Printer Friendly

Genetic variation and covariation for characteristics associated with cadmium tolerance in natural populations of the springtail Orchesella cincta (L.)

Abstract. - Heavy metals can be strong and stable directional selective agents for metal-exposed populations. Genetic variation for the metal-tolerance characteristic "cadmium excretion efficiency" was studied in populations of the collembolan Orchesella cincta from a reference- and a metal-contaminated forest soil. Previously it has been shown that "excretion efficiency" influences tolerance through midgut-mediated immobilization and excretion of toxic metal ions, and that an increased mean excretion efficiency is present in animals inhabiting metal-contaminated litter. In the present research, offspring-parent regressions showed that additive genetic variation for cadmium excretion efficiency was present in the population from the reference site. The heritability estimate was 0.33. In the natural population exposed to heavy metals from an industrial source, additive genetic variation was not significantly different from zero. Differences in the heritability between the reference and the exposed population were not significant. Genetic variation for cadmium excretion efficiency allows for a response to selection in the reference population. Such a response has probably occurred in the metal-exposed population. Half-sib analysis with animals from the reference population was used to estimate genetic variation and maternal effects for excretion efficiency, relative growth rate and molting frequency, and to determine genetic correlations between these characteristics. Additive genetic variation was demonstrated for all three characteristics, heritability estimates were 0.48, 0.75 and 0.46, respectively. Maternal effects were low for excretion efficiency and molting frequency, but may be present for relative growth rate. Phenotypic and genetic correlations among these characteristics were positive. The environmental correlation between relative growth rate and molting frequency was positive, others were negative. Direct selection for any of the characteristics, or genetic correlations between tolerance characteristics and growth characteristics, or both may have caused the responses previously observed in field populations.

Key words. - Adaptation, body growth, Collembola, genetic correlation, genetic variation, heavy metal, Orchesella cincta, tolerance.

The emergence of heavy-metal tolerant ecotypes of species has been called one of the best explained examples of natural selection in action (e.g., Bradshaw et al., 1990; Brandon, 1991). In contrast to many other environmental factors, metal contamination is considered to represent a stable, permanent, and often intense selection pressure. Natural populations showing increased metal tolerance have been reported frequently, especially for liverworts (Briggs, 1972), mosses (Shaw, 1990), angiosperms (Antonovics et al., 1971; Macnair, 1987) and aquatic plants and animals (Klerks and Weis, 1987). Only recently, genetic differentiation concerning metal tolerance in natural populations,of terrestrial animals has been observed, namely in springtails (Posthuma, 1990) and isopods (Donker and Bogert, 1991). Data on within-population genetic variability for metal tolerance as a measure of the potential to respond to selection have been reported for various species, but are lacking for liverworts and terrestrial animals.

Metal tolerance, defined as the ability to prevent, decrease or repair adverse effects of metals that have entered the body (Levitt, 1980), can develop during individual exposure (acclimation), or through selection acting upon genetic variation in tolerance (adaptation), or from a combination of both. In addition, exposure of females may influence tolerance of offspring by non-genetic maternal effects. The distinction between these phenomena is important, since tolerance will be maintained in future generations only in the case of adaptation.

For natural populations, various ecological and genetic parameters must be known before one can causally explain differences in tolerance (or other features) (Brandon, 1991; see, e.g., Brakefield, 1987). Comparisons of field populations allow for conclusions on the results of selection, without assigning causes (Endler, 1986), i.e., conclusions on selection of characteristics (sensu Sober, 1984). Selection for increased metal tolerance (sensu Sober, 1984) pertains only to populations where natural selection operated on tolerance. Comparisons of offspring raised in a clean environment may help to identify selection as the cause of inter-population differences, as argued by Klerks and Levinton (1989a). Quantitative genetic parameters, moreover, may provide information about the mechanism generating such differences (see, e.g., Hoffmann and Parsons, 1991) and for prediction of future performance (Willis et al., 199 1).

The genetic parameters studied here were heritability and maternal effect for each characteristic, and genetic correlations among characteristics. The variation for metal tolerance must be heritable for natural selection by heavy metals to have evolutionary consequences. Genetic correlations between tolerance characteristics and other characteristics imply that correlated responses may occur as a consequence of direct selection for increased tolerance. If maternal effects are present, differences between tolerant and non-tolerant populations may be modified by these effects, otherwise one can exclude them as an alternative explanation for interpopulation differences.

Selection may alter the existing pattern of additive genetic variation and covariation in metal-exposed populations. It is to be expected that genetic variation for tolerance characteristics will reduce as a consequence of directional selection (Falconer, 198 1; see, e.g., Stearns, 1984; Shaw, 1988). The term costs of tolerance" has been widely used in pollution studies (Brakefield, 1987; Wilson, 1988; and references therein) to indicate that tolerant genotypes perform less well than sensitive genotypes in the absence of metals. In most cases, however, results do not allow the distinction between phenotypic effects of exposure and consequences of selection processes. Wilson (1988) first demonstrated "costs of tolerance" in the restricted, genetic sense, expressed as a reduced relative growth rate of metal-tolerant clones of the grass Agrostis capillaris. According to this application of the cost principle, population comparisons of offspring reared in a common environment may indicate the presence of costs, while costs can be explained and predicted using genetic correlations between tolerance characteristics and other ecologically relevant characteristics. Bell and Koufopanou (1986), in their review on the methods to measure costs of reproduction," concluded that the latter approach resulted in the strongest evidence for the presence of costs; this approach was also used here to assess "costs of tolerance."

