biological control of rodents using sarcocystis singaporensis

Biological control of rodents using Sarcocystis singaporensis

T. JaÈ kel a, b, *, Y. Khoprasert c, S. Endepols d, C. Archer-Baumanna,K. Suasa-ard c, P. Promkerd c, D. Kliemt a, P. Boonsong c, S. Hongnark c

aDivision of Parasitology, Department of Zoology, University of Hohenheim, Emil Wol� Str. 34, 70599 Stuttgart, GermanybGerman Technical Cooperation (GTZ), 65726 Eschborn, Germany

cAgricultural Zoology Research Group, Entomology and Zoology Division, Department of Agriculture, 10900 Bangkok, ThailanddBayer AG, 51368 Leverkusen, Germany

Received 20 January 1999; received in revised form 31 May 1999; accepted 31 May 1999

Abstract

Parasites have been identi®ed as an important factor in regulating vertebrate populations. In replicated ®eldexperiments (plots up to 4 ha) performed in Thailand we tested whether commensal and ®eld rodents could bearti®cally infected and controlled with the host-restricted apicomplexan protozoon Sarcocystis singaporensis which is

endemic in Southeast Asia. When bait-pellets containing high numbers of these parasites were consumed by rodentsof three species (Rattus norvegicus, Rattus tiomanicus, Bandicota indica) in di�erent agricultural habitats (chickenfarm, oil palm plantation, rice®eld), we observed a parasite-induced mortality ranging from 58% to 92%. Detection

of merozoites of S. singaporensis in lung tissue samples of rats collected dead at the experimental sites using aspecies-speci®c monoclonal antibody con®rmed that S. singaporensis was the causative agent of mortality. Asobserved with brown rats, the parasite's e�ect on the host was not related to sex. These experiments demonstratefor the ®rst time that arti®cial infection of rodents with an endemic protozoon has the potential for e�ective

population control. # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rightsreserved.

Keywords: Biological control; Sarcocystis singaporensis; Southeast Asia; Wild rats

1. Introduction

The cause of cycles of vertebrate populationshas for a long time interested biologists. It isnow considered by many that trophic interactionsrather than intrinsic mechanisms are the principalcause, and recent evidence suggests that parasitesare an important factor [1]. Research on rodent±

parasite interactions can be viewed from two

principal perspectives. These are: ®rst, analysis

and manipulation of the host±parasite relation-

ship in the wild to better understand the natural

situation; and second, exploitation of pathogenic

e�ects of parasites for a practical purpose in

rodent control. The latter will be discussed here.

New approaches in the biological control of

rodents which aim at a reduction of

fecundity [2, 3] or a reduction of survival of

rodents [4] are under development. Although

International Journal for Parasitology 29 (1999) 1321±1330

0020-7519/99/$20.00 # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved.

PII: S0020-7519(99 )00081-8

* Corresponding author: Tel: 49-711-459-3072: fax; 49-711-

459-2276, e-mail; [email protected]

microparasites have been associated with declinesof rodent populations [5±7] direct demonstrationof an e�ect on survival is lacking with theexception of certain rabbit viruses which, how-ever, were introduced into the environment asalien organisms [8, 9]. Recently, the potentialof an endemic virus was highlighted [10], andin the present study we examined whether anendemic apicomplexan protozoon, Sarcocystissingaporensis, could be used for populationcontrol.

Sarcocystis singaporensis frequently occurs inrodents in Southeast Asia [11, 12] and has beenfound to be host-restricted [4, 13, 14]. It usessnakes (Python reticulatus) and rodents of thegenera Rattus and Bandicota to maintain its lifecycle. Sporocysts containing sporozoites, thestages which are infective for rats, can beobtained in large quantities from the snakehost [4]. Its potential as a biocontrol agent wasrecognised [13, 15] because this normally apatho-genic parasite induces a fatal pneumonia inrodents once infection with sporocysts exceeds acertain threshold. We then provided preliminaryevidence that S. singaporensis increases mortalityof wild rats in the ®eld [4].

