing slopes at 820 m compared with 520 m. For these three codominants that averaged 58 % of the total ground cover, the key for the relative frequency of E. farinosa was apparently its greater root growth on the warmer slopes during the winter. The key for the other two species was most likely soil water availability, especially during the seedling stage for A. deserti. The wetter soil conditions on 20° north-facing slopes at 820 m apparently led to individual plants of P. rigida that were twice as large as on south-facing slopes. Thus root properties may exert the primary influence on relative plant frequency in this desert ecosystem for which soil temperature and water availability are crucial. # 1997 Annals of Botany Company Key words : Agae deserti, Encelia farinosa, Pleuraphis rigida, rooting patterns, soil temperature, Sonoran Desert, water availability. INTRODUCTION Ground cover and biomass productivity in deserts, which occupy approximately a quarter of the earth’s land area, are generally low compared with other ecosystems because of limited, and often sporadic, rainfall (McGinnies, Goldman and Paylore, 1968 ; Noy-Meir, 1973 ; Leith and Whittaker, 1975 ; Beadle et al., 1985 ; Walter, 1985). As a result, the interplay of environmental factors such as air temperature and soil moisture on relative plant performance can be more apparent in deserts with their extreme conditions than in ecosystems where rainfall is more plentiful and temperatures are more moderate (Beatley, 1974). These environmental factors can affect desert plants by influencing seedling establishment, net CO uptake and biomass productivity of # the shoots, or seasonal growth and deployment of roots (Rundel and Nobel, 1991). Desert plants exhibit a wide range of growth forms (Smith and Nobel, 1986 ; Gibson, 1996), although the influence of growth form on relative species frequencies has rarely been related to local environmental conditions accompanying different slopes or elevations. Microclimatic conditions associated with global climate change (Adams et al., 1990) may also have a major impact on species composition in deserts, as certain species may currently be on the verge of elimination based on water availability or temperature. Analysing effects of microclimate on plant distribution may therefore enable predictions of future species composition of deserts to be made. Approximately 92 % of the nearly 300 000 species of * For correspondence. Fax 310 825-9433, e-mail psnobel!biology.ucla.edu 0305-7364}97}06073109 $25.00}0 vascular plants use the C pathway (Nobel, 1991 a ; Winter $ and Smith, 1996). Approximately 1 % of vascular plants use the C pathway, most of which are grasses that have evolved % in tropical or subtropical regions with high temperatures and often limited water availability (Monson, 1989). C % species tend to replace C species as aridity increases in $ various ecosystems (Ehleringer and Monson, 1993) ; in a mixed prairie in the Northern Great Plains, the relative biomass of the C species increases in midsummer con% comitant with higher irradiation, higher temperature and water stress, and then decreases during the less arid autumn (Ode, Tieszen and Lerman, 1980). Crassulacean acid metabolism (CAM) plants, which constitute about 6 % of vascular plant species (Nobel, 1991 a ; Winter and Smith, 1996), tend to open their stomates at night when temperatures and hence evaporative demand are lower, leading to a high water-use efficiency (CO fixed per unit water # transpired). Hence CAM plants are favoured in waterlimited environments, such as in deserts and as epiphytes in tropical and subtropical regions (Nobel, 1988 ; Andrade and Nobel, 1996). At a site in the northwestern Sonoran Desert where the annually bimodal rainfall pattern leads to winter and summer wet periods, representatives of each of the three photosynthetic pathways constitute the three codominant species among the 33 common perennials (Drennan and Nobel, 1997), an unusual photosynthetic diversity. Specifically, the deciduous C brittle-bush Encelia farinosa Torrey $ and A. Gray (Asteraceae), the perennial C bunchgrass % Pleuraphis rigida Thurber (Poaceae), and the monocarpic CAM leaf succulent Agae deserti Engelm. (Agavaceae) bo970508 # 1997 Annals of Botany Company 732 Nobel and Linton—Frequencies on North- vs. South-facing Slopes together constitute over half of the local ground cover (Nobel and Jordan, 1983 ; Ehleringer, 1985 ; Nobel, 1988 ; Hickman, 1993 ; Drennan and Nobel, 1996, 1997). These three perennials all have relatively short shoots, minimizing interspecific shading, and have relatively shallow roots, facilitating examination of root deployment. Although similar in stature, the three species have different growth forms. In particular, E. farinosa has seasonally dimorphic leaves that are relatively large and dark green in winter, but greyish and smaller in summer (Ehleringer and Bjo$ rkman, 1978), P. rigida has short-lived, thin leaves with photosynthetically active sheaths (Nobel, 1980) and A. deserti has massive succulent leaves that remain photosynthetically active for many years (Nobel, 1976, 1988). Although similar in rooting depth, the optimal temperatures for root growth are 25–27 °C for E. farinosa, about 5 °C higher for A. deserti, and an additional 5 °C higher