We previously inferred selection of metal tolerance in the springtail species Orchesella cincta (L.) from population comparisons of first generation laboratory animals (Table 1). For the present research, we chose to investigate genetic variation and covariation for characteristics associated with cadmium tolerance in this species, viz. cadmium excretion efficiency (EE) and body growth characteristics. Laboratory-cultured animals were used throughout the experiments, to avoid acclimation effects which may be present in animals from contaminated field sites. Direct toxic effects were considered to be insignificant or absent in the experiments, since the characteristic "excretion efficiency" is not influenced by the cadmium concentration in food (Van Straalen et al., 1987), and since weight parameters were measured before individuals were exposed to cadmium. For these reasons, costs, if present, were considered to be (at least partly) genetically determined, rather than to be toxic effects.

Specifically, this study aimed to 1) quantify genetic variation and covariation for cadmium excretion efficiency (EE), body growth (RGR) and molting frequency (MF) in a natural population from a reference site; 2) quantify maternal effects for EE, RGR and MF in a reference population; and 3) assess the effects of selection for genetic variation for cadmium EE in a metal-exposed population.

Table 1. Previously reported significant differences between characteristics of populations of Orchesella cincta originating from the reference site (R, Roggebotzand) and the metal-exposed site (S, Stolberg), measured in field captured animals (P field or first generation laboratory animals (F1). Comparisons demonstrate a higher mean value in the population from S compared to R(S > R), or the opposite.


Materials and Methods

Springtails (Collembola) are thought to be among the oldest groups of the insects. Fossils of the species Rhyniella praecursor have been found in lower Devonian strata (Hirst and Maulik, 1926; Scourfield, 1940; Whalley and Jarzembowski, 1981), whereas fossils of winged insects are known only from more recent strata (Bergstrom, 1979). Extensive morphological evolution has been demonstrated in the cave species Pseudosinella hirsuta (Christiansen and Culver, 1968); some present-day existing taxa, in contrast, have oral appendages similar to R. praecursor (Massoud, 1967). Genetic variation for any continuously variable characteristic, however, has not been demonstrated for Collembola. Springtail species mainly inhabit soil and litter. They inhabit a wide range of habitat types, including the polar regions, dry environments, tidal zones and polluted habitats (reviewed by Joosse, 1983), which indicates a high variability among Collembolan species. Apart from morphology, within-species variability has been demonstrated for allozymes (Dallai et al., 1986) and karyotypes (Hemmer, 1990).

Springtails may suffer from metal toxicity in industrial areas, since the organic layer of the soil tends to accumulate heavy metals (Jones et al., 1988). Many sites are, furthermore, locally contaminated by mining activities or surface ores. The history of selection at a site is determined by site characteristics (viz. age and kind of metal source, concentration, and the combination of metals) and species characteristics (viz. presence, exposure and sensitivity). In comparison with other soil arthropods, springtails seem to represent a group with a high degree of inherent metal tolerance and a potential for metal adaptation (Hagvar and Abrahamsen, 1990).

Selection may occur at sites where metals reach a toxic concentration for several genotypes. O. cincta inhabits sites where a toxic concentration of cadmium is present, i.e., the ambient concentration exceeds the No Observed Effect Concentration (NOEC) for growth or survival (Van Straalen et al., 1989), which indicates the probable presence of selection. Tolerance at such sites is expressed as a low growth reduction upon cadmium exposure and as increased immobilization and excretion of cadmium (Posthuma et al., 1992a). The degree of tolerance is associated with the cadmium concentration in the top soil (Posthuma, 1990), which makes drift less plausible as a major explanation of the observed differences. Juveniles from tolerant populations, moreover, show a high body growth rate, and maturation is early compared to reference populations (Posthuma et al., 1992b). In view of these data it is probable that adaptation has occurred in O. cincta, and that both tolerance and growth characteristics have been influenced by selection.

Springtails lack Malpighian tubules, and regulate their ionic uptake and excretion in the midgut epithelium, where excess ions may be immobilized in granules (Humbert, 1978). Springtails molt throughout their life, and after each molt the midgut epithelium is renewed, and the old epithelium, containing the waste products, is shed as "gut pellet." The efficiency of excreting waste products is expressed as the "excretion efficiency" (EE), for which we assume polygenic inheritance in the case of cadmium: it has a continuously variable expression, probably since midgut-mediated excretion involves morphological (Dallai, 1966) and physiological features (Humbert, 1978). We do not have information on the presence of a major gene for tolerance, e.g., for metallothionein, in Collembola.

The ecology and life-history of O. cincta are well-known (Joosse, 1981, 1983). Indirect fertilization via spermatophores and iteroparous reproduction allow controlled mating. In view of these species characteristics, O. cincta is suitable for quantitative genetic studies based on the comparison of relatives.

Sites and Capture. - Specimens of O. cincta were collected in two forest areas: Roggebotzand (R), The Netherlands (reference site, latitude 52 [degrees] 34'N, longitude 5 [degrees] 47'E), and Stolberg (S), Germany (200 m downwind from the lead smelter at Binsfeldhammer, latitude 50 [degrees] 46'N, longitude 6 [degrees] 14'E).

At least 1,000 animals were collected randomly from litter samples with an aspirator. Animals from R were captured in stands of Pinus nigra without undergrowth. The stands are intermixed with stands of Betula sp., Populus sp. and Fagus sp., in which O. cincta also occurs. Animals from S were captured in a stand with several tree species, on a slope facing the lead smelter. The landscape is characterized by small forest stands and industrial and agricultural activities. O. cincta also occurs in forest stands along a 6 km NE cline of decreasing metal concentrations.

Small-scale metal smelting at S was started in the sixteenth century, large-scale industrial smelting operates since the middle of the nineteenth century (Schwickerath, 1954; Schneider, 1982). Surface lead and zinc enriched ores occur in patches and have been exploited from the Roman ages till the first decades of this century. Small, abandoned mining tips are present at circa 700 m from the sampling location. Large-scale mining is known at distances of circa 2 km in the NE and S directions (Wei[BETA]enberghalde and Breinigerberg). Zinc-tolerant vegetation of the Violetum calaminariae association is present within 50 m from the capture site in S. The litter contains ([mu]umol.[g.sup.-1] dry wt) 0.56 Cd, 23.9 Zn, 41.1 Pb and 20.2 Cu (Posthuma, 1990); apart from cadmium, other metals may also be toxic, dependent on actual exposure regimes of O. cincta. The history of selection is dominated by the large-scale industrial emissions, since the species appears to be absent from a nearby mining area (i.e., Breinigerberg).

