EVO'LUTIONARY ASPECTS OF THE ECOLOGY OF ARBOVIRUSES IN BRAZILIAN AMAZONIA, SOUTI-1

AMERICA. )I

s I Nicolas DI'XjAi,I,IIN'

8 Ainélia P. A. TI<AVASSOS DA IIOSA' Pedro F. C. VASCONCIMX'

Gregorio C. SA Fii,iio?

Elisabeth S. TIWVASSOS DA ROSA? Sueli G. ROI)KI(~UIIS'

Jorge E S. TI~AVASSOS DA ROSA^

1 ORSTOM B.P. 5045, 34032 Moiitpcllicr Ccdcs 0 I Francc. 2 Instituto Evaiidro Chagas, Fundação Nacioiial de Saiidc - MS, C. P. 1128, 66090-000 Bcldm l " 5 Brazil.

42

Y

ìi

INTRODUCTION

I

k ,

#

Y

The Amazonian rcgion of Brazil is probably tlic world’s richest rescrvoir of arboviruscs. The 183 different types of arboviruses so far detcctcd in this rcgion (Tablc 1; Hervé d., 1986; Pinhciro aal., 1986; Travassos da Rosa d., 1986, 1987, 1989 and unpublishcd data), account for ncar onc third of thc world’s arboviruscs: a total of 533 (Karabatsos, 1985). Ofthcsc, I36 (83.4 %) arc cndcmic in this part of thc ncotropical zoogcograpliical rcgion.

Of iiiaiii iiitcrest in relation with thc man-made modifications of thc natural environment is tlie study of the ecological factors which allow the coexistclice of such a great nuiiibcr and variety of viruses and their coniplcs transmission cycles (Pinheiro u., 1977; Dison dal., 198 1). Tlie virological and serological data, obtained betwccn 1954 and 1992 at the Evandro Chagas Institute / National Health Foundation (Belém), will be analysed from two complemcntary points of view: (i) the definition ofthe ecological niche of each arbovirus and (ii) the ecological factors which have possibly constrained their evolution.

The ccological iiichc conccpt has bccn uscd rccciitly by Calishcr (1 994) in rclation with tlic dcfinitioii of the virus spccies; “A viriis species is a polvflie fic class qf viruses îhat constitiiîes a replicaling lineage and occupies apnrticzilnr ecological niche .‘ (emphasis by LIS). Othcr definitions have been proposed by ecologists. One which best suits tlie arboviniscs is as follows: “A niche is a multi-dimcnsiolial hypcrvolunic of resourcc ases” (Colinvaux, 1986: 3 1). In tlie case of arboviruses, each host or altcrnatively, each conipoiieiit of the hosts’ cnviromiient inay reprcsent one of thcse resource ases or variablcs, allowing for tlie quantification of tlie niche. Multifactorial data analysis methods, which havc been uscd here for tlie first time with arboviruses (to our knowledge), sccni particularly well suited to (i) the nuincrical definition of the niche of cach virus species and (ii) the study of the ecological grouping of the viruses.

The question undcrlying our prcsent work relatcs to tlie ecological factors which prevent arboviruses from multiplying randomly in all available hosts. There are probable constraints of various origins and located at various levels, from inside the ccll to the ccosystem: genetical (or physiological) (Dubois, 199 I), eco- ethological and historical (or biogcobraphical) (Barbault, 199 1) .

M . m ” AND RIETIIODS

SEIWLOGICAL AND VIROLOGICAL RIETIIODS

The crude data were reprcseiited by the isolated strains and the results of haemagglutination inhibition (HI) tests (the laboratory techniques are described in detail in Shope & Sather, 1979). Tlie HI tests liave been considered positive for a particular arbovirus if it showed a titer at least four-fold higher than any other tested antigen in the same scrological (cross-reacting) group. This “four-fold titer criterion7’ has been established either froin unpublished results of experuiients conducted on mice, rats and fowls, or from follow-up serological studies in naturally infected people and sylvatic animals. This criterion inay appear overrigorous but the main objective was to prevent the introduction of falsc positives in the data, cvcn at cost of sonie false negatives Wicn sera were availablc, HI rcsults which appearcd doubtful havc bccn chcckcd by scrolieutralizatioli test (except in the case of turtle sera which do not contain iieutralizing antibodies. APA Travassos da Rosa, unpublished data). A contingency table has been constriictcd, of which three subsets have been submitted to the treatmiits described below.

I

I

NIJMERICAL AND srmsmxi., MC I r w D s The structure of tlie data was first explored qualitatively, using presence (1) vs. absence ( O ) of each virus

in each host as characters (annexes 1 & 2). The hosts havc been grouped at the order or genushbgenus levels for vertebrates and arthropods, rcspe&ely. The subsets wverc as follows: (a) arboviruses known from arthropods only; (b) arboviruscs known from vertebrates only and (c) arboviruses known froin both types of hosts. These analysis were bascd on tlic phylogenetic method (Wilcy, 198 1; d’Udckem-Gevers, 1990; Hennig in Goujet al., 1988; Janvier in Goujct d., 1988) and have bccii conducted with the PAUP software (Swofford, 1993). For each subset, an unrootcd coiisensus tree was computed, adopting tlie “50 %-majority rule” (Margush & McMorris in Swofford, 1993).

Quantitative treatments were done on the uacoded data matrix (contingency table) or subsets of it (Annexes 4-6). Two methods of multifactorial analysis, the factorial analysis of correspondances and the hierarchical

43

ascendent classification (AnaMul and ADDAD packages, rcspectively) have provcd useful for this approach (Fénelon, 198 1; Febvay & Bonnot, 1990). In order to study thc ecological rclationships csistiiig betwecn the arboviruses more specifically, the hosts have been grouped accordin2 to thcir terrestrial/ arboreal and diurnal/ nocturnal habits (variables nuniber 32 to 37 in Anncs 1). A more detailed study of the bird-associated arboviruses, based on data about habitat and lcvel prefcrcnces of the hosts has bccn done in a separate papcr by Dégallier et. ( 1992b) and will only bc summarized bclow.

RESULTS The following sample sizes (nuniber of specinlens or pools) formed the basis of the present study:

haematophagous Diptera, more than 5 15 O00 pools; marsupials, 6,427; bats, 9,276; wild monkeys, 2,428; rodents, 18,741; edentates, 861; carnivorous, 36 1: ungulates, 3,374; birds, 12,423; reptiles, 6.052; amphibians, 1,509.

