POPULATION AND COMMUNITY ECOLOGY Spatial Structuring of Triatoma infestans (Hemiptera, Reduviidae) Populations from Northwestern Argentina Using Wing Geometric Morphometry

JUDITH SCHACHTER-BROIDE, JEAN-PIERRE DUJARDIN,1 URIEL KITRON,2 3 AND RICARDO E. GU¨ RTLER

Laboratorio de Eco-Epidemiologõ´a, Dpto. de Ecologõ´a, Gene´tica y Evolucio´n, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires. Ciudad Universitaria, C1428EHA, Buenos Aires, Argentina

J. Med. Entomol. 41(4): 643Ð649 (2004) ABSTRACT Wing geometric morphometry was used to study the spatial structuring of populations of Triatoma infestans from different villages, ecotopes, and sites within a village in northwestern Argentina. A total of 308 male and 197 female wings of T. infestans collected from peridomestic and domestic ecotopes in March 2000 was analyzed. On average, female bugs had a signiÞcantly larger wing size than males. Triatomines collected from domiciles or structures associated with chickens had larger wings than bugs collected from goat or pig corrals. The wing size of bugs did not differ signiÞcantly between villages. Discriminant analyses of wing shape showed signiÞcant divergence between villages, ecotopes, andindividualcollectionsites. The study of metric variationof males betweensites belonging to the same ecotope also revealed signiÞcant heterogeneity. Indeed, within the same section of the village the differencebetweentwo goat corrals was sometimes greater thanthat between neighboring goat and pig corrals. Thus, morphometric heterogeneity within villages may be the result not only of ecotope and host associations, but also of physical isolation between subunits. The strong structuring of T. infestans populations in the study area indicates that recolonization could be traced back to a small geographic source.

KEY WORDS Triatoma infestans, geometric morphometry, populationstructure, isometric size, shape

Triatoma infestans (Klug) is the principal vector of ticides, but at least in Argentina they were not a source Trypanosoma cruzi, the causative agent of ChagasÕ of T. cruzi-infected bugs (Cecere et al. 1999, 2002). disease. T. infestans occurs almost exclusively indo- Most morphometric studies on T. infestans have mestic or peridomestic habitats, with very rare sylvatic beenbased ontraditionalmorphometry to compare populations (Carcavallo et al. 1988, Dujardin et al. insects grown under laboratory conditions or coloniz- 1987). The transmission of T. cruzi to humans mostly ing human dwellings and sylvatic habitats (Dujardin et occurs in human habitations (the domestic environ- al. 1997a, b). Triatomines collected from human dwell- ment or domiciles) as a result of the interaction be- ings tended to show a reduced size relative to sylvatic tween domestic triatomines, infected dogs or cats, populations (Dujardin et al. 1997a, 1997b, 1998a, 1999; people, and chickens (Cohen and Gu¨ rtler 2001). The SchoÞeld et al. 1999). None of these studies compared peridomestic rural environment usually includes a T. infestans from domestic and peridomestic ecotopes wide array of structures, such as goat, sheep, and pig withinandbetweenvillages from a well-deÞnedarea. corrals; chickencoops; andstorerooms. Innorthern Traditional morphometry does not allow recovering Argentina, these ecotopes are very frequently infested the geometry of the original form from the usual dis- by T. infestans and other triatomine species (Cecere et tance measurements (Rohlf and Marcus 1993). In con- al. 1997, 2002; Canale et al. 2000). Peridomestic sites trast, geometric morphometry preserves the informa- were the most important source of T. infestans that tion on spatial arrangement of the organism and reinvaded the house after residual spraying of insec- reduces the effects of differences in growth, which generally have environmental causes. When growth effects are properly removed from shape and signiÞ- 1 Unite´ Mixte de Recherche, Institut de Recherches pour le De´- cant differences between groups are still detected, veloppment-Centre National de Recherche ScientiÞque, Montpellier, shape variables are supposed to reßect adaptive (or France. 2 College of Veterinary Medicine, University of Illinois at Urbana- genetic) causes rather than mere environmental in- Champaign, IL 61801. ßuences. Geometric morphometry has been applied 3 E-mail: [email protected]. to Triatominae for taxonomic purposes (Matias et al.