Culture Conditions and Observations. - Animals were kept in a climate room (20 [degrees] C, 75% relative humidity, LD 12 hr: 12 hr) in PVC jars with a layer of moist plaster of Paris. Weekly, mass cultures were offered twigs overgrown with green algae from the reference site (which contain circa 5.4 nmol Cd.[g.sup.-1] dry wt), to prevent acclimation to heavy metals. Animals and eggs present on the twigs were killed prior to feeding by drying at 40 [degrees] C for four days. Juveniles (F1) were separated from their parents with a sieve. Juveniles hatched in the same week were used in the experiments.

Depending on the age of animals, 1, 2 or 10 of them were kept in small culture boxes. Weekly, fresh food was offered as a thick paste of green algae, sampled from bark of Acer sp. trees at the reference site (R). Algae were sieved (0.25 mm), air-dried, and stored at 4 [degrees] in darkness. Boxes were checked daily for egg production, mortality, molting and presence of gut pellets.

Cadmium food was prepared by addition of a calculated volume of cadmium nitrate stock solution to reference algae, to obtain a nominal concentration of 0.356 [mu]mol.[g.sup.-1] dry wt (O/P experiment) 0 or 0.178 [mu]mol.[g.sup.-1] dry wt (half-sib experiment). Exposure to such food results in detectable amounts of cadmium in animal samples. Metals are rapidly absorbed by the algal cells, which have a large binding capacity (Joosse and Verhoef, 1983). Graphite furnace atomic absorption spectrophotometry (GFA; for details see Posthuma et al., 1992a) was used to determine actual cadmium concentrations for every food preparation from the offspring-parent experiment. Mean actual cadmium concentrations in the food were 0.470 [mu]mol.[g.sup.-1] dry wt (O/P experiment); variation (circa 10%) was independent of the age of algae.

Cadmium EE was determined for individual animals according to the procedure described in Van Straalen et al. (1987). Animals were exposed for three days after a molt. After production of a gut pellet in a clean box, boxes were frozen until analysis. Animals with gut pellets produced later than seven days after molting, or with an incomplete or broken gut pellet, were discarded. Body (C) and gut pellet (P) burden (nmol. [animal.sup.-1]) were determined pairwise with GFA. Exuviae were not sampled, as they do not contribute significantly to metal excretion (Van Straalen et al., 1987). Dry weights (Dw) were determined prior to metal analysis (accuracy: 1 [mu]g). The EE was calculated as EE = 100*P/(P + C) (%), i.e., the ratio of the excreted to the assimilated amount of cadmium. Accuracy of analytical procedures was determined with a standard (bovine liver, Community Bureau of References no. 18 5, Commission of the European Communities), run for every 70 samples, which made up a series. Mean recovery was 112%; variation of recovery did not depend on series. Cadmium concentrations in a series were not corrected for recovery. The procedural blank over all series was 0.8 pmol.[ml.sup.-1] (SD = 0.1).

Experiment Design. - The experiments consisted of offspring-parent regression analyses and half-sib analysis. These genetic models require several assumptions for correct interpretation, as summarized by Mitchell-Olds and Rutledge (1986). The circumstances under experimental control were designed so as to minimize violations of assumptions (e.g., large sample sizes); others were present through species characteristics [diploidy, 2n = 12 (Hemmer, 1990)], or were assumed to be present. Experiments were started with just hatched first-generation laboratory animals. Offspring families were made up of equal numbers of hatchlings. Individuals were discarded if an animal had assimilated less than 1 pmol (C + P), or if no gut pellet was produced; families were discarded if they consisted of one or two offspring; the (mean) EE cannot reliably be measured in these cases. Data of both experiments were unbalanced for family size.

Offspring-parent Regression. - Offspring-parent regression analyses were used to estimate the narrow-sense heritability for cadmium EE for both populations, according to Falconer (1981, p. 152). We aimed to measure the maximum number of families which can be handled in a single experiment, i.e., two populations of 80-100 pairs with five offspring per family, which is reasonable compared to data from Klein (1974) when true heritability is low.

Specimens were collected in the reference area R and the contaminated area S, on subsequent days in July 1990. Newly hatched F1 animals from eggs produced in the fourth week after capture were used as the "parent" generation; thus, animals from S had lost most of their metal burden (Posthuma et al., 1992a). Animals were reared in groups of 10 (age 2-28 days after hatching) or singly (age 28-63 days). Between the age of 63 and 119 days they were kept in pairs in new rearing boxes. Cadmium EE was determined for individuals of reproducing pairs, other pairs were discarded. After hatching, the offspring generation was cultured and analyzed in a similar way, the EE procedure was started at the age of 63 days. Heritabilities were calculated separately from regressions on female, male and midparent values. Within-population tests for similarity of regression on male and female parents, respectively, were executed according to Neter et al. (1990). If maternal effects are present the slopes of both regressions differ; if so, the heritability estimated from the regression on the father is considered the most reliable.

Half-Sib Analysis. - Half-sib analysis was performed to estimate components of variation and covariation in the reference population, according to Falconer (1981, p. 155) and Sokal and Rohlf (1981), and to estimate upper limits for maternal effects. Juveniles (F1, weight about 0.4 mg), reared individually, were taken as the parent generation. Controlled crosses were made by transfer of spermatophores on twigs to boxes with a receptive Oust molted) female. Ten males were crossed with four females each. Females were separated from their clutch before hatching. Juveniles were put in groups of ten animals 10 days after hatching. At the ages of 24 and 52 days, individual fresh weight was determined. Molting frequency (MF) was determined between 24 and 52 days. The cadmium EE procedure was started at the age of 52 days. Relative growth rates (RGR) of individuals were calculated from the fresh weight data. Pearson correlation coefficients were calculated to determine phenotypic correlations among characteristics.

