parasitic nematodes in domestic ruminants

Functional Ecology 2001 Massive use of chemotherapy influences life traits of Blackwell Science, Ltd parasitic nematodes in domestic ruminants V. LEIGNEL and J. CABARET INRA, PAP, 37380 Nouzilly, France Summary 1. The size of nematodes is an operational indicator of fecundity. The variation in size due to chemotherapy was studied with a benzimidazole anthelmintic in susceptible and resistant nematodes. Teladorsagia circumcincta circumcincta (Stadelmann 1894), a nematode endoparasite of sheep and goats, was investigated as many components of its fitness are already known. 2. Susceptible worms submitted to increasing selective pressure by anthelmintics increased in size (by 6 10%); it was hypothesized that this was partly under the control of sheep, as treated lambs may mount and maintain a better response when infected (premunition). 3. The resistant worms, whatever the anthelmintic pressure, were always 3% larger than susceptible ones. 4. Thus, size may increase in susceptible worms and resistant worms for very different reasons. If resistance does not emerge, massive chemotherapy should lead to larger nematodes, and hence to more fertile worms. Key-words: Anthelmintic, resistance, size, teladorsagia Functional Ecology (2001) Ecological Society Introduction Chemotherapy has been used worldwide to control the parasitic nematodes of domestic ruminants. Such a massive use of anthelmintics is likely to select for changes in life histories. Traits other than those associated with anthelmintic resistance (Conder & Campbell 1995) are also likely to be under strong selection. In recent papers, it has been suggested that when anthelmintics reduce adult life expectancy, natural selection will favour parasites that mature earlier (Medley 1994; Poulin 1998): selection would then be age-specific. Conversely, Skorping & Read (1998) argued that parasitic nematodes with larvae in host tissues ( poorly susceptible to treatments) and adults in the gastrointestinal tract (highly susceptible to treatments), should postpone maturity, become larger and more pathogenic: selection by drugs would be stagespecific. Anthelmintic selection should then favour either smaller (earlier maturity) or larger ( postponed maturity) parasites, depending on authors. In mammalian nematodes, developmental period, body size and fecundity are positively correlated (Skorping et al. 1991; Morand 1996), so that body size is an operational indicator of fecundity. Size in the nematode Teladorsagia circumcincta is strongly Author to whom correspondence should be addressed. E-mail: leignel@univ-lemans.fr correlated to egg numbers in utero (L. Gruner, unpublished data; Spearman correlation = 0 77 as established on 680 worms harboured by 17 lambs). Size among trichostrongyle nematodes of ruminants also varies according to several environmental factors such as season, species and morph (Teladorsagia circumcincta morph trifurcata is larger than the major morph T. circumcincta circumcincta) (Cabaret 1984). It depends also on density (females in large infections lay fewer eggs per capita). Size cannot be interpreted in terms of fecundity without control of the above-mentioned factors (season and morph). Selective pressure on nematodes of the trichostrongylid family submitted to anthelmintic treatment differs strongly according to their resistance status. The parasitic stages (intramucosal stages; third-stage infective larvae, fourth-stage infective larvae, juveniles and adults in the lumen of organs) are equally susceptible to benzimidazole anthelmintics at normal dosage of the drug (Cabaret et al. 1979). We may expect that resistant worms will not be under selective pressure and that their size will be stable, whereas susceptible worms should respond to selection by modifying their size, whatever their parasitic stage or age. Usually under temperate climate, three overlapping generations occur during the grazing season, and the number of generations may increase owing to anthelmintic treatments that remove the existing susceptible worms. Resistance to one group of anthelmintics, the benzimidazoles, is 569

