And Sarcobatus Vermiculatus (Chenopodiaceae)1

AmericanJournal of Botany 83(12): 1637-1646. 1996.

WATER RELATIONS AND LEAF CHEMISTRY OF CHRYSOTHAMNUS NAUSEOSUS SSP. CONSIMILIS (ASTERACEAE) AND SARCOBATUS VERMICULATUS (CHENOPODIACEAE)1

LISA A. DONOVAN,2 JAMES H. RICHARDS, AND MATTHEW W. MULLER

Departmentof Land, Air, and WaterResources, Universityof California,Davis, California95616-8627

At Mono Lake, California,we investigatedfield water relations,leaf and xylem chemistry,and gas exchange for two shrubspecies thatcommonly co-occur on marginallysaline soils, and have similarlife historiesand rootingpatterns. Both species had highestroot lengthdensities close to the surfaceand have large tap roots thatprobably reach groundwater at 3.4-5.0 m on the studysite. The species differedgreatly in leaf waterrelations and leaf chemistry.Sarcobatus vermiculatus

had a seasonal minimumpredawn xylem pressurepotential (Wpd) of -2.7 MPa and a midday potential(Wind) of -4.1 MPa. These were significantlylower thanfor Chrysothamnus nauseosus, whichhad a minimumWpd of -1.0 MPa and Wind of -2.2 MPa. Sarcobatus had leaf Na of up to 9.1 % and K up to 2.7 % of dry mass, and these were significantlyhigher than for Chrysothamnuswhich had seasonal maxima of 0.4% leaf Na and 2.4 % leaf K. The molar ratiosof leaf K/Na, Ca/Na, and Mg/Na were substantiallylower forSarcobatus than for Chrysothamnus.Xylem ionic contentsindicated that both species excluded some Na at the root,but thatChrysothamnus was excluding much more thanSarcobatus. The higherNa content of Sarcobatus leaves was associated with greaterleaf succulence, lower calculated osmotic potential,and lower xylem pressurepotentials. Despite large differencesin waterrelations and leaf chemistry,these species maintainedsimilar diurnal patternsand rates of photosynthesisand stomatalconductance to watervapor diffusion.Sarcobatus Wpdmay not reflectsoil moistureavailability due to root osmotic and hydraulicproperties.

Key words: Asteraceae; Chenopodiaceae; halophyte;nutrition; photosynthesis; root distribution; salinity; sodium; water relations.

In cold desertsof westernNorth America, plant distri- summer,maintain high photosynthetic rates through sum- bution on the landscape is often closely related to soil mer,flower in mid-latesummer, and senesce in late fall chemistry(Billings, 1949, 1951; West, 1988; Comstock afterproducing large numbersof wind-dispersedseeds and Ehleringer,1992). In general,saline and alkalinesoils (Roundy, Young, and Evans, 1981; Robertson, 1983; at lower topographicpositions in basins between moun- Romo and Haferkamp,1989; Donovan and Ehleringer, tain ranges are occupied by shrub species in the family 1994a; Fort and Richards, 1995). The consimilissubspe- Chenopodiaceae, while nonsaline,less alkaline soils are cies of Chrysothamnusis thoughtto be somewhattolerant occupied by shrubsfrom the familyAsteraceae (Gates, of alkaline and saline soils (Robinson, 1958; Rollins, Stoddart,and Cook, 1956; West, 1988). At many loca- Dylla, and Eckert, 1968; Roundy, Young, and Evans, tions, however, species from these families co-occur 1981), whereasSarcobatus is very salt and alkali tolerant along broad gradients,and one commonlyfound associ- (Shantz and Piemeisel, 1940; Fireman and Hayward, ation is Chrysothamnusnauseosus (Palla.) Britt.ssp. con- 1952; Gates, Stoddart,and Cook, 1956). Both shrubspe- similis (E. Greene) H.M. Hall & Clements (Asteraceae; cies are phreatophytic,indicating that they are generally common name, salt rabbitbrush)and Sarcobatus vermi- rooted down to the capillary fringeof groundwaterat culatus (Hook.) Torrey(Chenopodiaceae; commonname, depthsof 2-12 m, and hence should have similaraccess greasewood) (Shantz and Piemeisel, 1940; Robinson, to water (Robinson, 1958; Rollins, Dylla, and Eckert, 1958; Wiebe and Walter, 1972; Roundy, Young, and 1968; Harr and Price, 1972; Branson, Miller, and Mc- Evans, 1981; Robertson,1983). These two species have Queen, 1976; Rickard and Warren,1981). However,both similar life histories.They are winterdeciduous shrubs species have been shown to respond to severe drought withthe C3 photosyntheticpathway. Both produce leaves with eitherleaf drop, reduced canopy size, or increased in early spring,have highestshoot growthrates in early mortality,indicating that access to wateris not unlimited (Rickard and Warren,1981; Toft,1995). Relationships to soil chemistryand water potentials 1 Manuscriptreceived 5 January1996; revisionaccepted 3 July1996. The authorsthank E. J. Schaber,G. Kyser,G. Besne, K. Fort,and J. have been investigatedfor both Chrysothamnusand Sar- Hurrell for assistance with field work and with soil, plant, and water cobatus, but only fornon-co-occurring populations. Sar- chemistry measurements.This research was supported by USDA cobatus can accumulateNa in its leaves, and createa salt- NRICGP grants 92-37101-7419 and 94-37101-1144. Access to the enrichedmicrosite under its canopy due to leaching of study site was provided by the USFS througha Special Use Permit. salt fromshed leaves. In contrast,soils underthe canopy Cooperationof theMono Lake RangerDistrict and LarryFord is greatly appreciated. of Chrysothamnusand other shrubs in the Asteraceae, 2 Authorfor correspondence, current address: Departmentof Botany, such as Artemisiatridentata, do not differin salt content Universityof Georgia, Athens,GA 30602-7271. from soils in interspacesbetween shrubs (Fireman and 1637 1638 AMERICAN JOURNAL OF BOTANY [Vol. 83

