Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1598-5504(Print)
ISSN : 2383-8272(Online)
Journal of Agriculture & Life Science Vol.53 No.1 pp.93-104

Optimization of Fecal Calprotectin Assay for Pig Samples

Paul Bogere1, Yeon Jae Choi2, Jaeyoung Heo2*
1Department of Agricultural Convergence Technology, Chonbuk National University, Jeonju, 54896, Korea
2International Agricultural Development and Cooperation Center, Chonbuk National University, Jeonju, 54896, Korea
Corresponding author: Jaeyoung Heo Tel: +82-63-270-5925 Fax: +-63-270-5927 E-mail:
September 7, 2018 November 16, 2018 January 3, 2019


Fecal calprotectin is a noninvasive marker of gut inflammation and has been widely utilized in human gastrointestinal diagnostics. This marker, however, has not been extensively utilized in porcine samples. The aim of this study was to optimize a protocol for the extraction of porcine fecal calprotectin, and to the best of our knowledge this is the first study to be conducted in this regard. Freshly collected swine fecal samples were used in this study. We determined the variability of three commercial ELISA assays in the recovery of porcine fecal calprotectin. We further studied the effect of dilution factor and roller shaker homogenization on the yield of calprotectin from swine fecal samples. Calprotectin recovery was significantly different(p<0.05) across the three commercial assays with MBS033848 having a greater recovery compared to DAEF-012 and calprest. Fecal calprotectin yield increased with an increase in dilution factor with maximum recovery at 1:250. Furthermore, homogenization of fecal sample extracts using a roller shaker for tubes for 30 min led to a 30.75% relative increase in calprotectin yield. Further increase in shaking time(at 60 min) led to a reduced calprotectin recovery. Calprotectin recovery ratio was 130.8% and 101.4% at 30 min and 60 min homogenization respectively. In our conclusion, we observed that various factors affect the recovery of porcine fecal calprotectin, and therefore the researcher should double check certain parameters in regard to the type of kit, the dilution factor and homogenization time if reliable and reproducible results are to be obtained. Results of the present study provide useful information on a non-invasive protocol to veterinarians and researchers in examining and monitoring swine gut healthusing the fecal calprotectin.


    Rural Development Administration


    Calprotectin is an inflammatory protein released from neutrophils and activated macrophages(Lallès & Fagerhol, 2005;Lallès et al., 2007). Originally, calprotectin was discovered in the cytoplasm of neutrophil granulocytes as a protein with antimicrobial properties(Stříž & Trebichavský, 2004). It is a calcium and zinc binding protein belonging to the S100 family(S100 A8/A9) with a molecular weight of 36.5k Da(Fagerberg et al., 2003;Li et al., 2015). This protein is an inflammatory marker found in various systems and organs of the body including; respiratory system(lungs), gastrointestinal system, the skin, kidneys, oral cavity, udder, and joints(Stříž & Trebichavský, 2004;Lallès & Fagerhol, 2005). Increased concentrations of calprotectin can be measured in plasma, urine, feces and synovial fluid during active inflammation accompanied by neutrophil recruitment(Fagerberg et al., 2003). In human diagnostics, fecal calprotectin has been extensively used as a non-invasive marker of gut inflammation(Savino et al., 2010;Xiao et al., 2014). Fecal calprotectin detection has been used as a valuable screening tool for the identification of patients with inflammatory bowel disease, and has been shown to dependably differentiate inflammatory bowel disease(IBD) from irritable bowel syndrome due to its outstanding negative predictive value in ruling out IBD in undiagnosed, symptomatic patients (Konikoff & Denson, 2006;Van Rheenen et al., 2010). Among other reasons for the wide use calprotectin detection in stool as an inflammatory marker is the fact that the protocol is simple, non-invasive and relatively cheap compared to other diagnostic methods (Van Rheenen et al., 2010). Due to the relevance of this marker in diagnostics, different companies and hospitals have developed various Enzyme-linked immunosorbent assay-ELISA protocols to aid in the detection of fecal calprotectin levels in the various body tissues and organs. However, researchers have studied the different protocols and suggested improvements that lead to a maximum yield of the protein. Tøn et al.(2000) developed an improved assay procedure for human fecal sample preparation that results in higher calprotectin yield and reduced contamination. In their paper, they observed that increasing dilution factor between feces and extraction buffer increased the calprotectin yield. Several other patents also exist outlining the preferable quantities of fecal sample to be used in human fecal calprotectin analysis(Weber et al., 2017).