570 V. Leignel & J. Cabaret controlled by a point mutation on the beta-tubulin gene (Elard et al. 1996), and Teladorsagia circumcincta worms are easily genotyped into resistant homozygotes rr, and susceptible heterozygotes rs or homozygotes SS (Humbert & Elard 1997). We therefore have a tool to evaluate the susceptibility status to benzimidazoles for each worm, and thus establish a correlation between susceptibility (worms prone to anthelmintic selection)/ resistance (worms not susceptible to anthelmintic selective pressure) and size of worms with the same genetic background within one isolate. We investigate the following questions: (i) are susceptible worms submitted to increasing selective pressure likely to evolve with regard to size and (ii) are susceptible worms (subjected to anthelmintic selective pressure) larger than resistant ones (not subjected to anthelmintic selective pressure)? Materials and methods WORM ORIGIN AND CHARACTERIZATION Teladorsagia circumcincta circumcincta (Stadelmann 1894) males were recovered from one isolate (Touraine, France) known to be resistant (Elard et al. 1998) from 1991. This isolate was previously characterized as susceptible in 1987 (Cabaret 1991). This isolate consisted of 25% resistant worms (homozygote rr) and 75% susceptible worms (25% homozygote SS and 50% heterozygote rs) at the onset of the experiment in 1998. At least 50 randomly chosen males in each host were measured and genotyped on each occasion in relation to benzimidazole resistance in the various experimental batches. Males were investigated rather than females, as their morph (Cabaret et al. 1986) or line ( sheep and goat and goat lines: Gasnier et al. 1997) is easily determined. Our isolate belonged to the sheep and goat line. It was established from the experimental data of Denham (1969) that the size of T. circumcincta adult males and females was highly correlated (r 2 = 0 97; P = 0 0004) from day 10 ( juveniles) to 20 (adult) after infection, thus males represent their corresponding females well. Although the relationship decreased after day 20 after infection, it remained highly significant up to day 40 after infection. The males were measured using semiautomatic image analysis software (Esilab V. 3 10 ; Prolabo, Paris, France). The accuracy of each measure was 0 1 mm. Worms were measured first and then genotyped for resistance to benzimidazoles. They were placed at 41 C for 15 h in a digestion solution of 15 µl of 0 1 M Tris-50 mm EDTA buffer, 10 µl Proteinase K (5 mg ml 1 ), 0 5 µl Tween 20. The tubes were then incubated at 95 C for 20 min to destroy the proteinase K and 6 µl of this solution were processed using an allele-specific PCR ( polymerase chain reaction) method (described by Humbert & Elard 1997). This PCR method detects the presence of a recessive mutation in the isotype 1 β-tubulin gene. Amplification of three fragments can be generated: two allele-specific (susceptible or resistant) and one non-specific allele. For example, resistant homozygotes will present two alleles: alleleresistant and allele non-specific, and heterozygotes will present three alleles. EXPERIMENTAL DESIGNS Worm length in relation to age of infection in lamb Thirteen lambs (immunosupressed with long-acting corticoids in order to avoid variability due to host) were infected orally with 4000 infective third-stage larvae (L3) each. The lambs were maintained in experimental conditions and were killed on day 35 (10 lambs) or on day 60 (3 lambs) after infection. These dates corresponded approximately to the age of worms we expected to recover from tracer lambs in single species infection of lambs. 548 male T. c. circumcincta were examined (445 and 103, respectively, at D35 and D60 after infection). Length difference of T. circumcincta in single species infections of lambs Four newly seeded (rye-grass and white clover) paddocks (7000 m 2 each) were infected with the same T. circumcincta isolate. Ten lambs were grazed on each paddock during all the grazing season (mid-april to mid-november) and in order to reseed the pasture with larvae that underwent different anthelmintic pressure. They were called permanent lambs. Four different anthelmintic treatment regimens were experienced by the lambs in each paddock. Lambs were treated bimonthly in 1998 and monthly in 1999. Lambs in paddock 1 acted as non-treated controls, and those in paddock 2 were treated with levamisole [an imidazothiazole efficient against this isolate, Anthelsol (Rhone-Merieux, Lyon, France)] and were considered treated controls. Alternative treatment with levamisole or fenbendazole [a benzimidazole, Panacur (Intervet SA, Beaucouze, France)] was performed on lambs in paddock 3, whereas fenbendazole treatment only was performed on lambs in paddock 4. All doses were in agreement with the manufacturer s recommendations and no cross-resistance has been recorded previously between imidazothiazoles and benzimidazoles. Different selective pressures were obtained on resistant and susceptible worms, which in turn determined a different population turnover. A slow turnover was expected in paddock 1 for susceptible and resistant worms, and a rapid one in paddock 2 for susceptible and resistant worms. The selective pressure was high on susceptible worms exposed to benzimidazole treatment in paddock 4, the benzimidazole levamisole treatment in paddock 3 and the levamisole treatment in paddock 2. The selective pressure was low on worms resistant to benzimidazoles from lambs in paddocks 3 and 4. Tracer lambs were used to obtain adult T. circumcincta. Every month or every 2 months, one of these

