Great Basin Naturalist

Volume 55 Number 2 Article 3

4-21-1995

Growth and reproduction in an alpine cushion plant: Astragalus kentrophyta var. implexus

Wayne R. Owen University of California, Davis and White Mountain Research Station, University of California, Las Angeles

Follow this and additional works at: https://scholarsarchive.byu.edu/gbn

Recommended Citation Owen, Wayne R. (1995) "Growth and reproduction in an alpine cushion plant: Astragalus kentrophyta var. implexus," Great Basin Naturalist: Vol. 55 : No. 2 , Article 3. Available at: https://scholarsarchive.byu.edu/gbn/vol55/iss2/3

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Great Basin Naturalist 55(2), © 1995, pp. 117-123

GROWTH AND REPRODUCTION IN AN ALPINE CUSHION PLANT: ASTRAGALUS KENTROPHYTA VAR. IMPLEXUS

Wayne R. Owen1

ABSTRACf.-A two-year field experiment was conducted to investigate factors hypothesized to affect the reproduc­ tive potential ofAstragalus kentrophyta var. implexus and to test the importance oftrade-offs between growth and repro­ duction in this species. Levels of mineral nutrients, water, herbivory, and competition were manipulated. Seed output R."1d growth of individuals in treatment groups were compared against control plants. Neither water nor mineral nutri­ ents alone were shown to affect growth or reproduction. Herbivory was shown to be similarly unimportant in affecting growth and reproduction. Competition with other species influenced growth but not reproduction. No significant trade­ offs between growth and reproduction were detected within years. However, there did appear to be a trade-offbetween these major fitness components when compared between years.

Key words: Astragalus, alpine, competition,!ecundUy, trade-off, White Mountains.

The impact of resource availability on the limited. Experiments involving other organ­ reproductive output of plants is well estab­ isms from this habitat have shown that avail­ lished (Harper 1977, Schoener 1983, Fowler ability of resources influences the competitive 1986, Welden and Slausen 1986). Plants may ability and distribution of species (Wright and experience resource limitation as a result of Mooney 1965, Mooney 1966, Marchand 1973), competition (inter- or intraspecific) or poor though this is not generally true of all alpine hahitat quality. Resource limitations can also habitats (Korner 1989). Second, standing bio­ occur when a portion of a plant's photosyn­ mass and percent cover are substantially lower thetic organs are removed (e.g., by herbivory), on dolomitic soils than on adjacent sandstone­ damage which clearly interferes with the plan(s and granite-derived substrates, suggesting that ability to provision its offspring (Marquis plants on the dolomite barrens might be rela­ 1991). A number of authors (Cody 1966, tively resource limited (Mooney 1966, Owen MacArthur and Wilson 1967, Harper 1977, 1991). Third, A. kentrophyta plants routinely Grime 1979, Tilman 1982, Weiner 1988, 1990) abort the majority of flowers they produce have considered the ecological consequences each year (Owen 1991), a pattern that has been of resource limitation for individuals and pop­ attributed to resource limitations in a broad ulations and have described various strategies spectrum of species (Lovett Doust and Lovett that plants might be expected to pursue to optimize the allocation oflimited resources. Doust 1988). This study tests wbether tbe availability of An experiment was designed (1) to test resources limits the fecundity of Astragalus whether there are resource constraints on the kentrophyta Gray var. implexus (Canby) reproduction and growth of A. kentrophyta Barneby (hereafter, simply A. kentrophyta) and and (2) to assess the interactions between two to what extent trade-offs between growtb and major components of fitness (Le., growth and reproduction might influence patterns of reproduction) under different regimes of reproduction observed in this species. A. ken­ resource availability. To do this, a factorial trophyta is an alpine cushion plant indigenous field experiment was established in which sep­ to high elevations throughout the Intermoun­ arate groups ofplants would receive either (1) tain West ofNorth America (Barneby 1964). water or (2) nutrient supplements, (3) protec­ Many lines of evidence suggest that repro­ tion from herbivory, or (4) relief from the duction in A. kentrophyta might be resource potentially competitive influence ofneighbors.

lUniversity ofCalifOrnia. Davis. and White Mountain Resem:ch Station, University ofCalifornia, Los Angeles. Present addres~: Boise National Forest, 1750 Front Sh'eet, Boise, 10 83702.