Hayward, 1952; Rickard, 1964, 1965; McNulty, 1969; mined. On 20 May, 7 July,17 August, and 20 September,stems were Rickard,Cline, and Gilbert,1973; Wallace, Romney,and also measured for predawn xylem pressurepotential (Wpd). Additional Hale, 1973; Cline and Rickard, 1974; Glenn and diurnal ' measurementswere made on 7 Julyand 17 August, to ac- O'Leary, 1984; Eddleman and Romo, 1987). Leaf ele- company measurementsof gas exchange. Rates of photosynthesis(A) mentalcontents for Chrysothamnuson saline soils have and stomatalconductance (g) were determinedfor stems with leaves, not been reported,but other species in the same family witha LI-COR 6200 portablephotosynthesis system (LI-COR., Lincoln, have been shown to accumulate some Na (Glenn and NE) and a quarter-litrechamber. The leaf-to-airvapor pressuredeficit O'Leary, 1984). Field values of xylempressure potentials was calculatedfrom air temperatureand relativehumidity assessed with for adult Chrysothamnusrange from-1.0 to -4.0 MPa the chamber open to the atmosphere,and leaf temperatureswere as- (Branson,Miller, and McQueen, 1976; Donovan and Eh- sumed to be the same as air temperaturebecause both shrub species have small linear leaves. For individualplants, the maximumrate of A leringer,1994b), and values of xylem pressurepotentials on each date was used to calculate Alg and instantaneouswater-use for Sarcobatus are generallymore negative,reported to efficiencyor WUE (AIE, mmol C02/molH20, whereE is transpiration), range from -1.5 to -6.5 MPa (Detling and Klikhoff, and photosyntheticnitrogen-use efficiency or NUE (A/N,pLmol C02 mol 1973; Branson, Miller, and McQueen, 1976; Romo and N-'-s-', where N is leaf nitrogencontent per unit area) (Field, Merino Haferkamp,1989). Although xylem pressure potentials and Mooney, 1983). Leaf areas were measured with a Delta-T system forwoody plants are generallyrelated to minimumwater (Decagon Devices, Pullman,WA). availability in the soil column (Branson, Miller, and Leaf chemistrywas determinedfor mature leaves thatwere collected McQueen, 1976; Hinckley,Lassoie, and Running,1978; fromeach plant at the time of Wpd samplingin May, July,and August. Davis and Mooney, 1986), xylem pressurepotentials of Afterair dryingat 60?C, a portionof groundleaf materialwas extracted Chrysothamnusand Sarcobatus appear to be independent with acid dissolutionand microwavedigestion (Sah and Miller, 1992). of moisturein surface soil layers, presumablybecause Extractswere analyzed forNa, Ca, K, P, B, S, and Mg on the ICP-AES. they are phreatophytic.When co-occurringshrubs are The remainingleaf tissue was analyzed for percentageN by Dumas phreatophytic,differences in water relations and leaf combustionwith a Carlo Erba NA1500 elementalanalyzer (Milan, It- chemistrymay reflectphysiological responses to ground- aly), and forpercentage ash by combustionfor 6 h at 520?C. Leaf water water and soil chemistry,not differencesin rooting and mass per area characteristicswere determinedfor leaves priorto depthsand soil moistureavailability. senescence in early October.Ten leaves were collected at sunrisefrom In this studywe comparedthe seasonal waterrelations each shrub,and freshmass, area, and dry mass were determined.Es- and leaf chemistryof matureindividuals of Chrysotham- timatesof leaf osmotic potentialswere calculated for each species on nus and Sarcobatus where they co-occur on dunes near the basis of leaf Na, K, H20, and dry mass (Storey,Ahmad, and Wyn Mono Lake, CA. Although these shrubs have similar Jones, 1977). Xylem sap was extractedfrom 0.3 m long, 0.5 cm diameter,woody rootingzones and similarlife histories,we expectedthem stem segmentscollected at sunriseand at midday on 17 August from to differin seasonal water relations and leaf chemistry four shrubs of each species. Upon collection, stem segments were due to theirdifferent physiological mechanisms for salin- wrapped in parafilmand stored cold until extractionof sap within6- ity tolerance.We assessed the causes and consequences 48 h aftersevering. For extraction,bark and phloemwere removedfrom of differentxylem pressurepotentials by examiningdi- each end of the stem.Plastic tubingwas attachedto the top of the stem, urnal gas exchange, xylem chemistry,root distribution, and a 2-m column of water and acid fuchsindye solution (0.5%) was and soil and groundwaterchemistry. used to displace xylem sap. Volume collected varied from2 to 9 mL. Evaporative loss duringcollection was minimizedby exposing sap to MATERIALS AND METHODS a small volume of air in parafilm-sealed,15-mL vials, and by keeping collectionvials on ice. Xylem extractswere analyzed forNa, Ca, K, P, The studysite (380 5' N, 1180 58' W, 1958 m elevation) is located B, S, and Mg on the ICP-AES. For each species, morningand midday -2 km fromthe northeastshore of Mono Lake, CA. The climate is samples did not differsignificantly for any element(P > 0.05 for all arid, with annual precipitationat the site averaging 160 mm (Toft, comparisons),and data fromthe two collection times were combined 1995). From October 1993 to March 1994 precipitationtotalled 93 mm for subsequentstatistical analysis. and an additional 54 mm fell at the beginningof the growingseason Root densityand depth distributionwere determinedin pits at the in April and May 1994. No precipitationoccurred during the summer canopy edge of each shrubat the end of the growingseason. Soil cores (June-August1994) when the studywas conducted.The site was a 100 (180 cm3) were taken at 0.25, 0.50, 0.75, and 1.0 m depths and roots x 100 m area of late Holocene semistabilizeddunes (describedby Toft, were elutriatedonto fine screens. Root lengthswere determinedfor a 1995). At the site 2.0-3.6 m of aeolian sand overlies lacustrineclay, subset of cores with a Comair root lengthscanner (Hawker De Havil- with the free groundwatersurface at depths of 3.4-5.0 m, depending land VictoriaLimited, Melbourne, Australia). Root dry masses after48 on the thicknessof the overlyingsand deposit. Six interspecificpairs h at 60?C, were determinedfor all samples. Root mass per volume of of Chrysothamnusand Sarcobatus shrubswere chosen withinthe site. soil (rwd; g/m3)was convertedto root lengthdensity (rld; m/m3)from Each shrub was a discreteindividual, separate fromother individuals the regressions:for Chrysothamnusrld = rwd-165.87 - 1.82 (N = 12, by at least 2 m, and shrubswere paired for canopy size, measured as r2 = 0.86, P < 0.05), and for Sarcobatus rld=rwd 146.52 - 0.93 (N heightand canopy volume. In spring 1994, soils were collected at 50 = 13, r2 = 0.87, P < 0.05). cm depths froma hole augured to groundwaterin a shrubinterspace. Data were analyzed with repeated-measuresanalysis of variance for Electricalconductivity (EC, dS/m)and pH were determinedfor ground- variables thatinvolved repeatedsampling of individualplants through water and saturatedsoil extracts.Groundwater was also analyzed for the season, throughthe day, or at several depths,and for which there Na, Ca, K, P, B, S, and Mg on an inductivelycoupled plasma-atomic were few missing values, i.e., seasonal xylem pressurepotentials and emission spectrophotometer(ICP-AES; Thermo JarrellAsh Corp., leaf chemistry,diurnal xylem pressure potentials, rates of gas exchange, Franklin,MA). and root lengthdensity (Gurevitch and Chester,1986; SAS, 1989; Pot- Shrub stem xylem pressurepotentials were determinedwith a pres- vin, Lechowicz, and Tardif,1990; von Ende, 1993). Between-subject sure chamber(PMS Inc., Corvallis, OR). When plantswere stillleafless effectswere block (shrubspaired forcanopy size) and species. Within- on 22 April, only midday xylem pressurepotentials (WJid) were deter- subject effectswere repeated interval(either month,hour, or depth), December 1996] DONOVAN ET AL.-DESERT SHRUB WATER RELATIONS AND LEAF CHEMISTRY 1639

