velickovic 2010

8/13/2019 Velickovic 2010

1/9

Reduced developmental stability inTilia cordata

leaves: effects of disturbed environment

Abstract

Background and Purpose:Developmental stability (DS) or homeosta-sis refers to the ability of an individual to produce a consistent phenotype ina given environment. Reduced DS can result from a wide variety of envi-ronmentally (or genetically) induced perturbations. The main aim of the

presented paper is to highlight the importance of the differences in ability ofTilia cordata leaves to buffer their development under contrasting environ-

mental conditions and points to the concept that developmental stability ischaracter specific.

Materials and Methods:Three different techniques were performed inthis study: fluctuation asymmetry (FA) and leaf size as integrative measuresof environmental stress during leaf developmental processes and within--plant variance in leaf morphology, presented as coefficient of morphologi-

cal variation (CV). The study tested the hypothesis that the population froma chronically polluted area would express greater developmental instabilityin leaf traits. Two bilateral, linear dimensions on each leaf: leaf width (LW)

and lobe length (LL) were analyzed.

Results: The three different measures of developmental stability allshowed a trend for T. cordata leaves in the polluted area to be developmen-tally less stable than leaves from the reference area. Leaves in the referencearea were significantly larger compared with those from the polluted site.

Although leaves tend to be larger on the outside of a trees crown, the patternfound here was the reverse. Both, outside and inside leaves from the pollutedarea had significantly higher FAs than leaves from the same position sam-pled in the reference area for both traits. Within-tree variance assessed as CVshowed that LL was a more variable measure than LW. Moreover, the datasuggest that LL is under more selective pressure to adapt to current environ-mental conditions than LW.

Conclusions:Obtained data suggest that T. cordata leaves may representa reliable indicator for developmental stability evaluation studies based onan assay using a combination of end-points. Furthermore, my results high-lighted the differences in ability of leaf morphometric characters to buffer

their development under contrasting environmental conditions.

INTRODUCTION

Developmental stability or homeostasis refers to the ability of an in-dividual to produce a consistent phenotype in a given environ-ment (1). By studying developmental stability it might be possible to as-sess the synergetic effects of toxic compounds or interactions between

MIROSLAVA V. VELI^KOVI]

Institute for Biological ResearchSini{a Stankovi}despot Stefan Blvd. 14211060 Belgrade, SerbiaE-mail: [email protected]

Received May 12, 2007.

PERIODICUM BIOLOGORUM UDC 57:61VOL. 112, No 3, 273281, 2010 CODEN PDBIAD

ISSN 0031-5362

Original scientific paper


2/9

pollutantsand other stressful factors (2), which are either

difficult or impossible to study by other means.Reduced developmental stability can result from awide variety of environmentally (or genetically) inducedperturbations(3, 4, 5).

Fluctuating asymmetry or FA(5),as random differ-ences in the development of both sides of a bilaterallysymmetrical character(6),has been proposed as an indi-cator of environmentalas well as genetic stress (3,7,8,9).Stress is considered to be a significant and lasting devia-tion from favorable conditions that leads to abnormal de-mands and destabilization of vital processes (10). Asstress during development may influence developmentalprecision, FA has been proposed as a potentially usefultool for monitoring stress levels in natural populations(1).Thus FA has been used to estimate developmentalinstability (DI), inability of a bilateral organ or organismto buffer its development against disturbances and toproduce a predetermined phenotype(11).However, themain drawback of the use of FA as a diagnostic tool is thedifficulty in discriminating the genetic from the environ-mental components producing FA in the field (12).Al-though developmental instability, measured as fluctuat-ing asymmetry, is expected to be positively related tostress and negatively to fitness, empirical evidence is of-ten lacking or contradictory when patterns are compared

at population level(13).Also, Parsons(8)has suggestedthat only extreme stressors, which are to be expected inmarginal or disturbed environments, can increase FAunder field conditions (may lead to detectable FA alter-ations in organisms).