for P. rigida (Drennan and Nobel, 1996). This study investigated whether the photosynthetic pathway, growth form, or other morphological and}or physiological property could explain the preliminary observations of different frequencies among the three codominants on various slopes in the northwestern Sonoran Desert. The populations studied were found on north-facing and south-facing slopes, whose microclimates were quantified at the two solstices and an equinox. To help interpret influences of air temperature and water availability, relative ground cover was additionally examined for north-facing slopes of greater inclination at 820 m and at a 300-m-lower elevation. Leaf temperatures, light levels in the planes of the leaf surfaces, shoot size and the locations of roots of all three species at various depths in the soil were determined in the late winter, a time following the seasonally most reliable wet period (Nobel, 1988). Production of new roots was also examined for all three species in the late winter. Seasonal variations in soil temperature can affect root growth and hence water uptake from various soil layers, which in turn influences plant productivity and thereby possibly ecological success. An explanation was thus developed as to why the more arid south-facing slopes favoured E. farinosa over P. rigida compared with north-facing slopes. Optimal temperatures for root growth coupled with the effect of different slopes on water availability were also used to interpret the distribution of A. deserti, leading to the conclusion that environmental effects on root properties can be the primary factor determining the relative frequency of these three species in a desert ecosystem. MATERIALS AND METHODS Sites The main field site was a series of small hills with different slopes centred on 33° 38« N, 116° 24« W at a mean elevation of 820 m in the University of California Philip L. Boyd Deep Canyon Desert Research Center about 8 km south of Palm Desert, California (a site known as Agave Hill). Additional line transects were made about 2 km north of the main site at a lower elevation of 520 m. The entire region is characterized by small hills and numerous granitic outcroppings, with a relatively low plant ground cover averaging about 30 % and a loamy-sand soil texture (Nobel 1976, 1980, 1988 ; Drennan and Nobel, 1997). The three codominant species with respect to ground cover are Agae deserti Engelm. (Agavaceae), Encelia farinosa Torrey and A. Gray (Asteraceae), and Pleuraphis rigida Thurber (Poaceae) [formerly Hilaria rigida (Thurber) Scribner] (Hickman, 1993). Measurements at 820 m were made on 20³2° northfacing slopes, 36³3° north-facing slopes (the latter for line transects only) and 20³2° south-facing slopes. At the 520m site, line transects were examined on 20³2° north-facing slopes. Angles for slopes were judged along approx. 3-m ground intervals using an inclinometer consisting of a large protractor and a plumb bob. The aspects of all slopes were within 12° of the indicated compass direction [readings with a Brunton 37215 Pocket Transit (Forestry Suppliers, Jackson, Mississippi, USA) were corrected for the 13° local difference between true and magnetic north]. The mean annually-bimodal precipitation at the main site is 266 mm, with winter-early spring precipitation averaging 20 % more than summer-early autumn precipitation and being less variable from year to year (Young and Nobel, 1986 ; Nobel, 1988). Plant frequency and size Non-overlapping line transects 60 m in length using the line intercept method (Greig-Smith, 1983 ; Bonham, 1989) were measured in the late winter (1–6 Mar. 1996) at random locations and in random directions on a series of hills having the designated slopes. The length of intersection of any living plant of A. deserti, E. farinosa, P. rigida or all of the other species combined was noted ; such distance intervals were summed along each transect and then divided by 60 m to estimate the percent ground cover for the three species considered individually and for all the other species combined. Ten m¬10 m quadrats were located randomly on 20³2° north-facing or south-facing slopes at 820³10 m and subdivided into four 5 m¬5 m subsections to facilitate plant counting. All plants of the three species considered were counted, including all individual rosettes of A. deserti, which occurs predominantly in clonal groups. Canopy area was also determined on 1–3 and 26–27 Mar. 1996 for randomly selected plants of A. deserti, E. farinosa and P. rigida in the same region as the quadrats ; areas of the approximately elliptical canopies were defined as major axis¬minor axis¬π}4. Only living (green) leaves were considered ; this was the most difficult to ascertain for P. rigida, many of whose leaves were brown. Slope and leaf microclimate To help characterize the microclimate of a slope, air and soil temperatures, photosynthetic photon flux (PPF, wavelengths of 400 to 700 nm) and soil water potential were measured at the centre of bare-ground patches approx. 