    Animals are able to survive and stay health in an environment full of pathogenic microbes due to their immune system that is endowed with strong defense mechanisms particularly the innate and adaptive immunity(Yu et al., 2013). In the innate immunity pathway, higher invertebrates including mammals secrete several antimicrobial peptides including calprotectin which has ability to bind zinc, an essential element in microbial growth(Stříž & Trebichavský, 2004;Linde et al., 2008). In the canines, calprotectin complex S100A8/9 and the S100A12 have been demonstrated to be closely linked with acute and chronic inflammation and malignant transformations. They have been further shown to have regulatory functions in cell proliferation and metastasis, and their release into the extracellular space is to function as alarmins(Heilmann et al., 2018). Calprotectin has been suggested to have potential as a marker of inflammation in dogs and has been demonstrated to be a useful surrogate marker of disease severity in Chronic inflammatory enteropathies(Heilmann et al., 2018). Furthermore, calprotectin expression in mucosal cells reduces invasion by pathogens including Listeria monocytogenes, Salmonella Typhimurium, and also inhibit growth of Escherichia coli(Nisapakultorn et al., 2001;Stříž & Trebichavský, 2004). In a study by Šplíchal et al.(2005), germ-free piglets infected with enteropathogenic E. coli O55 registered increased levels of calprotectin in the bronchoalveolar lavage coupled with a septicemia while E. coli Nissle strain elevated calprotectin concentration in the small intestinal lumen. However, the use of porcine fecal calprotectin as a gut inflammatory marker in swine remains minimal with only a few studies done. In 2005, Lallès and Fagerhol modified a human calprotectin assay to suit conditions for extraction of porcine fecal calprotectin, and thus published the very fast data on pig fecal calprotectin at various ages. They however, revealed that the conditions for pig fecal calprotectin extraction are as critical as is for humans. The purpose of this study was to establish the porcine fecal sample and assay protocols. We also compared the use of three commercially available kits: for human fecal calprotectin, pig calprotectin and porcine fecal calprotectin.

    Materials and Methods

    1 Fecal sample collection, handling and storage

    Healthy lively three-way crossbred piglets(Landrace ×Yorkshire×Duroc), 28 days of age and weighing approximately 8 kg body weight were used in this study. The piglets were selected randomly irrespective of their sex and fresh fecal samples were collected from them. The method of collection was by rectal palpation and evacuation using a gloved finger. In piglets where rectal evacuation was impossible, pigs were observed and a freshly dropped fecal sample was collected immediately and placed into clean sample tubes. All sample tubes were then placed on ice and transported to the laboratory for analysis. All samples were then stored at either 4℃ or -80℃ and analyzed within five days or one month respectively.

    2 Assay variability of commercial calprotectin ELISA kits in the recovery of porcine fecal calprotectin

    Three commercially available calprotectin assay kits were selected and their performances in the recovery of porcine fecal calprotectin were assessed and compared. Two fecal samples were assayed using MBS033848 assay(Mybiosource, San Diego, CA, USA), DAEF-012(Creative Diagnostics, Shirley, NY, USA) and Calprest(Eurospital, Trieste, Italy). Table 1 summarizes the key properties of these three ELISA assay protocols. Extraction of the analyte using Calprest was performed using the EasyCal devise for stool collection, ETS9062 performed according to manufacturer’s instruction (Eurospital, Trieste, Italy). The procedure involved dipping the devise grooved stick into previously collected fecal sample, repeatedly rotating the stick to homogenize the sample and also fill the grooves with fecal material(about 56 mg). Then inserted the stick with fecal material into the tube containing extraction solution(2.8 ml) and sealed completely. Each sample device was then vortexed for 60 seconds, roller shaken for 20 minutes(min) and later centrifuged for 10 min at 5000 rpm to obtain the supernatant. Accordingly, Phosphate buffered saline PH 7.4 and concentration 1% formulated without calcium and magnesium chloride was used as extraction buffer for MBS033848 and DAEF-012 kits. For all assay protocols, the dilution factor was set at 250. For each Kit, the standard assay procedures were carried out on the extracted samples according to manufacturer’s instruction. All incubation steps and Photometry was done using Multiskan GO without cuvette 100 to 240 V(Thermo Fisher Scientific Oy, Finland) spectrophotometer.