Heritabilities for RGR and MF were estimated from sire components of variation in the denominator, since the dam component of variation exceeded the sire value; for EE the heritability was estimated from sire and dam components together. Genetic correlations were calculated from the sire component of covariation and the pertinent variances. Furthermore, the upper limit of the variance due to common environment, the value resulting from subtracting the sire from the dam component, was calculated. Common environment variance was assigned the value zero if the subtraction result was negative. Upper limits of maternal effects were calculated as the ratio of the common environment variance and the total phenotypic variance, assuming that the former consists of maternal effects only (Falconer, 1981; Janssen, 1985).

Unbalanced designs may introduce bias in estimates for genetic parameters. Coefficients of variation (CV) for full and half-sib family size were 39 and 23%, respectively. Bias introduced at the present CV values is considered to be low (Caro et al., 1985). Standard errors of heritability estimates were approximated according to Funkhouser and Grossman (1982) and of genetic correlations according to Falconer (1981); standard errors of maternal effects and environmental correlations were not calculated.

Statistical Analyses. - Normality was tested according to Wilk and Shapiro (1968), homogeneity of variances according to Sokal and Rohlf (1981). Data transformations were executed if assumptions of the pertinent tests were violated. Calculations were done with the integrated SPSS program which was implemented on a Cyber 170-750 computer.


Offspring-Parent Regression. - Survival and reproduction characteristics were similar in both populations (survival 2-28 days: R 94%, S 92%; 28-63 days: R and S both 100%; number of clutches: total R = 105, S = 103; mean clutch size: similar). Undeveloped clutches (R = 5, S = 18) consisted mainly of one or two eggs, and in a few clutches the eggs were overgrown with fungi at the first day of observation. Offspring mortality differed slightly between the populations (S > R), but mortality occurred before metal treatment, so that the difference appears not to be directly related to tolerance.

Three-way ANOVA (population, sex, generation) for EE demonstrated a main effect of population only. The mean [EE.sub.R] was higher than the mean [EE.sub.S] (circa 1.6%, P < 0.01; see also Fig. 1), in contrast to the direction and magnitude of cadmium EE differences repeatedly determined in previous experiments (circa 10-15%, R < S, for P field and F1 animals; see Table 1).

Genetic variation for cadmium EE was demonstrated only for population R (Fig. 1; Table 2). Within-population comparisons demonstrated that regressions on father and mother were similar (for population R: F = 0.01, P > 0.05; for population S: F = 0.05, P > 0.05). Apparently, significant maternal effects were absent for EE. Also, significant interpopulation differences were absent (Table 2). Homogeneity tests to assess interpopulation differences of regression lines, however, were not powerful. For example, the probability of correctly rejecting the null hypothesis (populations have similar heritabilities), given the alternative hypothesis [h.sup.2.sub.R] = [h.sup.2.sub.s.] + 0.2 (at [alpha] = 0.05), is only 0.29 for regression on midparent EE (Klein, 1974; Sokal and Rohlf, 1981). At least 400 families of both populations will be needed to increase the power above 0.75 in this test. Heritability estimates based on regression on male parents demonstrated the most prominent, but not significant trend of interpopulation difference. If true heritabilities are similar, and both populations are considered subsamples of a large homogeneous population, then a pooled heritability estimate can be calculated. The heritability estimated from pooled data, using midparent regression, was 0.17 (SE = 0.04, P < 0.001).


It is concluded that additive genetic variation for cadmium EE is present in population R; a response is to be expected if selection operates on this characteristic. Such variation was not found in S. This probably indicates a decrease of additive genetic variation for EE in population S in comparison with R. However, this interpretation is not firmly supported by a significant difference between the regression lines of both populations.

Half-Sib Analysis. - Variance components and narrow-sense heritability estimates derived from these variance components, are summarized in Table 3 for RGR, MF and EE. Only for EE the dam component of variation was lower than the sire component, and the heritability estimate was calculated with the sum of the sire and the dam components. Additive genetic variation was significant for all characteristics. Upper limits for maternal effects indicate the probable absence of maternal influences for EE and MF; for RGR a maternal contribution to the offspring's phenotype may be present.


Genetic, phenotypic and environmental correlations are shown in Table 4. Phenotypic and genetic correlations all showed a positive sign, and were significantly different from zero, except for the correlation between EE and R GR ([t.sub.s] = 1.86, P < 0. 1). The standard error of the genetic correlation between RGR and MF is undefined, as the estimate exceeded one. This strong correlation indicates that both characteristics are influenced in the same way by the same set of polygenes. From the genetic correlations it is predicted that selection for one of the characteristics may influence the other characteristics in the same direction. A positive environmental correlation was found between MF and RGR, the other two were negative. Molting frequency and RGR were similarly influenced by the environment, whereas EE and RGR, and EE and MF responded differently to environmental variation.


Evolution of Metal Tolerance. - The present research has demonstrated additive genetic variation and covariation for body growth and a metal tolerance characteristic in a Collembolan species. The findings imply that a reference population may show a response to selection upon heavy-metal exposure. Evolutionary responses may result from direct selection for the characteristic, or from genetic correlations, or both. A decrease of genetic variation in a metal-exposed population may, furthermore, be present as a consequence of selection.

For O. cincta populations from the sites Stolberg (S), Plombieres (an abandoned lead mine), and Budel (a zinc smelter), we previously inferred selection of (sensu Sober, 1984) cadmium EE, growth rate of animals fed uncontaminated food, age at maturity and maintenance of body growth during cadmium exposure (Table 1). Evidence consisted of toxicity data, and of correlations between cadmium concentrations in the litter and population characteristics. Genetic variation for cadmium EE, RGR and MF, moreover, indicates that selection for these characteristics (sensu Sober, 1984) may have occurred in populations with increased tolerance, including site S, provided that the genetic parameters apply to any reference population.