571 Life traits of parasitic nematodes influenced by chemotherapy lambs was introduced into each paddock for 1 month, withdrawn for 2 weeks in a stable to let the lateingested larvae mature into adult, and then necropsied to obtain the parasites. Twenty-four tracer lambs were slaughtered in 1998 and 1999 and 875 T. circumcincta morph circumcincta were examined. Teladorsagia circumcincta length differences in natural multispecies infections in goats Seven isolates of T. circumcincta (known to harbour more than 90% sheep and goat line) were collected from dairy goat farms (three intensive farms in central France and four extensive farms in southwestern France). The nematode community consisted principally of trichostrongylid nematodes such as Haemonchus contortus, Trichostrongylus colubriformis and T. circumcincta. Seven to 46 males of the latter species were examined depending on the isolate. STATISTICAL ANALYSES The analyses of male length were conducted using a general linear model (GLM) with the Simstat program (Péladeau & Lacouture 1993). GLM allows much greater flexibility than standard analysis of variance procedures by allowing one to freely combine quantitative and categorical factors and to statistically control for covariates. Linear least-square regressions were also used to assess the influence of densitydependence on size. The distribution-free Jonckheere test (Jonckheere 1954) was used to test the order of group means (homozygote resistant > heterozygote susceptible > homozygote susceptible) as the data did not conform to the normality of distribution required by one-way analysis of variance. Results WORM LENGTH IN RELATION TO AGE OF INFECTION AND WORM RESISTANCE GENOTYPE Thirty-five days after infection worm length was evaluated. In 10 lambs, 445 male worms were genotyped (107 rr, 238 rs and 100 SS) and no difference was shown between the genotypes (P = 0 71). Sixty days after infection 103 male worms were collected from three lambs (28 rr, 30 rs and 45 SS). Worm lengths were different between the three genotypes: resistant rr were larger than susceptible SS or rs (P < 0 02). This was only evidenced in older worms aged 60 days. WORM LENGTH IN RELATION TO SEASON, INTENSITY, GENOTYPE AND TREATMENTS In tracer lambs, worms were aged approximately 50 days so that the conditions were good in order to evaluate their size, since differences were recorded only in older worms. The actual length of worms is presented in Table 1. The estimated length of worms was based on a GLM model, with the following factors: treatment (P < 0 001), season (P < 0 001), genotype (P < 0 03), interaction between treatment and season (P < 0 001), and number of worms (P > 0 05). The interaction expressed the fact that the influence of treatment regimen on size was different according to season. The GLM was highly significant (P < 0 001; F = 121; n = 874) and the estimated values are in Table 1. Worm length of the three genotypes was higher in spring than in summer (P < 0 0001) (Fig. 1). Worm length was partially controlled by intensity (P < 0 0001), although it was not possible to differentiate the influence of Table 1. Observed and estimated (obtained from general linear model) average length of T. c. circumcincta for the three genotypes (resistant homozygote rr, susceptible heterozygote rs and homozygote SS) obtained in natural infection in singlespecies infections of lambs (SD in parentheses) Period Genotype Observed length (cm) Estimated length (cm) Number of worms examined May 1998 rr 1 033 (0 073) 1 015 (0 023) 79 rs 1 035 (0 067) 1 006 (0 023) 35 SS 1 031 (0 064) 0 994 (0 023) 37 June 1998 rr 0 749 (0 084) 0 837 (0 067) 34 rs 0 775 (0 088) 0 828 (0 067) 80 SS 0 776 (0 071) 0 816 (0 067) 29 September 1998 rr 0 746 (0 085) 0 847 (0 008) 45 rs 0 837 (0 074) 0 838 (0 008) 55 SS 0 788 (0 095) 0 826 (0 008) 26 October 1998 rr 0 845 (0 096) 0 855 (0 061) 67 rs 0 833 (0 096) 0 845 (0 061) 36 SS 0 820 (0 100) 0 834 (0 061) 35 May 1999 rr 1 027 (0 049) 0 991 (0 026) 95 rs 1 004 (0 069) 0 982 (0 026) 37 SS 0 958 (0 137) 0 970 (0 026) 22 June 1999 rr 0 985 (0 053) 0 955 (0 037) 54 rs 0 957 (0 125) 0 946 (0 037) 83 SS 0 956 (0 091) 0 935 (0 037) 26