In the present study, we tested in manipulativereplicated ®eld experiments whether rodent popu-lations in Thailand could be arti®cially infectedby high dosages of S. singaporensis. If so, wouldthis provide a basis for control inside the parasi-te's natural distribution range in Southeast Asia?Changes in rat population size were measuredmainly using comparative techniques such as baitconsumption, footprint cover of tracking plates,or observation of activity at burrow entrances.Comparative indices can be related to actualnumbers [16] and are a good indicator of popu-lation size [17].

2. Materials and methods

2.1. Parasites and bait

Sporocysts of S. singaporensis were originallyisolated from a wild-caught reticulated python

from Thailand and produced in large numbers asdescribed previously [4].

For application of parasites in the ®eld, 20 mlof a sporocyst suspension in PBS were injectedinto a 1 g bait-pellet. A bait was used which hadproven to be highly attractive for wild rats inpreliminary studies (unpublished). It consisted ofwheat ¯our mixed with broken corn, talcum,corn oil and various other ingredients includingattractants like ®sh extract (for brown rats,Rattus norvegicus) or coconut extract (forMalaysian wood rats, Rattus tiomanicus). To pro-tect against ants, pellets were sealed into smallpolyethylene sacchets or, for wood rats, intowax-coated paper. They were stored at 48C untiluse.

2.2. Detection of parasites in rats

A mAb (SS11D5/H3) reactive against sporo-zoites and merozoites of S. singaporensis wasgenerated according to procedures describedelsewhere [18]. The antibody did not crossreactwith acetone-®xed air-dried stages of closely re-lated parasites (unpublished observation) some ofwhich frequently occur in wild rats [12, 19]. Theseincluded Sarcocystis zamani (sporozoites andmerozoites; life cycle maintained at theUniversity of Hohenheim), Sarcocystis cymruensis(bradyzoites; obtained from muscles of wild-caught brown rats from Thailand), Frenkeliaglareoli (bradyzoites isolated from cysts in thebrain of a wild-caught European bank vole,Clethrionomys glareolus), Toxoplasma gondii(sporozoites and tachyzoites of the strain GTP1;WHO Centre VPH, Hannover, Germany), andCryprosporidium parvum (sporozoites; PublicHealth Center of the State of BadenWuÈ rttemberg, Germany).

Immunostaining of merozoites in lung imprints(200 mm2) of rats collected dead at the exper-imental sites was performed by indirect immuno-¯uorescence. Brie¯y, imprints were ®xed ontoglass slides with acetone for 10 min, washed inPBS, and incubated with undiluted SS11D5/H3supernatant for 1 h at 378C. Primary antibodieswere detected in two consecutive steps (30 min in-cubation time each at 378C) using a biotin-conju-

T. JaÈkel et al. / International Journal for Parasitology 29 (1999) 1321±13301322

gated anti-mouse IgG antibody (Kirkegaard andPerry) diluted 1:100 in PBS and ¯uorescein-con-jugated streptavidin (The Binding Site) diluted1:50 in PBS containing 0.01% Evans Blue. Theywere examined with a microscope equipped withepi¯uorescence.

2.3. Field experiment 1

An experiment with brown rats, R. norvegicus,was performed at three feed storage buildings onchicken farms in Bang Nam Prieo district,Chachoengsao province, Thailand. The buildingswere surrounded by ponds which de®ned the bor-ders of the study areas; they were 630 m2 (con-trol), 645 m2 (site 1) and 820 m2 (site 2) large. Asa ®rst index of rodent activity, consumption ofdry oat ¯akes before and after treatment of ratswith parasites or placebos, respectively, wasmeasured. Bait stations containing trays wereplaced at a density of 1/37 m2, 1/28 m2, and 1/36 m2, respectively. As a second index, the per-centage of footprint cover on tracking plates cov-ered with sand was measured using a transparentgrid of 16 squares [20]. Plates were placed inde-pendently from bait stations on rat runs at a den-sity of 1/24 m2, 1/26 m2, and 1/27 m2, respectively(>90% of plates showing rodent tracks beforetreatment). During treatment, parasite-pelletscontaining 1�105 sporocysts each or placeboswere o�ered to rats in excess inside bait stations.The time-scale of the experiment, including lagphases (no oat ¯akes and measurement ofconsumption), to counter conditioning e�ectsof census baiting is displayed in Fig. 1. Allrodents found dead inside the study areas wereweighed, inspected externally for the reproductivestatus, and dissected to take lung samples andinspect reproductive organs. To recover remain-ing rats from the sites after the experiment, theywere kill-trapped for two consecutive nights fol-lowed by treatment with coumatetralyl-contain-ing bait.