0022-2585/04/0643Ð0649$04.00/0 ᭧ 2004 Entomological Society of America 644 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 41, no. 4

Table 1. Numbers (n) of male and female T. infestans wings examined for geometric morphometry according to site of capture

Males (n ϭ 308) Females (n ϭ 197) A M T S Total A M T S Total Domiciles 15 020172200022 Chickencoops 39 12 4 0 55 44 82054 Goat corrals 59 0 36 12 107 20 8 18 8 54 Pig corrals 81 0 6 20 107 37 04849 Wood piles 0 22 0 0 22 0 18 0 0 18 Total 194 34 48 32 308 123 34 24 16 197

Amama´ and nearby villages, March 2000: A, Amama´; M, Mercedes; T, Trinidad; S, San Pablo.

2001, Villegas et al. 2002); to distinguish laboratory- were those that did not share a roof with bedroom reared and Þeld specimens (Jaramillo et al. 2002); and areas, such as storerooms and chicken coops (where to assess the chromatic variationof T. infestans across chickens usually nested; both structures hereafter will several countries (Gumiel et al. 2003). As part of a be called chicken coops for convenience), corrals broader project onthe eco-epidemiology of ChagasÕ (goat and pig corrals), and wood piles. disease in northern Argentina, we report in this work Three skilled bug collectors from NCS searched for the Þrst attempt to use geometric morphometry to triatomines in all bedroom and peridomestic areas study the spatial structuring of T. infestans populations using 0.2% tetramethrin dislodgant (Icona, Buenos at a Þner geographic scale. The metric variation of Aires, Argentina) in 117 houses that included 525 wings was investigated within and between rural com- identiÞed peridomestic sites from 14 to 27 March 2000. ϫ munities in a 10 km 10 km area with documented The bugs were identiÞed to species and stage at the control interventions. Þeld laboratory, as described elsewhere (Canale et al. 2000), and stored at Ϫ20ЊC. Materials and Methods Metric Data. In total, 308 male and 197 female wings of T. infestans were examined (Table 1). The wings Study Area and Insects. The insects were captured were mounted between microscope slides and cover- inthe rural villages of Amama´, Trinidad, Mercedes, slips using a commercial adhesive. Photographs of and San Pablo (27.18Њ S, 63.08Њ W), Moreno Depart- ment, Province of Santiago del Estero, Argentina, in each pair of wings were taken using a digital camera March 2000. The area and its history of infestation by and a stereo-microscope. We identiÞed a total of 11 T. infestans were described by Gu¨ rtler et al. (1994). type I landmarks (venation intersections) and two Amama´ had 88% of houses with bedroom areas in- type II landmarks, according to Bookstein (1990) (Fig. fested by T. infestans before being sprayed with 2.5% 1). The geometric coordinates of each landmark were suspension concentrate deltamethrin at 25 mg active digitized using tpsDig version1.18 (Rohlf 1999a). ingredient/m2 by the National Chagas Service (NCS) Size Variation. For comparisonof overall wingsize inSeptember 1985, andagaininOctober 1992 (Gu ¨ rtler betweensexes andbetweenecotopes withineach sex, et al. 1994, Cecere et al. 2002). Reinfested sites were we used the isometric estimator known as centroid selectively sprayed with deltamethrinby NCS (1993Ð size (CS) derived from coordinates data. CS is deÞned 1995) or householders themselves (since 1996). Other as the square root of the sum of the squared distances rural areas at the department level were Þrst sprayed between the center of the conÞguration of landmarks with residual insecticides against triatomine bugs in and each individual landmark (Bookstein 1990). It was 1994Ð1995 and irregularly thereafter. Most domiciles extracted from each matrix using tpsRegr version1.18 had adobe walls and thatched roofs; peridomestic sites (Rohlf 1999b).

