Competition, Herbivory, and Reproduction of Trichome Phenotypes of Datura Wrightii

UC Riverside UC Riverside Previously Published Works

Title Competition, Herbivory, and Reproduction of Trichome Phenotypes of Datura wrightii.

Permalink https://escholarship.org/uc/item/8xt0j3bm

Journal Ecology, 86(2)

Authors Hare, J Daniel Smith, James L., II

Publication Date 2005

Peer reviewed

eScholarship.org Powered by the California Digital Library University of California Ecology, 86(2), 2005, pp. 334±339 ᭧ 2005 by the Ecological Society of America

COMPETITION, HERBIVORY, AND REPRODUCTION OF TRICHOME PHENOTYPES OF DATURA WRIGHTII

J. DANIEL HARE1 AND JAMES L. SMITH II Department of Entomology, University of California, Riverside, California 92521 USA

Abstract. The trichome dimorphism of Datura wrightii is intriguing because glandular trichome production has a high ®tness cost. Plants producing glandular trichomes (``sticky'' plants) are resistant to many insect herbivores that attack plants producing nonglandular trichomes (``velvety'' plants). When protected from herbivores, sticky plants initially produce fewer seeds than velvety plants but grow to a larger size. In a three-year ®eld experiment, we tested the hypothesis that sticky plants acquire a competitive advantage through greater vegetative growth. In the absence of herbivores, sticky and velvety plants grew to similar sizes in their ®rst year, but sticky plants grew larger in the second and third years. Seed production of sticky plants was 46±60% less than velvety plants in their ®rst year, and this caused a 13% reduction in their ®nite rate of increase overall, even though sticky plants produced more seeds than velvety plants did in later years. The impact of herbivory varied with plant density, and herbivores reduced plant ®tness more at low plant density than at high plant density. The differences in growth associated with trichome morphology occurred too late to provide a competitive advantage to sticky plants and probably contribute little to the maintenance of D. wrightii's trichome dimorphism. Key words: competition; Datura wrightii; herbivory; polymorphism; resistance; trichomes.

INTRODUCTION synergistic interactions are expected when resistance It has been dif®cult to develop generalizations con- and competitive ability are positively associated (Sie- cerning the combined effects of competition and her- mens et al. 2002). When resistance occurs at the ex- bivory on plant ®tness despite extensive research on pense of competitive ability, a number of outcomes are each factor individually. One hypothesis is that com- possible; competition may either amplify the costs of Reports petition and herbivory should affect plant performance resistance, especially on early growth under asymmet- independently (Reader and Bonser 1998), and many ric competition with low levels of damage, or amplify studies indeed ®nd such independence (Parker and the bene®ts of resistance under symmetric competition Salzman 1985, Rees and Brown 1992, Karban 1993, at high plant density and high levels of damage to Maron 2001). Independence is not universal, and dam- susceptible plants (Weis and Hochberg 2000). age by herbivores also can exacerbate (Lee and Bazzaz We studied competition and herbivory using Datura 1980) or mitigate the effects of competition (Parmesan wrightii Regel (Solanaceae). In this species, herbivore 2000, Haag et al. 2004). resistance is costly and is correlated with plant growth The lack of pattern may be in¯uenced by details of traits that may be bene®cial in a competitive environ- the experimental design. The effect of herbivory on ment. All plants produce leaf trichomes, and two phe- competition will depend upon how herbivory is dis- notypes occur (van Dam et al. 1999). ``Velvety'' plants tributed over the superior and inferior competitors. If are densely covered with short, nonglandular trichomes damage were concentrated on the inferior competitor, while ``sticky'' plants are less densely covered with then herbivory should increase the asymmetry of com- glandular trichomes. Glandular trichomes secrete acyl- petition favoring the superior competitor. If damage sugars (van Dam and Hare 1998) that confer resistance were concentrated on the superior competitor, then her- to many of D. wrightii's insect herbivores (Hare and bivory should reduce the superior competitor's advan- Elle 2002). Trichome phenotype is controlled by a sin- tage. If herbivores attack superior and inferior com- gle Mendelian gene, the sticky condition is dominant petitors similarly, then herbivory may mitigate the ef- (van Dam et al. 1999), and the frequency of sticky fects of competition overall (Gurevitch et al. 2000). plants ranges from 0 to 93% among populations (Hare Similarly, no interaction is expected between compe- and Elle 2001). tition and herbivory if competition and resistance The production of glandular trichomes is correlated mechanisms are independent (Uriarte et al. 2002), and with increased vegetative growth at the expense of early seed production (Hare et al. 2003). Although sticky Manuscript received 15 June 2004; revised 13 September plants initially produce fewer seeds than velvety plants, 2004; accepted 15 September 2004. Corresponding Editor: A. R. Zangerl. sticky plants may accrue a competitive advantage over 1 E-mail: [email protected] velvety plants by virtue of greater size at high densities. 334 February 2005 PLANT COMPETITION AND INSECT HERBIVORY 335