Age and fresh weight at maturity, clutch size and survival probability to maturity are characteristics associated with fitness in O. cincta (Janssen, 1985; Van Straalen, 1985). In exposed populations, moreover, an increased EE will improve fitness, because it is positively correlated with maintenance of body growth during cadmium exposure (Posthuma et al., 1992a). In exposed populations selection is likely to operate on EE, but also on body growth, to balance exposure-mediated growth reduction and delayed reproduction. A response to selection for these characteristics is expected from theoretical considerations of life-history evolution, and has been shown, e.g., in populations subject to size-specific predation (Reznick et al., 1990). A genetically determined reduction of the age at maturity has been demonstrated in metal-exposed field populations of O. cincta (Posthuma et al., 1992b) and of the isopod Porcellio scaber (Donker et al., 1992). The increase of both RGR and EE may also result from the positive genetic correlation between these characteristics (P < 0.1). This correlation will tend to cause correlated response in a similar direction if selection operates on one of the characteristics.

The midgut determines the major fluxes of metal ions through the body in many terrestrial invertebrate species (Hopkin, 1989). Studies on the role of the midgut in tolerance have focused on metallothioneins (MTs) and metal-binding granula. Metallothioneins, proteins directly involved with metal tolerance, are present in many invertebrate species (Maroni, 1990). In Drosophila melanogaster, specimens from industrialized regions carry duplications of the MT gene Mtn, and exhibit a higher metal tolerance than flies from pristine areas (Maroni et al., 1987). Expression of the Mtn-gene is induced by cadmium exposure, predominantly in midgut cells, where metals are eventually fixed in granula. These data suggest the presence of a major gene for metal tolerance in Drosophila. Results of artificial selection experiments have, however, indicated that genetic variation for metal tolerance in Drosophila is, at least partly, of a quantitative kind (Nassar, 1979; Magnusson and Ramel, 1986). In O. cincta, the midgut-mediated cadmium EE also appears to be continuously variable, so that it can be studied with the applied methods, despite the possibility that a major gene may partly determine tolerance.

At the start of exposure, maternal effects do not directly influence the response to selection for EE. For this characteristic, by way of theoretical example, a population is expected to show an increase for mean EE of 10% within eight generations, if there is a constant heritability of 0.35 and if 50% of the population contributes to the next generation. Such calculations show that the increased EE of the population at site S may have developed within the period of major industrial emissions (circa 150 yr, 300 generations), even if selection intensity is low, or if genetic variation decreases upon selection. A fast increase of tolerance has also been observed in plant populations near metalworks (Wu and Bradshaw, 1972; Gartside and McNeilly, 1974; Ernst, 1976). Maternal effects may, however, influence the response to selection of several characteristics in O. cincta. Strong negative maternal effects have been reported for age at maturity (Janssen et al., 1988). The observation that age at maturity is reduced in F1-animals originating from industrially contaminated sites or former mines indicates that negative maternal effects may have affected the magnitude of the expected response rather than its very occurrence or direction.

To our knowledge, one other estimate of additive genetic variation for metal tolerance is available for a reference population of an animal species, namely the benthic oligochaete Limnodrilus hoffmeisteri. A high genetic variation was present for survival time under metal exposure (Klerks and Levinton, 1989b). The prediction that control worms would be able to develop tolerance within a few generations matched with results of an artificial selection experiment and of population comparisons of laboratory-reared worms. Genetic variation for metal tolerance has also been reported for angiosperms (Macnair, 1990) and mosses (Shaw, 1988). Usually it is intermediate or high for species which may develop metal tolerant populations (Gartside and McNeilly, 1974; Bradshaw and McNeilly, 1981). It is also high for a species that has not been found on soil containing heavy metals, which is probably due to other edaphic factors (Gartside and McNeilly, 1974; Wu, 1990). These examples show, in addition to the results for the animal species O. cincta and L. hoffmeisteri, that additive genetic variation for metal tolerance is an important factor, but not the soie one, determining the evolution of tolerance. In O. cincta, genetic variation for life-history characteristics may contribute to successful adaptation to metal pollution, so that metal adaptation in this species can be characterized as a "complex adaptation strategy" sensu Ernst (1983).

Consequences of Metal Tolerance. - Individual tolerance results from genotype and environment, mediated by physiological characteristics. In 0. cincta tolerance has been changed through natural selection, and costs of tolerance can (at least partly) be attributed to genetic differences between populations.

O. cincta has an inherent degree of tolerance, since the excretion mechanism for toxic ions is associated with normal ion regulation. The phenotypic expression of the excretion mechanism is, however, highly variable, but with relatively large sample sizes ecologically relevant differences of EE between populations have repeatedly been demonstrated (Table 1). It is not clear whether all variation is interindividual, since intraindividual variation of EE cannot be determined due to the destructive nature of the measurement method. Intraindividual variation, or phenotypic plasticity, may account for the apparent absence of the expected difference between the mean EE values of R and S (Fig. 1). This phenomenon may have been caused by the method of rearing: animals were reared on plaster of Paris since hatching, whereas in previous experiments animals were kept on this substrate for a single moulting interval. Cadmium EE is influenced by the duration of exposure to cadmium (Posthuma et al., 1992a); similarly, it may be forced to a certain level in both populations, through acclimation of the ion regulation mechanism to excess calcium in the substrate, since calcium ions are involved in granule formation (Humbert, 1978). This problem, however, bears only on the EE, whereas differences between populations were consistent for other characteristics, e.g., body growth.