Besides the FA there are other indicators of develop-mental instability such as the frequency of asymmetrictraits, the frequency of phenodeviants, the within-indi-vidual variance and the coefficient of variation (CV) orthe relative variation in size of a trait, both among andwithin populations (11). In this century, investigatorsdelving into the nature of species have collected data onthe variance of vegetative and floral characters in both

natural populations and common gardens (14, 15).Be-cause the plant body consists of repeated nodal units, avariety of within plant measures of variation can be usedto estimate stability (1, 16, 17). Studies using within--plant variance as a measure of stability suggest that vari-ation in developmental stability exists between lines andpopulations(18).Paxman(16)found differences in sta-bility among several lines of tobacco in leaf and flowertraits. However, care must be taken that the units used toestimate variance are indeed developmentally compara-ble andnot confounded by the effects ofprogrammed de-velopmental changes along the shoot(16)or by environ-mental changes over time (18). Under the accepted

theory, if there was strong developmental stability withinthe organisms growth processes, then morphologicalvariance would be small. Hence, this had led authors topredict that when developmental processes are highlystabilized, the coefficient of variation is small (11).The-refore, the presence of developmental stress is expected topositivelyrelatetothelevelsofFAandthephenotypicco-

efficient of variation for analyzed morphological traits.

Insupportofthis,(19) and (20) found a positive relation-ship between fluctuating asymmetry and the coefficientof variation inDrosophilaandPlantago majorL. respec-tively. However, there are sometimes (occasionally) pro-blems in applying and interpreting CV, especially whentesting for genetic differences among populations, for ex-ample, if trying (when attempting) to elucidate the rela-tionship between heterozygosity and developmental sta-bility. If the genes under investigation determine theexpression of the trait that is being assessed in terms ofdevelopmental stability, then it may be expected thatmorphological variance will rise, rather than fall, withincreased heterozygosity (21, 22, 23). In this case a stabil-

ity index of morphological variance could only be ap-plied if genes involved did not directly influence the de-velopment of the trait(24).

In the present study, developmental stability in twopopulations of the small-leaved lime Tilia cordataMill.(Tiliaceae), under different environmental conditions,was estimated using an assay based on a combination ofend-points. Three different techniques were performedin this study (leaf size and FA as integrative measures ofenvironmental stress during leaf developmental proces-ses and within-plant variance in leaf morphology pre-sented as coefficient of morphological variation). Thestudy tested the hypothesis that the population from a

chronically polluted area would express greater develop-mental instability in leaf traits.

MATERIALS AND METHODS

Study areas

Two sites were selected as sampling areas in Serbia.The Pan~evo site is a site close to a large industrial settle-ment (estate/zone). The industrial area (covers 290 hect-ares) is the site of a large petrochemical complex and fuelstorage site in Serbia and includes an ammonia plant(Azotara, founded in 1962), a factory for chemical fer-

tilizers (HIP Petrohemija, founded in 1975), and acrude oil refinery (Rafinerija Pan~evo, founded in1968). The data obtained from a long-term monitoringstudy of the Pan~evo site, based on the UNMIX Version2.4/MATLAB Version 6.5(25),at three measured loca-tions (Vojlovica, Star~evo and Vatrogasni dom) showedthat in the Pan~evo site extremely high concentrations oftoxic compounds were detected (26).Also, Pan~evo is ex-tremely windy and is exposed to the effects of two kindsof winds (southeastern and northwestern) which directlybring pollutants from the industrial zone. For this kindof investigation, however, it is important to emphasize/stress certain chemical compounds, their metabolites

and unwanted by-products because they represent themost dangerous chemical pollutants with long-term ne-gative effects on the environment, human health and liv-ing organisms. Results showed that the main contami-nants released from Azotara are: NH3, NH4+, NOx,SOx, CO(30).Results also indicated that Azotara, to-gether with other sources from the industrial zone have

274 Period biol, Vol 112, No 3, 2010.

Miroslava Velickovi} Developmental stability inTilia cordataleaves


3/9

released SO2 (9%), benzene (9%), NO2 (21%) and

TNMHC (7%) compared with the total amount of toxiccompounds found in the air. An investigation performedin HIP Petrohemija showed that the bio-filter is oneof the most dangerous contaminants of the air (withemissions of benzene of about 64kg h1 and of xilene ofabout 34kg h1). Furthermore, there are other pollutantsreleased in huge concentrationsfrom HIP Petrohemijasuch as: benzene, xilene, hydrocarbons from oil and oilderivatives, S, HS. Interestingly, up to 4.5% in oil deriva-tives is representedby sulphur. Also, in natural gas, HS ispresent (ed) in the range 0.865.40 mg/L. Moreover,Rafinerija Pan~evo is the dominant source of environ-mental contaminants such as volatile organiccompounds,