1 m in diameter for 20³2° north-facing and south-facing slopes at 820³10 m near the winter solstice (16 Dec. 1995), the spring equinox (2 Mar. 1996) and the summer solstice (15 Jun. 1996). Using copper-constantan thermocouples 0±51 mm in diameter, temperature was measured over 24-h periods at 30 cm above the ground, which represents the Nobel and Linton—Frequencies on North- vs. South-facing Slopes 733 T 1. Percent ground coer for arious slopes Slope Ground cover (%) Direction faced Angle (°) Elevation (m) Agae deserti Encelia farinosa Pleuraphis rigida Other Total North South North North 18–22 18–22 33–39 18–22 810–830 810–830 800–830 510–530 8±6³0±5 7±7³1±0 7±2³0±9 3±1³1±6 1±5³0±4 11±9³1±4 1±1³0±3 8±9³1±0 8±8³0±9 3±7³0±6 4±0³0±6 0±9³0±4 13±1³1±9 11±4³1±7 13±6³1±7 9±8³1±3 32±2³1±3 34±7³1±9 25±9³1±4 22±6³1±2 Data are based on 60-m line transects measured on 1–6 Mar. 1996 and are presented as means³s.e. (n ¯ 10 transects). upper part of the canopy where most leaf area occurs for the three species considered (the thermocouple was shielded from direct solar irradiation), at the soil surface and 10 cm below the soil surface, which is approximately the centre of the rooting zone for the three species (Nobel and Jordan, 1983 ; Drennan and Nobel, 1996). Daytime PPF in the plane of the slopes was determined with an LI-191SB line quantum sensor (Li-Cor, Lincoln, Nebraska, USA). Soil moisture content was determined gravimetrically using 200–300 g samples that were collected 9 to 11 cm below the soil surface and immediately weighed, followed by drying at 80 °C until no further weight change occurred (generally 3 d) ; water content by mass was converted to water potential using a moisture release curve for soil from the field site at 820 m (Young and Nobel, 1986). Leaf temperatures and PPF in the planes of the leaves were also determined over a 24-h period at 820³10 m at the end of the winter wet period (2 Mar. 1996). Measurements were made on medium-sized plants based on their contribution to ground cover (approx. 18 leaves unfolded from the central spike for A. deserti, 25 branches emanating from the main stem near the plant base for E. farinosa and 60 living culms for P. rigida) using four leaves at mid-canopy occurring in each of the four cardinal directions. Temperature was measured at mid-leaf with copper-constantan thermocouples 0±51 mm in diameter, and PPF was measured on both abaxial and adaxial leaf surfaces with a Li-Cor LI190S quantum sensor. Rooting depth Excavations of roots of the three species occurring on 20³2° north-facing or south-facing slopes at 820³10 m were also performed in the late winter (1–3 Mar. 1996) by carefully removing sequential 4-cm-thick layers of soil with a geologist’s pick, trowel and various brushes ; roots were cut by scissors and then collected. In the succeeding year (28 Feb.–1 Mar. 1997), the entire root system was excavated without dividing into layers, and old roots were distinguished from new, whitish roots that grew in response to winter rainfall (Nobel, 1997) ; the total rainfall for the winter wet period preceding the excavations was 52 mm, 71 % occurring on 13 Jan. 1997. In all cases, multiple-yearold, but relatively small isolated plants of similar size (approx. nine unfolded leaves for A. deserti, 13 main branches for E. farinosa and 35 culms for P. rigida) were selected. Excavations extended up to 50 cm radially from the shoot and to a depth of 36 cm, thereby also ensuring recovery of essentially the entire root system (Nobel and Jordan, 1983 ; Nobel, 1988). The surface area of living roots (verified by the presence of turgid cells within the stele for representative roots) in each layer was the product of root length, mean root diameter and π ; roots of the monocotyledonous A. deserti and P. rigida were fairly uniform in diameter, whereas those for the dicotyledonous E. farinosa differed in diameter and so were divided into central tap root, primary lateral roots, and secondary or higher-order lateral roots before such multiplication. The mean root depth was the root surface area for each layer multiplied by the depth at mid-layer summed over all layers and then divided by the total root surface area. Data were analysed using ANOVA and Student’s t-test. RESULTS Distribution The average ground cover for Agae deserti, Encelia farinosa and Pleuraphis rigida between 20³2° north-facing and 20³2° south-facing slopes at 820³10 m was 6–8 % for each species (Table 1). Under these conditions the percent cover did not vary significantly with the slope for A. deserti, was eight-times higher on south-facing slopes for E. farinosa (P ! 0±001), and was twice as high on north-facing slopes for P. rigida (P ! 0±01). For north-facing slopes at this elevation with a mean slope of 36°, the ground cover for A. deserti and E. farinosa was not significantly different than for 20° northfacing slopes, but the ground cover for P. rigida was halved (Table 1 ; P ! 0±01). For 20° north-facing slopes at an elevation that was 300 m lower (520 m), the percent cover for A. deserti and P. rigida was 64 to 90 % lower (P ! 0±01), whereas the ground cover for E. farinosa was six-times higher (P ! 0±001 ; Table 1). Plant frequencies (Table 2) showed somewhat different patterns than plant cover (Table 1) for the 20° north-facing s. south-facing slopes at 820 m. In particular, A. deserti was 40 % more frequent on north-facing slopes (P ! 0±05), E. farinosa was seven-times more frequent on south-facing slopes (P ! 0±001), and P. rigida was not statistically different on the two slopes (P " 0±5). The mean canopy area per plant (Table 2) was 26 % larger for A. deserti on southfacing slopes (P ! 