TIIE QUALITATIVE APPROAClI

NUMBER OP IIOS~S BY VIRUS

Tlie number of different spccies of hosts for each virus group (tablc 2) givcs a crudc indication of thcir host spectrum, and consequcntly, of the iiitciisity of their adaptivc radiation. It is obvious l’rom thcsc data (coniparcd with thosc in tablc 1) that tlic virus faniilics with grcatcr niumbcr OF spccics arc not ncccssarily thosc wliich were found in the greater varicty of hosts. For csamplc, thc Flaviviridac and Togaviridac with only 8 viral species each in the Brazilian Amazon region, have bcen found associated with at lcast 57 and 56 diffcrcnt hosts, respectively. This compares with 54 hosts known for the Bunyaviridae which includes 45 species. On the other hand, the Reoviridae, ayounting for 63 different viruses, arc known from only 14 differcnt spccics of hosts. Thus, the ecological diversification, in tcrnis of number of hosts involved in transmission cycles, seems to be independant from the $$teniatic divcrsity of thc viruses (= nuniber of spccies). However we will sec in later discussion thc importan& of sampling bias on this pattcrn.

b

Table 1: Numbcrs of genera, serological groups and species for each family of arboviruses present in the amazonian region of Brazil; their order of enumeration follows dccrcasiiig nuiiibcr of spccics.

Family genera groups species

Bunyaviridae Biinynvirus (a)

Phlebovirus Reoviridae Rliabdoviridae Togavi ridae Flaviviridae Coronaviridae Poxviridae Unclassified Arenaviridae (0

Herpesviridae(0 Paramyxoviridae(0

2

, .

70 45 25 63 15 8 8 1 1 12 3) 1) 1)

~

(*) with 1 Bicnyavinis-like virus included (b) with 4 ungrouped viruses included

with 3 ungrouped viruses included (d) with 3 ungrouped viruses included (e) with 1 ungrouped virus included (0 probably not arboviruses

44

i

NUMUEH OF VIRUSES I ~ Y IIOS‘I’

Table 3 shows the nuniber of diffcrcnt spccics of viruses found in each ecological typc and/or systcinatic group of hosts.

Some types of hosts seem to be more favorablc to the speciation of arboviruses than others. Aniong the arthropods, sandflies arc almost the solc hosts €or the majority of Reoviridae hiown from our region. As these viruses do not forni agglutinins in vertebrates, it is not yct possible to know if they are diversified in this respect. Due to their minutencss and tlie lack of an idciitification key for fresh females, the phlebotoniine sandflies wcrc not identified and thus, may contain niany species with various habits. On average, nocturnal mosquitoes harbor more diffcrcnt viruscs than diurnal ones. This difference rcsults mainly from the nuniber of Bunyaviridae transmitted by nocturnal mosquitoes. Among tlie vcrtcbratcs, tlic samc may be said, i . e., that tlic nocturnal oncs harbor a larger variety of viruses, duc cspccially to tlic prcdominancc of the bunyaviruscs. The Flaviviridac sccm to bc as “diurnal” than ‘Ìiocturnal” but the Togaviridae niay be more “diurnal” if wc consider tlieir vertcbratc hosts.

Nevertheless, thc data discussed in tlie two previous paragraphs are very Cnide and will be examined in more details in tlic following sections.

The phylogenetic assumptions Any phylogenetic study niay ideally need assumptions about (i) tlie transformation sequence of the states

of tlie characters, and (¡i) tlie states wvliich niay be considered as apomorphic (= dcrived or specialized) or plcsiomorphic (= anccstral or primitive).

However, in the casc of arboviruses, fcw authors hnvc ndrcsscd this qucstion. Mattingly ( 1960) and Calishcr (19S8) supposed that thc CuIm - bird cycle of niany arboviruscs inay be primitive (plcsiomorphic). I n fact, if the arboviruscs can be considered! to have originatcd in arthropods before they become adaptcd to vcrtcbratcs (Goldbach & Wellink, 1988) havc<si!ggcsted the samc for inscct viruses which adapted to plant), it would be reasonable to associate ancestry of the vcrtebratc group with diversity of arboviruscs. The phylogeny of mosquitoes is even worse hiown and cannot give any indication to solve this qucstion although culicincs were first recordcd from the Oligoccnc. Thus, the prcsciit study lias been donc, without doing my apriori polarization of tlie characters.

L

a

Results of the phylogenetic analysis Figures 1, 2 and 3 show hypothctical pliylogcnics of tlic arboviruscs whosc arthropod and vcrtcbratc, only

arthropod, or only vcrtcbratc hosts arc known, rcspcctivcly. Fifty, 34 and 29 viruses liave been included in these rcspcctivc subsets. Sisty diffcrcnt viruses are known

only from phlcbotomine sandflies (Travassos da Rosa aal., 1983, 1954; see also anncx 3) and, as such, liave been grouped under tlic denomination: “sandfly bornc”.

The first consensus tree (Figurc I ) was computed from a total of 500 trccs, each of 14 I steps. Its rcsolutioii is not perfect as it shows yct sonic polychotomies. Thus, the viruscs Acara to Ncpuyo, Bcnfica to Guajara, Bcnevidcs to Itaqui, lrituia to Tapara and Jurona nced to be studied in more dctail with rcspect to their hosts andior the possible sampling bias. Some ecological groups seem better established if we consider tlie value of the majority-rulc indcs: Moju to Una, Marituba to SLE. Turlock to Triniti, and Guaroa to Oropouchc.

Thz second consciisus trce (Figure 2), obtained from 500 most parsimonious trees, each 27 steps-long, and \\hich shows tlic hypothcticai rc1;ationships of the viruscs isolatcd only fron! liaematophagous arthropods,

and TaiassuiNyeoniyk, th:: g-oups: ,arc cither polychotomous or with poor conscnsus rates. However, despite its pol~~chotomy, the Anlniateua to %ucunli group was prcsent in more than half of the trees.

Kcmarks of the sanie order than tlie preceding can be made about tlie tree presented in Figure 3, where viruses only kn0n.n from vertebrate hosts have been considered (consensus tree obtained from 1 O2 distinct trees, each 2.6-steps-long). The relationships between the viruses from Agua Preta to Parisa and Beleni to Jatobal are poorly resolved. 011 tlic other hand, some associations appeared at greater rates: AnliangalJari; Araguari/Itaituba/Piry and Bocas to Timbó. One group seems to be fairly resolved: Urucuri to Utinga.

