Soil compression influences the avoidance behavior of Allonychiurus kimi(Collembola)to cadmium and copper

Jino SONYun-Sik LEEYongeun KIM and Kijong CHO∗

1OJeongEco-Resilience Institute,Korea University,Seoul02841(Republic of Korea)

2Department of EnvironmentalScience and EcologicalEngineering,Korea University,Seoul02841(Republic of Korea)

3Institute of Environment and Ecology,Korea University,Seoul02841(Republic of Korea)

ABSTRACT The behavior of soil organisms inhabiting soil pore spaces can be influenced by soil compression,which can affect their avoidance behavior to pollutants.In this study,we aimed to evaluate the effect of soil compression on the avoidance behavior of Allonychiurus kimi(Collembola)to heavy metals cadmium and copper.Initially,to assess the applicability of the avoidance test guideline of the International Organization for Standardization(ISO)developed for Folsomia candida,we investigated the avoidance behavior of A.kimi to cadmium and copper in an artificial soil with a loose structure(bulk density of 0.25 g cm−3),the porous texture of which was sufficiently loose to enable A.kimi to move between pore spaces.The effect of soil compression on the avoidance behavior of A.kimi to both metals was evaluated in compressed soil(bulk density of 0.64 g cm−3)with a uniformly compressed soil surface,and avoidance behavior was investigated at 24-h intervals over a 120-h period.Given that A.kimi is unable to burrow into compressed soil,the compressed soil test can minimize the effects of differences in soil factors,such as soil porosity and bulk density,on the behavior of this collembolan.In the artificial soil,a statistically significant avoidance behavior of A.kimi was observed at cadmium and copper concentrations greater than 50 and 200 mg kg−1,respectively,thereby indicating the applicability of the ISO avoidance test guidelines for A.kimi.When compared at the same exposure time point,the avoidance response to both metals in compressed soil was less sensitive than that in uncompressed soil.In addition,we observed differences in the effects of metals on avoidance response in the compressed soil over time,with the effect of cadmium increasing with time and the effect of copper showing the opposite trend.Overall,we found that soil compression can affect the avoidance behavior of A.kimi to cadmium and copper,and we discussed the advantages and limitations of using compressed soil for assessments of pollutant toxicity.

KeyWords: behavioral endpoint,ecotoxicity test,habitat function,heavy metal,pollutant,springtail

INTRODUCTION

Ecotoxicological tests using soil organisms such as earthworms, springtails, and enchytraeids (ISO, 1999, 2012;OECD,2004,2009)have become a useful tool to complement conventional chemical analysis for assessing the effects of pollutants in contaminated soils (Jensenet al., 2017).However,in cases where large numbers of chemicals or soil samples need to be evaluated in a short period of time,and a trade-offbetween time and cost is desirable,a rapid and cost-effective test with ecological relevance and sensitivity may be necessary.

In this regard,avoidance tests can provide a useful complement to chronic tests,given that an initial evaluation of the toxicity of a pollutant or a soil sample can be made within a few days.Moreover, shifts in the abundance and distribution of soil organisms in a local ecosystem due to avoidance are ecologically relevant, as they can lead to a disruption or loss of community integrity and ecosystem functions associated with soil organisms.A basic assumption with respect to avoidance behavior tests is that the test organisms have chemoreceptors that are highly sensitive to the pollutants,and thus these organisms can avoid contaminated soil during testing.Subsequent to the standardization of avoidance testing forFolsomia candida,a collembolan species commonly used in soil ecotoxicological studies(ISO,2011),several studies have been reported on the avoidance behavior of collembolans to various pollutants (Buchet al., 2016;Niemeyeret al.,2018; Gaineret al.,2019; Serafiniet al.,2019).In this regard,use of avoidance tests in the screening phase of ecological risk assessments of contaminated soils has been gaining in popularity(Niemeyeret al.,2010;Da Silvaet al.,2018).This is mainly because avoidance tests are more cost-effective in terms of experimental duration and workload compared to the collembolan reproduction test(Natal-da-Luzet al.,2004;Loureiroet al.,2005;Jensen and Mesman,2006),and are as sensitive as the reproduction test,whilst also providing information on the habitat suitability of soils.