TABLE 1. One-way analysis of variance comparisons between Chrysothamnusnauseosus and Sarcobatus vermiculatusgrowing at Mono Lake, CA in 1994 forplant canopy size, leaf ash content,leaf mass and area (N = 6), and xylem sap chemistry(N = 4). For each variable: mean + 1 SE, F statistic and significance. * P < 0.05, ** P < 0.01, *** P < 0.001, and NS is not significant.

Chrysothamnus Sarcobatus F Canopy size (18 August) Canopy height(m) 1.3 + 0.2 1.3 + 0.4 0.1NS Canopy volume (m3) 3.6 + 1.0 4.7 + 2.8 0.6NS Leaf ash content 20 May (%) 16.6 + 6.5 22.4 + 4.2 0.3NS 7 July(%) 11.9 + 3.4 29.1 ? 1.1 23.3*** 17 August (%) 9.9 + 0.6 33.3 + 0.6 694.3*** Leaf mass and area (10 October) Fresh mass/area(g/m2) 492.3 + 16.1 1125.7 + 61.3 99.9*** Dry mass/area(g/m2) 207.0 ? 6.0 212.3 ? 8.0 0.3NS Organic dry mass/area(g/m2) 186.5 ? 5.5 142.3 + 7.7 48.4*** Water/area(g/m2) 283.0 t 16.0 905.0 + 58.0 196.1*** Fresh mass/drymass 2.4 ? 0.1 5.3 + 0.2 235.7*** Xylem chemistry(17 August) B (mg/L) 3.2 ? 0.4 3.7 + 0.6 0.6NS Ca (mg/L) 66.2 ? 5.7 21.2 + 6.1 29.2*** K (mg/L) 219.2 + 18.6 227.5 ? 46.8 0.3NS Mg (mg/L) 21.5 + 1.73 12.8 + 1.1 18.0*** Na (mg/L) 21.6 ? 4.2 141.9 + 16.2 83.0*** P (mg/L) 13.3 ? 1.4 17.3 ? 4.4 1.1NS S (mg/L) 5.9 ? 0.9 17.7 + 3.5 22.1***

intervalby species interactions,and intervalby block interactions.Uni- for percentash on each sampling date because it had many missing variate analyses and Huynh-Feldt(H-F) adjusted P values were used values and hence was inappropriatefor repeated-measuresanalysis for testingwithin-subject effects. Mauchly's criterionwas used to test (SAS, 1989). Variables were log transformed,when necessary,to meet for sphericityand indicated no significantdeparture from compound assumptionsof normalityand equivalentvariance of residuals.For these symmetryfor most variables. For the variables where therewas a sig- variables, back-transformedmeans and standarderrors (averaged ) are nificantdeparture (photosynthetic rates in July and August, and leaf presented. chemistrymolar ratios), the results of the H-F adjusted P concurred Relationshipsbetween photosyntheticrates and conductances were with unadjusted P values (von Ende, 1993). For all variables except determinedwith a linearregression and subsequentF testfor each spe- root lengthdensity, the block effect(between subject), and block by cies on each samplingdate. Differencesin regressioncoefficients were intervalinteraction effect (within subject) were notsignificant (P > 0.05 determinedwith the Tukey-Kramermethod (Sokal and Rohlf, 1981). for all analyses) and these block effectsare not presentedin the ANO- VA tables. RESULTS One-way analysis of variance was used to determinespecies differences for variables sampled only on one date, i.e., canopy size, xylem The depth of the free groundwatersurface did not chemistry,and leaf mass characters.One-way ANOVA was also used change throughthe 1994 growingseason, remaining3.4- 5.0 m below the sand surface,depending on the sand depthat the sampled shrubs.Groundwater contained 475 n^ I . ~~I ~ ~I ~ ~I E mg/L (ppm) Na, 3 mg/L B, 5 mg/L Ca, 47 mg/L K, 1 mg/LP, and 51 mg/LS, and had an EC of 1.5 dS/mand 10 Sarcobatusvermiculatus pH of 8.6. EC of saturatedsoil extractsranged from1.0 to 1.5 dS/m and pH ranged from7.8 to 9.2. Since most of the salt in the soil at this site is NaCl (J.H. Richards and L.A. Donovan, unpublisheddata), the soil solutions D5100 at saturationhad a range of Na concentrationsfrom 230 to 345 mg/Land osmoticpotentials from -0.05 to -0.07 MPa. The average heightand canopy volume of the repro- ductivelymature Chrysothamnus and Sarcobatus did not 0 Chrysothamnusnau differ(Table 1). The species differedin root lengthdensity,with Sarcobatus having46% higherroot length den- sity than Chrysothamnusin the 0-1.0 m rooting zone 25 50 75 100 (betweensubject, species effect),but both species showed a similarpattern of decrease in root lengthdensity with Depth(cm) increasingdepth (withinsubject, species by depth interaction) [(Fig. 1, Table 2)]. Roots of both species penetrate Fig. 1. Profile of root length densityin the 0-1 m soil layer for Chrysothamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, much deeper than 1 m and were occasionally found in CA at the end of the 1994 growingseason; mean ? 1 SE, N = 6 (for augered holes at 3-5 m depths. statisticssee Table 2). Chrysothamnusand Sarcobatus had differentseasonal 1640 AMERICAN JOURNAL OF BOTANY [Vol. 83

TABLE 2. Results of repeated measures analysis of variance for comparisons of characteristicsof Chrysothamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, CA in 1994. Between-subjecteffect is species, and within-subjecteffects are the repeatedinterval (month, time of day, or depth) and intervalby species interaction.Presented for each effectare an F statistic,significance (* P < 0.05, ** P < 0.01, *** P < 0.001), and degrees of freedomfor numeratorand denominator(df). Means and SEs forthe variables comparedhere are in Figs. 1-4.