L appcars evcn lcss resolvcd than tlil previous one. With tlic csccption of little groups likc Buritirana/Itacaiuiias

45

Table 2: Minimum numbers of different specics of vertebrate and arthropod hosts for each family and/or genus of arboviruses prcsent in the amazonian rcgion of Brazil; thcir order of ciiumcration follows decreasing total number of species.

Virus/Min. Nr. of different species

Family/Genus Vertebrates Arthropods Total

17 56 Flaviviridae 37 20 57 Togavi ridae 39 . Bun yaviridae

Bunynviriis Phlebovirits

Reovi r idae Rhabdoviridae Coronavi r idae Poxviridac Unclassified

15 9 2 3 1 1 -

39 3

12 6

2

54 12 14 9 1 1 2

TIIE QUANTITATIVE APPROACH

The quantitative analysis have becn donc with tlic viruses whose cycles arc bcttcr known, i .e. which havc bcen found in both vertebrate and arthropod hosts. A spccial analysis has bceii donc prcviously with thc viruscs known from birds (Degallicr d., 1992b).

The multidimensional ecological niche Tlie first four factors which resulted froni the corrcspondance analysis, account for 91.9 96 of tlic total

inertia of the dot cloud. Tlie representation of the variables is best along the factors 1,2, 3 and 4 (Figure 4). The fomier seems to be oriented from diumal to nocturnal canopy vcrtebrates, from diurnal to nocturiial ground vertebrates and from diunial to noctunial mosquitoes. The third factor is less discriminant for these saine ecological variables. The fourth factor separates wcll thc mosquito-relatcd variablcs (diuriiaynocturIla1). However, in order to interpret graphically the relationships between these variablcs and the viruses, we havc to consider the relative contribution of each of the latter to the same factors. Figurcs 5 and 6 show simultaneously tlic projcctions of thc variablcs and viniscs on thc plaics fornicd by thc factors 2 and 3, and 2 and 4, rcspcctivcly. Only the elements which contributed for at least 20 % of thc total inertia of at least one of the two factors have

Table 3: The hosts of sylvatic arboviruses in Brazilian Aniazonia, grouped according to their ecological characteristics. For each type of host is indicated the number of virus species associated with.

Mosquitoes Sandflies Midgcs Ticks Vertebrates Arbovirus Nocturnal Diuqpai Diurnal Nocturnal Families i, b terr. canopy terr. canopy

I . Flaviviridae Togaviridae Bunyaviridae Reoviridae Rhabdoviridae Coronaviridae Poxvi ridae Unclassified

4 6

37 8 3 O O 4

4 6

19 5 3 O O 2

O O 2

51 3 O O 2

6 6

13 O 4 1 O O

6 7

17 1 2 O O O

6 5

21 2 1 O 1 O

6 5

17 O 1 O O 3

~~ ~~

Total 63 40 69 4 3 36 40 45 38

Y

46

L

i

been retained. For clarity, the labcls of tiic elcnicnts which coiitributcd significantly to the two factors are in boldface and those \dich contributed only to tlie ordinate factor are in italics,

The following associations may be deduced from tlie Figurcs 5 and 6: - BEN, ICO, BSB, MOJU, ACA, BSQ, CAR, CAR-likc, CATU, MUC, NEP, CAP, GMA, BVS, ITQ,

MUR, ORI, BIM, GJA, GAM, AURA, Trombetas: nocturnal tcrrestrial vertcbrates/noctunial mosquitoes; - UNA, MCA, ILH, TNT, KRI, MAG (+ KWA-like, ANU I? ) : diumal terrestrial vertebrates/diurnal

mosquitoes; - YF, GRO, MAY, SLE, TCM, TUR, WEE, ORO, TCM, JUR: diurnal canopy vertebrates/diurnal

mosquitoes; Thc viruscs EEE and PAC-likc scciii to localizc at intcrnicdiatc positions, bctwccn diurnal and nocturnal

and bctwecn canopy and ground-dwcliing hosts. I n tlic following scction, w c will ‘-80 back” to thc wcodcd data to csaminc thc csact ecological coniponcnts of cach of thcsc groups.

I

T l I E ECOLOCICAI~ (NURIEIIICAL,) CI,/\SSIIWAIION

Thc ecological groups obtained,,by the ascendclit hicrarchical classification algorithm arc sliowi in Figures 7 and 8 for the viruses and thcir serological groups, respectively. The proportions representing the different types of hosts have been put on thc ordinate scalc. Scrological group A and buiiyaviruscs may contain sonic viruses which do not cross react inside thesc groups, thus they have becn indicated by an asterisk.

Aciiru Ncpupo Ikrifica I%l,SIl I%IISIl I Caranaru like

1 O0 83

95 7 “C

1 O0

1 O0

1 O0

1 O0 1 O0

1 O0 !.W 1 O0 L

Guajhra fkr1cvitlcs Capim ilaqui Mwju Orihoca Guallla Magu;tri u na hhri Lu bil Apcu Murulucu Ananindcua LISI . Et]. 13nc. c;ltu Ciiriiparu It;lporarlga Bussuquara St. Louis Enc. Bi miti Monle Dourado lriluia Pacui Santiircm Tapam Aura Kairi Tronibclas J II rollii Kwalta like l’ncoril l ike ‘ru r I oc k WCSI. Eq. EIIC. MaCJUJ Pixuna Triniti Gaciihoa Guama Tacaiu1rra Iltaus M;iy:iro Muciunho Ycllow Fcvcr Icoaraci Oropouchc

Figure 1. Unrooted, 50 % - majority rule consensus tree of the arboviruses known from both vertebrate and arthropod hosts in Brazilian Aniazonia, 1954-1992. The value at base of each group is the proportion (%) of the 500 most parsimonious trees which show thc group.

47 I

Some reinarks which reniain of preliminary naturc can be drawn from this analysis: ( i ) thcrc is a continuum (or gradient) from “diurnal” to “noctur~ial” viruses and from cbarboreal*’ to “terrestrial”; (ii) some groups like the GAM to WEE, TUR to ACA and Trombetas to GRO were mostly found in “diunial canopy vcrtebratcs” aiid “nocturnal mosquitoes”; (iii) the PIX/ TNTI MAG group is predoniinantly “diurnal” and “tcrrestrial”, opposing to the almost strictly “canopy-liking” viniscs MAY, ILH, YF and UNA; (iv) thc ccology of somc groups of viruscs like ORO/ ANU, Tapara to KRI and IC0 weds more inforination to bc gathered, cspecially about their vectors; (v) in fig. 8, we see that with some csccptions, each serological group lias but one virus in one ecological group; we need finer definitions of the nichcs of 2 I viruses pertaining to A, B, BUN, C, CAP, CGL and GMA serological groups.