Phenotypic correlations may, in principle, indicate physiologic costs of tolerance (Weis and Weis, 1989). In ). cincta, such costs may result from competition for binding sites between toxic and nutrient ions (see Posthuma et al., 1992a) and from an investment in organic molecules for granule formation (Humbert, 1978). At the phenotypic level, cadmium EE appeared to be positively correlated with lead EE (Van Straalen et al., 1987), and with body growth in the reference population (Table 4). Opposite environmental and genetic correlations, however, obscured the phenotypic correlation between molting frequency and EE. A decreased body growth rate, which has been suggested as a cost of tolerance in other species (e.g., Wilson, 1988), or other costs, therefore, have not been uncovered with certainty by phenotypic correlations, or by the usual population comparisons approach. This supports the opinion expressed by Willis et al. (1991) that genetic parameters are needed to predict future performance.

Contrasting results have been presented from other studies that aimed to obtain genetic evidence for consequences of selection for non-tolerance characteristics. Relative growth rate of eight clones of Agrostis capillaris from various clean and contaminated sites, grown in clean culture, was negatively correlated with metal tolerance (Wilson, 1988). In contrast, phenotypic and genetic correlations were positive between growth rate and metal tolerance in Funaria hygrometrica (Shaw, 1988, 1990). For O. cincta, the genetic correlation between EE and RGR tended to be positive (P < 0.1). The ecological implications of this correlation can be evaluated in view of the phenology of a field population, which consists of two generations per year (Van Straalen, 1985). The ecological cost (or with a neutral description: consequence) of selection for cadmium EE is that the generation time of a tolerant population, in the absence of heavy metals, is short in comparison with that of a non-tolerant population under similar conditions, which may imply that reproductive activity becomes asynchronous with seasons.

Strong evidence for a decrease of genetic variation for EE as a consequence of selection was absent in the present experiments. A major experimental effort will be needed to demonstrate a significant reduction of additive genetic variation, given the range of heritability values found here for EE in a reference population (0.33-0.48). A comparison of regressions of offspring on fathers is the most reliable method to detect the effects of selection for genetic variation, since maternal effects do not influence the heritability estimates in this case. The test indicated that the expected trend ([h.sup.2.sub.s] < [h.sub.R]) tended to be present (P < 0.25). Decreased genetic variability associated with selection by metals has been demonstrated in natural populations of the moss Funaria hygrometrica for 10 out of 14 growth and tolerance characteristics (Shaw, 1988). In contrast, heritability estimates for copper tolerance for populations of Agrostis tenuis from a reference and a mine site were similar (Gartside and McNeilly, 1974). Judged by electrophoretic data, selection needs not be associated with a decreased overall variability for enzyme loci, as was shown for O. cincta (Frati et al., 1992), Silene cucubalus (Verkleij et al., 1985) and Funaria hygrometrica (Shaw, 1991). Verkleij et al. suggested that these observations may have resulted from the absence of tolerance encoding loci in the samples of loci scored. Further studies are needed to quantify the effect of selection for metal tolerance on genetic variation for tolerance characteristics, but also on genetically correlated characteristics, in O. cincta and other species.

Population and Site Characteristics. - Data on the population structure of tolerant and neighboring populations and on habitat heterogeneity are needed to indicate the relative contribution of processes other than natural selection to population differentiation (Brandon, 1991). In earthworms, for example, continuous immigration and patchy metal distribution may be the explanation for the absence of an adapted population at a metal-contaminated site (Bengtsson et al., 1992). In springtails, the spread of tolerance genes may result from passive dispersion by wind currents (Johnson, 1969), in particular of individuals that show tree-climbing behavior in response to habitat fluctuations (Bowden et al., 1976). The absence of tolerant populations near tolerant ones in Stolberg (S) and Plombieres (cf. Simon, 1978) indicates that gene flow from tolerant populations is low, or that costs of tolerance are high, or both. This cannot be evaluated from the present data.

Our findings for O. cincta have established the presence of a selection response in this species with respect to the criteria proposed by Brandon (1991). The response has occurred on a micro-evolutionary time-scale, which has not been reported before for Collembola. In the Funaria case, and probably also the Orchesella case, moreover, the genetic variation which is normally maintained in uncontaminated populations has been altered, and is probably reduced in an exposed population. This may constrain the possibility to respond to future environmental characteristics, such as further metal exposure, or a period of extreme drought in the case of springtails. More examples are needed to evaluate the occurrence and intensity of effects of pollution-mediated selection upon patterns of genetic variation and covariation in view of the duration and the intensity of selection events.