CO, SOx, Nox, polycyclic aromatic compounds (theirmetabolic transformations by aquatic and terrestrial or-ganisms into carcinogenic and mutagenic metabolites),benzene, xilene. The factory, therefore, represents thefirststepintheprocessofbenzeneandxileneproduction.The highest concentrations of benzene were detectedlate at nightand early in the morning, which is positivelycorrelated with the most intensive time of traffic flow. In2005 the average value for the amount of released ben-zene in the atmosphere was 5 mg m3 (26).Furthermore,recent papers have emphasized the presence of benzenein the air of urban environments caused by increasedtraffic flow(27).

Vince village is an area far from any known contami-nation and is practically closed to traffic, and is thereforeused as a control area. The two sites are approximately150 km apart (Figure 1).

Sampling and statistical treatment

Tiliaceae is a family of trees, shrubs or rarely herbs; itincludes 400 woody species, among those 3040 speciesofTilia, most of them found in the tropics. Ten speciesare found in the temperate region of the northern hemi-sphere. In Europe four species are present: Tilia cordataMiller, Tilia platyphyllosScop., Tilia tomentosaMoenchand Tilia dasystylaStev. (28). The core region for T.

cordatais central and Eastern Europe (29).In Serbia,T.cordatagrows in several types of mixed forest, i.e. com-monly in the Querco-Carpinetum forest type. The treesfavor good loamy site conditions, but they can also befound on sandy infertile soils. Climatic conditions andhuman impact have been a serious threat to the distribu-

tion ofTilia in most European countries (30). The leavesofT. cordataare simple, alternatively cordate or orbicu-late and the margins are serrate(31).

During July of 2005 leaves were obtained from 7 treesofT. cordatafrom two sites in Serbia.Trees were of simi-lar age/height ( 40 years/10m). Twenty-five fully devel-oped leaves with respect to the position of a leaf withinthe trees crown (inside, outside), thereby 50 leaves pertree (350 leaves per each site) were analyzed. All leaveswere sampled from the same height of approximately 2m. Leaves were washed, dried between sheets of filter pa-per, measured and then analyzed. Each analysis was per-

formed on the same leaves.Two bilateral, linear dimensions on each leaf: (1) leafwidth (LW) and (2) lobe length (LL) are presented inFigure 2. Both morphometric traits were measured witha digital caliper (0.01mm accuracy). To avoid measure-ment errors each measurement of the left and right sidesof each leaf was performed twice, during two independ-ent sessions. Additionally, all measurements were per-formed by the same person (M. V.). Trait size for leafwidthandlobelengthwascalculatedastheaveragevalueof right (R) and left (L) sides; (R+L)/2. Individual sig-ned asymmetry was calculated as (R-L) and absoluteasymmetry was calculated as the absolute value of the

(R-L).

Period biol, Vol 112, No 3, 2010. 275

Developmental stability inTilia cordataleaves Miroslava Velickovi}

Figure 1.Locations of the sampling sites.Figure 2.Drawings of T. cordata leaf showing lobe length (LL) andleaf width (LW) measurements taken.


4/9

To estimate FA values for the bilaterally symmetrical

character three indices of(6, 32)were used:1) FA1= meanR-L;

2) FA4 = variance (R-L);

3) As a better estimate of the true between sides vari-ance, si2 was calculated by partitioning measurement er-ror out of the side x individual mean squares of the twoway mixed-model ANOVAs. These ANOVAs were per-formed on all replicates and involved side (R or L) as afixed effect, individuals as random one, and their interac-tions. Additionally, this procedure allowed simultaneoustesting for the presence of DA (a significant mean squareof the side factor)(32).

Before proceeding with the asymmetry analysis, sta-tistical tests were carried out to detect the features con-founding analyses of FA(6, 32),suchasmeasurementer-ror, directional asymmetry (DA) and antisymmetry (AS).Firstly, the accuracy of the measurements was tested bycalculating a Pearson correlation coefficient (r2). Sec-ondly, deviation from normality of the (R-L) distribu-tions was assessedusing the Kolmogorov-Smirnov test ofnormality. These distributions were also tested for signif-icant skewness (g1) and kurtosis (g2) according to (33).Thirdly, a one-samplet-test for a departure of the meanof (R-L) from an expected mean of zero was done. Four-thly, to determine if asymmetry increases with leaf size, asimple linear regression of absolute asymmetry on leafsize;R-Lon (R+L)/2, was used(32).