74 Melao Scrra do Navio Trocara

.

60

48

Galibi - lcri Purus

72 Ilupi ranga

The bird-associated viruses We will summarize below the main results of this spccific study, alrcady published by Dégallier a.

(1 9924. The subset including thc bird-associated arboviruscs accountcd for 30 different arboviruscs. An ascciideiit hierarchical classification has bceii obtained, using as ecological variables five types of vegetation, two of which (igapo or inuiidatcd forest and “terra fimic” forest) lias becn subdividcd in two and five strata,

1 O0

Agua Preta

Caj,v,ci ras

Juruaca

Mapueri Mojui dos C‘anipos

Parixa

ltaituha

85 -

Bclem

Candiru

7 Cacipacorc

L Utinga

A

Figure 3. Unrooted, 50 % - majority rule consensus trec of the arboviruses luiowvn only from vertebrate hosts in Brazilian Amazoiiia, 1954- 1992. The value at base of each group is the proportion (YO) of the 102 most parsimonious trees wliicli show the group.

respectively. A gradient (contiiuum) has becn observed between bird species which prefer secondary vegetation or forest (= “capoeira”) and, those which are found mainly in priiiiary forest.

CDU virus lias been found iiiaiiily in canopy birds which are dwelling mainly (50 %) above 15 and 30 m. This virus, as for sonic others (CPC, MAY, ILH, TCM), lias an important sccoiidary forest component (25 %). The birds which are hosts of UTI, KWA, GAM and IC0 viruscs arc spccics living exclusively in the “terra-firme” forest.

49

EEE virus has been “found” in all but one rare type of vegetation or strata (forest on sandy ground) and is considered as ecologically versatile. This may be linked to a great potential of this virus to colonize new niches, including in urban environnient (strains isolated from mosquitoes in Fortaleza, Ceará, Brazil).

I 1 100%

80%

60%

40%

20%

0% Factor 1 Factor 2 Factor 3 Factor 4

Canopy diurnal Vert.

Ground diurnal Vert.

Canopy noct. vert.

Ground noct. Vert.

Diurnal mosquitoes

Noct. mosquitoes

Ticks

Culicoid midges

@¡ Phlebotomine sandflies _______-_ -

Figure 4. Contribution (% of explained incrtia) of tlic ninc ccological variables to tlic factors 1-4 of tlic correspondance aiialysis of the arboviruses known from both arthropod and vertebrate hosts in Brazilian Anazonia, 1954- 1992.

The viruses which niay be considered tlie more prone to infect huntan people in rural places are BSQ, GMA, MUC, MUR and APEU because they have been found in birds which arc living in the secondary (or degradated) enviromnents nixed with cultivated areas as is often the case in Aniazonia. They are actually fairly prevalent in human sera.

r

DISC~JSSION The ecological niche concept for arboviruses

Previous works have already dcscribcd the probablc sylvatic cycles of many amazonian arboviruses (Woodall, 1967; Dégallier, 1982; Hervé d . , 1986). Howcvcr, tlic grouping of ccologically siniilar viruscs was made mainly after the number and naturc of the hosts, i . e. tlic rclativc “complcsity” of the cyclcs. Wc have reexamined the sanie data in a phylogcnctic pcrspcctivc. In a quantitativc ccological study of tlic viruses

I

m

8 u. li

GAAß Trombetas AURA. ” ’ ’

CATU Muc

50

pertaining to the group C of Bzinymirzis, Woodall (1979) lias sliow~i a marked niclic separation between canopy- and ground level-transmitted viruses and, when two viruses appeared to share tlie same niche, that the vectors were distinct species. Thus, for establishing themsclves in a locally stable equilibrium, related viruses cannot share tlie same arthropod and/or vcrtebrate hosts. It niay even be said that the less they are serologically related, tlie iiiore they can share tlie same ecological niche.

Tlie complexity of the amazonian environment, cscmplified by an cstremc divcrsity of vcrtcbratc and arthropod spccics, and conscqucntly of nichcs, liavc favourcd thc diversification and sympatry OF many

=

UNA

MuGm m(“ ... GMÄ..cATÜ. MR ~ I T Q . B V S ... CAP

MAY

... !?!M.CAR-like PAC-like =..q$@o B . . . . . .

arboviruses of the same group. As tlie prcsent study has show, the presence of two or more different viruses of the saine serological group in what has bccii charactcrizcd as one nichc may bc duc to the coarseness of tlic ecological variables which have bcen considcrcd. As many ccologists havc shown, temporal and spatial variablcs may also be includcd in the multidimensional definition of the nichcs. Tlie isolatioiis of strains andor serological coilversions in sentinel animals should be iiiterpreted in the futurc for tlic “temporal” characterization of the niches.

Excepted for some viruses which are ecologically very distinct (PAC-likc, CDU), tlicre is no clear-cut separation between one group and the nest along thc “ecological transccts” defincd either by the preferred vegetatioli types and strata, or by the habits of thc hosts. This may mean that, with similar historical (= biogeographical + genetical) constraints, tlie arbovirus population in a dcfincd coniinuni ty forms a dynamic equilibrium. Subsets of this population may sliarc tlic same niche, at lcast at thc two lcvcls studicd hcrc, and each nichc is separated from the others by ecological constraints acting on thc host populatioiis. I-lowcvcr. two very differclit typcs of pcrturbations can modify this cquilibrium. Whcn soinc fluctuations arc going on in host populations, as occurs seasonally for non-iiiuiiunc hosts, only tlic arbovirus transmission icvcls arc affcctcd. On the contrary, whcn the host populations are permanently modified, the cquilibrium of arbovirus populations need to shift to a new statc. Eventually, new niches may appcar nhich would be filled aftcr a short time. These inay be colonized by new genotypes, obtained by recombination, reassortmcnt or introduction from adjacent communities (Calisher, 1988). These relatively rapid adjustments of cquilibrium liavc been noted in the case of the important perturbations iiiduccd by the filling of a dam rcscrvoir, where “ncw” arboviruscs appeared in tlie area, either as exogenous material or as autoclitonous speciation.