572 V. Leignel & J. Cabaret SUSCEPTIBLE WORMS LENGTH IN RELATION TO ANTHELMINTIC PRESSURE Worm length increased when anthelmintic efficacy increased (P < 0 0001). Linear regression indicated that parasites were smaller in the control group ( paddock 1) than in treated groups (paddocks 2, 3 and 4) (Table 2). This was particularly obvious in the susceptible SS, for which length increased in treated groups compared to the control group from 6 to 10% (P < 0 0001). Fig. 1. Estimated length (obtained from a general linear model) of T. c. circumcincta in the four regimen of anthelmintic treatment. season and density as both factors were related. Worm length decreased when worm number ( worms ) increased: length (µm) = 9724 8 0 132 worms with r = 0 57 and P < 0 0001. For each genotype (Table 2), the following significant regressions were determined: homozygote susceptible individuals (SS) length (µm) = 9508 2 0 118 worms, heterozygote individuals (rs) length (µm) = 9551 1 0 117 worms and homozygote resistant individuals (rr) Length (µm) = 9947 8 0 154 worms. The resistant worms were influenced more by intensity (P < 0 05) than susceptible ones, since higher values of the regression slope were found in the resistant individuals. Worm lengths were different between rr, rs and SS individuals: rr > rs > SS (Jonckheere test, P < 0 02) (Table 2 data). Resistant worms were 3% larger than susceptible worms. The same size hierarchy among genotypes was found in natural infections (Table 3). Discussion The effect of chemotherapy on the life traits of parasitic nematodes has not really been described until now in controlled experiments. We considered nematode length an operational criterion that could summarize fecundity of the worms. This criterion underestimates the number of eggs in utero, as length is linear and the uterus is better represented as a volume. We showed that an increase of 1% in length corresponded to a 3 3% increase in egg numbers in a susceptible T. circumcincta (data not reported). The observed difference between susceptible and resistant worms was approximately 3% and should thus correspond to an increase of 10% in fecundity in resistant ones. According to Skorping & Read (1998), larger parasites are expected when stage-specific selective pressure is imposed, whereas Medley (1994) and Poulin (1998) presented the opposite view when age-specific selective pressure is imposed. Gemmill et al. (1999) stated that the time between infection and the onset of reproduction Table 2. Estimated maximum length (µm) of the three resistance genotypes of T. c. circumcincta obtained by linear regression (length = x y worms) in natural infection in single-species infections of lambs under four treatments (r ranged from 0 49 to 0 83) Control Levamisole (LEV) Alternation (LEV/FDZ) Fenbendazole (FDZ) Treatment x y x y x y x y Resistant homozygote (rr) 9719 0 114 10206 0 347 10323 0 266 10025 0 137 Susceptible heterozygote (rs) 9698 0 098 9894 0 317 9865 0 150 9458 0 120 Susceptible homozygote (SS) 9065 0 088 9982 0 295 9678 0 155 9656 0 111 Number of worms examined 215 213 217 230 Number of worms/host 444 23432 403 9192 384 4290 533 13075 Table 3. Estimated average length (cm) of T. c. circumcincta for the three genotypes (resistant homozygote rr, susceptible heterozygote rs and homozygote SS) obtained by general linear model analysis from natural infection in dairy goat farms Farm Cuv Ser Pic Men Gro Gar Ech Resistant homozygote (rr) 0 941 1 005 0 946 0 982 0 939 0 962 0 979 Susceptible heterozygote (rs) 0 935 0 999 0 939 0 976 0 933 0 956 0 970 Susceptible homozygote (SS) 0 919 0 984 0 924 0 961 0 918 0 941 0 955 Number of worms examined 7 43 34 15 39 46 45 Number of worms/host 1268 587 7864 343 532 1357 1065

573 Life traits of parasitic nematodes influenced by chemotherapy (maturation time) is a key determinant of body size, fecundity and generation time in parasitic nematodes. Our experiment was devoted to a demonstration of the effect of chemotherapy on worm size in a case where selection pressure was equally high for all stages (even those in tissue). All our measurements were done after maturation time and we could not evaluate all the factors included in the Gemmill et al. model. Benzimidazole-resistant worms should remain the same size in benzimidazole treated groups and susceptible worms should be different in size in untreated controls when compared with treated groups. Similarly we expect that a change in susceptible worm size will occur in the treated groups when compared to the untreated control. The susceptible worms become larger at the same stage after infection (adult stage) when submitted to massive chemotherapy, in agreement with Skorping & Read (1998) hypothesis. The larger body size is attained at the expense of a slower growth rate according to Stearns (1992). Our experiments lead us to question this statement. No difference in size was observed at day 35 after