Fig. 1. Points of reference measured as coordinates of wings of T. infestans. Numbering on the points denotes the arrangement followed to obtain the coordinates using tpsdig version 1.18. Points 8 and 11 correspond to landmarks type II, and the rest to landmarks type I. July 2004 SCHACHTER-BROIDE ET AL.: SPATIAL STRUCTURING OF T. infestans 645

Fig. 2. Variationof CS betweenecotopes andsex of T. infestans. Box plots show the isometric size differences of wings between ecotopes for each sex. Each box shows the mean, standard error, and standard deviation of each group. Asterisks followed by different letters showed statistically signiÞcant differences.

Shape Variation. Shape variables (partial warps) The software used was: BAC (version23; http:// were obtained using the Generalized Procrustes Anal- life.bio.sunysb.edu/morph/) for principal component ysis superimpositionalgorithm (Rohlf 1996). They analyses; PAD (version40; http://life.bio.sunysb.edu/ were computed separately for each sex and tested for morph/) for discriminant analyses and permutations; their variationusing tpsRelw version1.18 (Rohlf PHYLIP (version3.5c.) and treeview (version1.6.6.) 1999c). for unweighted pair-group method with arithmetic Statistical Analysis. We explored the variationof average tree construction; and JMP (version3.2.2) for size, shape, or their combination at different spatial ANOVA onsize andKappa statistic. levels: 1) the village, 2) the ecotope, and 3) the indi- vidual site of insect collection. For the Þrst two levels Results of comparison, the interaction with sex was Þrst ex- ϭ amined using either analysis of variance (ANOVA) Size Variation. Female wing size (CSf 1017.1, (for size, which is a single variable) or multivariate SD ϭ 39.32) was signiÞcantly larger than male wing ϭ ϭ ϭ ϭ Ͻ analysis of variance (MANOVA) (for shape, which is size (CSm 977.3, SD 38.98) (F 66.35, df 1, P a set of many variables). For the third level, such 0.0001). Wing size signiÞcantly differed between ec- analyses could not be performed. otopes, but not between villages (F ϭ 7.18, df ϭ 4, P Ͻ Whenpossible, the problem of sample size lower 0.0001, and F ϭ 1.61, df ϭ 3, P Ͼ 0.1, respectively), with thanthe numberof shape variables (see discussionin no signiÞcant interaction effect of sex (F ϭ 0.43, df ϭ Bookstein 1996) was circumvented by using a re- 4, P Ͼ 0.5, and F ϭ 2.1, df ϭ 3, P ϭ 0.1, respectively). stricted representation of shape, i.e., a set of principal For each sex, triatomines collected from domiciles and components (relative warps) derived from the shape chickencoops had larger wingsize thanthose col- variables. This procedure reduced the amount of cap- lected from goat or pig corrals and wood piles (Fig. 2). tured shape, but the reductionneverexceeded 10%. Statistically signiÞcant differences were observed be- Because sample size for females did not reach the tweenmale bugs collected from domiciles or chicken minimum number of wings required per individual coops and corrals, and between females collected site, analysis of shape variation at site level was per- from chickencoops andcorrals or wood piles. The formed on male insects only. The analysis compared 8 wing size of the T. infestans collected from goat and pig units across three villages: three goat and three pig corrals did not differ signiÞcantly between villages. corrals from Amama´, one goat corral from Trinidad, Shape Variation. Wing shape showed signiÞcant and one pig corral from San Pablo. divergence between villages (P Ͻ 0.002) and ecotopes Mahalanobis distances were derived from these (P Ͻ 0.0001). Because of a highly signiÞcant interac- analyses, and their statistical signiÞcance was com- tionwith sex at both subdivisionlevels, this divergence puted by nonparametric permutation tests (1000 runs was examined by separate discriminant analyses in each) after Bonferroni correction. These distances each sex. The last level of subdivision(the site of were used inanunweightedpair-group method with capture) was examined for males only, producing arithmetic average (UPGMA) cluster analysis to pro- again a signiÞcant heterogeneity of size and shape. duce a dendrogram.ReclassiÞcationbased onthese The reclassiÞcation scores of the discriminant anal- distances was evaluated by the Kappa statistic (Fleiss ysis were the lowest for the villages that had a wider 1981). range of ecotopes (Amama´ and Trinidad for males, 646 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 41, no. 4