The phylogenetic study of arboviruses, based on the viruses-hosts associations is yet vciy tentative because of the lack of phylogenetic classifications of either group (Eldridgc, 1990). This author lias looked for sonic evidcncc of a host-parasitc coevolution in tlie cast of tlic California scrogroup viruscs What \ve havc dcfincd as historical constraints arc doubtlessly related with sonic effects of coevolution but it remains difficult to distinguish these from tlie constraints arising from virus competition aftcr horizontal traiisfcrs (Tabachnick, 199 1). These hypothesis need to be evaluatcd by molecular biologists.

L

t

51

ANU ORO

ILH MAY UNA YI:

CAP EEE GAMSLE WEE

ACA APEU ITQ MTB UKI TCMTUR

AURA GIA GRO KWA.likt! Trombetas

MAG PIX TNT

BEN BIM BSB BVS CAR- like RI NR KRI MCA

'lapan MD0 NEP KA-like STM

CAR BSQ OMA ITP MOJLl MUC

CAIU MUR

IC0

PAC

Figure 7. Ecological composition of the groups of viruses, obtained by the hierarchical ascendent classification analysis of the ecological data, available,for the agents known froin both arthropod and vertebrate hosts in Brazilian

U

hnazonia, 1954- 1992. 'p THE EVOLUTIONARY ASPECTS OF THE ECOLOGY OF ARBOVIRUSES

What happened with the viruses for which some evolutionary hypothesis. are available, based on protein sequencing ? Levinson d. (1990) fimished some interesting hypothesis which will be discussed from an

52

ri,

ecological point of vicw. MAY (with UNA, not studied by thcsc authors), EEEWEE, AURA and MUC (member of VEE coniplex) seem to have diverged in four different directions (f ig. 7). As Halm d. (1988) havc slio\vn, WEE virus arosc probably as a rccombinant bctwccn EEE and anothcr A Iphovi~us which may bc AURA. EEE sccnis to be a vcry vcrsatilc virus, cspccially in birds (fig. 9) whcrcas AURA lias been found only in monkeys. The ccological link betwccn the two may be the Melonoconion and Och/eroro/zw subgenera of Culex and Aedes mosquitoes, respectively.

E a h O0 O !2 O O O as 6 6 8 as as

i

E il

3 e P

L I

Figure 8. Graphic representation of the saine data as in fig. 7, but \villi the serological groups of viruses in place of their actual names; when more than one virus is included in a group, the nuinber of virus species is shown in parenthesis; asterisks refer to ungrouped viruses, i. e. virubs which do not cross react serologically inside tlieir group.

53 b

CONCLUSIONS Tlie present study was not dctailed enough to csplain why sonie viruses of the same serological group

coexist and othcr do not, but it allowed us to define tlic main ecological groups. Each of these groups further needs to be studied separately as it has been done with the bird-borne arboviruses.

Despite the great number of tested pools between 1954 and 1992, serological and virological data are lacking either for viruses which do not forni agglutinins, or about potential hosts which are difficult to collect. It has been seen that less than one third (28.9 %or 50/173) of the viruses known from sylvatic hosts have been found in both vertcbrate and arthropod hosts, 17.7 % (29/173) are knowii only from vcrtcbratc hosts and 54.3 % (94/173) only from arthropods. Among the lattcr, 63.8 O/o (60/94) are known only from phlebotominc sandflies and represent probably a very complex ecological system, even if their vertebrate hosts are vcry diverse.

Sampling bias may account for some distortions in the quantitative delineations of the niches. For example, ground dwelling rodents, marsupials and birds are much casier to trap than canopy frequenting hosts and among them those which cannot be attractcd by any type of baited trap. Thus, many species are poorly known, not only for the viruses they may harbour but also for their bioecology. An important ecological “axis” which has been yet neglected is the time or seasonal one. It is quite conceivable that sonie hosts may harbour different viruses of the sanie serological group at different times of tlic year. This niay be especially tlie case with bunyaviruses whose antibodies are not life-long lasting (Shope &., 1967). I n future studies, more precise serological tests may allow us to define each virus-host association better.

Nevertheless, the methodology followed in the above quantitative analysis revealed itself adequate to tlie study ofthe multidiniensional niche concept of arboviruses, and it will also be a usefùl tool to make prcdictions about the evolution of the arboviruses in response to modifications of the environment.

Acknowledgments We wish to thank all staff members, both from the Evandro Chagas Institute and other Institutions, for

their past and present participation in the field and laboratory work which has made this paper possible. The present work has beneficiated of the unvaluable financial and/or logistic help of Eletronorte (Eletrobras), SUDAM (Polos Agropecuários da Aniazoiia), CNPq, ORSTOM and Foundation SESP (now National Health Foundation, Ministry of Health). Thanks are also due to Dr. Brian H. Kay (CSIRO, Brisbane) who has kindly reviewed tlie manuscript.

54

1

Y

REFERENCES BARBAULT R - 1991 - Systématique et écologie: vers un renouveau de l’histoire naturelle - Le point de vue

d’un écologiste. Biosystemna, SociètP.Pançaise de Systéniatiqiie, Paris: NK 6 (Systénialiqiie et Ecologie) 1- 52.

CALISHER CH - 1988 - Evolutionary significance of the taxonomic data regarding buiyaviruses of the family Bunyaviridae. Intervirology 29: 268-276.

CALISHER CH - 1994 - Arbovirus evolution: 443-480. I n : ~’Virológica 91 II Sinipdsio Infernacional sobre Arboviriis dos Tropicos e Febres Hemorrdgicas. 17 a 23 de Novembro de 199 1 CENTUR (Fundação Cultural Tancredo Neves) Beldm - Pará - Brasil’’ (Travassos da Rosa APA & Isliak R, ed.) instituto Evandro Chagas - UFPA - Sociedade Brasileira de Virologia, 683 p.

COLINVAUX P - 1986 - ‘LEcolog)~”, Jolui Wiley & Sons: 707 p.

D’UDEKEM-GEVERS M - 1990 - “L‘analyse cladistique: problème et solutions hcuristiqucs informatisées”,

DÉGALLIER N - 1982 - “Les arboviriw selvatiqiies en Giijane.fiançaise et leiirs vectezirs”, Third Cycle

DÉGALLIER N, TRAVASSOS DE ROSA APA, HERVÉ JP, VASCONCELOS PFC, TRAVASSOS DA

relation with the construction of dams in Brazilian Amazonia. Ciência e Cziltzira (Journal qf the Brazilian Assacialioi~.fbr (he Advancenient ofScìence) 44 (213): 124- 135.