The cost of producing glandular trichomes might be reduced or eliminated when sticky and velvety plants are grown under suf®ciently high competition. We carried out a competition experiment in the ®eld over three years where sticky and velvety plants were grown in pure or mixed stands at three different densities (Marshall and Jain 1969, Gurevitch et al. 1990, Goldberg and Scheiner 1993), and exposed to, or protected from herbivores to test the following hypotheses:

1) Because sticky plants eventually grow larger than velvety plants, the sticky type may be the better competitor, and the advantage that velvety plants have over sticky plants should decline when plants are grown at increasing densities. FIG. 1. Distribution of nearest-neighbor distances be- 2) If the sticky type is the superior competitor, then tween individual Datura wrightii plants. Data were obtained from four populations (two sites from University of Califor- the seed production of sticky plants should be nia, Riverside, California, USA: Pictograph Trail and Mill relatively greater, and that of velvety plants should Creek; see van Dam et al. [1999] and Hare and Elle [2004] be relatively less, when grown in mixed than in for more detailed population descriptions). pure stands. 3) Herbivory and competition will act independently upon plant ®tness. competitors. Each data plant was in direct competition with six adjacent plants: the center plant, two adjacent METHODS data plants, and three nondata competitors in the out- The life history of D. wrightii in southern California side ring. We used three stand types: pure velvety, pure has been described previously (Elle et al. 1999). In sticky, and mixed. The mixed stand utilized three data brief, plants are found in sandy or gravelly disturbed plants of each type, each of which was surrounded by Reports sites in the southwestern United States and Mexico three sticky and three velvety competitors. To deter- (Avery et al. 1959). Seeds germinate during the winter mine relevant plant densities for our experiment, we rainy season, and plants ¯ower and produce seeds from surveyed 131 plants from four local populations. The April through December. The large white ¯owers are nearest-neighbor distances between plants yielded a open for one night, and, despite obvious characteristics skewed distribution ranging from 0.5 m to 12 m with favoring pollination by hawk moths (Grant and Grant a mean of 2.9 m but a mode of 1.0 m (Fig. 1). On the 1983), outcrossing averaged only 29% in previous ex- basis of this information, we replicated our stands at periments (Elle and Hare 2002). Leaves and stems se- three plant spacings: 3 m, 1 m, and 0.5 m between nesce during the winter, and new foliage is produced plants. from a storage root each spring. Plant leaf canopies We also included an herbivory treatment in which often reach 1±2 m in diameter and1minheight (Mira half of the plots were exposed to herbivorous insects and Bernays 2002, Hare et al. 2003). and the other half were protected from herbivory by Experimental conditions were similar to those used twice-monthly applications of acephate (0.63 kg/ha ac- previously (Hare et al. 2003), except that plants were tive ingredient), an insecticide that does not affect plant never irrigated after establishment. Seedlings were uti- growth (Elle et al. 1999). Applications were made be- lized from three randomly selected families originating tween 24 May and 27 September 2001, 3 April and 13 from crosses between a randomly selected velvety pol- October 2002, and 3 April and 16 October 2003. Weeds len donor and heterozygous sticky pollen acceptor from were controlled by hoeing and treatment with gly- two local populations. Because the sticky phenotype is phosate when needed. inherited as a dominant Mendelian character, the ex- There were a total of 54 plots (three densities ϫ three pected ratio of sticky and velvety offspring from these stand types ϫ three families ϫ two herbivore treat- crosses is 1:1. Seedlings were germinated in the green- ments), 324 data plants, and 1026 plants in total. Plots house in the winter of 2000±2001 and reared until adult were separated by at least 5 m of bare soil and the trichome type could be assessed. Plants were trans- whole ®eld site was 0.6 ha in size. planted to a ®eld at Agricultural Operations, University A fully expanded leaf from each data plant protected of California, Riverside, on 19 and 20 April 2001, after from herbivores was collected in July 2001. The num- the ®eld had been furrow irrigated for 20 h. bers of glandular and nonglandular trichomes on the Plants were arrayed in hexagonal plots of 19 plants abaxial midrib of each leaf were tallied under a dis- (Harper 1977: Fig. 8/1). The center plant was surround- secting microscope using standard procedures (van ed by a hexagon of six data, or focal, plants and the Dam et al. 1999). Densities of glandular and nonglan- data plants were surrounded by an outer hexagon of 12 dular trichomes were analyzed by analyses of variance Reports raetmtsoe ie nuiso cm of units in per- time, leaf over the weekly estimates and the integrating area size by canopies the leaf both of sistence total in plant. the variation per estimate herbivores combined by to We damaged area or leaf removed area total re- leaf the area by leaf mul- this and the herbivores tiplied by of branch selected proportion the from the moved her- as estimated We damage plant plot. bivore data protected each low-density from a each leaves in from expanded USA) fully Nebraska, of area sample Lincoln, de- leaf Inc., area 3000 (Li-Cor leaf model Li-Cor mean meter a the with by total monthly plant the termined per multiplying leaves by of estimated number was leaf total plant The per plant. of per area values leaves two number of these number total of the product the estimates The as plant. the well on the as branches on