The second experiment was conducted inthree rice ®elds located in Banglen district,

Thailand, which were mainly infested with ban-dicoot rats, Bandicota indica. Population sizebefore and after two rounds of treatment (twonights each) was estimated by the number of re-opened burrow entrances (active burrows, [21])which had been closed with earth two nightsbefore. Burrows were located on dikes whichsurrounded ®elds which were 4-ha-large each.Those ®elds had been selected which containedapproximately 100 active burrows out of a totalof 200 closed at a dike length of 800 m. Allentrances were marked and monitored through-out the study. As a second index, the totalnumber of footprints on tracking plates(20�20 cm) marked crosswise with stripes ofblack ink was counted. Plates in each ®eld(n=110) were placed independently from bur-row-entrances every 6±8 m on rat runs in analternating fashion at both sides of the dikes.Bandicoot rats received the same parasite-pelletsand placebos as described previously. Pelletswere evenly distributed on the dikes and placedclose to burrows. Rats found dead in the studyareas were dissected and tissue samples of thelungs taken for analysis by indirect immuno-¯uorescence.


The experimental plots were located inside anoil palm plantation in Thasae district,Chumporn province, Thailand. The Malaysianwood rat, Rattus tiomanicus, is the predominantrodent species in oil palm plantations inThailand [22]. Four oil palm ®elds were ran-domly grouped to two parasite-treated and twocontrol plots. Each ®eld was square, 3.24 halarge, and contained 400 palms which were 9 mapart. Parasite-bait or placebos were applied ineach entire ®eld, whereas wood rats were cap-tured and their activity measured inside a corearea comprising 100 palms (0.81 ha). Parasite-bait (a pellet contained 4�105 sporocysts) orplacebos were distributed in two rounds with a15 days interval. Three pellets were placedunder each palm tree during the ®rst round;two pellets per tree in the second round.Rodent numbers were determined before and

T. JaÈkel et al. / International Journal for Parasitology 29 (1999) 1321±1330 1323

after treatment using 100 drop-door live traps(one trap per tree) for three consecutive nights;the animals were released each following morn-ing. Rodent activity was measured before treat-ment, and after the ®rst and second applicationof parasites; this was done using tracking platesas in the ®rst experiment. A total of 150 plateswas used in each ®eld, at a density of 1/54 m2

(>90% covered with tracks before treatment).Eight to 15 days after each application of para-sites, the core areas were extensively examinedfor presence of dead rats.

2.6. Computation of rodent mortality andstatistical analyses

Rodent mortality was calculated using the for-mula of Henderson and Tilton [23]: M=100�[1-(t2� c1)/(t1� c2)], with M(%)=rodent mor-tality, t=treated population, c=control popu-lation, 1=population before treatment,2=population after treatment.

Where appropriate, the chi-square test usingthe Yates correction or the Mann±Whitney U-test were performed. Both tests are part of the

Fig. 1. Time-course of the activity of wild brown rats, Rattus norvegicus, on three chicken farms before and after treatment with

Sarcocystis singaporensis or placebos (control). Bait take (mean2S.E.M.) trayÿ1 dayÿ1 and tracking plates expressed as the mean

percentage2S.E.M of footprint cover dayÿ1. Values above the x-axis indicate the total number of parasite-pellets or placebos, re-

spectively, consumed by rats.


statistical computer programme SigmaStat2 ofthe Jandel Corporation.