Several studies have reported that soil properties,such as pH,organic matter,and clay content,alone or in combination,can affect the behavior of soil organisms such as potworms and earthworms(Amorimet al.,2008a;Natal-da-Luzet al.,2008b).In contrast, however, the avoidance behavior of collembolans appears to be less affected by soil properties than that of earthworms(Natal-da-Luzet al.,2008b).Notably in this regard,the combination of organic matter and clay contents, which can affect the texture, bulk density, and porosity of the soil,can influence the behavior of test species,as has been demonstrated by Natal-da-Luzet al.(2008b),who observed thatF.candidashowed a significant avoidance of fine-textured soil with a low organic matter content,even in the absence of pollutants.Furthermore,the findings of a recent study conducted by Bori and Riva(2015)indicated that slightly compressed soil could prevent collembolans from hiding in the soil, thereby facilitating the investigation of avoidance responses over time without destructive sampling.It is accordingly assumed that, to some extent, the use of compressed soil negates the underlying effects associated with soil porosity and bulk density on avoidance behavior.However,soil compression that reduces the effective pore diameter can also affect the behavior of the test species in the exposure matrix.Given that collembolans are incapable of excavating their own burrows and are thus dependent on the pore spaces provided by the soil,when soil is compressed they may be confined to inhabiting the soil surface where they can be continuously exposed to pollutants present in the thin layer of water that covers the compressed soil surface.Therefore,it can be hypothesized that the extent to which organisms are exposed to pollutants in compressed soils will differ from that in loose soils.

In this study, a collembolan species was selected as a test species.It has been maintained in the laboratory for over 20 years and was recently identified asAllonychiurus kimi(formerly known asParonychiurus kimi)(Leeet al.,2021).Given that collembolans are highly abundant and play an ecologically important role in soil ecosystem as consumers of bacteria,yeasts,and fungi(Rusek,1998),the avoidance behavior of collembolans induced by pollutants may cause an imbalance in local populations and,in turn,may result in the impairment of ecosystem functioning associated with this species.In addition,this species is listed as an alternative collembolan species toF.candidafor reproduction toxicity testing(OECD,2009).The objectives of this study were:i)to investigate the applicability of the avoidance test guidelines developed forF.candida(ISO,2011)toA.kimiafter 48-h exposure to heavy metals cadmium and copper in an artificial soil with a loose structure (bulk density of 0.25 g cm−3)(hereafter referred to as uncompressed soil)and ii)to examine the avoidance behavior ofA.kimito cadmium and copper in the artificial soil compressed to 0.64 g cm−3in bulk density(hereafter referred to as compressed soil)at 24-h intervals over a 120-h exposure period.The second objective was designed to minimize the effects of soil factors,such as soil porosity and bulk density,which may affect the behavior ofA.kimiand thereby mask the possible effects of pollutants on the avoidance response ofA.kimi.

MATERIALS AND METHODS

Test organism

Specimens ofA.kimicollected from Korean paddy soil were reared in a laboratory on a moist substrate of plaster of Paris and charcoal in an incubator under continuous dark conditions at 20±1°C and fed once a week with brewer’s yeast suspended in deionized water.Age-synchronized cohorts were obtained by transferring adults to a fresh charcoal substrate and allowing them to lay eggs,which were subsequently removed after 48 h.These eggs hatched after approximately 14 d under the aforementioned temperature and light conditions,and the hatched juveniles were maintained using the same procedures and conditions as described above.Age-synchronized adults between 42 and 44 d old were used in the avoidance test.