Between-subjecteffect Within-subjecteffects

Species Month Month X species F df F df F df Season variables

4jpd 69.5*** 1,5 65.8*** 3,15 39.6*** 3,15 qind 297.5 * * * 1,5 63.7*** 4,20 21.2*** 4,20 4jmd4qpd 15.2* 1,5 10.5*** 3,15 6.8** 3,15 Leaf N 16.6** 1,5 232.8*** 2,10 1,3 2,10 Leaf B 5.1 1,5 31.4*** 2,10 2.6 2,10 Leaf Ca 0.1 1,5 17.8*** 2,10 0.2 2,10 Leaf K 6.6* 1,5 3.5 2,10 1.8 2,10 Leaf Mg 7.4* 1,5 12.0** 2,10 2.2 2,10 Leaf Na 335.7*** 1,5 67.7*** 2,10 84.2*** 2,10 Leaf P 51.5*** 1,5 13.1** 2,10 1.3 2,10 Leaf S 3.5 1.5 7.5 2,10 8.0** 2,10 Leaf K/Na 23.8** 1,5 2.5 2,10 3.1 2,10 Leaf Ca/Na 19.6** 1,5 5.4* 2,10 5.4* 2,10 Leaf Mg/Na 26.2** 1,5 8.9** 2,10 9.2** 2,10

Species Time of day Time of day X species F df F df F df Diurnal variables July4 131.7*** 1,5 225.5*** 5,25 30.0*** 5,25 JulyA 11.9* 1,5 27.2*** 3,15 1.2 3,15 Julyg 9.0* 1,5 19.8*** 3,15 2.0 3,15 Aug 4 416.8*** 1,5 133.5*** 5,25 2.0 5,25 Aug A 0.2 1,5 37.8*** 4,20 0.7 4,20 Aug g 1.6 1,5 12.6*** 4,20 3.8* 4,20 Species Depth Depth X species F df F df F df Spatial variable Root lengthdensity 7.9* 1,5 9.0*** 3,15 0.8 3,15

courses of gPpd, gmd' and gImd-lpd (the differencebetween ratios of K/Na, Ca/Na, and Mg/Na in leaf tissues (Table midday and predawnwater potentials), as indicatedby a 2). Throughoutthe season, K/Na molar ratioswere > 3.8 significantspecies effectin the between-subjectanalysis forChrysothamnus and < 0.3 forSarcobatus. In addition, (Fig. 2, Table 2). Chrysothamnushad consistentlymuch Ca/Na molar ratios were > 1.6 for Chrysothamnusand less negative gPpd and gPmd values throughoutthe season. < 0.06 for Sarcobatus, while Mg/Na molar ratios were The firstsampling date was priorto leaf out, and repre- > 0.9 forChrysothamnus and < 0.4 forSarcobatus. Spe- sented late-winterdormant conditions. After leaves were cies differencesin leaf chemistrywere generallyparal- produced in May, both species showed significantsea- leled by differencesin xylem sap chemistryin August: sonal declines in both gPpd and kPmd (within-subject,time Chrysothamnushad much lower xylem Na and S than effect),but the shapes of the seasonal curves were dif- Sarcobatus, while the reverse was true for Ca and Mg ferentas indicatedby a significantspecies by timeinter- (Table 1). The species did not differin B, K, and P con- action in the within-subjectanalysis (Table 2). tentsin the xylem sap. The seasonal species differencesin xylempressure po- Sarcobatus leaves were more succulentthan those of tentialswere accompanied by seasonal differencesin leaf Chrysothamnus,based on a higherfresh mass/area ratio, chemistryfor N, P, K, Mg, and Na, but not for Ca, B, higher H20/area ratio, and lower organic dry mass per and S (Fig. 3, Table 2). Chrysothamnushad higherleaf unit area (Table 1). The greaterleaf succulence of Sar- N, P, and Mg, but lower leaf K and Na thanSarcobatus. cobatus is consistentwith the higher leaf Na and ash There was a significantseasonal (month) effectfor all contentsper unit leaf area. Calculated leaf osmotic po- elementsexcept K and S (even when the species effects tential,based on totalleaf Na, K, and waterper gramdry were not significant)and a significanttime by species mass and ignoringcompartmentation effects, were -1.8 interactionfor Na. Sarcobatus leaves accumulatedNa to and -4.6 MPa for Chrysothamnusand Sarcobatus, re- 9.1 % of dry mass while ChrysothamnusNa contentfell spectively. throughthe season to 0.2 % (Fig. 3). Na was theprimary Diurnal courses of 4, A, and g differedbetween Chrys- cation contributingto the significantspecies differencein othamnusand Sarcobatus in July,but in August only 4 ash contentby the end of the growing season (Fig. 3, was significantlydifferent (Fig. 4, Table 2). Both species Table 1). The species also differedsignificantly for molar had significantdifferences associated withtime forall of December 1996] DONOVAN ET AL.- DESERT SHRUB WATER RELATIONS AND LEAF CHEMISTRY 1641