117 118 GREAT BASIN NATURALIST [Volume 55

STUDY AREA as "Water," (2) Another 50 plants received sup­ plemental nutrients. These plants were given The study was conducted on the alpine approximately 17 g of a balanced general-pur­ dolomite barrens of Sheep Mountain Pass pose fertilizer (Scott's All-Purpose Builder, above the Patriarch Grove bristlecone pine 12:10:12 N:P:K), providing each plant with 2,0 g forest, in the White Mountains of Mono N (in the form of ammoniacal nitrogen, ureas, County, CA, Elevations at the site range from and water soluble nitrogen), 1.7 g P (from 3535 m (11,600 ft) to 3660 m (12,000 ft), and phosphoric acid, P20 S), and 2,0 g K (from sol­ topographic relief of the site is minimaL In the uble potash, K20), These quantities are equiv­ White Mountains A kentrophyto occurs only alent to application rates of 13,8, 11.7, and on dolomitic soils (Lloyd and Mitchell 1973, 13.8 kg ha-1, respectively, A balanced fertilizer Hall 1991), was chosen because experiments by Chambers Weather data were obtained from the White et al, (1987) and Shaver and Chapin (1980) have Mountain Research Station, Mt. Barcroft shown that plants in cold environments re­ Laboratory, located 6 km north of the study spond most vigorously to resource augmenta­ site at an elevation of3800 m, Soils on the dolo­ tion with fertilizer containing a balance of mite barrens have a high cation exchange essential nutrients. The dry fertilizer was scat­ capacity and are depauperate in nitrogen, tered in an approximately 2-cm-wide ring phosphorus, and potassium (Mooney et al. around the perimeter of each test plant. 1962, Wright and Mooney 1965, Brayton and Summer seasonal precipitation in 1989 was Mooney 1966, Mooney 1966, Marchand 1973, apparently sufficient to solubilize the fertilizer 1974), The moisture-holding capacity of and deliver it to the soil profile, as the granules dolomite-derived soils is equivalent to that of had completely disappeared from the surface adjacent granitic soils (Mooney et al. 1962, in approximately one month. This treatment Wright and Mooney 1965, Marchand 1973), group will be referred to as "Fertilized," (3) A Vegetation of the White Mountains is general­ third treatment was designed to protect plants ly xerophytic; this trend is especially prevalent from herbivory and predation on flowers and on the dolomite barrens (Lloyd and Mitchell young fruits, Two locally common insects ha­ 1973), bitually consume the reproductive parts of A, kentrophyta, The more common of these in­ MATERIALS AND METHODS sects, a darkling beetle (Tenebrionidae: Coleop­ tera), consumes flowers. Larvae ofa locally com­ In June 1989, 195 healthy A kentrophyta mon Lycinid butterfly species (Lycaenidae: plants were selected randomly from within an Lepidoptera) occasionally consume immature area ofapproximately 0,2 ha, Decadent (senes­ A, kentrophyta fruits, "Tangle-foot" brand cent) plants were disqualified from inclusion sticky-trap was applied in a circle around each in this experiment. The specific location of the of 25 plants to exclude potential herbivores, site was chosen for its apparent homogeneity Tanglefoot barriers were repaired as needed. with respect to soil physical characteristics, This treatment group will be called "No vegetation, and topographic profile, Plants Predation," (4) The fourth treatment sought to were randomly allocated to five treatment relieve a group of 20 A kentrophyta plants regimes: (I) 50 plants were provided with from neighborhood competition, A 0,25-m­ three separate 1-L applications of water dur­ radius circle around a central target A. kentro­ ing the 1989 growing season, Plants were phyta plant was cleared of all other plants by watered during the driest part of the summer cutting them off at ground level. This method (4 July, 2 August, and 19 August) to maximize minimized ground surface disturbance. the beneficial impact of the treatment. Water Clearings