leaves recorded, of was number the branch and Elle random leaves at 10±20 and chosen with was branch a (Hare October date, census 23 methods each to On 2002). April nondestructive 1 using from and 2003 2002, April 3 November from 3 2001, November to 1 to May 10 from weekly tri- families. either plant of among densities varied the type chome if determine to (ANOVA) 336 eas ftehg uvvlo lns(see plants of survival production. high seed the adjusted pro- of calculate dies seed Because to unadjusted competitors 5/6 the six by multiplied duction the we of ®ve, one leaving when competitors. of example, death the For for adjusted on viable production analyses with seed plant compared per production we seed com- viable this, unadjusted so of evaluate intensity dies, To the competitor petition. affect a may when survival reduced Com- differential the produced. be had determined will each and that petition seeds plants viable recorded We of surviving ger- number plant. of the each and number from capsule, the seeds per of plant, seeds rate per of mination capsules number seed as of mean 2003) number the al. the et of (Hare product previously the described as year per to freedom. used of was degrees method calculate Satterthwaite the normality and ensure residuals, to of used was transformation root cube The two-way effects. ``®xed'' their were and interactions higher-order tri- type and stand treatment, and herbivory density, type, of chome factors ef- the ``random'' whereas were repeated interactions fects, the its as and year Family factor. with using model 2000) repeated-measures Institute (SAS a SAS of procedure MIXED Elle and (Hare time 2002). over integrating estimates by total damage herbivores weekly by season-long the days calculated area leaf also of We removal days. area leaf ifrne.Acrigy eol eotunadjusted report production. seed only viable we signi®cant of Accordingly, pattern the alter differences. not did and overall only 3% production by seed viable changed adjustments these aao ln ieadhrioywr collected were herbivory and size plant on Data eetmtdttlval edpouto e plant per production seed viable total estimated We PROC the by days area leaf on data analyzed We .DNE AEADJMSL MT II SMITH L. JAMES AND HARE DANIEL J. 2 d hereafter, ´d, Results ), l df 1.794, df 0.348, P P wrightii D. vrgd86.97 averaged n y4%b euigpatsaigfo to m 3 from spacing plant reducing by 47% fro by spacing and plant m reducing by overall 30% fet fhrioy( herbivory of effects uae h nt aeo nrae( increase of rate ®nite the cal- We culated populations. birth-pulse and censuses breeding rwe rtce hnwe xoe oherbivores to exposed when 73.9%, vs. than (89.2% protected when high- was er Survival overall. 79.7% averaged years three x ihrsria hnvleypat 8.%v.78.4%, vs. (85.7% plants velvety than survival higher fdt lnsdfe rmta fnnaapat ( plants nondata survival of that did from plots differ exposed plants the data of in nor nei- plot protected In each the estimates. within survival ther precise plants more 19 within obtain all to plants order of survival velvety the and from survivorship sticky age-speci®c determined We the stands. mixed separate for constructed tables We family. life and stand, type, den- treatment, sity, herbivory of combination each in plants inldmg oec ln,hwvr i o aywith vary Propor- ( not did overall. density however, plant, 15% each by to damage days tional area leaf reduced vores eto aiy est ramn,adterinteraction their and (all treatment, density family, of dent fnnlnua rcoe nvleypat averaged plants velvety 211.02 on trichomes nonglandular of A SSIsiue20) eaaye the of analyzed IML We PROC 2000). using Institute (SAS matrix SAS Leslie the of eigenvalue est ( density pc® erdcin( reproduction speci®c ebvr c neednl,wudb eetdb a density. and by herbivory rejected between be interaction would signi®cant independently, act and herbivory competition that hypothesis, third Our interaction). rdniis(.. rcoetype high- trichome at (e.g., be- especially densities type, er interaction stand and an type with trichome tween consistent be asym- is would Our competition metrical, that density. hypothesis, increasing working with second declined ®tness in types differences trichome the between that between showing interaction density and signi®cant type a sup- by be that would ported competitor, hypothesis, superior ef- the working main are ®rst plants the sticky our were than example, we interactions For analyses, the fects. in these interested In the more residuals. applying ho- of and after normality mogeneity ensure above, to transformation as root square MIXED, PROC using n i o ayaogfmle,patdniytreat- density (all plant interaction their families, or among ments vary not did and ϭ ecluae g-pc® uvvrhp( survivorship age-speci®c calculated We and ϭ h est fgadlrtihmso tcyplants sticky on trichomes glandular of density The efae asdfee in®atydet h main the to due signi®cantly differed days area Leaf P .0 aantson.La rady eereduced were days area Leaf shown). not data 0.60, 0.023). Ն m Ϯ x F F .2.Teetihm este r yia of typical are densities trichome These 0.12). aus efre elemtie o post- for matrices Leslie formed we values, ϭ ϭ ,144 2, ,2.29 2, .7trichomes/mm 9.27 vnDme l 99.Patsria after survival Plant 1999). al. et Dam (van 1, 1, ϭ ϭ P P Ϯ 1.50, ϭ ϭ 111.16, .4trichomes/mm 1.84 P .8frepsdpos.Fo the From plots). exposed for 0.18 .6frpoetdposand plots protected for 0.56 Ͻ F m P ,284 1, R .01,adsik lnshad plants sticky and 0.0001), x ESULTS ϭ o aho h 2gop of groups 72 the of each for ) P .3 rtp (F type or 0.23) ϭ ϭ 2 26.61, P .02 i.2.Herbi- 2). Fig. 0.0052, ϫ n lowsindepen- was also and Ն clg,Vl 6 o 2 No. 86, Vol. Ecology, tn type stand ␭ .6.Tedensity The 0.26). stedominant the as ) P 2 (mean Ͻ m3mto1 .01 and 0.0001) ,136 1, l x n age- and ) ϫ ␭ Ϯ ϭ density values 1 0.27, G G SE ϭ ϭ ) February 2005 PLANT COMPETITION AND INSECT HERBIVORY 337