    3 Effect of dilution factor on the recovery of porcine fecal calprotectin

    To determine the relatively appropriate dilution factor for maximum recovery of porcine fecal calprotectin, two fecal samples(technical replicates) were each made into three biological replicates. About 10mg(8 to 12 mg) of each biological replicate were mixed with PBS solution, vortexed, centrifuged and subjected to four different serial dilutions in the range from 1:100 to 1:1000. The diluted sample extracts were then assayed using Mybiosource ELISA Kit Cat. No MBS033848 and the assay procedure was followed as per manufacturer's instructions. Using the standard curve generated, the levels of fecal calprotectin(ng/ml) at different dilution factors were determined using the standard equation.

    4 Effect of roller shaking time on porcine fecal calprotectin yield

    Three fecal samples were each divided into three biological replicates, mixed with PBS solution to a dilution of 100, vortexed for about 40 to 60 seconds and later subjected to three levels of roller shaker timing before assay i.e. No roller shaker(0 min), roller shaker for 30 min and roller shaker for 60 min. The roller shaker used operates in a seesaw motion and was set to a speed of 60 rpm. Consequently, all samples were centrifuged at 5000 rpm for 20 min at 4℃. The supernatants were collected into clean Eppendorf tubes and further diluted to 250 before assay. Samples were assayed using Mybiosource ELISA Kit. Using the standard equation, the concentration of fecal calprotectin(ng/ml) was determined. Recovery ratio(R) and relative increase/decrease in calprotectin yield at the different time intervals of roller shaking were thus calculated according to the equations below.

    R= (Calprotectin yield at time T2 ÷ Calprotectin yield at time T1) × 100%; where T2= 30 min or 60 min and T1= 0 min(no roller shaker).

    Relative Change= [(Calprotectin yield after Tx1- Calprotectin yield after Ty)/calprotectin yield after Ty] × 100%; Where Tx1 is either 30 min or 60 min, and Ty is 0 min(No roller shaker)

    5 Statistical analysis

    Data were evaluated by analysis of variance to compare mean differences between all treatments. Duncan’s multiple range test was used to test for differences between the specific treatments and p<0.05 was considered to indicate statistical significance. Analyses were performed using SAS software(version 9.1, 2004; SAS Institute, USA).

    Results and Discussion

    1 Assay variability of the selected commercial ELISA Kits

    Two fecal samples were each divided into three biological replicates and run in the three selected commercial kits with assay procedures according to the manufacturer's instructions. Using the standards provided along with the kits and their corresponding optical densities, standard curves, equations, standard error(s) and correlation coefficient(r) were obtained using Curve Expert 1.3 represented in(Fig. 1). Correlation coefficients were 0.9979, 0.9982 and 0.9967 for DAEF- 012, Calprest and MBS033848 assays respectively. Calprotectin recovery was determined using the standard equations and compared across the three kits. Table 2 shows calprotectin recovery(μg/g) of the samples across the three assays. There was statistical significance(p<0.05) in calprotectin recovery across the three commercial kits with assay MBS033848 having a greater recovery(14.95 and 19.31 μg/g) compared to (0.13/0.12 μg/g and 2.09/2.99 μg/g)from DAEF-012 and Calprest for sample one and sample two respectively. The variability in calprotectin recovery from the same samples using the three commercial kits in this study is in agreement with results of a similar study by Whitehead et al.(2013) who also demonstrated intra and inter-batch imprecision, linearity and recovery. Although there were differences in recovery of fecal calprotectin between the assays, recovery values(ng/ml) of MBSO33848 and DAEF-012 were in the range as outlined by the manufactures.

    Recovery values for Calprest assay were too low compared to the detection range stated by the manufacturer(Table 3). Since there exists no standard reference material and method for porcine fecal calprotectin recovery, it was hard for the research to determine which assay produced accurate results. However, since MBS 033848 assay produced more recovery than the other two from the same samples(Fig. 2), it was convincing to conclude that this assay would be suitable for porcine fecal calprotectin determination. Lallès & Fagerhol(2005) published swine fecal calprotectin concentrations range in various age groups, and calprotectin results obtained using MBS033848 assay are consistent with this range. The low recovery of calprotectin using Calprest and DAEF-012 assays could be because of the specificity and lack of cross reactivity of the antibodies coated onto the plates to the calprotectin antigen in the pig fecal samples(TIP, 2010;Wilson, 2013). Owing to the varied results in the three assays, the subsequent calprotectin tests were performed using Mybiosource Porcine Calprotectin(CP) Elisa Kit.