were 0.2 m2 in area. The average was applied slowly (to maximize infiltration) in number of neighbors (ramets) removed was 63 a radius of 12,5 cm around each plant. This (mostly tillers of Poa rupicola), covering an treatment supplied 6, I cm ofmoisture to each average of 15% of the ground surface, plant, Expected precipitation for the three­ Excavations ofA. kentrophyta plants show that month growing season is 8,7 cm (Pace et al. its roots grow straight downward into the soil 1968), The 1989 summer precipitation was I.I with minimal lateral root spread (Owen 1991), cm, This treatment group will be referred to Roots of the target plants were therefore 1995J ASTRAGALUS GROWTH AND REPRODUCTION 119 thought to be well isolated from interactions production, but previous experience (Owen with actively assimilating roots ofother plants. 1991) had shown that seed production is a sig­ Plants clipped in the cleared areas were nificant function of flower production (Owen trimmed ifthey resprouted. Plants in tlris treat­ 1991). Flowers, when aborted, are dropped at ment group are referred to as the "Target" a very early age (Owen 1991) and probably group. (5) A final group of 50 unmanipulated represent a minimal per-unit cost in resources plants was marked as a "Control" group. Size to the plant (Bookman 1983, Stephenson of the experimental groups was based on an 1984). Therefore, the cost offlowers should be analysis of expected variances in responses to proportional to a plant's seed output and can the treatments; lower expected variances re­ safely be disregarded for the purpose of this quire smaller necessary samples (Sokal and work Fruits and seeds were cleaned and sepa­ Rohlf 1981). rated in the laboratory, counted, and weighed. Plant sizes (cushion area) were measured and recorded on 23 June 1989, shortly after RESULTS initiation of growth for the season. Treatments were initially applied on 4 July 1989. In Sep­ Weight of individual reproductive struc­ tember 1989 all plants were remeasured, and tures (seeds and fruits) was independent of the entire fruit and seed crop produced by total numbers of those items produced per each of the 195 plants was harvested. Since A. plant in both years (Table 1). Average seed and kentrophyta forms a tight cushion that never fruit weights were significantly correlated (R exceeds 1 em in height and seeds are not ~ .429 in 1989, R ~ .443 in 1990). There were released from the plant before the end of the no significant differences between treatment growing season, there was great confidence groups for the weight of individual seeds or that the entire seed crop of each individual fruits (resuIts not presented). Because seed was retrieved. In early June 1990 I again mea­ production is well conelated with other possi­ sured the area of all plants just as they were ble measures of fitness in A. kentrophyta and initiating growth for the season. Fertilized and weights of those seeds are independent of the Water treatments were not repeated in 1990 numbers of reproductive structures produced so as to evaluate the potential for lags in the on a plant (Table 1), seed output was used as effectiveness of resource supplementation. an index oftotal reproductive effort. Tanglefoot barriers were maintained during In a comparison of slopes of regression 1990 to test for interannual variation in the analyses, growth was a significant function of effects ofherbivores and predators. Clear zones plant size in both 1989 and 1990 (Table 2), around Target plants were maintained in 1990. though the relationship was weaker in 1990. All plants were allowed to grow through the The weight of individual seeds and fruits was season, and in September 1990 all 195 plants independent of seasonal growth (Table 2). The were remeasured and all fruits and seeds har­ amount of growth across years was significant­ vested. No attempt was made to quantify flower ly but poorly correlated.