3. Results

3.1. Field experiment with brown rats

Two colonies of brown rats on chicken farmswere infected with the parasite. Rodent activitymarkedly declined at these sites 10±13 daysafter parasite-bait had been disseminated, indi-cating an increase of rodent mortality (Fig. 1).Activity remained unchanged in the controlwhich had received placebos. In contrast to thecontrol, food consumption at the parasite-trea-ted sites was signi®cantly greater in the pre-treatment phase than in the post-treatmentphase (Table 1) (Mann±Whitney U-test; site 1,U[n1= 20,n2= 23]=628, P<0.0001; site 2,U[23,23]=732, P<0.0001; control, U[17,17]=299,P=0.97). Similarly, footprint frequency wassigni®cantly decreased during the post-treat-ment phase when compared to the control(Mann±Whitney U-test; site 1, U[24,24]=871,P<0.0001; site 2, U[29,29]=456, P<0.0001;

control, U[25,25]=494, P=0.15). The obser-vation of numerous dead rats at the parasite-treated sites between days 20 to 30 (Fig. 1) con-®rmed rodent mortality (Table 1). No dead ratswere noticed at the control site. Intriguingly,most moribund rats left their burrows shortlybefore death in search for food or water; thus,the majority of individuals could be recoveredat the treated sites. Application of a species-speci®c mAb, SS11D5/H3, on impression smearsof lung tissue con®rmed that merozoites of S.singaporensis were present in all out of 20samples taken from dead rats at the parasite-treated sites (Fig. 2). There were >20 to severalhundreds of merozoites per sample. Twentysamples from rats trapped at the control sitewere negative.

There are no statistically signi®cant di�erencesbetween: (i) parasite-treated sites and the controlsite with regard to proportions of males andfemales that succumbed to infection or wererecovered from the site by trapping and treat-ment with coumatetralyl, respectively (statisticalanalysis refers to numbers of recovered rats inTable 1, excluding some animals of which sexcould not be determined unequivocally due to

Table 1

Comparison of footprint cover of tracking plates and bait take before and after treatment of brown rats with Sarcocystis singapor-

ensis

Location CensusaFootprint cover

(mean %2S.E.M.)

No.

tracking

plates

Bait

taken

(g)

No.

trays

Mean amount

eaten trayÿ1

(g2S.E.M.)

No.

rats/tray

(estimate)b

Total

no. rats

(estimate)

No. rats

recoveredcPercentage

mortalityd

Control Pre-treatment 82.325.5 25 879.9 17 51.828.2 3.54 60

Post-treatment 73.425.7 25 931.2 17 54.8211.2 3.74 64 38

Site 1 Pre-treatment 92.523.2 24 1256.4 20 62.8218.7 4.23 85 70

Post-treatment 13.824.2 24 212.9 23 9.323.2 0.63 14 17 83, 86

Site 2 Pre-treatment 82.324.5 29 2163.8 23 94.1214.6 5.27 121 93

Post-treatment 23.525.2 29 458.1 23 19.925.5 1.11 26 17 68, 80

a Refers to Fig. 1; days of reference are 13 (footprint cover) and 6 (food consumption) for the pre-treatment census, and day 28

for the post-treatment census.b It was assumed that each rat consumes 5% of its body mass dayÿ1 [24]. Mean body masses of rats were 293 g (n=28; control),

297 g (n=62; site 1), 357 g (n=83; site 2).c After the experiment, remaining rats were recovered from all sites by kill-trapping and treatment with coumatetralyl. Post-treat-

ment values include these animals, and pre-treatment values are the sum of the latter plus animals that were collected dead after

treatment.d The ®rst value refers to mean footprint cover, the second value to mean bait take.


an advanced stage of decay; 17 males/21 femalesat control site, 25/24 site 1, 29/36 site 2;w 2=0.54, df=2, P=0.76), (ii) proportions ofmales and females that survived within the trea-ted sites (site 1; eight males out of 33 males andnine females out of 33 females, w 2=0.003,df=1, P=0.96, site 2; eight out of 37 andnine out of 45, w 2=0.015, df=1, P=0.90).This indicates that the e�ect of the parasite onthe rats was not related to sex. When rats thathad died inside the treated plots were groupedaccording to their body mass (0age) and thechronology of their appearance in the ®eld,it became apparent that adult females withmore than 300 g body mass succumbed ®rst toinfection followed by males and juveniles ofthe recent breeding season (weight group 0±100 g) (Fig. 3). Necropsy revealed that mostof these females were pregnant or lactating(Fig. 3).