    2 Dilution factor

    The fecal preparation and extraction procedure performed before the actual analyte quantification using ELISA has a huge influence on the overall imprecision of the calprotectin assay(Whitehead et al., 2013). Various factors including; fecal sample amount, type of extraction buffer and the ratio between the stool sample and extraction buffer affect calprotectin recovery(Tøn et al., 2000). In this study, we determined the dilution ratio corresponding to the maximum yield of fecal calprotectin using two fecal samples with 3 biological replicates each. Using 10 mg sample amounts, all biological replicates were diluted with extraction buffer to dilutions of 100, 250, 500, and 1000. The sample extracts were consequently assayed for calprotectin quantification using MBS033848 kit, and results were compared across the dilution levels. There was no statistically significant difference(p>0.05) in recovery across the dilution factors in sample one but the difference (p<0.05) was observed in sample two(Table 4). However, general recovery of calprotectin tended to increase with increasing dilution factor with maximum yield peaking around 250 to 500 dilutions, while further increase in dilution reduced the calprotectin yield(Fig. 3). The maximum calprotectin yield was taken to be the concentration of calpro tectin(ng/ml) directly read from the standard curve. The increased fecal calprotectin recovery at dilution 250 could be due to the increased ratio between the extraction buffer and the fecal sample(Tøn et al., 2000). However, the reasons for the observed reduction in calprotectin yield with a further increase in dilution remain obscure and further research is necessary to establish this. Increasing the ratio between extraction buffer and samples upsurges the accessibility and solubilization of the sample’s molecular components including proteins like calprotectin(Fagerhol, 2001). Less concentrated homogenates have greater uniformity in the liquid phase and the final composition will vary less in parameters like acidity/alkalinity, metal iron concentration which could be influenced by the feces (Fagerhol, 2001). Lallès & Fagerhol(2005) in their study revealed that porcine fecal calprotectin levels decreased with increasing concentration of feces in the extraction buffer. This could explain the observed lower calprotectin recovery at lower dilution factor (1:100) in the present study. Extractive dilution ratios in the range of 1:250-1:1000 have been cited to increase extraction recovery of proteins from gastrointestinal samples especially those containing higher protein levels(Heilmann et al., 2011). More still, this dilution range is believed to produce extraction results that are more accurate, reliable and reproducible. In contrast to the study by Lalles and Fagerhol who used 100 mg samples and 5 ml extraction buffer as recommended for the human protocol, we used 10 mg(8 to 12 mg) fecal samples for all our dilution tests. The advantages for using small sample size is that it allows for a much greater dilution of the sample, and it also makes handling of volumes in the laboratory simple since you do not need to use large quantities of the extraction buffer to dilute the fecal sample.