TABLE 1. Correlation matrix for selected demographic traits. Values above the diagonal are correlation coefficients (R) based on 1990 data; those below the diagonal are derived from 1989 data.

Seed Seed Fruit Fruit Reproductive Seeds weight weight Fruits weight weight weight produced (average) (total) produced (average) (total) (total) Seeds producecl"'* 1 .003 ,976* .964* .143* .920* .966* Seed weight (average) .042 1 .139* -.001 .433* .081* .115* Seed weight (total) ,977* .200* 1 .945* .229* .937* .987* Fruits produced** .963* .024 .033* 1 .106 .963* .968* Fruit weight (average) .136* .429* .215* .074 1 .289* .260* Fruit weight (total)** .943* .120* .949* .952* .284* 1 .981* Total reproductive weight** .973* .1656* .989* .954* .249* .985* 1

·Kendall RanK Correlation is significant at P < .05. UTreatment differences noted with one-way ANOVA. These diff""'ellCes do not affect the magnitude ofsignificance ofthe correlations. 120 GREAT BASIN NATURALIST [Volume 55

TABLE 2. Slopes ofregressions for selected demographic TABLE 3. Result ofan ANCOVA on seed production and traits on growth in 1989 and 1990 using the total data set growth by treatment group. The covariate is plant size. (i.e., not partitioned by treatment). Where the overall The treatments are those listed in the text (see also Table 4). regressions are not significant, there were also no treat­ ment differences. Covariate Treatment p Growth in 1989 Growth in 1990 F F P Growth in 1990 .168* 1989 Seed production 37.164 <.001 1.358 .25 Plant size .340* .110* 1990 Seed production 39.818 <.001 1.854 .12 Seed weight -.038 -.054 1989 Growth 27.207 <.001 0.822 .583 Fruit weight .035 .036 1990 Growth 0.893 .346 2.453 .047

~Rcgrcs~i()ns "r", .\ignifkuntly positive (P <; ,O.S), Qoo,way AKOVAs sugge,t djIJer(Jn<.:()~ hetw

A series of simple linear regressions was Seed production (square root transformed) used to compare seed production with growth was a positive linear function of plant size. to test for the presence of a trade-off between Overall values of R2 for regressions of seed these two primary components of fitness. production on plant size were .206 in 1989 When the data are corrected for the fact that and .182 in 1990. Slopes of individual regres­ larger plants are inherently more capable of sions for each treatment for seed production producing more flowers and fruits, the analy­ on plant size did not differ from the slope for sis finds no significant differences among control plants. treatment groups (by virtue of overlapping Plant size was a minor but important factor 95% confidence intervals); and, therefore, no influencing both growth and reproduction in trade-off between growth and reproduction A. kentrophyta and indicates that size should within a given year was detected. be considered as a covariate in an analysis of To compare trade-offs across years, the variance of treatment effects in this experi­ ratio of 1990 to 1989 data was used (Table 5). ment. Analyses of covariance (ANCOVA) and This provides a number>1.0 when 1990 data experimental results are presented in Tables 3 values exceed 1989 values; the converse is and 4, respectively. Plant size was a significant true when results are < 1.0. Seed production covariate in three offour analyses. There were was greater in 1990 than in 1989 regardless of no differences among treatment groups in treatment group. In contrast, growth in 1990 seed production (reproduction) for either year. was less than that experienced in 1989 with Growth did not differ among treatment groups the notable exception of Target plants. The in 1989, but there was a significant difference results can be interpreted as evidence for a between groups in 1990 (P ~ .047). A protect­ trade-off between growth and reproduction. ed least-significant-difference (LSD) test indi­ They indicate that, in general, increased seed cates that growth in the Target group was production is associated with decreased growth. greater than that of individuals in other treat­ Furthermore, plauts may be relieved of trade­ ment groups (Table 4). off constraints by removing competitors, Table 5 gives the results oftwo-tailed t tests which should increase availability of mineral comparing mean reproduction and growth resources to the remaining (target) plant. across years within treatment groups. There were no significant differences for seed pro­ DISCUSSION duction among treatment groups between 1989 and 1990. Average size for plants in 1990 Resource supplementation or alleviation of was consistently significantly greater than the resource competition did not significantly size of the same plants the previous year (ie., influence the reproductive output ofA. kentro­ on average, plants grew larger over the course phyta. Instead, seed production was more close­ of the experiment). The No Predation treat­ ly related to the individual's past record of seed ment grew significantly less in 1990 than output (Tables 1, 3, 5). Plants that produced 1989, whereas plants in the Target group grew many seeds in 1989 tended to produce many significantly more in 1990. There were no sig­ seeds in 1990, regardless of treatment. Growth, nificant differences in growth across years for while similarly unresponsive to the addition of plants in the Control, Fertilized, or Water single resources, increased significantly when groups. potential competitors were removed (Tables 4, 1995] ASTIlAGAWS GROWrn AND REPRODUCTION 121

TABLE 4. Treatment means (SD) in both 1989 and 1990 for important demographic trail~.

Control No bugs Fertilized '-Vater Target 1989 Seed production 25.8 (25.2 16.1 (11.8) 30.6 (24.5) 25.1 (22.2) 44.2 (41.4) 1990 Seed production 32.2 (32.2S) 20.5 (16.7) 39.7 (37.3) 30.7 (27.9) 54.5 (58.4)

1989 Plant size 5997.1 (2851.7) 4594.6 (1871.8) 6833.9 (2892.7) 6333.2 (2891.4) 7683.2 (3683.8) 1990 Plant size 7247.3 (3128.8) 5596.3 (2156.6) 7934.0 (3242.6) 7418.2 (3627.4) 8393.0 (4159.9)