Fig. 2. Detection of a merozoite (arrowhead) of Sarcocystis

singaporensis in an imprint of the lungs of a brown rat which

was recovered dead in the study area. Indirect immuno¯uores-

cence with monoclonal antibodies of the mouse hybridoma

cell line SS11D5/H3. Host cells (hc) were counterstained with

Evans Blue.

Fig. 3. Demography of the population decline of Rattus norvegicus. Shown are the numbers of dead rats recovered from the two

parasite-treated sites of experiment 1 (upper and lower panel, respectively) grouped according to sex, body mass, and the chronol-

ogy of appearance in the ®eld. Numbers adjacent to bars indicate numbers of pregnant or lactating females.


3.2. Field experiment with bandicoot rats

Parasite-bait was o�ered to bandicoot rats, B.indica, on two occasions inside two 4-ha rice®elds(Table 2). After the second round, activity indiceswere signi®cantly decreased in both parasite-trea-ted ®elds when compared with the control(Table 2; burrows, w 2=24.5, df=1, P<0.0001[control/site 1], w 2=52.5, df=1, P<0.0001[control/site 2]; footprints, w 2=133.1, df=1,P<0.0001 [control/site 1], w 2=579.6, df=1,P<0.0001 [control/site 2]). Additionally, the

data show a signi®cant di�erence between thetwo parasite-treated sites (burrows, w 2=5.1,df=1, P=0.024; footprints, w 2=127.0, df=1,P<0.0001) indicating that there was variation inthe course of treatments at the di�erent sites.

In rice ®elds where parasites had been released,a smell of decay con®rmed the presence of deadrodents 12±19 days after the ®rst round. Due tothe dense vegetation and high ambient tempera-tures recovery of dead rats was minimal (20 adultbandicoot rats from both treated sites). All outof six lung samples obtained from these rats were

Table 2

Activity of bandicoot rats before and after treatmenta with Sarcocystis singaporensis

No. active burrowsb No. footprints

Location February 3±5

(before treatment)

February 24±26

(after 1st treatment)

March 20±22

(after 2nd treatment)February 3±5 February 24±26 March 20±22

Percentage

mortalityc

Control 93 87 125 489 298 640

Site 1 98 50 42 548 481 233 65, 68

Site 2 112 71 24 1092 250 121 84, 92

a Dissemination of 3.5�107 and 3.8�107 sporocysts haÿ1 on two 4-ha experimental plots, respectively, at February 5±7 and

March 5±7, 1997, each. A bait pellet contained 1�105 sporocysts. Bandicoot rats accepted 65.3 and 50.3% of the parasite bait, re-

spectively, during the ®rst treatment round, and 15.2 and 28.2% in the second round. Placebos were o�ered to rats in the control

®eld.b Approximately one entrance per burrow system was closed. It was reported about a closely related bandicoot species that on

average 1.5 adults (range 0±4) inhabitate a burrow system [25].c The data of February 3±5 and March 20±22 were compared; the ®rst value refers to the burrow data, the second to footprint

data.

Table 3

Numbers of Malaysian wood rats and their activity before and after treatmenta with Sarcocystis singaporensis

No. rats capturedb Footprint cover (mean %2S.E.M., n=150)

Location 28±31 January

(before treatment)

25±28 March

(after 2nd treatment)

5±7 February

(before treatment)

1±2 March

(after 1st treatment)

24±25 March

(after 2nd treatment)

Percentage

mortalityc

Control 1 31 72 51.122.5 25.122.1 31.121.7

Control 2 51 58 38.422.3 33.822.2 51.322.2

Site 1 45 21 52.922.8 9.621.5 17.221.9 71, 65

Site 2 50 23 59.022.7 9.821.3 23.021.8 71, 58

a Rodents inside the parasite-treated core areas accepted 94 and 100%, respectively, out of 300 bait-pellets during the ®rst round

at February 15, 1998, and 100% out of 200 bait-pellets on both plots during a second application at March 3. Each bait-pellet con-

tained 4�105 sporocysts. Placebos were applied in the control plots.b Numbers captured in 300 trap-nights.b The ®rst value refers to rat numbers, the second value to footprint data. Values of sites 1 and 2 were compared with the mean

of the controls.


positive for S. singaporensis when probed withmAb SS11D5/H3. In the control ®eld, no smellof decay was noticed nor were dead rats seen.