0±05), 57 % larger for E. farinosa on south-facing slopes (P ¯ 0±08) and 104 % larger for P. rigida on north-facing slopes (P ! 0±001). The plant frequency multiplied by mean plant canopy area (Table 2) predicted 734 Nobel and Linton—Frequencies on North- vs. South-facing Slopes T 2. Plant frequency and canopy area on north-facing and south-facing slopes Frequency (number of plants per 100 m#) Slope A. deserti E. farinosa P. rigida North South 51±0³4±8 36±6³3±2 17±4³3±2 115±2³8±8 39±4³3±6 36±0³4±2 Microclimate Plant canopy area (cm#) North South A. deserti E. farinosa P. rigida 1774³63 2231³92 693³47 1085³77 2216³169 1088³107 Slope angles were 20³2° and the elevation was 820³10 m. Frequency was based on 10 m¬10 m quadrats, and plant canopy area (major axis¬minor axis¬π}4) was based on individual plants. Data were obtained 1–3 and 26–27 Mar. 1996 and are presented as means³s.e. (n ¯ 11 quadrats for frequency and 100 plants for area). ground cover (Table 1) within 0±5 % of the measured cover (also in %) for both slopes for all three species. Plant height was not significantly affected by slope for A. deserti 16 Dec. 1995 40 Air temperatures 30 cm above the ground were generally similar for 20° north-facing s. south-facing slopes near the winter solstice (Fig. 1 A), the spring equinox (Fig. 1 B) and the summer solstice (Fig. 1 C). Temperatures were higher for south-facing compared with north-facing slopes at the soil surface (P ! 0±05 for most times during the daytime) and 10 cm below the soil surface (P ! 0±05 for most times during the day- and night-time) near the winter solstice and the spring equinox, but were similar near the summer solstice (Fig. 1 D–I). Near the centre of the rooting zone, 10 cm below the soil surface, the temperatures for north-facing s. south-facing slopes averaged over a 24-h period were 11±5 s. 18±0 °C for the winter solstice (Fig. 1 G), 13±2 s. 18±3 °C for the spring equinox (Fig. 1 H) and 35±8 s. 36±0 °C for the summer solstice (Fig. 1 I). Differences in PPF in the planes of 20° north-facing s. 2 Mar. 1996 A B 15 Jun. 1996 C 30 cm North-facing South-facing 30 (45±0³1±2 cm on 20° north-facing slopes and 46±9³1±3 cm on 20° south-facing slopes ; means³s.e. for n ¯ 100 plants) or P. rigida (40±2³1±2 cm on north-facing slopes and 38±3³1±1 cm on south-facing slopes), but E. farinosa was taller on south-facing slopes (32±8³1±2 cm on north-facing slopes and 40±1³1±4 cm on south-facing slopes ; P ! 0±01). 20 Air or soil temperature (°C) 10 0 D E F 0 cm 50 40 30 20 10 0 40 G H I –10 cm 30 20 10 0 6 12 18 0 6 12 18 Solar time (h) 0 6 12 18 24 F. 1. Air and soil temperatures 30 cm above the ground (A–C), at the soil surface (D–F) and at 10 cm below the soil surface (G–I) for 20³2° north-facing (D) and south-facing (^) slopes at 820³10 m in the northwestern Sonoran Desert. Measurements were made near the winter solstice (A, D and G), the spring equinox (B, E and H) and the summer solstice (C, F and I). Data are means (n ¯ six slopes) ; s.e. averaged 0±3 °C at 30 cm, 0±5 °C at the soil surface and 0±3 °C 10 cm below the soil surface, essentially all within the symbol size. Nobel and Linton—Frequencies on North- vs. South-facing Slopes 16 Dec. 1995 2000 2 Mar. 1996 A 735 15 Jun. 1996 B C North-facing South-facing PPF (µmol m–2 s–1) 1500 1000 500 0 6 9 12 15 6 9 12 15 6 9 12 15 18 Solar time (h) F. 2. Photosynthetic photon flux (PPF) near the winter solstice (A), the spring equinox (B) and the summer solstice (C) in the planes of 20³2° north-facing (D) and south-facing (^) slopes at 820³10 m. Data are means³s.e. (n ¯ six slopes). Leaf temperature (°C) 30 Encelia farinosa Pleuraphis rigida B C 20 10 North-facing South-facing 0 Leaf PPF (µmol m–2 s–1) Agave deserti A D E F 1500 1000 500 0 6 12 18 0 6 12 18 0 6 12 18 24 Solar time (h) F. 3. Leaf temperatures (A–C) and photosynthetic photon flux (PPF ; D–F) in the late winter for Agae deserti (A and D), Encelia farinosa (B and E), and Pleuraphis rigida (C and F) on 20³2° north-facing (D) and south-facing (^) slopes at 820³10 m. Measurements on 2 Mar. 1996 are averages for four leaves at mid-canopy, one in each of the four cardinal directions, for medium-sized plants. Data for temperature are means (n ¯ 4 plants) and s.e. averaged 0±4 °C ; data for PPF represent totals for both leaf surfaces and are means³s.e. (n ¯ six plants). south-facing slopes were also greatest near the winter solstice, intermediate near the spring equinox and least near the summer solstice (Fig. 2). The total daily PPF integrated over the daytime for north-facing s. south-facing slopes on clear days was 10 s. 32 mol m−# d−", respectively, near the winter solstice (Fig. 2 A), 27 s. 41 mol m−# d−" near the spring equinox (Fig. 2 B) and 55 s. 57 mol m−# d−" near the summer solstice (Fig. 2 C). The water potential 10 cm below the soil surface at various times during the spring was higher for 20° north- 736 Nobel and Linton—Frequencies on North- vs. South-facing Slopes Encelia farinosa Agave deserti 0 Pleuraphis rigida B A C 4 Depth in soil (cm) 8 12 16 20 24 28 North-facing South-facing 32 36 0 10 20 30 0 10 20 30 0 Root surface area per layer (% of total) 10 20 30 F. 4. Root distribution with depth for A. deserti (A), E. farinosa (B) and P. rigida (C) excavated in the late winter on 20³2° north-facing (D) and south-facing (^) slopes at 820³10 m. Plants on the two slopes were similar in shoot height and diameter and relatively isolated from other plants : A. deserti had eight–ten leaves unfolded from the central spike of folded leaves, E. farinosa had 12–14 main branches bearing leaves and P. rigida had 33–36 living culms. Data obtained on 1–3 Mar. 1996 are plotted at the middle of the 4-cm-thick soil layers and are means³s.e. (n ¯ five plants). facing compared with south-facing slopes (P ! 