Table 2. Reclassification of samples of T. infestans according to sex and the village in which they had been captured, after a discriminant analysis using size and shape variables

Males Females Village Observed Observed Observed Observed AMTS AMT S total agreement total agreement A 124 16 32 22 194 64 68 13 25 17 123 55 M 1 28 5 0 34 82 7 20 4 3 34 59 T 10 5 31 2 48 65 2 2 18 2 24 75 S 2 2 1 27 32 84 0 1 0 15 16 94 Total 308 68 197 61

A, Amama´; M, Mercedes; T, Trinidad; S, San Pablo.

64Ð65%; or Amama´ and Mercedes for females, 55Ð grouped together with bugs collected from pig corral 59%) (Table 2). Whensize andshape variables were A44 (Figs. 3 and 4). combined together to perform the discriminant anal- ysis, the Kappa statistic was always improved (data not Discussion shown). Therefore, the rest of the analyses was per- formed by combining size and shape variables. The All species of Triatominae grown under controlled observed agreement between the observed and ex- laboratory conditions or colonizing human dwellings pected classiÞcationfor bugs accordingto the village tendto show a reductioninsize relative to their or ecotope where they had beencaptured was mod- sylvatic populations (Dujardin et al. 1997a, 1997b, erate (according to Landis and Koch 1977). The third 1998a, 1999; Jaramillo et al. 2002; SchoÞeld et al. 1999). level of comparisons (between physical sites of cap- Natural selectionmay favor larger phenotypesinsyl- ture) gave the most satisfactory reclassiÞcationscores vatic habitats, possibly because of a greater capacity to (Table 3). resist temporary food shortages, while smaller indi- The results of the unweighted pair-group method viduals apparently survive better under laboratory or with arithmetic average dendrogram (Fig. 3) showed domestic conditions where food availability is less that nearly all Amama´ sites (group 1) were classiÞed restricted (Dujardinet al. 1997a, 1997b). Dujardinet together as similar sites. These Þve sites were all lo- al. (1999) suggested that the morphological changes cated inthe southwesternsectionof the village and observed between sylvatic and laboratory populations within200 m from each other (Fig. 4). However, one withina giventriatominespecies might parallel those site from Amama´ (A14) was grouped separately existing between sylvatic and domestic populations. (group 2) with another site (T12) from Trinidad vil- Extrapolating to peridomestic versus domestic triato- lage, both being goat corrals (Fig. 3). It is noteworthy mines, we would expect to Þnd bigger specimens in that site A14 was located Ͼ1.5 km northeast of the peridomestic rather thandomestic ecotopes. How- other Amama´ sites (group 1) and separated from them ever, triatomines collected from the more peridomes- by a canal. tic goat corrals, pig corrals, or wood piles had smaller Withingroup 1 (southwest Amama´), the similarity wing size than those collected from domestic sites or of wing shape from bugs in structures that were closest chicken coops. These differences may be explained by to each other was higher thanfrom bugs from the same the poorer nutritional status and larger abundance of ecotope. Thus, bugs collected ingoat corral A34 were T. infestans ingoat or pig corrals compared with bugs grouped together with bugs collected from pig corral collected in structures in which chickens nested (un- A46, and bugs collected in goat corral A45 were published data). Goat or pig corrals also experienced more marked variations in microsite temperature and