Test soiland spikingwith test metals

An artificial soil consisting of quartz sand(70%),kaolin clay(20%),and finely ground Sphagnum peat(10%)(OECD,2009) was used as the test soil.Soil pH was adjusted to 6.0±0.5 using CaCO3.Stock solutions of cadmium and copper were prepared by dissolving cadmium chloride salt(CdCl2·2.5H2O,≥98%purity;Sigma-Aldrich,St.Louis,USA)and copper chloride salt(CuCl2·2H2O,≥99%purity; Sigma-Aldrich, St.Louis, USA) in deionized water,respectively.The artificial soil was spiked with diluted stock solutions in deionized water to obtain a series of cadmium and copper concentrations at 50%soil water-holding capacity(WHC).The test concentrations used in the uncompressed soil were 25,50,100,200,and 400 mg kg−1for cadmium and 50,100,200,400,and 800 mg kg−1for copper,with four replicates assessed for each concentration.As no significant avoidance behavior was observed at the lowest concentrations for the two metals in the uncompressed soil test,these concentrations were not included in the subsequent compressed soil test.The test concentrations used in the compressed soil were 50,100,200,and 400 mg kg−1for cadmium and 100,200,400,and 800 mg kg−1for copper,each with four replicates.We also prepared four replicate untreated control soil samples for each metal by adding the same amount of deionized water corresponding to 50%WHC for dual control tests.

Avoidance behaviors to cadmium and copper in uncompressed soil

Avoidance tests in uncompressed soil usingA.kimiexposed to cadmium and copper were conducted in accordance with the ISO 17512-2(ISO,2011).A round-shaped polystyrene container(100 mm in diameter and 40 mm in height)was divided into two equal compartments by inserting a removable plastic divider vertically.For each treatment(control soil+metal-spiked soil),30 g of control soil(wet weight basis)was added to one of the compartments of each container and the same amount of the metal-treated soil was added to the other compartment.Thereafter,the container filled with soils was gently tapped,such that the soil surface on both sides of the divider was at a comparable level.As a control,a dual control test was also performed in a container containing the same control soil in both compartments to confirm the homogenous distribution of collembolans in the two compartments in the absence of metals.After the soils were introduced, the plastic divider was removed, and 20 adultA.kimiwere carefully placed onto the midline of the soil surface in each test container.The containers were then covered with lids to prevent evaporation and maintained in an incubator under continuous dark conditions at 20±1°C for 48 h.During the test period,the animals remained unfed.After 48 h,the plastic divider was carefully re-inserted into the midline between the two soils,and the number of collembolans present within each soil compartment was counted separately after floating with water.Missing organisms were considered dead and were excluded from further data analysis.

Avoidance behavior to cadmium and copper in compressed soil

With the exception of soil surface compression and exposure times, all experimental procedures used for the compressed soil were the same as those described for the uncompressed soil.For soil compression,the soil surface of each compartment containing 30 g of each of the control and treated soil was flattened to the same height by applying a unidirectional force using a cylindrical rod prior to removing the plastic divider.This procedure prevents collembolans from moving deeper into the soil, thereby confining their distribution to the soil surface during the periods of exposure.The depth and bulk density(calculated from the oven-dried mass per unit volume of soil)of the compressed soil were approximately 0.6 cm and 0.64 g cm−3,respectively,whereas those of the uncompressed soil were approximately 1.5 cm and 0.25 g cm−3, respectively.Twenty containers, each populated with 20A.kimiadults, were prepared for each treatment at the beginning of the experiment.All containers were maintained under the same conditions as described above.At 24-h intervals over the 120-h exposure period,four containers were randomly selected,and the number of collembolans present in each compartment at each exposure time was counted in the same manner as described above.

Data analysis

For each replicate, the avoidance response, expressed as a percentage of avoidance net response (NR, %), was calculated as follows(ISO,2011):