1989 C1"0\...1h 1478.4 (1329.7) 1530.0 (987.7) 1772.1 (1634.2) 1797.9 (1486.9) t503.1 (988.6) 1990 Crowth· 1156.1 (1529.9) 808.4 (1000.4) 1587.8 (2044.5) 1395.0 (1760.3) 2433.2 (1749.0)

*Gmwth in 1990 VlU'!ed significantly amOllg trutmenb (lit:<.' 11l!>le 3). 1lK: Target grouP$ yew mere. on avCTll{:<'. INn did plants in :my Of"'~r Ill:>l.fm.,.llt ~p. ~o other differellCeJ were signiflC:ll.11.

5). These results differ from those of Wright 1989), is most sensitive to the proximity of its and Mooney (1965), Mooney (1966), and neighbors. It is unclear, however, why repro­ Marchand (1973), which show that mineral duction among such species is rarely similarly nutrients were the primary factors limiting influenced (as is the case with A. kentrophyta). other species that occur on dolomite in the The buffering of fitness components against ~'hite Mountains (Artemisia tridentata. two environmental stochasticity is characteristic of Erigeron species, and Lupinus argenteus. density-vague demographics as described by respectively). Komer (1989) repOlis that the Strong (1986). Under density-vague condi­ effect of fertilization on the growth of species tions, selection favors demographic functions from nutrient-poor environments is often diffi­ with indetenninate functional thresholds. That cult to detect. He does not cite studies that is, current allocation decisions are only loosely address the relationship between growth and linked to current environmental conditions reproduction in nutrient-supplementation (Strong 1986). experiments. Trade-offs between growth and reproduc­ The addition of mineral nutrients or water tion within years were not observed in this alone may have been insufficient stimuli for A. experiment under any conditions. A weak kentrophyta to increase either reproduction or h'adc-off between growth and reproduction growth if both liactors were limiting. Multiple was identified in most treahnent groups when limiting factors have been reported in a vari­ data were compared acrOss years (Table 5). It ety of species (Harper 1977) and are specifi­ is of great interest that tbe Target group alone cally predicted by Tibnan's (1980, 1982) mod­ experienced an increase in both seed produc­ els of optimal resource consumption. That tion and gl'Owth in 1990 compared to 1989 val­ there may be multiple resource limits to A. ues (and thus did not experience a trade-of!). kentrophyta growth and reproduction is sup­ The absence ofwell-defined trade-offs between ported hy the response of A. kentrophyta to primary components of fitness could be due to the removal ofcompetitors in this study. one of several reasons, Lack of a discernible Tanglefoot barriers were very effective at trade-off would be noted if resources were not excluding ground-moving herbivores and truly limiting. It mal' also be that growth and predators. This was evidenced by the lack of reproduction are not co-limiting for this foliar damage or partially eaten fruit and the species in this environment. If this were true, capture of many insects in the traps. Flowers factors that influence growth and reproduction of A. kentrophyta are produced in sulTicient are likely to be iudependent (e.g., one fitness excess to buJfer individuals against the levels component might be canalized and the other of flower and fruit predation observed in this dependent on environmental conditions). population. Finally, a trade-off' between growth and repro­ Growth in A. kentrophyta, as bas been re­ duction would not he detected if a resource ported for a number ofspecies from arid regions other than one provided in this experiment throughout the world (Fonteyn and Mahall were limiting. 1981, Robberecht et a!. 1983, Ehleringer Adult A. kentrophyta mortality at the Sheep 1984, Parker and Salzman 1985, Shaw 1987, Mountain study site is low, juvenile mortality Manning and Barbour 1988, and Chapin et aI. is extremely high (even though germination 122 GREAT BASIN NATURALIST [Volume 55

TABLE 5. Cross-year comparisons of fitness components. 1990 values represented as a fraction of 1989 trait values. Values of t and the associated probabilities (P) represent results of two-tailed t tests for differences in values between years. Refer to Table 4 for raw data.