    3 Roller shaker homogenization time and calprotectin yield

    The effect of homogenization time using a roller shaker revealed that continuous uniform shaking of the samples for 30 min on a tube shaker increased the extraction yield after assaying. Total calprotectin extraction at the three levels of homogenization was statistically different(p<0.05) with recovery in calprotectin at 30 min higher compared to no roller shaking (Table 5). The relative increase in calprotectin yield after 30 min of homogenization was 30.75% higher than recovery obtained when samples were not subjected to roller shaking(Table 6, Fig. 4). In developing an improved assay for human calprotectin recovery by Tøn et al.(2000), it was also revealed that the yield of the protein depended on the duration of homogenization of the fecal extracts. They observed an increased recovery after 20 min of homogenization by shaking. Their findings thus agree with our study which revealed increased calprotectin after 30 min of roller shaking. This notwithstanding, our results also revealed that further increase in homogenization time beyond 30 min to about 60 min led to a reduced recovery of calprotectin from the fecal extracts. In proteomic analysis, making the whole protein compliment available is important in attaining maximum extraction and solubilization of the protein from the tissue. This further aids in minimizing protein losses and to increase reproducibility (Ericsson & Nister, 2011). Moreover, for efficient release of the desired protein from any tissue, it is required to thoroughly mince and or homogenize the tissue to enlarge the surface exposed to the solvent. Homogenization and cell lysis serve to disrupt tissues and cells to release the proteins and other biomolecules to be collected in soluble form for further processing and analysis(Shao et al., 2016). This could, therefore, explain the variation in calprotectin recovery observed at zero roller shaking and 30 min shaking. It follows that shaking the fecal extracts for 30 min led to enough release, homogenization and solubilization of the calprotectin into the extraction buffer hence the increased recovery. However, in as much as homogenization happens to increase recovery of the target analyte, the technique used should just provide enough force to disrupt the tissues to release it but should not cause extensive damage to the tissues and analytes(Skehel, 2004). Our study showed that extending homogenization time beyond 30min reduced the concentration of the recovered calprotectin after analysis though we cannot conclusively state that the longer shaking period caused calprotectin degradation. Fecal calprotectin has been reported to be stable for 7 days at room temperature(Whitehead et al., 2015), and still viable for over 3 months at -80℃, however, no study has yet been done on the stability of porcine fecal calprotectin to shear force and homogenization time. Therefore, since currently there isn’t any method to help quantify the initial amount of calprotectin in porcine feces, we suggest the use of a 30 min shaking of the fecal extract on a roller shaker for tubes as a way of achieving complete extraction, solubilization and homogenization of pig fecal calprotectin.

    Owing to the results of our study, we draw the following conclusions. First, the assay variability observed in this study points to the need for assay standardization. Secondary, the researchers need to establish the suitable dilution factors and homogenization time in a way of investigating how transferable the manufacturer’s recommendation apply in obtaining maximum yield of swine fecal calprotectin. Thirdly, since MBS033848 kit produced the highest recovery of calprotectin with slight modification of manufacturer’s extraction protocol, we conclude that it is a suitable kit for porcine fecal calprotectin determination.


    This research was supported by a grant from the Next-Generation Bio Green 21 Program(PJ01322302), Rural Development Administration, Republic of Korea.



    Standard curves showing the correlation coefficients(r2) for the three assay kits.


    Variability in fecal calprotectin recovery using three different assay protocols. Each graph represents the mean±standard error of three replicates per assay protocol. Different letters above each graph indicate significant difference among assay protocols by Duncan’s multiple range test at p≤0.05.


    Effect of dilution factor on calprotectin yield. Each graph represents the mean±standard error of three replicates per dilution factor. Asterisks(*) represent highest recovery with corresponding dilution factor.


    Effect of homogenization time on fecal calprotectin recovery. *Higher relative increase in yield(%).


    Key properties of the three ELISA assay protocols

    Calprotectin recovery(μg/g fecal matter) across the three assay protocols

    Recovery of calprotectin(ng/ml) compared with manufacturer stated ranges

    Mean fecal calprotectin yield at 4 levels of sample dilution

    Effect of roller shaker duration on fecal calprotectin recovery

    Relative increase in fecal calprotectin yield at 30 min and 60min roller shaker intervals