whereCis the number of collembolans in the control soil,Tis the number of collembolans in the metal-spiked soil,andNis the total number of collembolans.A positive NR value indicates an avoidance (a larger number of collembolans are found in the control soil), whereas a zero or negative value indicates no avoidance or attraction toward the test soil.If on average more than 70%of the organisms are found in the control soil (i.e., NR>40%), the habitat function of the metal-treated soil is considered to be limited(ISO,2011).An equal distribution between the two compartments in the dual control tests was evaluated using a two-tailed Fisher’s exact test (Zar, 2010); avoidance was evaluated with a one-tailed Fisher’s exact test,as only the avoidance response was being evaluated(Natal-da-Luzet al.,2008a).Prior to conducting one-and two-way analyses of variance(ANOVA), Levene’s test and the Shapiro-Wilk test were performed to verify equal variance and normality of the residuals using the R packages“car”and“agricolae”(version 1.2.5033,RStudio,Boston,USA),respectively.When the data did not conform to either homoscedasticity or normality,even after data transformation, data were analyzed using Welch’s ANOVA followed by the Games-Howellpost hoctest using the R package“userfriendlyscience”.Data obtained for the uncompressed soils were analyzed using one-way ANOVA followed by Tukey’s honestly significant difference(HSD)post hoctest to compare the mean NR values of each treatment.Data obtained for the compressed soils were analyzed using two-way ANOVA followed by Tukey’s HSDpost hoctest to compare the difference in mean NR values between treatments and exposure times.Values of median effective concentration(EC50),concentration causing a 50%NR, were determined using the NLMIXED procedure in SAS version 9.4(SAS Institute Inc.,Cary,USA).If negative NR values (i.e., attraction to the contaminated soil) were observed,these data were considered to have an NR value of 0 for estimation of the EC50 value, as recommended by the ISO(2011).For each metal,the ratio test proposed by Wheeleret al.(2006)was performed to compare EC50 values between exposure times and between uncompressed and compressed soils using the R package“ecotox”.

RESULTS

Avoidance behavior in uncompressed soil

After 48-h exposure,there was no significant difference in the distribution of collembolans in the dual control tests for either cadmium or copper.On average,the percentage of collembolans in each compartment was within the range of 40%–60%in the dual control tests.This range met the validity criterion of the ISO(2011)guideline,implying a homogenous distribution of collembolans in the dual control tests.A statistically significant avoidance response was observed in the treatments in which soil was spiked with cadmium and copper at concentrations greater than 50 and 200 mg kg−1,respectively,which was confirmed using the one-tailed Fisher’s exact test.On average,>70%of collembolans were only found in the control soil when copper concentration was 400 mg kg−1,indicating limited habitat function at this copper concentration.In general,the NR values increased as the concentrations of both metals increased(Fig.1),ranging from 8.75%to 85%.The exception in this regard was copper at 50 mg kg−1, for which a negative NR value(−8.75%)was recorded.When comparing NR values between the dual control and treatment groups,significantly higher NR values were observed at concentrations greater than 200 mg kg−1for both metals.The EC50 value of cadmium(159 mg kg−1)was significantly lower than that of copper(210 mg kg−1)(Table I),indicating that cadmium has higher toxicity than copper,and accordingly promotes a more pronounced avoidance response.

Fig.1 Avoidance net response(NR)of Allonychiurus kimi after 48-h exposure to an uncompressed artificial soil treated with different concentrations of cadmium and copper.Error bars are standard errors(n=4).*indicates a statistically significant NR according to one-tailed Fisher’s exact test at P <0.05.The dashed line(NR=40%)represents the threshold NR value for limited habitat function(ISO,2011).

Avoidance behavior in compressed soil

In the dual control tests, an equal distribution of the collembolans was observed for both metals at all exposure times (P >0.05).On average, the percentage of collembolans in each compartment was within the range of 40%–60%,which met the validity criterion for the dual control test of the ISO guideline.In each treatment,the number of dead or missing collembolans was less than 20%,which met the validity criterion of the ISO guideline.

The avoidance response ofA.kimito cadmium varied with the treatments and the exposure time(Fig.2).With the exception of the two lowest cadmium concentrations at an exposure time of 96 h,significant avoidance response was observed in all cadmium treatment groups after exposure for 72 h or longer.Two-way ANOVA of the mean NR values showed that there was no significant interaction between concentration and exposure time(F(16,75)=0.822,P=0.657).In contrast,the main effects of metal concentrationand exposure time on avoidance response were found to be significant(F(4,75)=13.660,P <0.0001 andF(4,75)=2.668,P=0.039,respectively).The avoidance response was higher at longer exposure time than at shorter time and was higher at higher metal concentrations than at lower concentrations.