Table 3. Degree of correct classification among T. infestans bugs according to village, ecotope, and site of capture for both sexes

Observed Kappa ClassiÞcationSex agreement (standard error) By village Male 68 0.51 (0.04) Female 61 0.43 (0.05) By ecotope Male 56 0.43 (0.04) Female 61 0.51 (0.04) By site of capture Male 72 0.68 (0.04)

Observed agreement: proportion of individuals that have been Fig. 3. Unweighted pair-group method with arithmetic correctly attributed to their respective group. Kappa: measure of average dendrogram derived from Mahalonobis distances. A, agreement estimated between observed and expected classiÞcation scaled from 0 to 1, a score below 0 is considered poor; between 0.00 Amama´; T, Trinidad; S, San Pablo; the number after the letter and 0.20, slight, between 0.21 and 0.40, fair; between 0.41 and 0.60, of the village is the house number; pg, pig corral; gt, goat moderate; between 0.61 and 0.80, substantial; and between 0.81 and corral; the number between parentheses is the sample size 1, perfect (Landis and Koch 1977). for each site. July 2004 SCHACHTER-BROIDE ET AL.: SPATIAL STRUCTURING OF T. infestans 647

Fig. 4. Map of southeasternAmama´ indicatingthe locationof the collectionsites inwhich bugs from group 1 of the unweighted pair-group method with arithmetic average dendrogram were collected. relative humidity, which would affect growth rates villages and between unique sites of capture of insects (Vazquez-Prokopec et al. 2002). Host species effects such as individual (and neighboring) goat or pig cor- may also be involved, because hosts may also differ in rals. The lower classiÞcationscores betweenvillages their nutritional characteristics (Guarneri et al. 2000). or ecotopes versus individual sites of capture could Experiments using the typical host species found in indicate that villages or ecotopes were actually sub- the Þeld would help clarify whether there were any divided into discrete units, whereas the higher scores host species effects on wing shape and size. obtained for sites of capture suggested that they rep- An alternative, although not exclusive, explanation resented the discrete units in which metric differen- to size differences is associated with bug dispersal tiationtook place. betweenecotopes, resultingintriatominescaptured The scores of village reclassiÞcationwere the lowest inanecotope inwhich they had notdeveloped. Char- for villages presenting a wider range of ecotopes acteristics of these migrant bugs would be those that (Amama´ and Trinidad for males; Amama´ and Mer- ensure success in dispersing to and colonizing the new cedes for females) (Table 2). This is probably caused site. Parallel studies onreinfestationdynamicsinour by the fact that bugs collected from different ecotopes study area suggested that such process was directional varied signiÞcantly in size and shape; therefore, pool- and progressed from peridomestic to domestic sites (Cecere et al. 1997, 2002). The higher local abundance ing different ecotopes into a single group (i.e., a vil- and lower nutritional status of T. infestans from goat lage) could confound metric identiÞcation. However, corrals suggested that adult triatomines from the latter the study of metric variationbetweensites belonging would be more likely to disperse by ßight thantriato- to the same ecotope also revealed signiÞcant hetero- mines from chicken coops (unpublished data). Such geneity. Indeed, the difference between two goat cor- selective dispersal would cause a decrease inmean rals was sometimes greater thanbetweenneighboring bug size ingoat corrals andanincreaseinstorerooms, goat andpig corrals withinthe same sectionof the chickencoops, anddomestic sites. Light trap collec- village. Distance per se, the canal or both could be tions revealed signiÞcant ßight dispersal of T. infestans acting as barriers to genetic ßow between sites in the insummer (Vazquez-Prokopec et al. 2004). same village. This suggests that morphometric heter- The notable feature of morphometric variation dis- ogeneity within villages may be related not only to closed by this study is the signiÞcant metric hetero- ecotope and host associations, but also to some degree geneity at a very Þne spatial resolution, i.e., within of physical isolationbetweensubunits. 648 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 41, no. 4