0O I I - I 1 1- 1 out the season. The large differencesin 4pd, especially 0~~~~~ late in the summer (Fig. 2), between Sarcobatus and Chrysothamnuswould conventionallybe interpretedas -1.0 -i- an indicationof differentialaccess to moist soil layers. This does not appear to be the case in thissituation, however. Rather,evidence fromthis study,the literature,and CL-2.0\ stable hydrogen isotope analyses suggests that these co-occurringshrub species have root systemsthat are exposed to a similarbelowground environment in termsof -3.0 water availability (and soil chemistry,discussed later) and have access to the waterin the capillaryfringe when soil moistureis not available in the overlyingsands. Both species have theirhighest root lengthdensity in the shal- ~ -4.0 lower soils (Fig. 1) but are reportedto root to ground- 01? I --I I 1 1- waterwhen it occurs withinthe depthrange foundin this 0 study ._. (Robinson, 1958; Rollins, Dylla, and Eckert,1968; Harr and Price, 1972; Branson, Miller, and McQueen, E Ch-0 Chaysothamnus nauseosus 1976; Rickard and Warren,1981). As at otherlocations, both species have massive tap roots at the studysite, and we have found roots in soil cores frommoist lacustrine -2.0 clay layersjust above the groundwaterlevel. Woody Sar- cobatus roots (> 4 mm diameter) in such cores were identifiedby color and a distinctivewhite starchylayer CL -3.0 beneaththe bark. Other woody rootsfound at depthcould E notpositively be identifiedas Chrysothamnus,but no oth- 0 Sarcobatus vermiculatus at Moa,A er shrubs occur at the sites where the deep holes were augered. Psychrometricmeasurements of soil water po- >< -4 .0 ______tential and neutronprobe determinationsof soil water contentat this and othersites on the dunes indicatethat, LM ~ ~ ~ ~ ot while moistureis available in the shallowersoils earlyin _0. the growing season, duringJuly and August soil water potentialsare -4.0 to -7.0 MPa or lower fromthe surface to 2 m depth(Muller, Richards, and Donovan, 1995). -2.0~ Soil moisture,however, is available in deeper soil layers (- -2.2 MPa at 2 m or deeper) or from groundwater E Apr May Jun Jul Aug Sep throughoutthe growingseason. In addition,preliminary Month determinationsof the hydrogenisotopic compositionof water in the stem xylem of both species in midsummer Fig. 2. Seasonal courses of xylem pressurepotentials for Chryso- are not distinguishablefrom the isotopic compositionof thamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, CA in the groundwaterat the site. All of this evidence,coupled 1994: predawn(4pd), and the differencebetween midday (qjmd), midday with the high A and g of Sarcobatus, suggests that the and predawn(0Pmd-4Ppd); mean + 1 SE, N = 6 (for statisticssee Table 2). low Ppd of thatspecies is not indicativeof limitedwater availabilitynor limitedroot access to moist soil layers. the diurnalvariables on both dates, and the shapes of the The discrepancyof Sarcobatus 4Jpd and groundwater/ curves differedfor species (indicatedby significanttime soil moistureavailability requires other explanations. One by species interaction)for 4 in Julyand forg in August. possibilityis a low hydraulicconductivity of Sarcobatus There was a significantpositive linear relationshipbe- deep roots duringperiods of low flow,such as at night. tween A and g for each species on each sampling date This would mean that the roots of Sarcobatus, which (Fig. 5; F > 19.6 and P < 0.05 foreach test).Comparing based on our xylem and seasonal leaf chemistrydata are the slopes of the relationshipsbetween A and g indicated relativelyhigh in salt, would not approachequilibrium as thatthe species did not differin July(P > 0.05), but did rapidlyas Chrysothamnus.Recent analyses of the com- differsignificantly in August (P < 0.05), with Chryso- posite transportproperties of tree root systems where thamnushaving the steeper slope. Chrysothamnusand flows were affectedby osmotic ratherthan hydrostatic Sarcobatus did not differfor instantaneousWUE (A/E) gradientshave shown long halftimes of waterflow equil- nor instantaneousNUE (A/N) in eitherJuly or August ibrationand low hydraulicconductivities (Hallgren, Ru- (Table 3). dinger,and Steudle, 1994; Steudle, 1994). Those experiments were done with well-wateredroots subjected to DISCUSSION externalosmotica, so extensionto thelow xylempressure potential conditions present in our study will require The co-occurringshrub species, Chrysothamnusnau- more detailed data on the xylem pressures,osmotic con- seosus and Sarcobatus vermiculatus,differ substantially tents,and waterfluxes and pathwaysin the roots of Sar- in theirdiurnal and seasonal water relations,with Sar- cobatus underconditions similar to those we observed in cobatus having much more negative1Jpd and lmd through- the field,i.e., with shoot 4Ppd as low as -2.7 MPa. 1642 AMERICAN JOURNAL OF BOTANY [Vol. 8 3

4.0 I I3.0 - Chrysothamnusnauseosus 3.0 - .-,2.5 -~ Z 2.0-20-

1.0 - Sarcobatusvermiculatus -. 1.5 _

1.5 I I E- C 20o150_X10 -1- 1.0

m50 0.5 0.60 I I I 0.4-

0.3o CL~ 0.2- 0.2 - 0.0J 0.4 -9.0

Wl ~3.0- o - I . May Jun Jul Aug May Jun Jul Aug Month Month Fig. 3. Seasonal courses of leaf element contentsfor Chrysothamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, CA in 1994: nitrogen(N), boron (B), phosphorus(P), sulfur(S), potassium (K), calcium (Ca), magnesium(Mg), and sodium (Na); mean ? 1 SE, N = 6 (for statisticssee Table 2). Total leaf ash contentthrough the season is given in Table 1.

A second possible cause of the large discrepancybe- transportduring midsummer (Flanagan, Ehleringer,and tween Sarcobatus Ppd and groundwater/soilmoisture Marshall, 1992; Donovan and Ehleringer,1994b). This availabilitymight be continuednighttime water leakage would effectivelyeliminate shallow soil capacitance and (i.e., hydrauliclift) by an extensive,active, shallow root allow much more rapid equilibrationwith the moist soil system,which preventsthe shrubfrom rapidly reaching layers. Thus differencesbetween the species in shallow an equilibriumwith water availability in the deeper,wet- root systemdevelopmental characteristics, such as phe- tersoil layers.Here the dry soil layers act as an external, nology,suberization, and fineroot shedding,which affect linked capacitance that affectsthe rehydrationtimes for root systemhydraulic properties, may contributestrongly the entireroot/shoot system. Extended equilibration times to the species differencesin water relationscharacteris- were demonstratedexperimentally in hydraulicallylift- tics. With this hypothesis,both species would thenhave ing, nonphreatophyticArtemisia tridentata (Richards and access to the groundwater,be able to maintainrelatively Caldwell, 1987). Hydrauliclift, and thusactive fineroots high rates of daytimewater use, as reflectedby our A in shallow (< 2 m depth), dry soil layers, have been and g measurements,yet operate at quite differentlevels demonstratedfor at < Sarcobatus a site 500 m fromthis of water potential.The advantage for Sarcobatus to op- study site (Muller, Richards, and Donovan, 1995). The erate at much lower waterpotentials relates to its salinity effectof hydrauliclift on thenighttime recovery of xylem pressurepotential will be dependenton the relativedis- tolerance,which requireslarge accumulationsof osmot- tributionand hydraulicproperties of rootsin shallow dry ically active salts in the leaves. soil and deep moist soil. In Sarcobatus the proportionof The two possibilitiesthat we have proposed to explain roots in the deep soil is probablyquite small, given the the discrepancybetween Sarcobatus and soil wateravail- observed high root lengthdensity in the upper metreof abilitycould certainlyoperate together, and could also be soil, the extensivespread of major lateralsthat we have joined by a thirdpossibility. It may be thatshoot xylem observed in field excavations (>7 m), and the difficulty pressurepotentials in Sarcobatus do not reflectroot xy- that roots would face proliferatingin the physicallyre- lem pressurepotentials. This would requirelow axial hy- strictivelacustrine clay layersabove the groundwater.Al- draulic conductancein the taproot-stemsystem of Sar- though Chrysothamnusroots would face the same con- cobatus. This seems somewhatunlikely given the appar- straintsnear groundwater,the stable hydrogenisotope ent capability of the system to sustain high water flux data fromtwo Utah sites, one a sand dune site, suggest rates duringthe daytime,allowing relativelyhigh A and thatChrysothamnus shallow roots are not active in water g. The role of xylem cavitation and embolism repair December 1996] DONOVAN ET AL.-DESERT SHRUB WATER RELATIONS AND LEAF CHEMISTRY 1643