3.3. Field experiment with Malaysian wood rats

As in the previous experiment, parasite-baitwas o�ered to rats twice. Compared with thesituation before treatment, signi®cantly less woodrats were captured in the parasite-treated ®eldsafter the second round, whereas numbers in thecontrol ®elds had increased (Table 3). There isno signi®cant di�erence between parasite-treatedsites regarding rodent numbers before and afterapplication of parasites (w 2=0.02, df=1,P=0.89), in contrast, data are signi®cantlydi�erent when compared to the controls (e.g.control 2/site 1, w 2=6.8, df=1, P<0.01; con-trol 2/site 2, w 2=7.5, df=1, P<0.01).

A similar trend is apparent for the footprintdata, although there was considerable variationamong the controls with control ®eld 1 showinga signi®cant drop of rodent activity comparedwith pre-treatment values (Mann±Whitney U-test, U[150,150]=26975, P<0.0001, after secondtreatment round) despite a marked increase inrodent numbers (Table 3). Footprint cover incontrol ®eld 2 was similar to pre-treatment valuesafter the ®rst baiting round (U[150,150]=23571,P=0.19), but was signi®cantly increased afterthe second round (U[150,150]=19495, P<0.0001).Inside the parasite-treated ®elds, a signi®cantdecrease of rodent activity after the ®rst baitinground (site 1, U[150,150]=30870, P<0.0001; site2, U[150,150]=31432, P<0.0001) translates intomortality rates of 72 and 74%, respectively. Thiswas paralleled by the observation of 54 and 34dead wood rats, respectively, 11 to 17 days afterdissemination of sporocysts. After the secondtreatment, another 10 and 12 rats, respectively,were found dead at these sites. At that time,footprint frequency also was signi®cantly lowerthan pre-treatment values (site 1,U[150,150]=29388, P<0.0001; site 2,U[150,150]=29297, P<0.0001). No dead rodentswere seen at the control sites.

4. Discussion

The results from three independent exper-iments indicated that an endemic protozoon, S.singaporensis, signi®cantly increased mortality of®eld populations of rodents when o�ered at arti-®cially high doses. Rodents in treated popu-lations had a level of mortality that iscomparable to chemical rodenticides tested undersimilar environmental conditions [26]. It was pre-viously suggested that control of eruptive pestsmay require tactically released control agents [27].Our data indicate the feasibility of such a con-cept: the parasite's e�ect is ampli®ed bymanipulation [15] and can be terminated if notrequired.

Sarcocystis singaporensis was e�ective againstthree di�erent rat species in di�erent agriculturalhabitats proving that the observations on viru-lence obtained in the laboratory [4, 15] are alsovalid in the wild. Non-susceptible hosts [4, 13, 14]appeared not to be a�ected by this parasite.Although, the experimental sites had beenselected to be largely inhabitated by single targetspecies only, some other rodents such as Musspp. were present but no dead specimens of thesewere found. With respect to the ethical questionsinvolved in killing of vertebrates, it is noteworthythat rats do not appear to be ill until shortlybefore death because food consumption is unal-tered most of the time during acute infection [4].The present results conform with this; food con-sumption remained stable until a sharp declineoccurred (Fig. 1).

To measure changes in population size, weused various indices of abundance or activity. Allproved to be useful because estimates of mor-tality within single sites were quite similar inmost cases. Occasionally, indices obtained withtracking plates showed considerable variationand did not conform with results of robust tech-niques like live-trapping (e.g. experiment 3). Thisindicates that tracking plates should not serve asa sole source of information in the ®eld. Besidesthat, di�erential host±parasite interactions mayalso account for the variation observed amongparasite-treated ®elds (e.g. experiment 2). Withinthe endemic area of S. singaporensis, di�erent


populations of a particular rodent species mayexpress genetical di�erences with regard to sus-ceptibility to infection (role of immunity, seebelow). We have shown that brown rats, R. nor-vegicus, living outside the endemic area are moresusceptible to infection with this parasite thanthose living within; additionally, there also existdi�erences between rodent species [28].