    1. EricssonC and NisterM . 2011. Protein extraction from solid tissue . Methods Mol Biol.675: 307-312.
    2. FagerbergUL , LoofL , MerzougRD , HanssonLO and FinkelY . 2003. Fecal calprotectin levels in healthy children studied with an improved assay . J. Pediatr Gastroenterol Nutr.37: 468-472.
    3. FagerholMK . 2001. Process for the extraction of proteins, U.S Patent 6, 225, 072 B1.
    4. HeilmannRM , BerghoffN , MansellJ , GrutznerN , ParnellNK , GurtnerC , SuchodolskiJS and SteinerJM . 2018. Association of fecal calprotectin concentrations with disease severity, response to treatment, and other biomarkers in dogs with chronic inflammatory enteropathies . J. Vet. Intern. Med.32: 679-692.
    5. HeilmannRM , PaddockCG , RuhnkeI , BerghoffN , SuchodolskiJS and SteinerJM . 2011. Development and analytical validation of a radioimmunoassay for the measurement of alpha1-proteinase inhibitor concentrations in feces from healthy puppies and adult dogs . J. Vet. Diagn Invest.23: 476-485.
    6. KonikoffMR and DensonLA . 2006. Role of fecal calprotectin as a biomarker of intestinal inflammation in inflammatory bowel disease . Inflammatory bowel diseases.12: 524-534.
    7. LallèsJP and FagerholMK . 2005. Faecal Calprotectin: A non invasive marker of inflammation in pigs? Paper presented at the Proceedings of the XIIth International Congress of the International Society for Animal Hygiene, Warsaw, Poland.
    8. LallèsJ , FormalP and FagerholM . 2007. Changes in calprotectin concentration in sow’s milk throughout lactation. >Paper presented at the Animal health, animal welfare and biosecurity. Proceedings of 13th International Congress in Animal Hygiene, Tartu, Estonia. 1: 17-21.
    9. LiF , MaJ , GengS , WangJ , LiuJ , ZhangJ and ShengX . 2015. Fecal calprotectin concentrations in healthy children aged 1-18 months . PLoS One.10: e0119574.
    10. LindeA , RossC , DavisE , DibL , BlechaF and MelgarejoT . 2008. Innate immunity and host defense peptides in veterinary medicine . Journal of Veterinary Internal Medicine.22: 247-265.
    11. NisapakultornK , RossKF and HerzbergMC . 2001.  Calprotectin expression inhibits bacterial binding tomucosal epithelial cells . Infection and Immunity.69: 3692-3696.
    12. SavinoF , CastagnoE , CalabreseR , ViolaS , OggeroR and MinieroR . 2010. High faecal calprotectin levels in healthy, exclusively breast-fed infants . Neonatology.97: 299-304.
    13. ShaoS , GuoT , GrossV , LazarevA , KohCC , GillessenS , JoergerM , JochumW and AebersoldR . 2016. Reproducible tissue homogenization andprotein extraction for quantitative proteomics using micropestle-assisted pressure-cycling technology . Journal of proteome research.15: 1821-1829.
    14. SkehelJM . 2004. Preparation of extracts from animal tissues . Methods Mol. Biol.244: 15-20.
    15. ŠplíchalI , FagerholMK , TrebichavskýI , ŠplíchalováA and SchulzeJ . 2005. The effect of intestinal colonization of germ-free pigs with Escherichia coli on calprotectin levels in plasma, intestinal and bronchoalveolar lavages . Immunobiology.209:681-687.
    16. StřížI and TrebichavskýI . 2004. Calprotectin-a pleiotropic molecule in acute and chronic inflammation .Physiol Res.53: 245-253.
    17. TIP T.2010. ELISA technical guide and protocols. Thermo Fisher Scientific Inc.USA.
    18. TønH , BrandsnesØ , DaleS , HoltlundJ , SkuibinaE , SchjønsbyH and JohneB. 2000. Improved assay for fecal calprotectin . Clinica Chimica Acta.292: 41-54.
    19. Van RheenenPF , Van de VijverE and FidlerV . 2010. Faecal calprotectin for screening of patientswith suspected inflammatory bowel disease: diagnostic meta-analysis . BMJ. 341: c3369.
    20. WeberJ , UberschlagME , PricaM and JermannT . 2017. Method for determination of a protein, European Patent Specification, EP 2 947 459 B1.
    21. WhiteheadS , BrookesM , FrenchJ , FordC and GamaR . 2015. PTH-075 The effect of storage conditions on the stability of faecal calprotectin. In: BMJ Publishing Group.
    22. WhiteheadS , FrenchJ , BrookesM , FordC and GamaR . 2013. Between-assay variability of faecal calprotectin enzyme-linked immunosorbent assay kits . Annals of clinical biochemistry.50: 53-61.
    23. WilsonR. 2013. Sensitivity and specificity: twin goals of proteomics assays. Can they be combined? Expert RevProteomics.10: 135-149.
    24. XiaoD , WangY , LiuG , HeJ , QiuW , HuX , FengZ , RanM , NyachotiCM and KimSW . 2014. Effects of chitosan on intestinal inflammation in weaned pigs challenged by enterotoxigenic Escherichia coli .PLoS One.9: e104192.
    25. YuJZ , WenCM , YangCMJ and TsaoPH . 2013.  Antimicrobial peptides characteristics and theirapplications as feed additives in domestic animals . J. Chin. Soc. Anim. Sic.42: 1-13.
    오늘하루 팝업창 안보기 닫기