0±01). In particular, Ψsoil on north- s. south-facing slopes was ®0±41³0±11 and ®0±96³0±10 MPa, respectively, on 22 Apr. 1995, ®0±97³0±14 and ®2±32³0±33 MPa on 2 Mar. 1996 and ®1±56³0±23 and ®3±64³0±39 MPa on 26 Mar. 1996 (data are means³s.e. for n ¯ 6 slopes). At other times (1 Jun. 1995, 16 Dec. 1995, 15 Jun. 1996), the soil was extremely dry, with a water content of less than 1 % by mass, corresponding to soil water potentials below ®12 MPa. Even in these dry cases, the water content was always higher (average of 27 %) for the north-facing compared with the south-facing slopes (P ! 0±05). Leaf temperature and PPF The daily variations for leaf temperatures at the end of the winter wet period (Fig. 3) were generally similar to those for air temperatures (Fig. 1 B). However, temperatures for leaves of A. deserti (Fig.3 A) substantially exceeded those of the air (P ! 0±01) during the daytime, when leaves of E. farinosa (Fig. 3 B) on north-facing slopes were lower than those on south-facing slopes (P ! 0±05). The relatively small leaves of P. rigida were tightly coupled to air temperatures (cf. Fig. 1 B) throughout the day on both slopes (Fig. 3 C). At the end of the winter, slope had the greatest influence on the PPF in the planes of the leaves for basal rosettes of A. deserti (P ! 0±01 ; Fig. 3 D–F). In particular, the total daily PPF summed for both leaf surfaces at mid-leaf and mid-canopy for medium-sized plants on north-facing s. south-facing slopes was 32 s. 43 mol m−# d−", respectively, for A. deserti (Fig. 3 D), 37 s. 42 mol m−# d−" for E. farinosa (Fig. 3 E) and 34 s. 36 mol m−# d−" for P. rigida, whose culms were essentially vertical (Fig. 3 F). T 3. Mean rooting depth and total root surface area of relatiely small, isolated plants of similar size for 20³2° north-facing and south-facing slopes at 820³10 m Mean depth (cm) Slope A. deserti E. farinosa P. rigida North South 8±4³0±5 11±4³0±5 10±6³0±4 13±4³0±6 9±0³0±4 11±4³0±3 Root surface area (cm#) North South A. deserti E. farinosa P. rigida 548³33 714³44 225³18 291³21 368³18 464³27 Data were obtained as for Fig. 4 and are means³s.e. (n ¯ 5 or 6 plants). Rooting depth and new roots Roots of all three species extended deeper into the soil for small, similar-sized, isolated plants on 20° south-facing slopes (28–32 or 32–36 cm soil layers) compared with 20° north-facing slopes (20–24 or 24–28 cm soil layers ; Fig. 4). The mean root depth for north-facing slopes was about 8 cm for A. deserti (Fig. 4 A), 11 cm for E. farinosa (Fig. 4 B) and 9 cm for P. rigida (Fig. 4 C) but 35, 26 and 27 % greater, respectively, for south-facing slopes (P ! 0±01 ; Table 3). Root surface area was also greater for plants on southfacing compared with north-facing slopes, averaging 30 % for A. deserti, 29 % for E. farinosa and 26 % for P. rigida (P ! 0±05 ; Table 3). Nobel and Linton—Frequencies on North- vs. South-facing Slopes T 4. New root area on relatiely small, isolated plants of similar size for 20³2° north-facing and south-facing slopes at 820³10 m Area of new roots (% of total root surface area) Slope A. deserti E. farinosa P. rigida North South 0±20³0±06 2±16³0±37 1±84³0±29 4±88³0±42 0±00³0±00 0±18³0±04 Data are for plants of the size ranges in Fig. 4 and Table 3 were obtained 28 Feb.–1 Mar. 1997 and are means³s.e. (n ¯ 9 or 10 plants). The area of new roots produced in response to a winter wet period ranged from none for P. rigida on north-facing slopes to nearly 5 % of total root surface area for E. farinosa on south-facing slopes (Table 4). For the two slopes, new roots averaged 1±18 % of the total root surface area for A. deserti, 3±36 % for E. farinosa and 0±09 % for P. rigida. In all cases, the proportion of new roots was higher for southfacing compared with north-facing slopes (P ! 0±001 for A. deserti and E. farinosa, P ! 0±01 for P. rigida ; Table 4). DISCUSSION The frequency of the C subshrub Encelia farinosa sub$ stantially increased from 20° north-facing to 20° southfacing slopes at 820 m in the northwestern Sonoran Desert, whereas the C bunchgrass Pleuraphis rigida and the CAM % leaf succulent Agae deserti changed relatively little in frequency. The south-facing slopes had similar air temperatures but higher soil temperatures and lower soil water availability than the north-facing slopes. The change from a 20° to a 36° north-facing slope, which should lead to less water availability due to greater runoff and slightly lower soil temperatures, was accompanied by a major decrease in the ground cover only for P. rigida. Comparing 20° northfacing slopes at 820 m s. 520 m, the approx. 