The experiment with brown rats focused onthe temporal aspects of the disease. Because ac-tivity declined 10±13 days after the ®rst appli-cation of parasite-bait (Fig. 1) and most of thedead rats were seen in the ®eld after this interval,it can be concluded that the lethal dose had beenalready consumed within the ®rst 2 days. Theabove interval is exactly the time required formerozoites to appear in lung tissue [4]. The twoother rodent species tested also accepted theparasite-bait immediately. This underlines thehigh degree of palatability of the bait material. Italso indicates that aversion towards novel food,or neophobia (which is assumed to be generallytrue for brown rats, [29]), may not always applyto rodents in particular geographic locations orcontrol situations.

The ®rst experiment further demonstrates thatthe e�ect of the parasite is not related to the sexof the host. This is an important result because ifmales were predominantly a�ected then thepopulation consequences would be lesspronounced [30]. Interestingly, our data showthat pregnant or lactating females were the ®rstto be infected with the parasite. This is probablydue to increased energy requirements of femalesduring the reproductive period; thus, they prob-ably accept novel foods faster.

The experiments with bandicoot rats and woodrats di�ered from the previous one in that thestudy areas were considerably larger, no pre-bait-ing was used before application of parasite-bait(therewith avoiding a possible selection of ratsthat liked the food), and two treatment roundswere performed instead of one only. The ®rsttwo points outline that the experimental pro-cedures conformed with current control practices.The latter is important because the data showthat rodents were also a�ected by a second treat-ment round. Dead wood rats were found after

the second application of parasite-bait; similarly,activity of bandicoot rats decreased after the sec-ond treatment. In a previous study, we haveshown that consumption of a sublethal amountof sporocysts can immunise rats, and argued thatprotective immunity could be an obstacle to con-secutive treatments [4]. Obviously, this appearsnot to be a problem in the short term.Intriguingly, arti®cial infection with S. singapor-ensis is highly pathogenic for rats in Thailanddespite the fact that this protozoon is highlyprevalent in the wild [12]. We regularly observedrats with natural infections of S. singaporensisthat do not resist challenge infection in the lab-oratory (unpublished observation). The mechan-isms of immunity are still poorly understoodand, therefore, the subject of our currentresearch.

There is a promising prospect for an appli-cation of S. singaporensis within the endemicarea in Southeast Asia. Its application elsewheremay be considered, provided its high degree ofhost speci®city can be further con®rmed.

Acknowledgements

We thank B. J. Wood and two anonymousreferees for comments on an earlier version ofthe manuscript.

References

[1] Hudson PJ, Dobson AP, Newborn D. Prevention of

population cycles by parasite removal. Science

1998;282:2256±8.

[2] Singleton GR, Chambers LK. A large scale manipulative

®eld study of the e�ect of Capillaria hepatica on wild

mouse populations in southern Australia. Int J Parasitol

1996;26:383±98.

[3] Chambers LK, Singleton GR, Hood GM.

Immunocontraception as a potential control method of

wild rodent populations. Belg J Zool 1997;127(suppl

1):145±56.

[4] JaÈ kel T, Burgstaller H, Frank W. Sarcocystis singaporen-

sis: studies on host speci®city, pathogenicity, and poten-

tial use as a biocontrol agent of wild rats. J Parasitol

1996;82:280±7.


[5] Brunner-Korvenkontio M, Henttonen H, Vaheri A.

Hemorrhagic fever with renal syndrome in Finland: ecol-

ogy and virology of nephropathia epidemica. Scand J

Infect Dis 1982;38:88±91.

[6] Mihok S, Turner BN, Iverson SL. The characterization

of vole population dynamics. Ecol Monogr 1985;55:399±

420.

[7] Grenfell BT, Dobson AP. Ecology of infectious diseases

in natural populations. Cambridge: Cambridge

University Press, 1995.

[8] Fenner F. Biological control, as exempli®ed by smallpox

eradication and myxomatosis. Proc R Soc Lond B

1983;218:259±85.

[9] Lawson M. Rabbit virus threatens ecology after leaping

the fence. Nature 1995;378:531.