Fig.2 Avoidance net response(NR)of Allonychiurus kimi at different exposure time in a compressed artificial soil treated with different concentrations of cadmium and copper.Error bars are standard errors(n=4).*indicates a statistically significant avoidance response according to one-tailed Fisher’s exact test at P <0.05.The dashed line(NR=40%)represents the threshold NR value for limited habitat function(ISO,2011).

TABLE IMedian effective concentration(EC50)values, with corresponding 95%confidence intervals, of cadmium and copper to Allonychiurus kimi after 48-h exposure to an uncompressed artificial soil(OECD,2009)or after different exposure time to a compressed artificial soil spiked with cadmium or copper

In contrast to cadmium, copper induced an avoidance response over a shorter time frame(Fig.2).Notably,there was a statistically significant avoidance response in all treatment groups at an exposure time of 24 h or 96 h,but in none of the treatments at an exposure time of 120 h, indicating a high variability in avoidance response with exposure time.Comparison of the mean NR values indicated that there was no significant interaction effect (F(16, 75) = 1.441,P=0.146)between metal concentration and exposure time,whereas the main effects of both concentration and exposure time on avoidance response were found to be significant(F(4,75)=17.748,P <0.0001 andF(4,75)=6.537,P <0.001,respectively).In general,the avoidance response was higher at shorter exposure time than at longer exposure time,except on day 4,which had the second-highest mean avoidance response.In contrast,stronger avoidance responses to cadmium were observed with increasing exposure time.This indicates that copper induced a more immediate avoidance response compared with cadmium,and the avoidance response ofA.kimito copper became insensitive over time.In addition, the mean avoidance responses were higher at higher pollutant concentrations.

The estimated EC50 values for both metals at each exposure time are presented in Table I.An accurate estimation of EC50 values for cadmium at an exposure time of 48 or 72 h and copper at an exposure time of 120 h could not be obtained because of a weak concentration-response relationship or failure to produce 50%of the maximal avoidance response,even at the highest test concentration.For cadmium,the EC50 values decreased as exposure time increased,indicating that the sensitivity of avoidance response increased over time.The EC50 value at an exposure time of 120 h was significantly lower than those at the other exposure time and became equal to that obtained in the uncompressed soil.In contrast,A.kimishowed very different avoidance response to copper over time.The EC50 value for copper at an exposure time of 24 h was significantly lower than those at the other exposure times and was comparable but significantly lower than the value obtained in the uncompressed soil (P= 0.0301).These results suggest that the intensity of the avoidance response may vary with metals tested as well as soil compression.