2±6 °C higher temperature and 30 % lower annual rainfall at the lower elevation (Nobel and Hartsock, 1986) were accompanied by major reductions in the frequency of A. deserti and P. rigida but increases in E. farinosa for this more arid location. The average relative growth of new roots was consistent with observations for plants in flat ground, when a winter wet period with more rainfall (169 mm) led to new roots averaging 3±4 % of the total root surface area for A. deserti, 9±0 % for E. farinosa and 0±9 % for P. rigida (Nobel, 1997). The large size of the leaves of A. deserti and its nocturnal stomatal opening caused its leaf temperatures to be below ambient air temperatures at night and to be above near midday, consistent with previous results (Nobel, 1988). The higher soil water potential for 20° north-facing compared with south-facing slopes, and the resulting differences in transpiration, apparently accounted for the lower midday leaf temperatures for E. farinosa on north-facing s. southfacing slopes. Yet because the plants were relatively far apart, leading to a total ground cover averaging only 29 %, 737 interspecific competition for soil water and for light was probably less than for most other ecosystems (Gates, 1980 ; Monson et al., 1992). The increases in frequency, canopy size, plant height and ground cover for E. farinosa in going from a 20° north-facing slope at 820 m to a 20° south-facing slope or to a 20° north-facing slope at a lower elevation (520 m) were presumably not caused by less soil water or by slight changes in PPF. Because its leaf temperatures at midcanopy and the air temperatures 30 cm above the ground were not appreciably affected by the aspect of the 20° slopes, attention is focused on soil temperatures. Because few roots occurred near the soil surface except for approximately vertical ones directly under the plants, the soil surface temperature should be less crucial for root deployment than the soil temperature 10 cm below the surface, which is near the centre of the rooting zone for all three species. Near the beginning and the end of the winter wet period of 16 Dec. 1995 to 2 Mar. 1996, soil temperatures at this depth averaged 12 °C on north-facing slopes and 18 °C on south-facing slopes. Compared with root growth at the optimal temperature of 25 °C for E. farinosa under winter conditions, growth at 12 °C is only 11 % of maximal while that at 18 °C is 58 % of maximal (Drennan and Nobel, 1996). In this regard, approximately three-fold more new roots of E. farinosa were observed for south-facing compared with north-facing slopes at the end of the winter wet period in 1997. For the 2±6 °C higher ambient temperatures at the lower elevation, winter root growth at 10 cm would be 31 % of optimal, so plant success measured by ground cover—if dictated by winter root growth—would be expected to be intermediate to that for the two slopes at higher elevation ; this was only the case for E. farinosa. Using a model to predict soil temperature on 15 September as a function of depth and soil surface temperatures (Nobel, 1991 b), along with recorded air temperatures at 820 m, the properties of the local soil (Nobel and Geller, 1987) together with soil temperatures measured on 15 Jun. 1996, the soil temperature 10 cm below the soil surface during a typical summer wet period averaged 34±1 °C on 20° north-facing slopes and 36±0 °C on 20° south-facing slopes. Under summertime conditions, root growth of E. farinosa at this depth would be 43 % of maximal on the north-facing slopes and 38 % on the south-facing slopes (Drennan and Nobel, 1996) ; the mean duration of soil water availability and hence root growth is less during the summer (Young and Nobel, 1986 ; Nobel, 1988). Indeed, production of new roots per plant is five-fold greater in the winter than in the summer for E. farinosa in relatively flat ground at 820 m (Nobel, 1997). On the other hand, optimal temperatures for root growth are higher for the other two species, averaging 31 °C for A. deserti and 36 °C for P. rigida (Drennan and Nobel, 1996), so they can have greater root growth during summer wet periods than E. farinosa. In particular, for plants in flat ground A. deserti has approximately equal production of new roots in the winter and the summer and P. rigida has six-fold greater root production in the summer (Nobel, 1997). For plants of the same shoot size on the two slopes at 820 m, roots for A. deserti, E. farinosa and P. rigida 738 Nobel and Linton—Frequencies on North- vs. South-facing Slopes averaged 28 % greater in area and were deeper on 20° southfacing compared with 20° north-facing slopes. Root systems of Lupinus albus (Rodrigues, Pacheco and Chaves, 1995), Picea abies (Persson et al., 1995) and other species (Nye and Tinker, 1977) also tend to be more extensive vertically and larger in surface area when the plants are droughted. Thus the deeper roots with more surface area per unit shoot area for plants on south-facing slopes for the three species are consistent with the more arid conditions there than on north-facing slopes. The changes in distribution of P. rigida with slope and elevation may relate to the observation that most of its leaves are brown and apparently dead during much of the year at 820 m. Although it can exhibit an extremely high photosynthetic rate of up to 67 µmol m−# s−" (Nobel, 1980), it appears to be only marginally able to cope with the local conditions. Also, its leaves take approx. 