DÉGALLIER N, TRAVASSOS DE ROSA APA, SILVA JMC, RODRIGUES SG, VASCONCELOS PFC, TRAVASSOS DA ROSA JFS, SILVA GP, SILVA RP - 1992b - As aves como hospedeiras de arbovirus na Amazônia brasileira. Bol Miis Paraense Eniilio Goeldi sér Zoo1 8( 1): 69- 1 1 1.

Biosystenia No. 6, Socièlé.flançaise de S’istématique, Paris, p 114.

Thesis, Pierrc & Marie Curie Univ., Paris VI, France, ORSTOM multigr. p 42, anneses 1-5.

h ROSA JFS, SA FILHO GC, PINHEIRO FP - 1992a - Modifications of arboviruses transmission in

DIXON KE, LLEWELLYN CH, TRAVASSOS DA ROSA APA, TRAVASSOS DA ROSA JF - 1981 - Programa iiiultidiscipliiiario de vigilancia de las enfernicdadcs infecciosas en zonas colindantes con la carretera transamazonica en Brasil. II. Epidcmiologia de las infecciones por arbovirus. Bol qf sanit Panani 91(3): 200 - 18.

DUBOIS A - 1991 - Systéniatique et écologie: le point de vue d’un systématicien. Biosyster?ia, Société

ELDRIDGE BF - 1990 - Evolutionary relationships among California scrogroup viruses (Bunyaviridae) and

FEBVAY G, BONNOT G - 1990 - AnaMul: ensemble de logiciels d’analyses niultidiiiieiisionnelles (ACP,

FÉNELON JP - 198 1 - “Qu ’est-ce que 1 ‘analyse des données? ”, Lefonen, Paris, p 3 11.

GOLDBACH R, WELLINK J (1988) Evolution of plus-strand RNA viruses. Intervirology 29: 260-267.

GOUJET D, MATILE L, JANVIER P, HUGOT J-P - 1988 - “Systématique cladistique. Quelques testes fondamentaus. Glossaire”, Biosystema 2. DeitxiPnie édition révisée et augmentée, Société Française de Systéniatiqiie, Paris, p 195.

HAHN CS, LUSTIG S, STRAUSS EG, STRAUSS JH - 1988 - Western equine encephalitis virus is a recombinant virus. Proc natl Acad Sci USA 85: 5997-600 I.

.fiançaise de S’isténintiqiie, Paris, No. 6 (Systéniatiqiie et Ecologie): 7 1 - 106.

Aedes mosquitoes (Diptera: Culicidae). J nied Elit 27(5): 738-749.

AFD, FD et AFC) sur Macintosh@. Cal? . Tecl?n. INRA 24: 7 1-74.

HERVÉ J-P, DÉGALLIER N, TRAVASSOS DA ROSA APA, PINHEIRO FP, SÁ, FILHO GC - 1986 - Arboviroses. Aspectos ecológicos. In : “Institicto Evandro Chagas 1936 - 1986, 50 anos de contribuiçclb cis ciências biológicas e ci niedicina tropical vol. l”, Fundação SESP, Belém. p 409 - 37.

KARABATSOS N - 1985 - “International Catalogue of Arbovirus incliiding certain other Viruses of Vertebratesy’, 3rd ed, American Society of Tropical Medicine and Hygiene, Sai Antonio, Texas, p 1147.

55

LEVINSON RS, STRAUSS JH, STRAUSS EG - 1990 - Complete sequence of the genomic RNA of O’Nyong- nyong virus and its use in the construction of Alphavirus phylogenetic trees. Mrology 175: 110-123.

MATTINGLY PF - 1960 - Symposium on the evolution of arborvirus diseases. II. Ecological aspects of the evolution of mosquito-borne virus diseascs. Trans R Soc IropMed Hyg 54(2): 97-1 12.

8

PINHEIRO FP, BENSABATH G, TRAVASSOS DA ROSA APA, LAINSON R, SHAW JJ, WARD R, FRAIHA H, MORAES MAP, GUEIROS ZM, LINS ZC, MENDES R - 1977 - Public health hazards among workers along the Trans-Amazon Highway. J occiip Med IY(7): 490 - 7.

PINHEIRO FP, TRAVASSOS DA ROSA APA, FREITAS RB, TRAVASSOS DA ROSA JFS, VASCONCELOS PFC - 1986 - Arboviroses. Aspectos clínico-epidemiológicos. In: “Institzito Evandro Chagas 1936 - 1986, 50 anos de contribuiqiio Lis ciências biológicas e 13 medicina tropical vol.l”, Fundação SESP, Belém. p 375 - 408.

SHOPE RE, ANDRADE AHP, BENSABATH G - 1967 - The serological response of animals to virus infection in Utiiiga forest, Belém, Brazil. In: “Atas do Simposio sbhre a Biota Amazônica vol 6 (Patologia)”: 225-230.

SHOPE RE, SATHER GE - 1979 - Arboviruses: 767-8 14. In : “Diagnostic procediires.fÒr viral, rickel¿sial and chlamydial infèctions” (Lcnnette EH, and Schmidt NJ, ed.) 5th cd., American Public Health Association Washington, D.C., p i-ss + 1138.

SWOFFORD DL - 1993 - “PAUP:S~h)~logenetic Analysis Using Parsimony, Ersion 3. 1. 1 Marc11 1993”, Computer program distributed by the Illinois Natural History Survey, Champaign, Illinois, p vi f 257.

TABACHNICK WJ - 1991 - “To thesditors”. Jnied Ent 28(3): 297-298.

TRAVASSOS DA ROSA APA, DÉGALLIER N, HERVÉ JP, SA FILHO GC - 1987 - La recherche sur les arbovirus en Amazonie. In 1 “Connaissance d i i milieii amazonien. Actes du seminaire I 5 et I 6 octobre 1985 Paris”, coll. Colloqucs ct Sdminaircs ORSTOM, p 223 - 47.

TRAVASSOS DA ROSA APA, SHOPE RE, PINHEIRO FP, TRAVASSOS DA ROSA JFS, VASCONCELOS PFC, HERVÉ JP & DÉGALLIER N - 1989 -Arbovirus research in the brazilian Amazon. In: “Arhovirzis Resenrch in Aiistmlia. Proc. Flfih Syniposiiim, AiigiisI 28 - September I , 1989, Brisbane Aiislralin” (Uren MF, Blok J, and Manderson LH, cd.) CSIRO Tropical Animal Sciences Brisbane. p sviii + 393.