The three-way factorial MANOVAs were used in or-der to assess the effects of site, position, tree and their in-teractions on leaf size and absolute asymmetry values ofleaf width and lobe length. The Sequential Bonferonnicorrections(34)were applied to avoid false significantresults in the MANOVAs.

Furthermore, FA values (FA4and the si2) are vari-ances, so differences between samples can be detected bycomparing the heterogeneity of variances. To detect dif-ferences between pairs of samples,F-test(35)was calcu-lated forboth FA4 andsi2.Thistestissimplyaratioofthelarger over the smaller variance. The significance of thisratio need only be looked up in a statistical table for theappropriate degrees of freedom.

Within-tree variance and coefficient ofvariation

Univariate ANOVAs were carried out in order to re-veal the possible within-tree differences among leavesmeasured.TheseANOVAsofeachtraitforeachsitewerecarried out separately with trees within sites designatedas a random effect. Because there were no significant dif-

ferences among leaves within trees under a given envi-ronmental condition (see RESULTS), within-tree sizedifferences as well as coefficients of variation were calcu-lated for each tree separately by site and by trait.

Additionally, in order to assess differences in CVs be-tween sites it was necessary to calculate averages of CVsfor each character measured inT. cordatapopulations.

In order to compare the mean values of CVs, Stu-

dentst-test was performed.

RESULTS

Preliminary results on leaf size andasymmetry analyses

The Pearson correlation coefficient (r2),betweenorig-inal and repeated measurements, for each side of eachleaf (for left side: r2 =0.999 and for right side r2 =0.999)showed that the measurements are reliable.

Descriptive statistics for leaf size and asymmetry inTilia cordatapopulations are presented in Table 1. All(R-L) distributions were normal (Kolmogorov Smirnovtest results: 0.060


5/9

cordataof long-term exposure to environmental pollut-ants had significantly greater absolute FAs compared tothose from the reference site (for LL: F(1,672)=6.369,p=0.014, for LW: F(1,672)=3.929, p=0.049). ANOVAsresults also indicated that significant differences in FAvalues resulted from differences between positions in theeffects of the site. Thus, outside leaves in the polluted

area were more asymmetrical.Furthermore, comparisons of the si2 values indicated

that there were significant differences between sites withrespect to the leaf position within a trees crown, for bothleaf traits. Both, outside and inside leaves from the pol-luted area had significantly higher si2 values than leavesfrom the same position sampled in the reference area, forboth traits(F-test results for LL: 20.981


6/9

(F-test results for LL: 15.429


7/9

caliptus sp. The leaves that were collected from a polluted

area showed reduction in all parameters investigated. Inprevious years otherworkers (investigators/studies?) alsoshowed significant reduction in different leaf variables inthe polluted environment in comparison with clean (un-polluted) atmosphere. Ninovaet al. (39) intheirstudyon

Platanus acerifolia showed changes in leaf blade and peti-ole size in polluted air. Similarly, leaf anatomy of theabove mentioned species also showed reduction in cuti-cle, epidermis, hypodermis, palisade, parenchyma cellsin polluted leaves as compared to leaves collected froman unpolluted area. Significant results were particularlyobserved in spongy parenchyma and lower epidermis in

F. bengalensisandEucalyptus sp., respectively. Changes in

shape and structure of thin walled mesophyll cells havebeen widely reported. Mesophyll cells are thin walledand are in direct contact with the environment throughstomates. The parenchytamous cells of spongy paren-chyma become flattened due to continuous exposure topollutants. A significant reduction in spongy parenchy-ma was in the leaves ofF. bengalensisof a polluted area.Similarly, Iqbal (40)has shown significant reduction inpalisade and spongy parenchyma in the leaves of whiteclover of a polluted population. On the other hand God-zik and Halbwacks(41)have shown fine and irregularcuticular folding on each epidermal cell of both adaxialand abaxial sides ofAesculus hippacastanumin the vicin-

ity of air pollution sources.Leaf size and shape can change significantly within a

plant due to developmental age of the plant and the posi-tion of the leaf within the plant (heterobplasty) (42).