Control No bugs Fertilized Water Target

Seed production 90/89'" 1.25 1.16 1.32 1.15 1.18 t 1.41 1.71 1.80 1.39 0.71 p .17 .10 .08 .17 .49 Plant size t 7.06 5.02 4.90 5.05 3.50 P <.01 <.01 <.01 <.01 <.01 Growth 90/89* 0.85 0.98 0.56 0.86 2.07 t 1.13 2.50 0.40 1.42 2.12 p .26 .02 .70 .16 .05

*Values listed repn'

tests under controlled conditions show seed LITERATURE CITED viability of greater than 95%), and recruitment is low (Owcn 1991). These demographic attri­ BARNEBY, R. C. 1964. Atlas of North American Astragalus. Memoirs of the New York Botanical Garden 13: butes would certainly favor a strategy that 1-1187. routes resources away from the risky business BOOKMAN, S. S. 1983. Costs and benefits of flower abscis­ of reproduction toward growth. Thc small but sion and fruit abortion in A~clepias speciosa. Ecology consistent portion of A. kentrophyta's annual 64,264-273. BHAYTON, R., AND H. A. MOONEY. ]966. Population vari­ accumulation of biomass allocated to repro­ ability ofCercocarpus in the White Mountains ofCali­ duction guarantees that each plaut will proba­ fornia as related to habitat. Evolution 20: 383-391. bly produce at least a few seeds each year CHAMBERS, ]. C., J. A. MACMAHON, AND R. W. BROWN. while being able to dedicate most of each sea­ 1987. Response of an early seral dominant alpine son's accumulated resources to growth and grass and a late seral dominant alpine forb to Nand P availability. Reclamation and Revegetation Research survival. That the allocation of resources to 6,219-234. reproduction, but not growth, in this species is CHAPIN, F. S., J. B. McGRAW, AND G. R. SHAVEn. 1989. constant over a broad range of resource avail­ Competition causes regular spacing of alder in abilities is consistent with a bet-hedging life­ Alaskan shrub tundra. Oecologia 79: 41.2-41.6- CODY, M. L. 1966. A general theory of clutch size. history strategy (Kozlowski and Stearns 1989, Evolution 20: 174-184. Philippi and Seger 1989, Stearns 1989). EHLERINGER, J. R. 1984. Intraspecific competitive effects Resource limitations on organisms are rarely on water relations, growth, and reproduction in simple or solitary. While fruit and flower pre­ Enceliafarinosa. Oecologia 63: 153-158. dation can be an important limit on fecundity, FONTEYN, P J., AND B. E. MAHALL. 1981. An experimental such an effect was not noted here. Similarly, analysis of structure in a desert plant community. Joumal ofEcology 69: 883-896. the reproductive output of plants growing on FOWLER, N. 1986. The role ofcompetition ill plant commun­ the Sheep Mountain dolomite barrens would ities in arid and semiarid regions. Annual Review of appear to be resource limited, although single Ecology and Systematics 17: 443-464. resource augmentation had no direct effect on GRIME, J. P. 1979. Plant strategies and vegetation process­ seed production. In combination, however, es. John Wiley and Sons, New York, NY. resources can influence the amount ofrealized HALL, C. A., EDI'IDU. 1991. Natural history of the White­ Inyo Range. University of California Press, Berkeley. growth that in subsequent years will affect HAHPER, J. L. 1977. Population biology of plants. Academic reproduction. Press, New York, NY. KORNER, C. 1989. The nutritional status of plants from ACKNOWLEDGMENTS higher altitudes. Oecologia 81: 379-391. KOZWWSKI, J., AND S. C. STEARNS. 1989. Hypotheses for the production of excess zygotes: models of bct­ I would like to tbank the White Mountain hedging and selective ahortion. Evolution 43: Research Station for providing logistic and 1369-]377. financial support for this project, especially LLOYD, R. M., AND R. S. MITCHELL. 1973. A flora of the the crew at tbe Mt. Barcroft Laboratory. T. White Mountains, California and Nevada. University Holmes, E. Nagy, A. Fitter, and two anony­ ofCalifornia Press, Berkeley. LOVETT DOUST, J., AND L. LoVETT DOUST. 1988. Plant mous reviewers made significant improve­ reproductive ecology. Oxford University Press, New ments on earlier drafts ofthis manuscript. York, NY. 1995] ASTRAGALUS GROWfH AND REPRODUCTION 123