DISCUSSION

In order for the avoidance behavior ofA.kimito be used as a sensitive endpoint for assessing soil contamination,it must at least meet the validity criteria developed forF.candidaand also be as sensitive as other lethal and sublethal endpoints.In the uncompressed soil,a homogenous distribution with 50%±10%collembolans in any compartment of the dual control tests satisfied the guideline validity criterion proposed byISO (2011), indicating that the avoidance test guideline developed forF.candida(ISO,2011)is also applicable toA.kimi.The clear avoidance responses to cadmium and copper at concentrations higher than 50 and 200 mg kg−1,respectively,indicate that the avoidance behavior ofA.kimicould be employed as a rapid method for assessing the sublethal level effect of pollutants in soil.As observed by De Silvaet al.(2009), the sensitivity of a species to pollutants could be a crucial factor in the avoidance test.The avoidance behavior ofA.kimito cadmium showed a similar or more sensitive response in terms of EC50 value compared with those previously reported for other soil organisms, such as isopodPorcellionides pruinosus(523 mg kg−1),enchytraeidEnchytraeus albidus(362 mg kg−1)(Loureiroet al.,2009),and earthwormEisenia fetida(183 mg kg−1) (Hund-Rinkeet al., 2005), and was even more sensitive than that of another collembolan,Sinella communis,which showed no avoidance to cadmium up to a concentration of 1 000 mg kg−1(Greenslade and Vaughan,2003).With respect to copper,the EC50 value forA.kimiwas found to be similar to that reported for the earthwormsEisenia andrei(181 mg kg−1)(Loureiroet al.,2005)andE.albidus(133 mg kg−1)(Amorimet al.,2008b),whereasA.kimiwas far more sensitive to copper than the isopodP.pruinosus(1 060 mg kg−1)(Loureiroet al.,2005)and the mitesOppia nitens(4 265 mg kg−1) (Owojoriet al.,2011) andHypoaspis aculeifer(944 mg kg−1) (Owojoriet al.,2014).In contrast,however,the avoidance behavior ofA.kimito copper was found to be less sensitive than that of another collembolanF.candida(61.2 mg kg−1reported by Greenslade and Vaughan (2003) and 17–18 mg kg−1reported by Boiteauet al.(2011)).Moreover, when the EC50 values of both metals for avoidance behavior in the uncompressed soil were compared with those for survival and reproduction ofA.kimireported in the literature,A.kimidisplayed avoidance response to both metals at similar or lower concentrations than those causing 50%mortality(median lethal concentration (LC50)) and 50% reduction in reproduction(EC50 for reproduction).For cadmium,the avoidance behavior ofA.kimiwas as sensitive as mortality(7-d LC50 of 532 mg kg−1(Sonet al., 2007) and 28-d LC50 of 129 mg kg−1(Sonet al.,2011)),but less sensitive than reproduction (28-d EC50 values for reproduction of 60.0 mg kg−1(Sonet al., 2007) and 28.3 mg kg−1(Sonet al.,2011)).For copper,A.kimishowed a strong avoidance response at lower or similar concentrations than those causing 50%mortality and reproduction(28-d LC50 of 1 567 mg kg−1(unpublished data)and 28-d EC50 for reproduction of 277 mg kg−1(Sonet al.,2017)).

As in the uncompressed soil,a homogenous distribution ofA.kimiwas observed in the dual control tests of the compressed soil,indicating that the compressed soil also met the validity criteria proposed by the ISO(2011).However,the avoidance responses to both metals in the compressed soil were less sensitive than those in the uncompressed soil when evaluated at the same exposure time point(Table I).The degree of exposure to pollutants in soil organisms can be determined by the extent to which they are in contact with the local environment.In this regard,it is known that pore water-mediated uptake is, in general, the dominant route of exposure to metals in soil-dwelling organisms,such as collembolans, which are in close contact with the soil solution.Therefore,if this holds true forA.kimi,it would be expected that the avoidance response in compressed soil would be more sensitive than that in uncompressed soil.This is becauseA.kimiwould be continuously exposed to the cadmium or copper present in the thin layer of water on the surface of compressed soil.However,our observations for both cadmium and copper did not support this assumption.We suspect that the higher toxicity in uncompressed soil might be attributable to the higher likelihood of exposure to metals through dermal contact whilst collembolans move through pores with dimensions close to their body size.