3 weeks to develop a substantial surface area following a wet period (P. S. Nobel, unpubl. res.), so its net CO uptake capability does # not respond as quickly to rainfall as that of E. farinosa or A. deserti, the latter having evergreen leaves that persist for up to 6 years, with plant water uptake occurring within 24 h of rainfall (Nobel, 1976, 1988). Following light rainfalls, the soil may therefore dry before leaves of P. rigida can develop. During the winter wet period in 1996 s. the drier one in 1997 (169 s. 52 mm rainfall), P. rigida developed ten-fold more new roots compared with less than three-fold more for the other two species. Also, P. rigida has the lowest plant water storage capacity of the three codominants (Nobel and Jordan, 1983), so it may be the most vulnerable to low soil water availability. In this regard, its ground cover at 820 m decreased at least two-fold in going from a 20° north-facing slope to a 20° south-facing slope, to a steeper (36°) northfacing slope, or to one 300-m-lower in elevation, all of which represent drier soil conditions. Individual clumps of P. rigida had over twice the canopy area for the wetter 20° north-facing compared with south-facing slopes at 820 m, again suggesting that small changes in soil water availability may be crucial. The percent ground cover for A. deserti was similar for the 20° and 36° north-facing slopes and the 20° south-facing slopes at 820 m but decreased for 20° north-facing slopes at 300-m-lower elevation. Because the orientation of the rosette of rigid massive leaves was dependent on the local slope, leaves of A. deserti on 20° south-facing slopes intercepted 34 % more PPF in the late winter than those on 20° north-facing slopes (PPF interception varied only 6 to 14 % for the two slopes for the other two species, whose flexible leaves are not as spatially affected by the slope). This may help to account for the larger mean shoot size for A. deserti on south-facing slopes despite lower water availability, as net CO uptake by this species is PPF-limited, # even in the high radiation environment of a desert (Nobel, 1984, 1988), so increased CO uptake capability may lead to # larger plants. The decrease in frequency of A. deserti from north-facing to south-facing slopes may be in response to the more arid conditions on the south-facing slopes, with the higher soil temperature playing a minor role. The lower ground cover for A. deserti at the lower elevation may also be in response to the more arid conditions. The effects of soil water might be expected to be least for the CAM species with its higher water-use efficiency than the C or C species, $ % but water is crucial for seedling establishment by A. deserti. In particular, only 1 year in a span of 17 years at 820 m had sufficient rainfall followed by short enough drought periods for its seedlings to become established (Jordan and Nobel, 1979), and over 99 % of the reproduction of A. deserti is vegetative, by means of ramets (Nobel, 1988). Thus its longterm success may be dictated by soil water availability at the seedling stage, making its frequency and hence its percent ground cover particularly vulnerable to local soil water content. In conclusion, E. farinosa at the sites considered was apparently most affected by soil temperatures and root growth during the winter, P. rigida by its inability to tolerate more arid conditions, and A. deserti also by soil water availability, probably during seedling establishment. Therefore, soil temperature and water availability are apparently more crucial than photosynthetic pathway or growth form for the relative success of the three codominants as measured by percent ground cover for various slopes and elevations in this desert ecosystem. Such responses emphasize the importance of root properties and can have important implications for predicting altered plant frequencies resulting from new environmental conditions, such as those accompanying global climate change. A C K N O W L E D G E M E N TS Preliminary data were collected in 1995 by Xilian Yue for microclimate and Dr Philippa Drennan for plant distribution. Much of the data from 1–3 Mar. 1996 was collected by Aaron De Hart, Jane George, Reid Hamamoto, Co Lai, Andrew Na, Dang Ngo, Deimkanh Nguyen, Maurice Pitesky, Sasan Sanandaji and Kiki Zaharopoulos ; Barry Greenhouse and Amir Lagstein collected data on that trip and also during a subsequent one. Much of the data on new roots was collected by Edward Bobich, Paul Chang and Yong Chung on 28 Feb.