MAcAR'rHULl, R. H., AND E. O. WiLSON. 1967. The theory SCHOi-:NER. T. w. 1983. rle1d experiments on interspecific of island biogeogntphy. Princeton University Press, competition. American Naturalist 122: 24~285. Princeton, NJ. SHAVF.R, G. R, AND F. S. CHAPIN. 1980. Response to fertil­ MANN LNG, S. J., AND M. G. BARBOUR. 1988. Root systems, ization by vwious plant growth forms in an Alaskan spatial patterns, and competition fOT soil moisture tundra: nutrient accumulation and growth. Ecology between desert subshrubs. American Journal of 61,662-675. Botany 75; 885-893. SHAW, R. C. WS7. Density dependence in Salvia /.yrala.: MARCHAND, D. E. 1973. Edaphic control orplant distrihu· experimental alterations of densities of established tion in the White Mountains, eastern Califomia. plants. Journal ofEcology 15: 1049-1063. Erology 54,2<13-250, SOKAL, R. R., AND F J. ROHLF. 1981. Biometry. W.H. ___" 1974. Chemical wentherin~ soil development, and Freeman and Company, New York. NY. p;eochemical fractionation in a part of the White STEAHNS, S. C. 1989. lradc-offs in life-history evolution. Mountains, Mono and lnyo counties, California. Functional Ecology 3: 259-268. USGS Professional Puper 352-J. STEPHENSON, A. G. 1984. The cost of over-initiating fruit. MARQUJS, R. J. 1991. Evolution of resistance in pla.nts to American Midland Naturalist 112: 379-386. herbivores. Evolutionary Trends in Plants 5: 23-29. STRONG, D. R. 1986. Density vagueness: abiding the vari­ MOONEY. H. A. 1966. Influence of soil type on the distribu· ance in the demography of real populations. Pages tion of two closely I"elll.tt:d species ofEri{!,eron. Ecology 257-268 in J. Diamond and T. J. Case. editors. 47,950-1)58. COllununity ecology. Harper and Row Puhlishers, MOONl5Y, H. A., C. ST. ANDttE. AND R. D. WRJCHT. 1962. New York. NY. Alpine and subalpine vegetation patterns in the TrLMAN, D. 1980. Resources, a waphiea.l~meehanistic White Mountains of California. American Midland approach to competition and predation. American Naturalist 68: 257-273. Nahmmst 116, 362--:393. OWEN, W. R. 1991. The reproductive crology of an alpine __;:-.' 1982. Resource oompetition and community shuc­ legume: A. kentrophyta var. impleros. Unpublished ture. Princeton Univel'sity Press. Princeton, NJ. dissertation, University of California, Davis. 226 pp. WElNER, J. 1988. Variation in the perfonnance of individ­ PACE, N., D. W. KIEPER1~ AND E. M. NISSEN. 1968. Clima~ uals in plant populations. P.dges 59--81 in A. J. Davey, tologicn.l data summalY for the Crooked Creek Laboru­ M. J. Hulchings, and A R. Wdtkinson, editors, Plant tory, 1949-1961, and the Barcroft Laboratory, 1953­ population ecolo~. Blackwell Scientific Publications, 1967. University of Califomia, White Moulltaill London. Research Station Puhlication, Berkeley. --c-' 1990. Resource competition and l:ommunity struc­ PARKER, M. A., AND A. G. SALZMAN. 1985. Herbivore ture. lrends in Evolution and Ecology 5: 360-364. exclosure and competitor removal: effects on juve~ WELDEN, C. w., ANn W L. SUUSEN. 1986. The intensity nile ~urvivorship and growth in the shrub of competition versu.') its importance: an overlooked Gutierrezia micrucephala. Journal of Ecology 73: distinction and some implications. Quarterly Review 903-913, of Biology 61: 23-44. PHILrpPI, 'J:, AND J. SEGEU. 1989. Hedging onc's evolution~ WRIGHT, R. D., AND H. A. MOONEY. 1965. Sub.')trate·ol'i­ ary bets, revisited. Trends in Ecology and Evolutjon ented distribution of bl'istlecoue pine in the White 4,41-44, Mountains ofCalifornia. AmeriC'aJl Midland Naturalist 'H.OBBERECHT, R.. B. E. MAHALI., AND P. S. NOBb:L. 1983. 73,257-284. Experimental removal of intraspecific cumpctitors­ effects on water relations and productivity of a R<=ived 21 J"'''"''y 1994 de~'ert bunchgrass HUaria rigida. Oecologia 60: A£cepred 28 O<;tober 1994 21-24,