July August o

a.X VPD 6 La

/ I-a a.> - / F . 40)

2Air temperature 2 20 L

-1 Chryothamnusnauseosus-

C,)m 2 X-2 X>t

Sarcobatus vermiculatus

20 I I

1460 15 0 E0

o-0. 3

02 (0 E01 A

6 8 10 12 14 16 6 8 10 12 14 16 Solar time(hour) Fig. 4. Diurnal xylem pressurepotentials (4i), rates of photosynthesis(A), and conductance(g) for Chrysothamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, CA on 7 July and 17 August 1994. Leaf-to-airvapor pressure deficit(VPD) and air temperatureat the time of measurementare also shown; mean ? 1 SE, N = 6 (for statisticssee Table 2). needs to be consideredbefore this possibility is discount- m depth). One exception to this uniformityis increased ed (Tyree and Sperry,1989). Na concentrationsimmediately under the Sarcobatus can- As mentionedabove, the differencesin xylempressure opies (Schaber and Richards, 1995). At our study site, potentials between Chrysothamnusand Sarcobatus are this Na (and B) enrichmentis small, and lateralroots of accountedfor, at least partially,by the species differences Sarcobatus extendmany metres beyond the canopy (J.H. in leaf chemistry,with greater leaf ion contentassociated Richardsand L.A. Donovan, fieldexcavations). The soils with more negative leaf osmotic potentials and xylem at the site are marginallysaline because some Na is avail- pressurepotentials in Sarcobatus. Soil cores at the study able in both the soil and groundwater,but cannotbe con- site indicate fairlyuniform soil chemistryconsisting of sidered saline as soil-saturatedpastes fromshrub inter- low ion concentrationsthrough both the aeolian sand and space areas usuallyhave EC values less thanthe technical buried lacustrineclay substrates(EC ranges from0.9 to saline soil guideline (1.5 dS/m; Richards, 1954). If the 1.7 dS/min saturatedsoil pastes of samples from0 to 4 two shrubspecies are exposed to a similarbelowground 1644 AMERICAN JOURNAL OF BOTANY [Vol. 83

TABLE 3. One-way analysis of variance comparisonsbetween Chrys- 20 - Chrysothamnusnauseosus othamnusnauseosus and Sarcobatus vermiculatusfor instantaneous July WUE (water-useefficiency) and NUE (nitrogen-useefficiency) (N = 6) at Mono Lake, CA in 1994. For each variable: mean ? SE, 15- F statisticand significance.* P < 0.05, ** P < 0.01, *** P < 0.001, and NS is not significant. e 10 Chrysothamnus Sarcobatus F InstantaneousWUE 0 5_ - l _ (AIE, mmol C02/mol H20) E Sarcobatusvermiculatus 7 July 1.49 + 0.04 1.44 + 0.06 0.6NS 18 August 2.06 + 0.13 1.73 + 0.14 2.9NS (00 InstantaneousNUE