Literature Cited

Antonovics, J., A. D. Bradshaw, and R, G. Turner. 1971. Heavy-metal tolerance in plants. Adv. Ecol. Res. 7:1-85. Bell, G., and V. Koufopanou. 1986. The cost of reproduction, pp. 83-131. In R. Dawkins and M. Ridley (eds.), Oxford Surveys in Evolutionary Biology, Vol. 3. Oxford University Press, UK. Bengtsson, G., H. Ek, and S. Rundgren. 1992. Evolutionary response of earthworms to long-term metal exposure. Oikos. 63:289-297. Bergstrom, J. 1979. Morphology of fossil arthropods as a guide to phylogenetic relationships, pp. 3-58. In A. P. Gupta (ed.), Arthropod Phylogeny. Van Nostrand Reinhold Company, N.Y., USA. Bowden, J., L. H. Haines, and D. Mercer. 1976. Climbing Collembola. Pedobiologia 16:298-312. Bradshaw, A. D., and T. McNeilly. 1981. Evolution and Pollution. Studies in Biology 130. Edward Arnold Publ., London, UK. Bradshaw, A. D., T. McNeilly, and P. D. Putwain. 1990. The essential qualities, pp. 323-334. In A. J. Shaw (ed.), Heavy-Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton, FL USA. Brakefield, P. M. 1987. Industrial melanism: Do we have the answers? Trends Ecol. Evol. 2:117-122. Brandon, R. N. 1991. Adaptation and Environment. Princeton University Press, NJ USA. Briggs, D. 1972. Population differentiation in Marchantia polymorpha L. in various lead pollution levels. Nature 238:166-167. Caro, R. F., M. Grossman, and R. L. Fernando. 1985. Effects of data imbalance on estimation of heritability. Theor. Appl. Genet. 69:523-530. Christiansen, K., and D. Culver. 1968. Geographical variation and evolution in Pseudosinella hirsuta. Evolution 22:237-255. Dallai, R. 1966. L'uitrastruttura dell'intestino di Orchesella villosa (Geoffroy) (Insecta, Collembola). Ann. 'Ist. Mus. Zool. Univ. Napoli 17:1-18. Dallai, R., P. P. Fanciulli, and R. Petrucci. 1986. Enzyme diversity in the genus Bilobella (insecta collembola). Rev. Ecol. Biol. Soc. 23:333-348. Donker, M. H., and C. G. Bogert. 1991. Adaptation to cadmium in three populations of the isopod Porcellio scaber. Comp. Biochem. Physiol. 100C: 143-146. Donker, M. H., C. Zonneveld, and N. M. Van Straalen. 1992. Early reproduction and increased reproductive allocation in metal adapted populations of the terrestrial isopod Porcellio scaber. Oecologia. In press. Endler, J. A. 1986. Natural Selection in the Wild. Princeton University Press, NJ USA. Ernst, W. H. O. 1976. Physiological and biochemical aspects of metal tolerance, pp. 115-133. In T. A. Mansfield (ed.), Effects of Air Pollutants on Plants. Cambridge University Press, UK. _____. 1983. Okologische Anpassungsstrategien an Bodenfaktoren. Ber. Deutsch. Bot. Ges. 96:49-71. Falconer, D. S. 198 1. Introduction to Quantitative Genetics, 2nd ed. Longman, N.Y., USA. Frati, F., P. P. Fanciulli, and L. Posthuma. 1992. Allozyme variation in reference and metal exposed populations of Orchesella cincta (L.) (Insecta, Collembola). Biochem. Syst. Ecol. 20:297-310. Funkhouser, E. M., and M. Grossman. 1982. Bias in estimation of the sampling variance of an estimate of heritability. J. Hered. 73:139-141. Gartside, D. W., and T. McNeilly. 1974. The potential for evolution of heavy metal tolerance in plants. II. Copper tolerance in normal populations of different plant species. Heredity 32:335-348. Hagvar, S., and G. Abrahamsen. 1990. Microarthropoda and Enchytraeidae (Oligochacta) in naturally lead-contaminated soil: A gradient study. Environ. Entomol. 19:1263-1277. Hemmer, W. 1990. Karyotype differentiation and chromosomal variability in springtails (Collembola, Insecta). Biol. Fertil. Soils 9:119-125. Hirst, S., and S. Maulik. 1926. On some arthropod remains from the Rhynie Chert (old red sandstone). Geol. Mag. 63:69-71. Hoffmann, P. A., and P. A. Parsons. 1991. Evolutionary Genetics and Environmental Stress. Oxford University Press, UK. Hopkin, S. P. 1989. Ecophysiology of Metals in Terrestrial Invertebrates. Elsevier Applied Science, London, UK. Humbert, W. 1978. Cytochemistry and X-ray microprobe analysis of the midgut of Tomocerus minor Lubbock (Insecta, Collembola) with special reference to the physiological significance of the mineral concretions. Cell Tissue Res. 187:397-416. Janssen, G. M. 1985. On the genetical ecology of springtails. Ph.D. Diss. Vrije Universiteit, Amsterdam, The Netherlands. Janssen, G. M., G. DeJong, E. N. G. Joosse, and W. Scharloo. 1988. A negative maternal effect in springtails. Evolution 42:828-834. Johnson, C. G. 1969. Migration and Dispersal of Insects by Flight. Methuen and Co., London, UK. Jones, K. C., S. A. Watts, A. F. Harrison, and J. Dighton. 1988. The distribution of metals in the forest floor of aged conifer stands at a plantation in northern England. Environ. Pollut. 51:31-47. Joosse, E. N. G. 1981. Ecological strategies and population regulation of Collembola in heterogeneous environments. Pedobiologia 21:346-356. _____. 1983. New developments in the ecology of Apterygota. Pedobiologia 25:217-234. Joosse, E. N. G., and S. C. Verhoef. 1983. Lead tolerance in Collembola. Pedobiologia 25:11-18. Klein, T. W. 1974. Heritability and genetic correlation: Statistical power, population comparisons, and sample size. Behav. Genet. 4:171-189. Klerks, P. L., and J. S. Levinton. 1989a. Effects of heavy metals in a polluted aquatic ecosystem, pp. 41-67. In J. R. Kelly and K. D. Kimball (eds.), Ecotoxicology: Problems and Approaches. Springer-Verlag, N.Y., USA. _____. 1989b. Rapid evolution of metal resistance in a benthic oligochaete inhabiting a metal-polluted site. Biol. Bull. 176:135-141. Klerks, P. L., and J. S. Weis. 1987. Genetic adaptation to heavy metals in aquatic organisms: A review. Environ. Pollut. 45:173-205. Levitt, J. 1980. Responses of Plants to Environmental Stress. Acadelpic Press, N.Y., USA. Macnair, M. R. 1987. Heavy-metal tolerance in plants: A model evolutionary system. Trends Ecol. Evol. 2:354-359. _____. 