Also, it has been reported that leaves tend to be larger ontheoutside of a trees crown, presumably because of greatlight intensity at this position(43).However, the patternfound here is the reverse; larger leaves are the insideleaves, both within and between sites. The significantposition by site effect on leaf size may therefore be ex-plained by the differences in intensity as well as in theavailability of the complex mixtures of environmentalstressors to affect exposed leaves, indicated that it is pos-sible that pollutants affect leaves differentially (differ-ently).

Additionally, the results of this study revealed that be-sides leaf size, leaf asymmetry also varies between sitesand within the crown ofTilia cordatatrees (Table 3). Al-thoughthereareexceptions,morerecentwork(research)has shown that developmental stability estimated by flu-ctuating asymmetry is predictable with respect to thehabitat quality. Developmental stability measured as fluc-tuating asymmetry is expected to be positively related tostress and negatively to habitat quality and fitness.

In this study leaves from the polluted area exhibited

significantly greatersi2 (FA) values(developmentally areless stable) than leaves from the unpolluted referencearea as expected (Table 1). The data presented here are inaccordance with those which reported that environmen-tal stresses such as pollution increase asymmetry inplants. For example, Sherhukova (44)estimated devel-opmental stability of small-leaved lime using fluctuating

asymmetry of the leaf traits. The authors concluded that

in the industrial, polluted region homeostatic develop-mental mechanisms are weakened and expressed in theincreased degree of leaf asymmetry than in control, re-served regions. Also, Veli~kovi}(20)pointed out the sig-nificantly higher FA levels (estimated as si2 values) forleaf width and vein distances within a leaf in Plantago

major L. in an urban, polluted area than in a control, un-polluted site. Grahamet al. (1)examined asymmetry ofblack locus leaves at several distances away from an am-monia production and storage facility in the Ukraine.They found that asymmetry declined with distance.Freemanet al. (45)report similar results for a variety ofplant populations around chemical production facilities

in Russia andthe Ukraine. Leaves of soybeans grown un-der high-voltage transmission lines had greater fluctua-tion asymmetry than those grown 100 m away (46).Kozlov et al. (47) found higher fluctuating asymmetrylevels of birch leaves near metal smelters.

In the opinion of the author differences in FAs be-tween sites may also be explained by differences in inten-sity as well as in availability of the complex mixtures ofchemical pollutants, their unwanted by-products andmetabolites to affect leaves. Additionally,si2 meansvaria-tion in the between-side differences among individuals,thus significant differences insi2 among sites implies ge-netic variations among individuals and/or environmen-tal variation in sites among individual plants. A polluted,artificially disturbed, area is often more homogeneousthananaturalareaandselectionduetopollutionmayre-duce genetic diversity (unpublished data). Thus, it is im-portant to consider that the possibility of genetic and/orenvironmental homogeneity causes this result ofsi2.

Many plants have sun and shade leaves and Mitton(48) suggested that these two kinds of leaves differ in de-velopmental stability. Based on results obtained for thefig (Ficus caricaL.), Cowart and Graham (43)found thatthe outside crown is a more stressful environment forleaves than the inner crown and that increased asymme-

try of outside leaves probably reflects stress rather thanplasticity. Although light intensity is less on the inside,leaves on the outside are subjected to greater cold, heat,ultraviolet light, visible light, desiccation, herbivore loadand environmental stresses(49).Thus, outer leaves, be-cause of their position, are more susceptible to the nega-tive effects of complex mixtures of environmental vari-ables, compared to inner leaves. Sakai and Shimamoto(50),however, showed that the effects of position on leafasymmetry inNicotiana tabacumdepended on the par-ticular genotype being examined.

Moreover, a large number of authors have used the

degree of morphological variation as an index of devel-opmental homeostasis, particularly for detecting differ-ences among populations(e.g. 18, 19, 20, 51, 52, 53).