s 20 - August (A/N,PIMOI C02-mol N-1-s-1) 7 July 57.0 + 4.6 52.6 + 5.6 0.7NS 1 5- 18 August 64.3 + 7.9 82.5 + 9.8 0.2NS > 15_ 0 0 X~ 10 son, 1980). Differencesin membranetransport have been associated withdifferential root membraneselectivity for K and Na (Greenwayand Munns, 1980; Osmond, Bjork- man, and Anderson,1980; Reimann and Breckle, 1993). (L ~ ~ ~2 Sarcobatus has high leaf Na levels and molar ratios of K/Na <<1, which are characteristicof the Chenopodia- 0.1 0.2 0.3 ceae and halophytesfrom other dicotyledonous families. The lower leaf Na values and K/Na >1 for Chrysotham- Conductance (mol*m2*s ') nus are more similarto values associated with nonhalo- Fig. 5. Relationshipof photosynthesis(A) to conductance (g) for phytes.These comparativeNa contentsand K/Na molar Chrysothamnusnauseosus and Sarcobatus vermiculatusat Mono Lake, ratios hold for halophytesand nonhalophytesco-occur- CA on 7 Julyand 17 August 1994; mean ? 1 SE, N = 6 (forregression ring in naturalhabitats in the Great Basin (Wiebe and statisticssee text). Walter,1972) as well as forthose growntogether exper- imentallyin standardnutrient media with added NaCl, althoughaddition of high concentrationsof K to the sub- chemical environment,in addition to the similar water stratecan increase K/Na ratios for both halophytesand availabilityenvironment discussed above, thenmany dif- nonhalophytes(Reimann and Breckle, 1993). Ca also ferencesin waterrelations and leaf chemistrymay reflect plays a role in membraneintegrity and abilityto maintain differencesin how the two species are dealing with the ion selectivityfor nonhalophytes(Cramer, Lauchli, and low ion concentrationsin the sands, lacustrineclays, and Epstein, 1986). In our study,leaf Ca was not significantly groundwater. differentfor these species at any timeduring the growing The Na concentrationsof saturated soil pastes or season, but xylem Ca in August was much higher for groundwaterprovide minimumestimates of the concen- Chrysothamnusthan forSarcobatus. trationsto which the roots of the two species would be When Na is taken up by halophytes,it is generally exposed. Based on a comparison of these soil and sequesteredin vacuoles in matureleaves, and balanced groundwaterNa concentrations(soil solution 230-345 withcytoplasmic K and organic solutes,particularly gly- mg/L; groundwater475 mg/L; see Results) and Na con- cinebetaine for many Chenopodiaceae (Flowers, Troke, centrationsin xylem sap (142 and 22 mg/L,respectively; and Yeo, 1977; Storey,Ahmad, and Wyn Jones, 1977; see Table 1), both species partiallyexclude Na prior to Gorham and Wyn Jones, 1983; but see Cheeseman, xylem transport.Chrysothamnus has much higherselec- 1988). This increased leaf ion concentrationcan be used tivityagainst sodium than Sarcobatus, as also reflected to maintainlower osmotic potentialsand thus maintain in leaf Na accumulationsthrough the season (Fig. 3). The wateruptake (Atkinsonet al., 1967; Flowers, Troke,and fact that field soil moisturecontent is generallymuch Yeo, 1977; Osmond, Bjorkman, and Anderson 1980; lower than that of saturatedpaste would increase esti- Glenn and O'Leary, 1984; Reimann and Breckle, 1993). mates of Na concentrationin the soil solution and thus Romo and Haferkamp(1989) concluded thatSarcobatus increase the estimatedextent of Na exclusion occurring reduced osmoticpotential in orderto maintainturgor and in the field. high stomatalconductance, as compared to an adjacent All plants take up some Na, and differencesin leaf nonhalophyte,Artemisia tridentata.In our study,large chemistryare the sum of differencesin salinityof sub- species differencesin leaf chemistry,succulence, and wa- strate,root uptake,transfer to the xylem and leaves, and ter relationsmight be expected to lead to differencesin leaf retention(Atkinson et al., 1967; Cheeseman, 1988; gas exchange characteristics.However, Chrysothamnus Wolf et al., 1990; Reimann and Breckle, 1993). Halo- and Sarcobatus showed similardiurnal gas exchangepat- phytes are distinguishedfrom nonhalophytes by the ex- terns,and the rates of photosynthesisand conductance tentto which theytake up and toleratelarger amounts of were differentonly by an average of 1.5 times in July Na in theirleaves (Flowers,Troke, and Yeo, 1977; Green- and were not differentin August(Fig. 4). NUE and WUE way and Munns, 1980; Osmond, Bjorkman,and Ander- also did not differ.This is surprisingsince greaterion December 1996] DONOVAN ET AL.-DESERT SHRUB WATER RELATIONS AND LEAF CHEMISTRY 1645 contentand morenegative osmotic potentials for Sarcob- accumulation and water content of dicotyledonous halophytes. atus mightbe expected to be associated with greaterN Plant, Cell and Environment7: 253-261. investmentin organic acids as cytoplasmic osmotica. GORHAM,J., AND R. G. WYN JONES. 1983. Solute distributionin Suaeda maritima.Planta 157: 344-349. These shrub species have similar patternsof gas ex- GREENWAY, H., AND R. MUNNS. 1980. Mechanisms of salt tolerancein change in Julyand August, despite the large underlying nonhalophytes.Annual Review of Plant Physiology31: 149-190. differencesin water relationsand leaf chemistry.Since GUREVITCH, J., AND S. T CHESTER, JR. 1986. Analysis of repeatedex- both species maintainlarge greenleaf areas throughsum- periments.Ecology 67: 251-255. mer and have similarlyopen canopies, the similaritiesin HALLGREN, S. W., M. RUDINGER, AND E. STEUDLE. 1994. Root hydraulic propertiesof spruce measured with the pressureprobe. Plant and gas exchange suggestthat they can maintaincomparable Soil 167: 91-98. whole-plantcarbon gain despite the large differencesin HARR,R. D., AND K. R. PRICE. 1972. Evapotranspirationfrom a grease- waterrelations and salinitytolerance. wood-cheatgrasscommunity. Water Resources Research 8: 1199- 1203. LITERATURE CITED HINCKLEY, T M., J. P. LASSOIE, AND S. W. RUNNING. 1978. Temporal and spatial variationsin the waterstatus of foresttrees. Forest Sci- ence Monograph20, Society of AmericanForesters, Bethesda, MD. ATKINSON, M. R., G. P. FINDLAY, A. B. HOPE, M. G. PITMAN, H. D. W. McNULTY, I. 1969. The effectof salt concentrationon growthand SADDLER, AND K. R. WEST. 1967. Salt regulationin the mangroves metabolismof a succulenthalophyte. In C. C. Hoff and M.L. Pie- Rhizophora mucronata Lam. and Aegialitis annulata R. Br. Aus- desel [eds.], Physiologicalsystems in semi-aridenvironments, 255- tralian Journalof Biological Science 20: 589-599. 262. New Mexico UniversityPress, Albuquerque,NM. BILLINGS, W D. 1949. The shadscale vegetationzone of Nevada and MULLER, M. W., J. H. RICHARDS, AND L. A. DONOVAN. 1995. Soil water easternCalifornia in relationto climate and soils. AmericanMid- availabilityto Sarcobatus vermiculatusalong a successional gra- land Naturalist42: 87-109. dient at Mono Lake, CA. Bulletin of the Ecological Society of . 1951. Vegetationalzonation in the Great Basin of western America 76: 192 (Abstract). North America. In Comptes Rendus du Colloque sur les Bases OSMOND, C. B., 0. BJORKMAN, AND D. J. ANDERSON. 1980. Physiolog- Ecologique de la Regenerationde la Vegetationdes Zones Arides, ical processes in plant ecology: toward a synthesiswith Atriplex. 101-122. InternationalUnion of Biological Sciences, Paris. Springer-Verlag,Berlin. BRANSON, F A., R. F MILLER, AND I. S. MCQUEEN. 1976. Moisture POTVIN, C., M. J. LECHOWICZ, AND S. TARDIF. 1990. The statistical relationshipsin twelve northerndesert shrub communitiesnear analysis of ecophysiologicalresponse curves obtainedfrom exper- Grand Junction,Colorado. Ecology 57: 1104-1124. imentsinvolving repeated measures. Ecology 71: 1389-1400. CHEESEMAN, J. M. 1988. Mechanisms of salinitytolerance in plants. REIMANN, C., AND S.-W. BRECKLE. 1993. Sodium relationsin Plant Physiology87: 547-550. Cheno- podiaceae: a comparativeapproach. Plant, Cell and Environment CLINE, J. F, AND W. H. RICKARD. 1974. Isotope uptakefrom halophyte- affectedsoil. NorthwestScience 48: 235-238. 16: 323-328. RICHARDS,J. H., AND M. M. CALDWELL.1987. Hydrauliclift: substan- COMSTOCK, J. P, AND J. R. EHLERINGER. 1992. Plant adaptationin the GreatBasin and Colorado Plateau. Great Basin Naturalist52: 195- tial nocturnalwater transport between soil layers by Artemisiatri- 215. dentata roots. Oecologia 73: 486-489. RICHARDS,L. A. 1954. and of saline and alkali CRAMER, G. R., A. LAUCHLI, AND E. EPSTEIN. 1986. Effectsof NaCl Diagnosis improvement and CaCl2 on ion activitiesin complex nutrientsolutions and root soils. United States Departmentof AgricultureHandbook Number growthof cotton.Plant Physiology81: 792-797. 60. Washington,DC. DAVIS, S. D., AND H. A. MOONEY. 1986. Water use patternsof four RICKARD,W. H. 1964. Demise of sagebrush throughsoil changes. co-occurringchaparral shrubs. Oecologia 70: 172-177. BioScience 14: 43-44. DETLING, J. K., AND L. G. KLIKHOFF. 1973. Physiological response to . 1965. Sodium and potassium accumulation by greasewood moisturestress as a factorin halophytedistribution. American Mid- and hopsage leaves. Botanical Gazette 126: 116-119. land Naturalist90: 307-318. , J. F CLINE, AND R. 0. GILBERT. 1973. Soil beneath shrub DONOVAN, L. A., AND J. R. EHLERINGER. 1994a. Carbon isotope dis- halophytesand its influenceupon growthof cheatgrass.Northwest crimination,water-use efficiency, growth, and mortalityin a natural Science 47: 213-217. shrubpopulation. Oecologia 100: 347-354. , ANDJ. L. WARREN. 1981. Response of steppe shrubs to the , AND . 1994b. Waterstress and use of summerprecip- 1977 drought.Northwest Science 55: 108-112. itationin a Great Basin shrubcommunity. Functional Ecology 8: ROBERTSON,J. H. 1983. Greasewood (Sarcobatus vermiculatus(Hook.) 289-297. Torr.).Phytologia 54: 309-324. EDDLEMAN, L. E., AND J. T RoMo. 1987. Sodium relationsin seeds and ROBINSON,T. W. 1958. Phreatophytes.United States Geological Survey seedlings of Sarcobatus vermiculatus.Soil Science 143: 120-123. Water-SupplyPaper Number 1423. Washington,DC. FIELD, C., J. MERINO, AND H. A. MOONEY. 1983. Compromisesbetween ROLLINS,M. B., A. S. DYLLA,AND R. E. ECKERT,JR. 1968. Soil prob- water-useefficiency and nitrogen-useefficiency in five species of lems in reseedinga greasewood-rabbitbrushrange site. Journalof Californiaevergreens. Oecologia 60: 384-389. Soil and WaterConservation 23: 138-140. FIREMAN, M., AND H. E. HAYWARD. 1952. Indicator significanceof RoMO, J. T., AND M. R. HAFERKAMP. 1989. Waterrelations of Artemisia some shrubsin the Escalante desert,Utah. Botanical Gazette 114: tridentatassp. wyomingensisand Sarcobatus vermiculatusin the 143-155. steppe of southeasternOregon. AmericanMidland Naturalist 121: FLANAGAN, L. B., J. R. EHLERINGER, AND J. D. MARSHALL. 1992. Dif- 155-164. ferentialuptake of summerprecipitation among co-occurringtrees ROUNDY, B. A., J. A. YOUNG, AND R. A. EVANS. 1981. Phenology of and shrubsin a pinyon-juniperwoodland. Plant, Cell and Environ- salt rabbitbrush(Chrysothamnus nauseosus ssp. consimilis) and ment 15: 831-836. greasewood (Sarcobatus vermiculatus).Weed Science 29: 448-454. FLOWERS, T J.,P. E TROKE, AND A. R. YEO. 1977. The mechanismof SAH, R. N., AND R. 0. MILLER. 1992. Spontaneous reactionfor acid salt tolerance in halophytes.Annual Review of Plant Physiology dissolutionof biological tissuesin closed vessels. AnalyticalChem- 28: 89-121. istry64: 230-233. FORT, K. P, AND J. H. RICHARDS. 1995. Seed dispersal onto recently SAS. 1989. SAS/STAT user's guide, Version 6, 4th ed., vols. 1 and 2. exposed playa at Mono Lake, CA. Bulletin of the Ecological So- SAS Institute,Cary, NC. cietyof America 76: 84 (Abstract). SCHABER, E. J., AND J. H. RICHARDS. 1995. Nutrientand toxic ion con- GATES, D. H., L. A. STODDART, AND C. W. COOK. 1956. Soil as a factor straintsto shrubestablishment on sand dunes at Mono Lake, CA. influencingplant distributionon salt-desertsof Utah. Ecological Bulletinof the Ecological Society of America 76: 237 (Abstract). Monographs 26: 155-175. SHANTZ, H. L., AND R. L. PIEMEISEL. 1940. Types of vegetationin GLENN, E. P., AND J. W. OLEARY.. 1984. Relationship between salt Escalante Valley, Utah, as indicatorsof soil conditions. United 1646 AMERICAN JOURNAL OF BOTANY [Vol. 83