Our Þndings are in general agreement with other (eds.), Proceedings, Michigan Morphometrics Work- genetic studies. Multilocus enzyme electrophoresis shop, 1988. The University of Michigan Museum of Zo- data have provided important information on gene ology, Special PublicationNo. 2, AnnArbor,MI. ßow and genetic structuring in domestic T. infestans Bookstein, F. L. 1996. Combining the tools of geometric populations. In Bolivia, the panmictic unit was sug- morphometrics: advances in morphometrics, pp. 131Ð152. gested to be a group of villages separated by 50 km in In L. F. Marcus, M. Corti, A. Loy, G.J.P. Naylor, and D. Vallegrande, and single villages separated only by few Slice (eds.), Advances in morphometrics NATO ASI, se- ries A, life sciences. Plenum Publication, New York. kilometers inthe Yungasvalleys (Dujardinet al. 1988, Brenie`re, S. F., M. F. Bosseno, F. Vargas, N. Yaksic, F. Noi- 1998b). Inother areas of Bolivia, however, the pan- reau, S. Noel, J. P. Dujardin, and M. Tibayrenc. 1998. mictic unit was represented by a single house or Smallness of the panmictic unit of Triatoma infestans chickencoop whenlarger numbersof bugs were an- (Hemiptera: Reduviidae). J. Med. Entomol. 35: 911Ð917. alyzed (Brenie`re et al. 1998). Such different levels of Canale, D. M., M. C. Cecere, R. Chuit, and R. E. Gu¨ rtler. spatial structuring can be caused by several factors, 2000. Peridomestic distributionof Triatoma garciabesi such as topography, landcover, human-assisted pas- and Triatoma guasayana in north-west Argentina. Med. sive dispersal, or active dispersal that may act as bar- Vet. Entomol. 14: 383Ð390. riers or corridors betweenvillages. Carcavallo, R. U., D. M. Canale, and A. Martı´nez. 1988. The fact that morphometric differences were de- Habitats de triatominos argentinos y zonas ecolo´gicas tected betweengeographically close populations donde prevalecen. Chagas 5: 8Ð17. raises the questionas to whether these differences Cecere, M. C., R. E. Gu¨ rtler, D. Canale, R. Chuit, and J. E. Cohen. 1997. The role of the peridomiciliary area inthe were genetically determined or whether they were a elimination of Triatoma infestans from rural Argentine consequence of sampling recently isolated triatomine communities. Pan. Am. J. Public Health 1: 273Ð279. populations affected by founder effects caused by Cecere, M. C., M. B. Castan˜ era, D. M. Canale, R. Chuit, and community-wide residual insecticide sprays in 1992 R. E. Gu¨ rtler. 1999 Trypanosoma cruzi infection in Tri- and selective sprays carried out in 1995Ð1998. Our atoma infestans and other triatomines: long-term effects study does not address the origin of morphometric of a control program in rural northwestern Argentina. differences in T. infestans populations. Regardless of Pan. Am. J. Public Health 5: 392Ð399. the underlying mechanisms of differentiation, the ev- Cecere, M. C., R. E. Gu¨ rtler, D. Canale, R. Chuit, and J. E. idence points to a strong structuring of T. infestans Cohen. 2002. Effects of partial housing improvement populations. Microsatellite molecular markers are ex- and insecticide spraying on the reinfestation dynamics of pected to give concluding evidence on the actual Triatoma infestans in rural northwestern Argentina. Acta extent of spatial structuring. Trop. 84: 101Ð16. Cohen, J. E., and R. E. Gurtler. 