–1 Mar. 1997. This research was financially supported by the US Department of Energy Office of Health and Environmental Research, Program for Ecosystem Research, grant DE-FG03-93ER61686 and US National Science Foundation grant IBN-94-19844. LITERATURE CITED Adams RM, Rosenzweig C, Peart RM, Ritchie JT, McCarl BA, Glyer JD, Curry RB, Jones JW, Boote KJ, Allen LH Jr. 1990. Global change and US agriculture. Nature 345 : 219–224. Andrade JL, Nobel PS. 1996. Habitat, CO uptake and growth for the # CAM epiphytic cactus Epiphyllum phyllanthus in a Panamanian tropical forest. Journal of Tropical Ecology 12 : 291–306. Beadle CL, Long SP, Imbamba SK, Hall DO, Olembo RJ. 1985. Photosynthesis in relation to plant production in terrestrial enironments. Oxford : Tycooly. Beatley JC. 1974. Phenological events and their environmental triggers in Mojave Desert ecosystems. Ecology 55 : 856–863. Bonham CD. 1989. Measurements for terrestrial egetation. New York : John Wiley. Drennan PM, Nobel PS. 1996. Temperature influences on root growth for Encelia farinosa (Asteraceae), Pleuraphis rigida (Poaceae), and Agae deserti (Agavaceae) under current and doubled CO # concentrations. American Journal of Botany 83 : 133–139. Nobel and Linton—Frequencies on North- vs. South-facing Slopes Drennan PM, Nobel PS. 1997. Frequencies of major C , C , and CAM $ % perennials on different slopes in the northwestern Sonoran Desert. Flora 192 (in press). Ehleringer J. 1985. Comparative microclimatology and plant responses in Encelia species from contrasting habitats. Journal of Arid Enironments 8 : 45–56. Ehleringer JR, Bjo$ rkman O. 1978. A comparison of photosynthetic characteristics of Encelia species possessing glabrous and pubescent leaves. Plant Physiology 62 : 85–190. Ehleringer JR, Monson RK. 1993. Evolutionary and ecological aspects of photosynthetic pathway variation. Annual Reiew of Ecology and Systematics 24 : 411–439. Gates DM. 1980. Biophysical ecology. New York : Springer. Gibson AC. 1996. Structure-function relations of warm desert plants. Berlin : Springer. Greig-Smith P. 1983. Quantitatie plant ecology. Studies in ecology, ol. 9. Oxford : Blackwell Scientific. Hickman JC. 1993. The Jepson manual : higher plants of California. Berkeley and Los Angeles, California : University of California Press. Jordan PW, Nobel PS. 1979. Infrequent establishment of seedlings of Agae deserti (Agavaceae) in the northwestern Sonoran Desert. American Journal of Botany 66 : 1079–1084. Leith H, Whittaker RH. 1975. Primary productiity of the biosphere. Berlin : Springer. McGinnies WG, Goldman BJ, Paylore P. 1968. Deserts of the World : An appraisal of research into their physical and biological enironments. Tucson, Arizona : University of Arizona Press. Monson RK. 1989. On the evolutionary pathways resulting in C % photosynthesis and crassulacean acid metabolism (CAM). Adances in Ecological Research 19 : 57–110. Monson RK, Smith SD, Gehring JL, Bowlan WD, Szarek SR. 1992. Physiological differentiation within an Encelia farinosa population along a short topographical gradient in the Sonoran Desert. Functional Ecology 6 : 751–759. Nobel PS. 1976. Water relations and photosynthesis of a desert CAM plant, Agae deserti. Plant Physiology 58 : 576–582. Nobel PS. 1980. Water vapor conductance and CO uptake for leaves # of a C desert grass, Hilaria rigida. Ecology 61 : 252–258. % Nobel PS. 1984. Productivity of Agae deserti : measurement by dry weight and monthly prediction using physiological responses to environmental parameters. Oecologia 64 : 1–7. Nobel PS. 1988. Enironmental biology of agaes and cacti. New York : Oxford University Press. 739 Nobel PS. 1991 a. Tansley review no. 32. Achievable productivities of CAM plants : basis for high values compared with C and C $ % plants. New Phytologist 119 : 183–205. Nobel PS. 1991 b. Physicochemical and enironmental plant physiology. San Diego, California : Academic Press. Nobel PS. 1997. Root distribution and seasonal production in the northwestern Sonoran Desert for a C subshrub, a C bunchgrass, $ % and a CAM leaf succulent. American Journal of Botany 84 : 949–955. Nobel PS, Geller GN. 1987. Temperature modelling of wet and dry desert soils. Journal of Ecology 75 : 247–258. Nobel PS, Hartsock TL. 1986. Temperature, water, and PAR influences on predicted and measured productivity of Agae deserti at various elevations. Oecologia 68 : 181–185. Nobel PS, Jordan PW. 1983. Transpiration stream of desert species : resistances and capacitances for a C , a C , and a CAM plant. $ % Journal of Experimental Botany 34 : 1379–1391. Noy-Meir I. 1973. Desert ecosystems : environment and producers. Annual Reiew of Ecology and Systematics 4 : 25–51. Nye PH, Tinker PB. 1977. Solute moement in the soil-root system. Berkeley and Los Angeles, California : University of California Press. Ode DS, Tieszen LL, Lerman JC. 1980. The seasonal contributions of C and C plant species to primary production in a mixed prairie. $ % Ecology 61 : 1304–1311. Persson H, Von Fircks Y, Majdi H, Nilsson LO. 1995. Root distribution in a Norway spruce (Picea abies (L) Karst) stand subjected to drought and ammonium-sulphate application. Plant and Soil 168–169 : 161–165. Rodrigues ML, Pacheco CMA, Chaves MM. 1995. Soil-plant water relations, root distribution and biomass partitioning in Lupinus albus L. under drought conditions. Journal of Experimental Botany 46 : 947–956. Rundel PW, Nobel PS. 1991. Structure and function in desert root systems. In : Atkinson D, ed. Plant root growth—an ecological perspectie. Oxford : Blackwell Scientific, 349–378. Smith SD, Nobel PS. 1986. Deserts. In : Baker NR, Long SP, eds. Photosynthesis in contrasting enironments. Amsterdam : Elsevier, 13–62. Walter H. 1985. Vegetation of the Earth and ecological systems of the geo-biosphere, 3rd edn. Berlin : Springer. Winter K, Smith JAC. 1996. Crassulacean acid metabolism. Biochemistry, ecophysiology and eolution. Berlin : Springer. Young DR, Nobel PS. 1986. Predictions of soil-water potentials in the north-western Sonoran Desert. Journal of Ecology 74 : 143–154.