(81.0% to 87.7% of the variation in total viable seed production within years over all plants pooled). Vari- ation in seeds per capsule and germination rate con- tributed little to the variation in total viable seeds per plant (seeds per capsule, 1.9±4.5%; germination rate, 1.0±2.5%). Germination rates ranged from 75.8% to 84.7% and did not differ between types or due to any other factors in any year. Because ␭ accounts not only for variation in total viable seed production but also for variation in survival and the timing of reproduction, ␭ provides a more revealing summary of overall plant reproduction than analyses on seed production in any given year. The ®nite rate of increase was strongly affected by three of the four main effects in our experimental de- ϭ ϭ sign: herbivory (F1, 46 7.09, P 0.0107), trichome ϭ ϭ ϭ type (F1, 31.9 5.41, P 0.0266), and density (F2, 46 43.96, P Ͻ 0.0001, Fig. 3). Herbivores reduced ␭ by 46% overall, and ␭ of velvety plants was 13% greater than that of sticky plants overall. Decreasing plant spacing from 3 m to 1 m reduced ␭ by 76% and decreasing plant spacing from 3 m to 0.5 m reduced ␭ by 89%. Our ®rst working hypothesis that the sticky type would be the superior competitor was not supported because there was no signi®cant interaction be- Ϯ ϭ ϭ FIG. 2. Least-squares mean ( 1 SE) leaf area days (log tween density and trichome type (F1, 46 0.31, P scale) of sticky and velvety D. wrightii exposed to, or pro- Reports tected from, herbivores at low, medium, or high density over three years after pooling over families and stand types.