Moreover,a comparison of the changes in toxicity over time between metals revealed that the EC50 value of cadmium in the compressed soil decreased with increasing exposure time and was comparable to that obtained in the uncompressed soil at an exposure time of 120 h,whereas the opposite trend was true for copper.The longer the collembolans exposed to copper,the greater the variability in the avoidance response that occurred over time;an exposure time of 24 h is sufficient to obtain a significant low-variability avoidance response inA.kimi.Given that behavior is the outcome of multiple complex developmental and physiological processes,evaluations of avoidance behavior at different exposure time can provide important information with regards to the irreversibility of toxic effects(Oliveiraet al.,2015).Natal-da-Luzet al.(2008a)reported consistent avoidance responses inF.candidaexposed to fungicides(benomyl and carbendazim)over time,whereas Kobetičováet al.(2009)reported an enhanced avoidance response in enchytraeidE.albidusexposed to the fungicide carbendazim at EC50.Similarly, Bori and Riva(2015)have also reported an increasing trend in EC50 values with increasing exposure time for the avoidance behavior ofF.candidato boric acid,phenmedipham, and petroleum hydrocarbons in a compressed soil.However,there is still limited knowledge with respect to the mechanisms involved in changes in the avoidance responses of soil organisms over time,although these may be partly related to the inherent high variability of avoidance test results(Bori and Riva,2015),the highly species-specific nature of avoidance responses to different pollutants(Lukkari and Haimi,2005;Loureiroet al.,2009),and the inability to escape from contaminated soils(Oliveiraet al.,2015).Wesuspect that in the case ofA.kimi,the different patterns of change in cadmium and copper toxicity over time might be also related to the homeostatic mechanisms associated with the regulation of the internal concentrations of these metals.Essential metals,such as copper,can be actively taken up and regulated inA.kimi,whereas non-essential metals,such as cadmium,are taken up indirectly or by chance.Given that cadmium is not essential for metabolic activity inA.kimi,these collembolans have not evolved mechanisms for active regulation of the internal concentrations of this metal,and consequently,it is conceivable that an increase in avoidance response would occur with increasing exposure time.In contrast, the apparent insensitivity of avoidance response to copper over time might reflect the fact thatA.kimican actively regulate internal concentrations of copper to some extent,and thus they are able to freely move into or out of the soil when the internal copper concentration is below or above the physiological requirement.A further possible explanation for the insensitivity to copper over time might be the neurotoxic effect of this metal,which has previously been observed in the nematodeCaenorhabditis elegans(Du and Wang,2009).If a similar neurotoxic effect occurs inA.kimiwhen inhabiting a contaminated compartment,this may impede the locomotion,thereby making it difficult to escape from contaminated soils.

In the present study,we used compressed soil to minimize the potential effects of soil factors,such as soil porosity and bulk density,which may affect the behavior ofA.kimi.In this regard, the findings of several previous studies have indicated that changes in soil porosity and bulk density due to soil compression can negatively affect the abundance and activity of collembolans(Heisler and Kaiser,1995;Dittmer and Schrader,2000;Larsenet al.,2004).Thus,changes in the pore spaces between soil particles may play an important role in determining the distribution and abundance of soil organisms.It is,of course,true that the exposure of organisms to pollutants in compressed soil contrasts markedly with that in uncompressed soil;however,the use of compressed soils has the advantage of not only minimizing the potential effects of soil porosity and bulk density on the behavior ofA.kimibut also minimizing disturbance of the test species during repeated investigations over time.Moreover,it can provide information about the actual effects of pollutants on the behavior of test species by continuously exposing the organisms to pollutants present in the thin water layer on the soil surface.

CONCLUSIONS

In this study, the applicability of avoidance test ISO guideline developed forF.candidatoA.kimiwas investigated in uncompressed soil.When the toxicity values of cadmium and copper for the avoidance behavior in the uncompressed soil were compared with those for survival and reproduction ofA.kimireported in the literature,A.kimidisplayed avoidance responses to both metals at similar or lower concentrations than those causing 50%lethal or sublethal effects on survival or reproduction.These results accordingly indicate that the avoidance behavior ofA.kimican be used as a rapid and sensitive endpoint for assessing the sublethal level effects of pollutants in soil.For both metals,less sensitive avoidance responses were observed in compressed soil than in uncompressed soil when compared at the same exposure time,indicating that soil compression can affect the effect of pollutants on avoidance behavior.Moreover,the avoidance response over time in the compressed soil differed between the two metals,presumably due to differences in the homeostatic mechanisms for regulating the concentrations of essential and non-essential metals.Overall,the findings of this study indicate that soil compression can affect the avoidance behavior ofA.kimito both cadmium and copper.Future studies should focus on the use of avoidance tests as an early screening tool in retrospective risk assessments for assessing soil contamination at lower tiers of risk assessment.

ACKNOWLEDGEMENT

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(No.NRF-2017R1D1 A1B03036474)and supported by a Korea University Grant.