TRAVASSOS DA ROSA APA, SHOPE RE, TRAVASSOS DA ROSA JFS, NAKAUTH C, VASCONCELOS PFC - 1986 - Arboviroses. Aspectos virológicos. In : “Institiit0 Evandro Chagas, 50 anos de contribiiiço íìs ciencias biolhgicas e 21 medicina iropical vol. l”, Fundação SESP, Belém. p 365 - 73.

TRAVASSOS DA ROSA APA, TESH RJ3, PINHEIRO FP, TRAVASSOS DA ROSA JFS, PETERSON NE - 1983 - Cliaractcrization of ciglit ncw Plilcbotomus fcvcr serogroup arboviruscs (ì~urt~~ovirirlie: Phlehovirus) from tlic Amazon rcgion of Brazil. Au7 J irop Mec/ Hyg 32 (5): 1164-7 1 .

r

TRAVASSOS DA ROSA APA, TESH RE, PINHEIRO FP, TRAVASSOS DA ROSA JFS, PERALTA PH & KNUDSON DL - 1984 - Cliaraqtcrization of flic Changuinola serogroup viruscs (Reoviriílíie: Orhivirus). Intervirology 21: 38-49. I’.* S.?

WILEY EO - 198 1 - “Phylogene/ics: ,/(:e ilieory andprac/ice ofj>hylogenelic sys~emalics”, Ed. John Wilcy

WOODALL JP - 1967 - Virus rcsearch in Amazonia. In: “A~ns do Simposio siibre n Hioln Aimzbnica vol 6

& Sons, Ncw York, p s v + 439.

(Patologia)“: 3 1-63. G

WOODALL JP - 1979 - Transmission of group C arboviruscs (~3ur~~~íiviri i l~ie). In: “Arc/ic and tropical Arhoviriises”: 123- 138.

56

i(

r

,”

I

Annex 1. Types of hosts used as variables in qualitative and quantitative ecological analysis of the sylvatic arboviruses found in Brazilian Aniazonia, 1954- 1992.

Anopheles (Nyssorhynchzis) spp. Anopheles (Anoplieles) spp. Anopheles (Stethoniyia) spp. Ciilex spp. (not Czilexhor Melanoconion)

Annex 2. Data matrix used in the qualitative aiialysis of the sylvatic arboviruses found in Brazilian Amazonia, 1954 - 1992; The three successive subsets correspond to viruses known from: arthropod and vertebrate hosts; only arthropods and ody vertebrates; the colunuis represent the variables 1-3 1 listed in annex 1; the “Phlc” (= sandfly-bornc) viruscs arc listcd in thc anncs 3.

II.

ACA ANU APEU AURA BEN BIM BSB

BVS CAP CAR CAN CATU EEE GAM GJA GMA

IC0 ILH IR1 ITP

JUR KRI KWAl MAG MAY MCA MDO MOJU MTB MUC MUR NEP ORI

BSQ

GRO 1.

ITQ

0001000000000000001100011000000 O00 1010001000 10000 10000000 100 10 0001010100000000001010000010010 0000010100000000000000000000010 0001000000000000000000010000000 0000010000000000000000010000000 0001000000000000010000010000000 O001 1 100010 10000 1 1 1 101 o 110 10 1 1 o 000l0l00000000000010000l0000000 O00 1 1 1000000000000 10000 10000000 O00 1 I1 O000 1 O0001 1 o 10 1 1 o1 O0 11 1 1 o

:, O00 10 100000000000000000 10000000 0011110001000010001010010000110 O00 I l 1 0 10 10000000 1 100001 O0 1 O0 1 1 1000001000000000000000000010000 0001000000010000000000010000000 000101 o 10 1 1 1 10000010000 1 O0 10 1 o 1 10000000000000000 100000000 1 1 1 1 o 000000000000 100010 1 I1 11 I1010 1 I1 0000000 1 1 o 1 O000 I 1000000 I 1 0 I 1 1 1 I 000000000000 10000000000 10000000 0001010001000000101010011011111 O00 1 o1 000000000000 1000 1 10000000 00000000100000000000000000l0000 0000000100110000000000000000010 0000100000000000000000000010000 10100101 101 10000001000010000001 00000000 I O0 100 I o 1 I I 1 100 1 1 o 1 1 I I O 00000000000 100000000000 1 10 I0000 000000000000 1000000000 I00000000 00010100010110000010l00l0000000 0001010000000000001000000010000 O00 1 o 10 1 1 1 o 1000000 10 1 1 o I 10 1 11 1 o 00010100000100100010100100100l0 0001000000000000001000010000010 00010101011100000010000l0000000

57

ORO PAC PCAl PIX SLE STM Tapa TCM TNT Troni TUR UNA WEE YF ANH AP ARA Belt BLM BOC BUJ Caja CDU CHO COT1 CPC CUI ITA Jacu JAN Jato Juru MCO MDC MOR MPR Pari PIRY SM TBT

URU UTI ACD ARU Arum Breu Buri CAN

TIM.

O000 100 100000 10001 1 1 1 1 o 1 10 1 o 1 1 o 0000000000001000000000010000000 0000100000000000000000000010000 001000000001000000101001l010100 O001 110 10 10 10000 10 10 10 11 10 10 1 10 0000000000001000000000010000000 0000000000001000001000010000000 10000000 100000000000000 1 I o 1 1 1 1 o 1010000000010000000000011010000 1000000000000000000000000000001 O000 1 1000000000000000000 10 100 1 o 00110001111100000000l0000010000 0000010100000000100000001010101 0000000110010000101011010000110 0000000000000000000010000000000 0000000000000000000000000000100 0000000000000000001000000000000 0000000000000000000000010000000 0000000000000000000000000010000 0000000000000000010000000000000 0000000000000000001010011000011 0000000000000000000000000000100 0000000000000000000000111010110 000000000000000 1 0 10000000000000 0000000000000000000000010000000 0000000000000001101010011010110 0000000000000000000000000100000 0000000000000000001000000000000 0000000000000000000000001000000 00000000000000000000 10000000000 0000000000000000000000000001000

’ 0000000000000000000000000000l00 0000000000000000010000000000000 0000000000000000000000000000 1 00 00000000000000000000000l0000001 0000000000000000000000000000100 0000000000000000000000000000 1 O0 0000000000000000001000000000000 0000000000000000010000000000000 0000000000000000000000010000000 00000000000000000 10000000000000 0000000000000000000000011000100 00000000000000001000l00100100l0 O000 10000000000000000000000000~ O000 1 I 0000000000000000000000000 1000000000000000000000000000000 1000000000000000000000000000000 0000000000000 100000000000000000 0000000000101000000000000000000