0.5 m. Leaf area days also differed due to the inter- ϭ actions of herbivory with trichome type (F1, 286 8.67, ϭ ϭ ϭ P 0.0035) and with density (F2, 284 6.358, P 0.0016, Fig. 2). The herbivory by type interaction occurred because sticky plants had greater leaf area days than velvety plants when protected from herbivores but not when exposed to herbivores (Fig. 2). The herbivory by density interaction occurred because the differences in leaf area days between protected and exposed plants declined with increasing plant density. Low-density plants had 20% fewer leaf area days when exposed to herbivores than when protected, medium-density plants had 14% fewer leaf area days, but high-density plants had only 4% fewer leaf area days when exposed to herbivores than when protected. Some of these differences varied among years (Fig. 2), but the differences were largely in magnitude and not in direction. A no- table exception was the difference in leaf area days between sticky and velvety plants over time as shown ϫ ϭ by the year trichome type interaction (F2, 270 9.85, P Ͻ 0.0001). In 2001, sticky plants had 9.9% fewer leaf area days than velvety plants, but sticky plants had 6.0 and 9.4% more leaf area days in 2002 and 2003. Neither stand type, family, nor any of their interactions were statistically signi®cant (all P Ն 0.13). FIG. 3. Least-squares mean (ϩ1 SE) ®nite rate of increase Viable seed production varied among years, and with of sticky and velvety D. wrightii exposed to, or protected several treatment effects and interactions with year, and from, herbivores at low, medium, or high density over three was strongly in¯uenced by the number of seed capsules years after pooling over families. Reports 4.49, in- its of any nor (all type, signi®cant stand statistically of was teractions, effect nei- main because supported the not ther was would also competition asymmetrical, that be hypothesis, second Our 0.58). 338 rtya flf.Teefcso ebvr lowere herbivory also study, prior herbivory the In of reduced studies. effects two the the The in in life. similar production and of seed (2003) year in al. ®rst reduction et study) (Hare (this 45% 46% a but suffered time over plants plants sticky became velvety plants than sticky larger studies, the progressively both for In study. replicates especially earlier spacing 2003), the plant al. whose et treatment, (Hare seen low-density study patterns prior to similar a was in experiment current the in zero. herbivory to toward due tended losses also ®tness additional toward the ®tness then plant zero squeezed competition lowest the As the at density. plants and for greatest however, was herbivory competition, of impact of af- com- Herbivory outcome a overall. the provide plants fected sticky to to late these bene®t study, too petitive the emerged of years differences third size and second the in plants h urn td,hrioyreduced herbivory study, current the cinwsntsgicn ( signi®cant not was action ioedmg stecueo u herbivory our her- of in cause the variation as density-related damage bivore exclude can den- we of independent sity, was level damage proportional the because Additionally, herbivores. by com- attacked were vs. competitors similarly and focal plants focal on because herbivory plants peting vs. differential of focal out rule effect can on any We densities. herbivory different at caused in plants competing were differences results systematic their competing by that and suggested her- focal They but between plants. high, varied at than densities more density bivore was 2000). plant herbivory low (Parmesan that at found unprecedented debilitating also (2004) not al. et is Haag density low plant ®nding high at at our debilitating than 2000), less al. is et herbivory at that Gurevitch than by competition being (reviewed of levels low herbivory high or at debilitating effects, more independent either show overall. 