[10] Feore SM, Bennett M, Chantrey J, Jones T, Baxby D,

Begon M. The e�ect of cowpox virus infection on

fecundity in bank voles and wood mice. Proc R Soc

Lond B 1997;264:1457±61.

[11] O'Donoghue PJ, Watts CHS, Dixon BR. Ultrastructure

of Sarcocystis spp. (Protozoa: Apicomplexa) in rodents

from North Sulawesi and West Java, Indonesia. J Wildl

Dis 1987;23:225±32.

[12] JaÈ kel T, Khoprasert Y, Sorger I et al. Sarcosporidiasis in

rodents from Thailand. J Wildl Dis 1997;33:860±7.

[13] Zaman V. Host range of Sarcocystis orientalis. Southeast

Asian J Trop Med Pub Hlth 1976;7:112.

[14] HaÈ fner U, Frank W. Host speci®city and host range of

the genus Sarcocystis in three snake-rodent life cycles.

Zbl Bak Hyg Orig A 1984;256:296±9.

[15] Wood BJ. Biological control of vertebratesÐa review,

and assessment of prospects for Malaysia. J Pl Prot

Tropics 1985;2:67±97.

[16] Quy RJ, Cowan DP, Swinney T. Tracking as an activity

index to measure gross changes in Norway rat popu-

lations. Wildl Soc Bull 1993;21:122±7.

[17] Cowan DP, Townsend MG. Field evaluation of rodenti-

cides. In: Buckle AP, Smith RH, editors. Rodent pests

and their control. Oxon: CAB International, 1994;181±

196.

[18] JaÈ kel T, Archer-Baumann C, Boehmler AM et al.

Identi®cation of a subpopulation of merozoites of

Sarcocystis singaporensis that invades and partially devel-

ops inside muscle cells in vitro. Parasitology

1999;118:235±44.

[19] Webster JP, Macdonald DW. Parasites of wild brown

rats (Rattus norvegicus) on UK farms. Parasitology

1995;111:247±55.

[20] Brunton CFA, Macdonald DW, Buckle AP. Behavioral

resistance towards poison baits in brown rats Rattus nor-

vegicus. Appl Anim Behav Sci 1993;38:159±74.

[21] Lund M, Lodal J. Current work at the Danish Pest

Infestation Laboratory. In: Dubock AC, editor. The or-

ganisation and practice of vertebrate pest control.

Surrey: ICI, 1984;253±264.

[22] Wood BJ. Rodents in agriculture and forestry. In:

Buckle AP, Smith RH, editors. Rodent pests and their

control. Oxon: CAB International, 1994;45±83.

[23] Henderson CF, Tilton EW. Tests with acaricides against

the brown wheat mite. J Econom Entomol 1955;48:157±

61.

[24] Kaukeinen DE. Activity indices to determine trends in

vertebrate pest populations. In: Dubock AC, editor. The

organisation and practice of vertebrate pest control.

Surrey: ICI, 1984;73±94.

[25] PocheÂ RM, Mian MY, Haque ME, Sultana P. Rodent

damage and burrowing characteristics in Bangladesh

wheat ®elds. J Wildl Manage 1982;46:139±47.

[26] Buckle AP, Rowe FP, Husin AR. Field trials of

Warfarin and Brodifacoum wax block baits for the con-

trol of the rice ®eld rat, Rattus argentiventer, in peninsu-

lar Malaysia. Trop Pest Manage 1984;30:51±8.

[27] Spratt DM. The role of helminths in the biological con-

trol of mammals. Int J Parasitol 1990;20:543±50.

[28] Boonsong P, Hongnark S, Suasa-ard K. et al. Rodent

management in Thailand. In: Singleton G, Hinds L,

Leirs H, Zhang Z, editors. Ecologically-based rodent

management. Canberra: ACIAR, 1999, in press.

[29] Macdonald DW, Fenn MGP. The natural history of

rodents: preadaptations to pestilence. In: Buckle AP,

Smith RH, editors. Rodent pests and their control.

Oxon: CAB International, 1994;1±21.

[30] Caughley J, Pech R, Grice D. E�ect of fertility control

on a population's productivity. Wildl Res 1992;19:623±7.


biological control of rodents using sarcocystis singaporensis

Documents