2001. Modeling household The source of reinfestations has been generally dif- ¨ transmission of American trypanosomiasis. Science 293: Þcult to determine, whether from local survivors of the 694Ð698. initial treatment or immigrants from untreated, syl- Dujardin, J. P., M. Tibayrenc, E. Venegas, L. Maldonado, vatic foci (Dujardinet al. 1996, 1998b). Our data give P. Desjeux, and F. J. Ayala. 1987. Isozyme evidence of conÞdence in the use of metric variation to tentatively lack of speciationbetweenwild anddomestic Triatoma identifythe originof reinfestationwithinavillage, infestans (Heteroptera: Reduviidae) inBolivia. J. Med. particularly when the comparisons involve contempo- Entomol. 24: 40Ð45. raneous generations of insects, or nearly so (all bugs used Dujardin, J. P., C. La Fuente, L. Cardozo, and M. Tibayrenc. inthis analysiswere collected withina 2-wk period). 1988. Dispersing behaviour of Triatoma infestans: evi- Under these circumstances, the use of size in combina- dence from a genetical study of Þeld populations in Bo- tionwith shape may improve the reclassiÞcation. livia. Mem. Inst. Oswaldo Cruz 88: 235Ð240. Dujardin, J. P., L. Cardozo, and C. Schofield. 1996. Genetic analysis of Triatoma infestans following insecticidal con- Acknowledgments trol interventions in central Bolivia. Acta Trop. 61: 263Ð 266. Special thanks to Elisa Angrisano, Ellen Dotson, Carla Dujardin, J. P., H. Bermu´ dez, and C. J. Schofield. 1997a. The Cecere, andthe EuropeanCommunityÐLatinAmericanNet- use of morphometrics in entomological surveillance of work for Research on the Biology and Control of Triatominae sylvatic foci of Triatoma infestans inBolivia. Acta Trop. 66: (ECLAT) network for support. This project was supported 145Ð153. by National Institutes of Health Research Grant R01 Dujardin, J. P., H. Bermu´ dez, C. Casini, C. J. Schofield, and TW05836 funded by the Fogarty International Center and M. Tibayrenc. 1997b. Metric differences between syl- the National Institute of Environmental Health Sciences to vatic and domestic Triatoma infestans (Hemiptera: Redu- U. K. and R.E.G., and by grants from the University of Buenos viidae) inBolivia. J. Med. Entomol.34: 544Ð551. Aires and Agencia Nacional de Promocio´nCientõ ´ÞcayTe´c- Dujardin, J. P., G. Forgues, M. Torres, E. Martinez, C. Cor- nica (Argentina) to R.E.G. J.S.-B. was supported by a schol- doba, and A. Gianella. 1998a. Morphometrics of domes- arship from Consejo Nacional de Investigaciones Cientõ´Þcas tic Pastrongylus rufotuberculaus in Bolivia. Ann. Trop. yTe´cnicas de Argentina (CONICET). R.E.G. is member of Med. Parasitol. 92: 219Ð228. CONICET ResearcherÕs Career. Dujardin, J. P., C. J. Schofield, and M. Tibayrenc. 1998b. Populationstructure of Andean Triatoma infestans: allo- References Cited zyme frequencies and their epidemiological relevance. Med. Vet. Entomol. 12: 20Ð29. Bookstein, F. L. 1990. Introduction to methods for land- Dujardin, J. P., M. Steindel, T. Cha´vez, C. Ponce, J. Moreno, mark data, pp 216Ð225. In F. J. Rohlf and F. L. Bookstein and C. J. Schofield. 1999. Changes in the sexual dimor- July 2004 SCHACHTER-BROIDE ET AL.: SPATIAL STRUCTURING OF T. infestans 649