Certain data from developmental stability analyses inthis study are similar to those reported by (20)for thecommon plantainP. major. Both,P. majorandT. cordataleaves from contaminated areas had increased FAs and




8/9


9/9

30. TURNER J 1962 The Tiliadecline: an anthropogenic interpreta-tion.New Phytologist 61: 328341

31. JOVANOVI] B 1972 TiliaceaeIn:Josifovic M (ed) Flora SR SrbijeIII p. 51132. PALMER A R 1994 Fluctuating asymmetry analysis: A primer In:

MarkowTA(ed) Developmental stability: Its originsand evolution-ary implications. Kluwer, Dordrecht, The Nederlands, p 335364

33. SOKAL R, ROHLF F 1981 Biometry: The principles and practiceof statistics in biological research. Freeman, New York.

34. RICE W 1989 Analyzing table of statistical tests. Evolution 43:223225

35. LEHMANN E L 1959 Testing statistical hypotheses. Wiley, NewYork, p 369

36. HEATH R L 2005 Air pollutants: Mode of action.In:Goodman B(ed)Encyclopedia of plant and crop science, p 15

37. KRAUSMAN P 1999Some basic principlesof habitat use.In: Laun-chbaugh K L, Sanders K L, Mosley J C (eds)Grazing behavior oflivestock and wildlife. Idaho forest wildlife and range exp. sta. bull.

70, Univ Idaho, Moscow ID38. JAHAN S, IQBAL Z M 1992Morphologicaland anatomical studies

of leaves of different plants affected by motor vehicles exhaust.Jour-nal of Islamic academy of sciences 5:2123

39. NIVOVA D J, DUSHKOVA P I, KOVACHEVA G V 1983 Anatom-ical, Morphological studies of Platanus acerifolia at various degreesof air pollution.Ekologiya (Sofia) 6:3547

40. IQBAL M Z 1985 Cuticular and anatomical studies of white cloverleaves from clean and air-polluted areas.Poll Res 4: 5961

41. GODZIK S, HALBWACKS G 1986 Structural alterations ofAescu-lus hippocastanumleaf surface by air pollutant.Z P Flanzekr Pflan-

zenchutz 93: 59059642. ALLSOPP A 1967 Heteroblastic development in vascular plants.

Advances in morphogenesis 6:12717143. COWART N, GRAHAM J 1999 Within- and among-individual

variation in fluctuating asymmetry of leaves in the fig (Ficus caricaL.).Int J Plant Sci 160: 116121

44. SHERZHUKOVA L V, KRIVTSOVA A N, MELUZOVA M, MI-SHALENKOVA Y N 2002 Estimation of developmental stability of

small-leaved lime on reserved and urbanized territories. Russianjournal of developmental biology 33:1112(2)

45. FREEMAN D, GRAHAMJ, EMLEN J 1993Developmental stabil-ity in plants: Symmetries,stress andepigenesist.Genetica89:97102

46. FREEMAND,GRAHAMJ,TRACYM,EMLENJ,ALADOSCL1999. Developmental instability as a means of assessing stress inplants: a case study using electromagnetic fields and soybeans. Int J

Plant Sci 160: 157166

47. KOZLOV M V, WILSEYB J, KORICHEVA J, HAUKIOJA E 1996.Fluctuating asymmetry of birch leavesincreasesunder pollution im-pact.Journal of Applied Ecology 33: 14891495

48. PALMER A R 1996Waltzingwithasymmetry.Bioscience46:518532

49. HALE M, ORCUTT D 1987 The physiologyof plants under stress.Wiley, New York.

50. SAKAI K, SHIMAMOTO Y 1965 Developmental instability inleaves and flowers ofNicotiana tabacum.Genetics 51: 801813

51.

MITTON J B 1978Relationshipbetweenheterozygosityfor enzymeloci and variation of morphological characters in natural popula-tions.Nature 273: 661662

52. KING D P F 1985Enzyme heterozygosityassociatedwith anatomi-cal character variance and growth in the herring (Clupea harengusL.).Heredity 54: 289296

53. YEZERINACSH,LOUGHEEDSC,HANDFORDP1992Mor-phological variability and enzyme heterozygosity. Individual andpopulation level correlations.Evolution 46: 19591964

54. FISHERR A 1930 The genetical theory of naturalselection. OxfordUniversity press, Oxford.

55. MAYNARD-SMITH J, BURIAN R, KAUFMANN S, ALBERCHP, CAMPBELL J, GOODWIN B, LANDE R, RAUP D, WOL-PERT L 1985 developmental constraints and evolution. Quarterly

Review of Biology 60: 265287

56. HOFFMANN A A, PARSONS P A 1993 Direct and correlated re-sponses to selection for desiccation resistance: a comparison ofDro-

sophila melanogaster andD. simulans.Journal of Evolutionary Biology6: 643657



velickovic 2010

Documents