infection, whereas highly different lengths were recorded on day 60, although they were mature females (Denham 1969) at both times. This suggests that, on the contrary, growth rate was accelerated in worms submitted to chemotherapy, and this only at the adult stage. Our results show unambiguously that susceptible worms (submitted at both dates to anthelmintic pressure) were smaller than resistant ones, which was not expected. We expected resistant worms within a group, not under to anthelmintic pressure, to behave like susceptible worms in the untreated control. The resistant worms should then be smaller than susceptible worms; however, we observed the reverse. We also observed from other isolates in natural conditions that resistant worms were larger than susceptible ones. This increase in size in relation to resistance status should prove general for the studied trichostrongylid nematode, T. circumcincta. We thus support the hypothesis presented by Medley (1994) and Poulin (1998): smaller parasites are expected when chemotherapy selection is imposed (susceptible worms are smaller than resistant ones). An increase in size may be linked to resistance through a random association (linkage disequilibrium between resistance and size increase at the adult stage?). The heterozygote worms behaved like susceptible ones under anthelmintic treatment at normal dosage (Elard et al. 1998), and presented a slight advantage over susceptible ones when hosts were underdosed (Silvestre et al. 2001); their size was intermediate between resistant and susceptible homozygotes. This argues strongly for a random association between the two traits (size and resistance). Body size has repeatedly been associated with fecundity or with total reproductive capacity in interspecific comparisons among nematodes (Morand 1996) or platyhelminths (Trouve et al. 1998). A similar finding of size and fecundity has been demonstrated in the nematode we studied, either in natural infections under various conditions (Cabaret 1984 in the Middle- Atlas of Morocco; Gruner et al. 1994 in France) or in experimental infections (Stear et al. 1995, 1997, 1999a). The between-group size increase of susceptible worms under high chemotherapy pressure may then result in an increase in fertility, and thus allows for their greater fitness in this unfavourable environment. The decreased size in the untreated group could be due to a better response of lambs, since they always harboured worms that could maintain their protective immune response (immunity of premunition). The treated lambs conversely behave like never uninfected, and thus were completely susceptible to infection. The size increase of resistant worms (not under anthelmintic selective pressure) within a treatment group will probably also result in higher fitness of resistant worms compared to that of susceptible ones and such an isolate should evolve rapidly to a fully resistant population. This was not recorded in experimental infections (Elard et al. 1998) and our unpublished data on natural infections do not support the hypothesis that higher fitness is demonstrated in resistant worms, as the percentage of resistant worms remained unaltered on our experimental control plot (not submitted to treatment). We may conclude that the putative increase in fitness of resistant worms due to an increase in size (owing to higher fecundity) does not result in higher production/survival of propagules or that it is outweighed by the disadvantage of an increase in pathogenicity. The increase in pathogenicity in larger T. circumcincta has previously been shown (Stear et al. 1999b). The higher production or survival of propagules is not really supported here, as the proportion of resistant larvae remained similar to that of adult worms collected at lamb necropsy (V. Leignel & J. Cabaret, unpublished observations). Full comparative fitness components (maturation time, survival during maturation time among others) of susceptible and resistant worms originating from the same isolate have not been completely investigated, and the overall similar fitness of both phenotypes of worms remains unexplained. Acknowledgements V.L. was funded by a PhD grant from the French Ministry of Research. Additional financial support was provided by Biotechnocentre 09/01/1998 and Hoechst Roussel Vet Company 18/02/1997. Fruitful discussions with J. F. Humbert and S. Morand, as well as comments of two referees helped to improve this manuscript. The English of the final version was checked by Anne Schwab. The work conforms to the legal requirements of the country in which it was carried out, including those relating to conservation and welfare, and to the journal s policy on these matters.