phism of Triatominae in the transition from natural to (eds.), Advances in morphometrics NATO ASI, series A, artiÞcial habitats. Mem. Inst. Oswaldo Cruz 94: 565Ð569. life sciences. Plenum Publication, New York. Fleiss, J. L. 1981. Statistical methods for rates and propor- Rohlf, F. J. 1999a. TPSDIG, version1.18. Departmentof tions, 3rd ed. Wiley, New York. Ecology and Evolution, State University of New York, Guarneri, A. A., M. H. Pereira, and L. Diotaiuti. 2000. In- Stony Brook, NY (http://life.bio.sunysb.edu/morph/). ßuence of the blood meal source on the development of Rohlf, F. J. 1999b. TPSREGR, version1.18. Departmentof Triatoma infestans, Triatoma brasiliensis, Triatoma sor- Ecology and Evolution, State University of New York, dida, and Triatoma pseudomaculata (Heteroptera: Redu- Stony Brook, NY (http://life.bio.sunysb.edu/morph/). viidae). J. Med. Entomol. 37: 373Ð379. Rohlf, F. J. 1999c. TPSRELW, version1.18. Departmentof Gumiel, M., S. Catala´, F. Noireau, A. Rojas de Arias, A. Garcı´a, Ecology and Evolution, State University of New York, and J. P. Dujardin. 2003. Wing geometry in Triatoma Stony Brook, NY (http://life.bio.sunysb.edu/morph/). infestans (Klug) and T. melanosoma Martinez, Olmedo & Rohlf, F. J., and L. F. Marcus. 1993. A revolutioninmor- Carcavallo (Hemiptera: Reduviidae). Syst. Entomol. 28: phometrics. Trends Ecol. Evol. 8: 129Ð132. 173Ð179. Schofield, C. J., L. Diotaiuti, and J. P. Dujardin. 1999. The Gu¨ rtler, R. E., R. M. Petersen, M. C. Cecere, N. J. Schweig- process of domesticationinTriatominae.Mem. Inst. mann, R. Chuit, J. M. Gualtieri, and M. C. Wisnivesky- Oswaldo Cruz 94: 375Ð375. Colli. 1994. ChagasÕ disease in north-west Argentina: Vazquez-Prokopec, G. M., L. A. Ceballos, M. C. Cecere, risk of domestic reinfestation by Triatoma infestans after and R. E. Gu¨ rtler. 2002. Seasonal variations of micro- a single community-wide application of deltamethrin. climatic conditions in domestic and peridomestic habi- Trans. R. Soc. Trop. Med. Hyg. 88: 27Ð30. tats of Triatoma infestans in rural northwest Argentina. Jaramillo, N., D. Castillo, and M. E. Wolff. 2002. Geometric Acta Trop. 84: 229Ð238. morphometric differences between Panstrongylus Vazquez-Prokopec, G. M., L. A. Ceballos, U. Kitron, and geniculatus from Þeld and laboratory. Mem. Inst. Oswaldo R. E. Gu¨ rtler. 2004. Active dispersal of natural popula- Cruz 97: 667Ð673. tions of Triatoma infestans (Hemiptera: Reduviidae) in Landis, J. R., and G. G. Koch. 1977. The measurements of rural northwestern Argentina. J. Med. Entomol. 41: 614Ð observer agreement for categorical data. Biometrics 33: 621. 159Ð174. Villegas, J., M. D. Feliciangeli, and J. P. Dujardin. 2002. Matias, A., J. X. de la Riva, M. Torrez, and J. P. Dujardin. Wing shape divergence between Rhodnius prolixus from 2001. Rhodnius robustus in Bolivia identiÞed by wings. Cojedes (Venezuela) and Rhodnius robustus from Me´rida Mem. Inst. Oswaldo Cruz 96: 947Ð950. (Venezuela). Infect. Genet. Evol. 2: 121Ð128. Rohlf, F. J. 1996. Morphometric spaces, shape components and the effects of linear transformations, pp. 117Ð129. In L. F. Marcus, M. Corti, A. Loy, G.J.P. Naylor, and D. Slice Received 31 December 2003; accepted 22 March 2004. Schachter-Broide J., Dujardin Jean-Pierre, Kitron U., Gürtler R.E. (2004) Spatial structuring of Triatoma infestans (Hemiptera, Reduviidae): populations from Northwerstern Argentina using wing geometric morphometry Journal of Medical Entomology, 41 (4), 643-649 ISSN 0022-2585