55% by aiyo n neato ihfml (all family with interaction any or family neato.Pat norotrrn fcmeioswere competitors of ring outer our in Plants interaction. ifrne eesmlrfrbt rcoetpsbe- types trichome herbivory both the for cause similar These density. me- were highest the the differences at at 7% 20% only density, and low density, dium the at 55% was herbivory hr a osgicn aito in variation signi®cant no was There observed herbivory we signi®cant because a herbivory rejected was and independent, competition were that hypothesis, third Our h iecus fpatgot n edproduction seed and growth plant of course time The velvety than larger grew plants sticky Although lhuhms tde ncmeiinadherbivory and competition on studies most Although P ϭ ␭ y6±3 Hr ta.20) hra in whereas 2003), al. et (Hare 69±83% by .15 i.3.Terdcinin reduction The 3). Fig. 0.0165, ϫ D ISCUSSION density ϫ est neato ( interaction density F 2,46 ϫ ϭ .DNE AEADJMSL MT II SMITH L. JAMES AND HARE DANIEL J. rcoetp inter- type trichome 0.94, ␭ ␭ fteeplants these of u oplant to due P P P ϫ Ն ϭ ␭ Ն density F 0.28). u to due 0.40). 0.11). 2,46 ϭ ie oee,peiu tde ntesria of longer survival a the for on run studies were previous However, experiment time. the the if in advantage an competitors their focal of the size of the treatments. size higher-density also the only but not plants competition reduced plant herbivores of because as intensity was the density of plant reduced effect high herbivores the to in low reduction from competing the herbivory and that speculate treat- Focal We low-density or the ment. six. in size three in not similar only most and were with plants plants competing other were ring four outer plants the because in probably medium treatments, the in high-density plants and data our than larger substantially vr,A,S aia n .Resm.15.Baele the Blakeslee: 1959. Rietsema. J. and Satina, S. A., Avery, rn,V,adK .Gat 93 eairo hawkmoths of Behavior 1983. Grant. A. K. and V., Grant, cec onain(S E 0859t .D Hare). D. J. to 00-89519 National DEB the (NSF by lab- Foundation funded and was Science Sig- ®eld research for This C. undergraduates assistance. and many oratory manuscript and the Schlax, of J. uenza, drafts previous on comments eena pce.Adtoal,bcueo h sensi- this the in of of competition tivity because rea- study Additionally, a to species. is time perennial seasons of three two so amount approximately 2004), sonable Elle is and conditions (Hare years natural in survival le . .M a a,adJ .Hr.19.Cs of Cost 1999. Hare. D. J. and Dam, van M. N. E., Elle, odeg .E,adS .Shie.19.AOAand ANOVA 1993. Scheiner. M. S. and E., D. Goldberg, lnswe ncmeiinwt evt lns our in tri- dimorphism plants, natural chome the velvety sticky understanding to with toward advantage contribute competition results an in confer when not plants did plant sticky of outcomes. would duration different Thus, longer qualitatively of years. study yield a later that in unlikely it advantages feel de®cit we with ®tness initial plants the sticky offset of to dif®cult more sively le . n .D ae 02 niomnal nue var- induced Environmentally 2002. Hare. D. J. and E., Elle, ntefeunyo tcypat netbihdpopu- 2004). established Elle in and plants (Hare sticky declines lations of of frequency observations the bene®ts recent in its with consistent exceeds is often and production glan- trichome of costs ®tness dular the that conclusion previous our re-emphasiz- es plants sticky to advantage growth-related ueic,J,J .Mrio,adL .Hde.20.The 2000. Hedges. V. L. and Morrison, A. J. J., Gurevitch, genus tmyb osbeta tcypat ih aehad have might plants sticky that possible be may It n¯wr of ¯owers on Hall, and USA. and Chapman York, design New experiments. York, The New editors. ecological Gurevitch, of J. analysis and Scheiner M. S. etakE le .Gsmn,adN .vnDmfor Dam van M. N. and Gassmann, A. Elle, E. thank We wrightii lnua rcoe,a`rssac' hrce in character ``resistance'' a trichomes, glandular NOA edcmeiineprmns ae 69±91 Pages experiments. competition ®eld ANCOVA: aini oa risafcstemtn ytmin system mating the affects traits ¯oral in iation USA. York, ept h atta h nrae eeaiegrowth vegetative increased the that fact the Despite neato ewe optto n rdto:ameta- a predation: and competition between interaction 280±284. wrightii wrightii Datura nte®l hwta h rbblt ffuture of probability the that show ®eld the in ucinlEcology Functional . ␭ ee Slnca) Evolution (Solanaceae). Regel oeryrpouto,i eoe progres- becomes it reproduction, early to h oadPesCmay e ok New York, New Company, Press Ronald The . auameteloides Datura A L CKNOWLEDGMENTS ITERATURE .wrightii D. 16 :79±88. C ITED oaia Gazette Botanical . clg,Vl 6 o 2 No. 86, Vol. Ecology, h bec fany of absence The . 53 :22±35. Datura Datura 144 D. in : February 2005 PLANT COMPETITION AND INSECT HERBIVORY 339