574 V. Leignel & J. Cabaret References Cabaret, J. (1984) Seasonal changes in the abomasal nematodes of naturally infected ewes in Moulay-Bouzza (Morocco). Veterinary Parasitology 15, 47 56. Cabaret, J. (1991) Efficacy of netobimin against Muellerius capillaris and resistant strain of digestive tract strongyles in dairy goats. American Journal of Veterinary Research 52, 1313 1315. Cabaret, J., Ouhelli, H. & Dakkak, A. (1979) Efficacité comparée du Fenbendazole et du Tétramisole sur les helminthes parasites du mouton au Maroc. II. Helminthes du tube digestif. Recueil de Médecine Vétérinaire 155, 785 793. Cabaret, J., Morales, G. & Durette-Desset, M.-C. (1986) Caractérisation de Teladorsagia circumcincta et de T. trifurcata. Annales de Parasitologie Humaine et Comparée 61, 55 64. Conder, G.A. & Campbell, W.C. (1995) Chemotherapy of nematode infections of veterinary importance, with special reference to drug resistance. Advances in Parasitology 35, 2 29. Denham, D.A. (1969) The development of Ostertagia circumcincta in lambs. Journal of Helminthology 43, 299 310. Elard, L., Comes, A.M. & Humbert, J.F. (1996) Sequences of β-tubulin cdna from benzimidazole-susceptible and -resistant strains of Teladorsagia circumcincta, a nematode parasite of small ruminants. Molecular Biochemical Parasitology 79, 249 253. Elard, L., Sauve, C. & Humbert, J.F. (1998) Fitness of benzimidazole-resistant and susceptible worms of Teladorsagia circumcincta, a nematode parasite of small ruminants. Parasitology 117, 571 578. Gasnier, N., Cabaret, J. & Durette-Desset, M.C. (1997) Sheep and goat lines of Teladorsagia circumcincta (Nematoda): from allozyme to morphological identification. Journal of Parasitology 83, 527 529. Gemmill, A.W., Skorping, A. & Read, A.F. (1999) Optimal timing of first reproduction in parasitic nematodes. Journal of Evolutionary Biology 12, 1148 1156. Gruner, L., Mandonnet, N., Bouix, J., Khang, V.T., Cabaret, J., Hoste, H., Kerboeuf, D. & Barnouin, J. (1994) Worm population characteristics and pathological changes in lambs after a single or trickle infection with Teladorsagia circumcincta. International Journal for Parasitology 24, 347 356. Humbert, J.F. & Elard, L. (1997) A simple PCR method for rapidly detecting point mutations. Technical Tips Online, http://www.tto.trends.com, T40076. Jonckheere, A.R. (1954) A distribution-free k-sample test against ordered alternatives. Biometrica 41, 133 145. Medley, G.F. (1994) Chemotherapy. Parasitic and Infectious Diseases: Epidemiology and Ecology (eds M. E. Scott & G. Smith), pp. 141 157. Academic Press, San Diego, CA. Morand, S. (1996) Life-history traits in parasitic nematodes: a comparative approach for the search of invariants. Functional Ecology 10, 210 218. Péladeau, N. & Lacouture, Y. (1993) Simstat: Bootstrap computer simulation and statistical program for IBM personal computers. Behavioral Research Methods Instrument Computers 25, 410 413. Poulin, R. (1998) Evolutionary Ecology of Parasites: from Individuals to Communities. Chapman & Hall, London. Silvestre, A., Cabaret, J. & Humbert, J.F. (2001) Effect of benzimidazole under-dosing on the resistant allele frequency in Teladorsagia circumcincta (Nematoda). Parasitology 123, 103 111. Skorping, A. & Read, A.F. (1998) Drugs and parasites: global experiments in life history evolution. Ecology Letters 1, 10 12. Skorping, A., Read, A.F. & Keymer, A.E. (1991) Life history covariation in intestinal nematodes of mammals. Oikos 60, 365 372. Stear, M.J., Bishop, S.C., Doligalska, M., Duncan, J.L., Holmes, P.H., Irvine, J., McCririe, L., McKellar, Q.A., Sinski, E. & Murray, M. (1995) Regulation of egg production, worm burden, worm length and worm fecundity by host responses in sheep infected with Ostertagia circumcincta. Parasite Immunology 17, 643 652. Stear, M.J., Bairden, K., Duncan, J.L., Holmes, P.H., McKellar, Q.A., Park, M., Strain, S. & Murray, M. (1997) How hosts control worms. Nature 389, 27. Stear, M.J., Strain, S. & Bishop, S.C. (1999a) How lambs control infection with Ostertagia circumcincta. Veterinary Immunology and Immunopathology 72, 213 218. Stear, M.J., Bairden, K., McKeller, Q.A., Scott, I., Strain, S. & Bishop, S.C. (1999b) The relationship between the number and size of nematodes in the abomasum and the concentration of pepsinogen in ovine plasma. Research in Veterinary Sciences 67, 89 92. Stearns, S.C. (1992) The Evolution of Life Histories. Oxford University Press, Oxford. Trouve, S., Sasal, P., Jourdane, J., Renaud, F. & Morand, S. (1998) The evolution of life-history traits in parasitic and free-living platyhelminthes: a new perspective. Oecologia 115, 370 378. Received 13 September 2000; revised 17 January 2001; accepted 10 April 2001