States Departmentof AgricultureTechnical Bulletin Number713. time-dependentmeasures. In S. M. Scheiner and J. Gurevitch Washington,DC. [eds.], Design and analysis of ecological experiments,113-137. SOKAL,R. R., AND F J. ROHLF. 1981. Biometry.W. H. Freeman,New Chapman & Hall, New York, NY. York,NY. WALLACE A., E. M. ROMNEY, AND V. Q. HALE. 1973. Sodium relations STOREY, R., N. AHMAD, AND R. G. WYN JONES. 1977. Taxonomic and in desertplants. 1. Cation contentsof some plant species fromthe ecological aspects of the distributionof glycinebetaineand related Mojave and Great Basin deserts.Soil Science 115: 284-287. compoundsin plants. Oecologia 27: 319-332. WEST, N. E. 1988. Intermountaindeserts, shrub steppes, and wood- STEUDLE, E. 1994. Water transportacross roots. Plant and Soil 167: lands. In M. G. Barbourand W. D. Billings [eds.], NorthAmerican 79-90. terrestrialvegetation, 213-230. Cambridge UniversityPress, New York, NY. ToFr, C. A. 1995. A 10-yeardemographic study of rabbitbrush(Chrys- WIEBE, H. H., AND H. WALTER. 1972. Mineral ion compositionof hal- othamnusnauseosus): growth,survival and waterlimitation. Oec- ophyticspecies fromnorthern Utah. AmericanMidland Naturalist. ologia 101: 1-12. 87: 241-245. TYREE, M. T, AND J. S. SPERRY. 1989. Vulnerabilityof xylem to cav- WOLF, O., R. MUNNS, M. L. TONNET, AND W. D. JESCHKE. 1990. Con- itation and embolism. Annual Review of Plant Physiologyand Mo- centrationsand transportof solutes in xylemand phloem along the lecular Biology 40: 19-38. leaf axis of NaCl-treatedHordeum vulgare. Journalof Experimen- VON ENDE, C. N. 1993. Repeated-measuresanalysis: growthand other tal Botany 41: 1133-1141.