58

Cara Coda Gali IACO IERI Iopa Itac ITU JAC LUK Majo MEL MIR MQO MUCU Prka Prni PUR SDN SOR Taia TME Trac Troc Tuci TUCU W O XIB Phle

1000000000000000000000000000000 0000000010000000000000000000000 0000000000100000000000000000000 00000000000 10000000000000000000 0000000000 100000000000000000000 O000 100000000000000000000000000 0000000000000 100000000000000000 0000000100100000000000000000000 0000010000000000000000000000000 1010000100000000000000000000000 0000000100000000000000000000000 0000000100000000000000000000000 0000010100100000000000000000000 0000010001010000000000000000000 1000000000000000000000000000000 O000 100000000000000000000000000 0000100000000000000000000000000 0000000000100000000000000000000 0000000100000000000000000000000 0000000000010000000000000000000 0111000000010000000000000000000 0010000000000000000000000000000 1000000000000000000000000000000 0000000100000000000000000000000 1000000000000000000000000000000 10 1 o 11 o 1 1 1 1 10000000000000000000 0100000000010000000000000000000 00000000000 1 0000000000000000000 0000000000001000000000000000000

59

Annex 3. Probable new arbovirus types, isolated only froni phlebotomine sandflies in the Brazilian hiazonia, 1954 - 1992; NC = not classified; NG = not grouped.

Genus Antigenic group Virus name Prototype Changuinola Acat inga AR 482250 Orbiviriis

Acurene AR 446985 Orbiviriis Changuinola Almeirini AR 389709 Orbiviriis Cliangu iliola

Changuinola Altamira AR 264217 Orbiviriis AR 40798 1 Phleboviriis Phebotomus Ambe Changu inola

Anapu AU 4960 I4 Orbivirils Orhiviriis Changuinola Aracai AR 425269

AR 4288 12 Orbivirirs Cliangu inola Aratau AR 505 172 Orbiviriis Cliangu iliola Arawete

Ariquenies AR 485678 Phleboviriis Phlebotomus AR 4288 15 Orbivirtts Chaiigu inola Aruaiia AR 482249 Ovbivirtis Cliangu inola Assuriiiis

Bacajai AR 482267 Orbiviriis Changu inola Bacuri AR 496008 Orbiviriis Cliangu iiio la Balbina AR 478620 Orbivirils Changu iiiola

AR 433317 Orbivinw Chang u i no I a Canoa1 AR 411391 ksict iloviriis v. s. v. Carajas

Catete AR 495605 Orbivirils Cliaiiguinola Coari AR 433343 Orbivirtrs Clianguinola Cupixi AR 502545 Orbiviriis Changuinola Gorotire AR 48225 1 Orbivirils Cliangu iiiola Gurupi Inhangapi AR 177325 ksiciiloviriis NG Ipixaia AU 490496 Orhiviriis Cliangu inola Iriri AR 4080.95 NC NG Iruana AU 496021 Orbivirils Cliangu inola Itaboca AR 496034 Orhiviriis Changu inola Janianxi AR 243090 Orhiviriis Cliaiigu inola Jandia AR 440489 Orb ivirits Clianguinola Jatuarana AR 440497 Orbiviriis Changuinola Joa AR 371637 I%leboviriis Phlebotomus Jutai AR 397374 Orbivirtrs Changuinola Kararao AR 447024 Orbivirus Cliangu inola Maraba Munguba AR 389707 Plilcboviriis Plilcbotomus Oriximina AR 385309 I’hleboviriis Phlcbotomus Oureni AR 4 1067 Orhiviriis Changuinoln Pacaja AR 440503 Orbivinrs Changu iliola Papura AR 450572 NC NG Parauapebas AR 415962 Orb iviri IS Cliangu inoh Paru AR 397370 Orbivirus Changuinola Pependana AR 440504 Orbiviriis Changuinola Pindobai AR 482675 Orhivirus C haiigu i no la Piratuba AR 47878 1 Orbivirils Changu inola Saraca AR 385278 Orbiviriis Changu iiiola Serra Norte AR 49,8935 Orhivirils Changuinola Surubim AR 440507 Orhiviriis Cliangu inola Tapirope AR 434080 Orhiviriis Changuinola Tekupeu AR 505(169 Orbivirus Chaiigu iiiola Timbozal AR 44054 1 Orbiviriis Cliangu inola Tocantins AR 486776 Orbiviriis Chaiigu iliola Tocam AR 505 170 Orbiviriis Changuinola Tuere AR 484704 Orbiviriis Changuinola Tumucumaque AR 397956 Orbivirta Changu iiiola Turuna AU 352492 IWebo virils Phlebotomus Uatuma AR 478626 Orbiviriis Changuinola Uriuraiia AR 479776 NC NG Uxituba AR 452652 Orbiviriis Changuinola Xaraira AR 490492 Orbiviriis Changu iliola Xiwanga AR 505171 Orbivirits Changu iriola

AR 35646 Orbioiriis Cliangu inola

AR 4 11459 fisiciiloviriis v. s. v.

60

b

” ”

f

An overview of Arbovirology in Brazil and

neighbouring countries.

Edited by: L ,.

Anzélia l? A. Travassos da Rosa Pedro I;: C. Vasconcelos

Jorge R S. Travassos da Rosa

BELÉM INSTITUTO EVANDRO CHAGAS

1998

. O 1998 by the Evandro Chagas Institute

Published by the Evandro Chagas Institute P.O. Box 1128 Belém, Pará, Brazil CEP 66090.000 0055 (091) 226-5262

All rights reserved. Printed in Brazil

This book was printed by Gráfica Santo aiitonio Tv. Rui Barbosa, 1080 Belém, Para, Brazil.

An overview of arbovirology in Brazil and neighbouring countries; edited by Ainélia P.A. Travassos da Rosa, Pedro F.C. Vasconcelos, Jorge F. S. Travassos da Rosa. Belém: Instituto Evandro Chagas, 1998. 29Gp.

Includes bibliographies and index ISBN 85-86784-01-X

1 ..Arboviruses 2. Arbovirus infections I. Travassos da Rosa, A.P.A. II. Vasconcelos, P.F.C. III. Travassos da Rosa, J.F. S.

CDD 576.64 6 16.925

. .. r

n