analysis of ®eld experiments. American Naturalist 155: Mira, A., and E. A. Bernays. 2002. Trade-offs in host use 435±453. by Manduca sexta: plant characters vs. natural enemies. Gurevitch, J., P. Wilson, J. L. Stone, P. Teese, and R. J. Stou- Oikos 97:387±397. tenburgh. 1990. Competition among old-®eld perennials at Parker, M. A., and A. G. Salzman. 1985. Herbivore exclosure different levels of soil fertility and available space. Journal and competitor removalÐeffects on juvenile survivorship of Ecology 78:727±744. and growth in the shrub Gutierrezia microcephala. Journal Haag, J. J., M. D. Coupe, and J. F. Cahill, Jr. 2004. Antag- of Ecology 73:903±913. onistic interactions between competition and insect her- Parmesan, C. 2000. Unexpected density-dependent effects of bivory on plant growth. Journal of Ecology 92:156±167. herbivory in a wild population of the annual Collinsia tor- Hare, J. D., and E. Elle. 2001. Geographic variation in the reyi. Journal of Ecology 88:392±400. frequencies of trichome phenotypes of Datura wrightii and Reader, R. J., and S. P.Bonser. 1998. Predicting the combined effect of herbivory and competition on a plant's shoot mass. correlation with annual water de®cit. MadronÄo 48:33±37. Canadian Journal of Botany 76:316±320. Hare, J. D., and E. Elle. 2002. Variable impact of diverse Rees, M., and V. K. Brown. 1992. Interactions between in- insect herbivores on dimorphic Datura wrightii. Ecology vertebrate herbivores and plant competition. Journal of 83:2711±2720. Ecology 80:353±360. Hare, J. D., and E. Elle. 2004. Survival and seed production SAS Institute. 2000. SAS for Windows. Version 8. SAS In- of sticky and velvety Datura wrightii in the ®eld: a ®ve- stitute, Cary, North Carolina, USA. year study. Ecology 85:615±622. Siemens, D. H., S. H. Garner, T. Mitchell-Olds, and R. M. Hare, J. D., E. Elle, and N. M. van Dam. 2003. Costs of Callaway. 2002. Cost of defense in the context of plant glandular trichomes in Datura wrightii: a three-year study. competition: Brassica rapa may grow and defend. Ecology Evolution 57:793±805. 83:505±517. Harper, J. L. 1977. Population biology of plants. Academic Uriarte, M., C. D. Canham, and R. B. Root. 2002. A model Press, London, UK. of simultaneous evolution of competitive ability and her- Karban, R. 1993. Induced resistance and plant density of a bivore resistance in a perennial plant. Ecology 83:2649± native shrub, Gossypium thurberi, affect its herbivores. 2663. Ecology 74:1±8. van Dam, N. M., and J. D. Hare. 1998. Biological activity Lee, T. D., and F. A. Bazzaz. 1980. Effects of defoliation and of Datura wrightii glandular trichome exudate against Man- duca sexta larvae. Journal of Chemical Ecology 24:1529± competition on growth and reproduction in the annual plant 1549. Abutilon theophrasti. Journal of Ecology 68:813±821. van Dam, N. M., J. D. Hare, and E. Elle. 1999. Inheritance Maron, J. L. 2001. Intraspeci®c competition and subterranean and distribution of trichome phenotypes in Datura wrightii. herbivory: individual and interactive effects on bush lu- Journal of Heredity 90:220±227. pine. Oikos 92:178±186. Weis, A. E., and M. E. Hochberg. 2000. The diverse effects Reports Marshall, D. R., and S. K. Jain. 1969. Interference in pure of intraspeci®c competition on the selective advantage to and mixed populations of Avena fatua and Avena barbata. resistance: a model and its predictions. American Naturalist Journal of Ecology 57:251±270. 156:276±292.