1990. The genetics of metal tolerance in natural populations, pp. 236-251. In A. J. Shaw (ed.), Heavy-metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton, FL USA. Magnusson, J., and C. Ramel. 1986. Genetic variation in the susceptibility to mercury and other metal compounds in Drosophila melanogaster. Teratogen. Carcinogen. Mutagen. 6:289-305. Maroni, G. 1990. Animal metallothioneins, pp. 216-229. In A. J. Shaw (ed.), Heavy-metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton, FL USA. Maroni, G., J. Wise, J. E. Young, and E. Otto. 1987. Metallothionein gene duplications and metal tolerance in natural populations of Drosophila melanogaster. Genetics 117:739-744. Massoud, Z. 1967. Contribution a l'etude de Rhyniella praecursor Hirst et Maulik 1926, Collembola fossile du Dbvonien. Rev. Ecol. Biol. Sol. 4:497-505. Mitchell-Olds, T., and J. J. Rutledge. 1986. Quantitative genetics in natural plant populations: A review of the theory. Am. Nat. 127:379-402. Nassar, R. 1979. Genetics of resistance to tetraethyllead. Aust. J. Biol. Sci. 32:127-132. Neter, J., W. Wasserman, and M. H. Kutner. 1990. Applied Linear Statistical Models. Regression, Analysis of Variance and Experimental Designs, 3rd ed. Irwin, Homewood, IL USA. Posthuma, L. 1990. Genetic differentiation between populations of Orchesella cincta (Collembola) from heavy-metal contaminated sites. J. Appl. Ecol. 27: 609-622. Posthuma, L., R. F. Hogervorst, and N. M. Van Straalen. 1992a. Adaptation to soil pollution by cadmium excretion in natural populations of Orchesella cincta (L.) (Collembola). Arch. Environ. Contam. Toxicol. 22:146-156. Posthuma, L., R. A. Verweij, B. Widianarko, and C. Zonneveld. 1992B. Life-history patterns in metal-adapted Collembola. Oikos. In press. Reznick, D. A., H. Bryga, and J. A. Endler. 1990. Experimentally induced life-history evolution in a natural population. Nature 346:357-359. Schneider, F. K. 1982. Untersuchungen uber den Gehalt an Blei und anderen Schwermetallen in den Boden und Halden des Raumes Stolberg (Rheinland). Geol. Jahrb. 53(D):3-31. Schwickerath, M. 1954. Die Landschaft und ihre Wandlung auf geobotanischer und geographischer Grundlage entwickelt und erldutert im Bereich des Me[BETA]tischblattes Stolberg. Veriag Dr. Rudolf Georgi, Aachen, Germany. Scourfield, D. J. 1940. The oldest known fossil insect. Nature 145:799-801. Shaw, J. 1988. Genetic variation for tolerance to copper and zinc within and among populations of the moss, Funaria hygrometrica Hedw. New Phytol. 109:211-222. _____. 1990. Metal tolerance in bryophytes,pp. 133-152. In A. J. Shaw (ed.), Heavy-Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton, FL USA. _____. 1991. The genetic structure of sporophytic and gametophytic populations of the moss, Funaria hygrometrica Hedw. Evolution 45:1260-1274. Simon, E. 1978. Heavy metals in soils, vegetation development and heavy-metal tolerance in plant populations from metalliferous areas. New Phytol. 81:175-188. Sober, E. 1984. The Nature of Selection. Evolutionary Theory in Philosophical Focus. MIT Press, Cambridge, MA USA. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman, San Francisco, CA USA. Stearns, S. C. 1984. Heritability estimates for age and length at maturity in two populations of mosquito-fish that shared ancestors in 1905. Evolution 38:368-375. Van Straalen, N. M. 1985. Comparative demography of forest floor Collembola populations. Oikos 45:253-265. Van Straalen, N. M., T. B. A. Burghouts, M. J. Doornhof, G. M. Groot, M. P. M. Janssen, E. N. G. Joosse, J. H. Van Meerendonk, J. P. J. J. Theeuwen, H. A. Verhoef, and H. R. Zoomer, 1987. Efficiency of lead and cadmium excretion in populations of Orchesella cineta (Collembola) from various contaminated forest soils. J. Appl. Ecol. 24: 953-968. Van Straalen, N. M., J. H. M. Schobben, and R. G. M. De Goede. 1989. Population consequences of cadmium toxicity in soil microarthropods. Ecotoxicol. Environ. Saf. 17:190-204. Verkleij, J. A. C., W. B. Bast-Cramer, and H. Levering. 1985. Effects of heavy-metal stress on the genetic structure of populations of Silene cucubalus. pp. 355-365. In J. Haeck and J. W. Wolendorp (eds.), Structure and Functioning of Plant Populations, Vol. 2. North-Holland Publishing Company, Amsterdam, The Netherlands. Weis, J. S., and P. Weis. 1989. Tolerance and stress in a polluted aquatic environment. The case of the mummichog. BioScience 39:89-95. Whalley, P., and E. A. Jarzembowski. 1981. A new assessment of Rhyniella, the earliest known insect, from the Devonian of Rhynic, Scotland. Nature 291:317. Wilk, M. B., and S. S. Shapiro. 1968. The joint assessment of normality of several independent samples. Technbrnetrics 10:825-839. Willis, J. H., J. A. Coyne, and M. Kirkpatrick. 1991. Can one predict the evolution of quantitative characters without genetics? Evolution 45:441-444. Wilson, J. B. 1988. The cost of heavy-metal tolerance: An example. Evolution 42:408-413. Wu, L. 1990. Colonization and establishment of plants in contaminated sites, pp. 269-284. In A. J. Shaw (ed.), Heavy-Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton, FL USA. Wu, L., and A. D. Bradshaw. 1972. Aerial pollution and the rapid evolution of copper tolerance. Nature 238:167-169.
COPYRIGHT 1993 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Posthuma, Leo; Hogervorst, Rene F.; Joosse, Els N.G.; Van Straalen, Nico M.
Date:Apr 1, 1993
Previous Article:Phenotypic correlation structure among elements of the color pattern in Precis coenia (Lepidoptera: Nymphalidae)
Next Article:Hybrid zone dynamics are influenced by genotype-specific variation in life-history traits: experimental evidence from hybridizing Gambusia species.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |