Bull. Org. mond. Sante 1963, 29, Suppl., pp. 127-133Bull. Wld Hlth Org.,

Ecology and Vector ControlD. H. S. DAVIS'

One of the primary functions ofecological work is to conduct surveys of those speciesof vertebrates and/or invertebrates that are involved directly or indirectly in vector-bornediseases ofman. The aim in southern Africa has been to contribute to a stable nomenclatureat the species level, to determine the range ofeach species and to make an attempt to singleout key environmental limiting factors. Once the taxonomic status and range of the specieshave been established with some degree of certainty, species may be studied in relation tothe distribution and prevalence of any particular disease in order to provide the basis forcontrol. The bio-geographical approach that has been adopted arose when the geographicaldistribution of human plague derived from wild-rodent sources came to be compared withthe distribution of the small mammals and their respective fleas. This threw much light onthe factors concerned in the limitation andpersistence ofplague. These comparative studieswere facilitated by plotting the distribution data on grid maps.

Ecology and vector control is a wide subject andI do not propose to discuss here more than oneaspect of it, based on experience in southern Africaparticularly in connexion with field work on plague,but also on other vector-borne diseases. In thecourse of these studies our knowledge of the taxo-nomy and geographical distribution of the smallmammals, the fleas and other arthropods of medicalimportance has steadily increased. The results ofthe flea survey in southern Africa have been pub-lished (de Meillon, Davis & Hardy, 1961). One ofthe objects of this paper is to stress the need foraccurate mapping of distributions of hosts andvectors for comparison with the distribution andincidence of a vector-borne disease, not only inorder to obtain indications of interrelations betweenhosts, vectors and environmental factors, but alsoas a necessary basis for more intensive investigationof the general ecology of disease agents and theircontrol.

I shall also refer briefly to the present status ofvector control in malaria, typhus and plague-thethree vector-borne diseases for which complete orpartial eradication is the aim.

1 Ecologist aud Chief Rodent Officer, Medical EcologyCentre, State Department of Health, P.O. Box 1038, Johan-nesburg, South Africa.

THE ECOLOGICAL APPROACH

In ecological research the ecologist relies heavilyon the taxonomist. Some ecologists, in order tolighten the burden on taxonomists and also toincrease their own insight, make a special study ofone group in which they have a particular interest.In our investigations of plague in southern Africataxonomic problems have played a decisive role indetermining the direction field work took. After arelatively short period devoted to intensive studiesof wild-rodent populations during the years 1939-41(Davis, 1939, 1953), the exigencies of plague controlduring the Second World War diverted field workfrom intensive research to extensive surveys of thedistribution of rodents and fleas in relation to thepersistence of plague.

Plague surveys

These surveys were planned and carried out byor under the direction of an ecological unit and indue course provided the basis for critically assessingthe role of the different species of rodents and oftheir fleas in the transmission cycle of plague; atthe same time this provided the material for taxo-nomic research. The identification of the fleas wasundertaken by a medical entomologist and that of

1325 - 127-

D. H. S. DAVIS

the rodents and other small mammals by theecologist, aided by a museum mammalogist. Fromthe start the principle was adopted of retaining themajority of specimens collected for future reference.

Mapping methods

In order to reduce this extensive material tomanageable proportions and to draw conclusionsfrom the results, a system of mapping was devisedwhich also provided a convenient means of integ-rating all relevant information on epidemiologyand ecology. Outline maps printed with a sub-divided degree-square graticule not only enablesymbols to be arranged uniformly on the map butalso provide convenient indexing units (in this caseone-sixteenth of a degree square). Each symbolrepresents the occurrence of one or more recordsof a species or event in the " locus ". (For a descrip-tion of the " locus system ", see de Meillon, Davis& Hardy (1961), p. 21.)

Since all the specimens are indexed and arrangedaccording to the degree-square locus, the recordsand reference collections are automatically cross-referenced. This makes it possible to extract allavailable information, including field notes, on aparticular geographical area in a relatively shorttime. It also provides a ready method of assessingthe degree of completeness of the recorded distribu-tion of any species by simple inspection of themaps; suspected errors in identification can quicklybe checked by noting the locus of any anomalousrecord and tracing, through the card index, thespecimen or specimens on which it was based.When it came to publishing the results of the flea

survey (de Meillon, Davis & Hardy, 1961) it wasfelt that the distribution data on most of the speciesof South African fleas gave a reasonably accurate,and in some species a very accurate, picture of theirdistribution. In that book an attempt was madeby the second author to interpret the distributionpattern of each species of flea in relation to host-association and general environmental factors inorder to assess the part played by each in the trans-mission cycle of plague. Although this was doneupon the available field evidence, experimentallaboratory confirmation, especially of the capacityof certain key species to act as vectors and harbourersof Pasteurella pestis, would, both from the pointof view of control and for a deeper understandingof the plague cycle, have been more convincing.However, the means to do this were beyond thereach of the ecological unit.

Small mammals

The taxonomic status of many of the smallmammals, especially the rodents, has been underconstant review and many problems still remainunresolved. However, the species more closelyassociated with the plague transmission cycle arenow reasonably well defined. Their distribution hasbeen plotted in as much detail as that of the fleas(see Davis (1962) for Muridae distributions). Takentogether, the distribution patterns of fleas, rodentsand plague have given many clues to the generalconditions which limited or permitted the extensionof the plague-infected area. Combined with thedetailed field reports of surveys they provided thebasis for drawing tentative conclusions about thecapacity of P. pestis to survive in different subregionsof the enzootic area. Since 1950 southern Africahas been comparatively free from human outbreaks(see Fig. 1). The years 1949-50 marked the occur-rence of the last major panzootic and at the timeof writing (1962) there is a lull, which, however,shows signs of being about to break.Whether the decline in plague is part of the general

world-wide recession or not, the present is a crucialmoment to re-institute intensive field studies andto intensify control.

VECTOR ERADICATION CAMPAIGN

Up to the present, no marked physiologicalresistance to residual insecticides has been detectedin southern Africa in the vectors of malaria, epidemictyphus or plague.

Malaria

Malaria has been greatly reduced, but still persistsalong international boundaries in the contact zonewith the endemic area. As malaria eradicationprogrammes get under way in the remaining endemiczones in southern Africa it is expected that infiltra-tion therefrom will cease. There is a need forresearch into the irritability response to differentinsecticides and into the precise nature of thechanges that seem to have taken place in the bitingand resting habits of Anopheles gambiae.

Epidemic typhus

As with malaria, epidemic typhus has recededfollowing the introduction of DDT. Unlike in thecase of malaria, however, the endemic area isgeographically isolated and is now restricted to

128

ECOLOGY AND VECTOR CONTROL

several districts in the eastern Cape Province. Aneradication campaign has been started to eliminatethe focus, but it is too early yet to gauge the outcome.In 1954 highly resistant strains of human lice werediscovered in two places, but subsequent tests(Steyn et al., 1960) have shown that this highresistance was localized and incidental and thatDDT properly applied is still effective. A factorcomplicating eradication is Brill-Zinsser disease orrecrudescent typhus; the epidemiological role ofthis disease in re-starting outbreaks has not beenestablished. Bedbugs and human fleas appear tohave developed physiological resistance and consti-tute a problem in that the villagers now have anantipathy to DDT. A switch to an alternativeinsecticide which will kill bugs and fleas as well aslice is being considered.

PlaguePlague is enzootic amongst wild rodents over the

greater part of the subcontinent lying in the under25-inch (635 mm) rainfall zone-that is, briefly, thesemi-arid subdesert regions comprising the Karoo,Kalahari and Highveld. A general decline in theprevalence of human plague in the longer-establishedfoci in the Orange Free State and eastern CapeProvince set in about 1940. Although outbreakscontinued to occur sporadically over a wide area,they became more concentrated in three fairly well-defined hyperenzootic areas (Davis, 1948). TheKalahari was invaded about 1928 and the epidemio-logical picture in this outer zone has been one ofepidemics at intervals, rather than scattered out-breaks, analogous to the epidemics that took placein the 1920's in the Orange Free State and easternCape Province.

Focal eradication. A tentative start has been madeto eradicate the rodents and/or fleas forming thewild-rodent reservoir in strictly limited foci. Whilethe intensity of control has varied from place toplace, these limited attempts to break the trans-mission cycle, combined with other processes suchas increased industrialization, more intensive farm-ing, the greater awareness of the public and a generalrise in the standard of rural building, hold somepromise of hastening the decline in such limitedareas. During the six-year period July 1955 to June1961 (see Fig. 1G), only 27 cases of plague werereported in the whole of southern Africa. SinceJuly 1961 there has been a further small outbreakin the eastern Cape and two epidemics in Ovambo-land on the north-west fringe of the Kalahari.

The species factor. It is difficult to see how muchimpression can be made on the enzooticity of sucha disease as plague, distributed as it is in wild-rodent populations over some half-million squaremiles. But with our knowledge of the part playedby the different species of rodents and their fleasit is possible to narrow the problem down to theessential maintaining host and vector species of thereservoir system and to concentrate on breakingthe transmission cycle (a) in parts of the above-mentioned hyperenzootic areas and (b) in otherareas as these are revealed through human out-breaks or by the isolation of P. pestis during routinesurveys. Straightforward vector control (DDT-dusting of the floors ofhuts and other rural buildings)is highly efficacious in arresting transmission to man.The extension of DDT-dusting to all gerbil

burrows, whether still occupied or not, in theimmediate vicinity of human habitations, disposesof the infected fleas of the primary reservoir host.Such local measures, applied at the time of a humanoutbreak or applied in an area threatened with anoutbreak following the demonstration of plagueinfection in rodents or fleas, are practicable.

Flea species complex. The basis for selectivevector control will be made clearer by reference toFig. 2, which shows the relative proportions of thedifferent species of fleas on the hosts in a hyper-enzootic area in the northern Orange Free State.The two key species of fleas to be attacked areXenopsylla brasiliensis in the hut environment whichtransmits plague to man, and X. philoxera, whichis primarily responsible, with its host Tatera brantsi,for maintaining P. pestis in nature.

Eradication experiment. An analysis of theaccumulated data from field surveys and obser-vations was done prior to the initiation of anexperimental eradication programme designed toeradicate P. pestis from the districts of Uitenhageand Port Elizabeth in the eastern Cape (see pointon south-east coast 33°S, 25°E in Fig. 1). This areais isolated from the main enzootic area and thepersistence of plague there is associated primarilywith the moderately resistant Namaqua gerbilDesmodillus auricularis. The analysis showed thathut-floor fleas consisted of Pulex irritans, a poorvector of plague, and that the Namaqua gerbil wasinfested almost exclusively by X. piriei, a flea rarelyencountered on other species of rodents in the area.The working basis was to test the hypothesis thatthe persistence of plague in this particular region

129

D. H. S. DAVIS

ILclN6

ZjX3

Z0

WZ

U.3

Z <

IL,i Z

ow10-

L6 0

o CO

r -

U_

o-a0.

0ICL- w

On

2o-

z

m )3--z

o 0

Ck X

_w

.J-

-mIL,.

a O

a-m

JZr

C Uw

U)

co,

aSLSp

_ hC?

I Iidci

U6-

m4'

130

LLN

ECOLOGY AND VECTOR CONTROL 131

CD

0l

U.

D. H. S. DAVIS

FIG. 2. THE FLEA-SPECIES COMPLEX IN THE NORTHERN ORANGE FREE STATEa

IXcryptondo

Ctconnatus

X.111oxera l

C1

,,0d

I Chross _

I..

XLbrasieIK- {d

,SqzULUS a

P irrton v~~~~~n

a Vector control is applied to Xenopsylla brasiliensis in the hut environment (bottom circle), to arrest transmission to man, andto Xenopsylla philoxera (and associated fleas) in the burrows and nests of gerbils (Tatera brantsi) (top centre circle), to break thetransmission cycle in the primary reservoir. Note that the flea-species composition of the multimammate mouse (Rattus natalensis)is intermediate between the primary reservoir and the hut environment. The arrows point to the host on which the different speciesof flea occur. The host groups (top left and top right circles) play an incidental role in persistence and are normally neglected invector control.

was dependent upon Desmodillus as the primaryreservoir host and that the remaining species ofrodents were incidental hosts but capable of actingas a secondary reservoir for a limited length oftime. The measures directed against human trans-mission in the hut environment are dusting thefloors with malathion (Pulex irritans have shownsome physiological resistance to DDT in the area),combined with malathion-dusting of Desmodillusburrows, poisoning and trapping.

CONCLUDING REMARKS

This outline serves to give some idea of the stepsin the process of narrowing down the field ofcontrol from an analysis of the broad epidemio-logical and ecological facts and factors of wild-

rodent plague. The gradual clarification of thetaxonomic status of the species of fleas and rodents,together with a knowledge of their geographicaldistribution and relative abundance, has enabled us

to concentrate on those species which appear to beessential for the survival of P. pestis. Vector controlper se is applied not only to the species of fleasresponsible for human infections in the domesticenvironment, in order to stop secondary infectionof man, but also to the species of fleas associatedwith the primary reservoir in wild rodents, in orderto disrupt the transmission cycle. These measurescan be applied only in strictly localized areas,singled out on the degree of enzooticity as reflectedby the frequency of past human outbreaks androdent epizootics.

, nsAd CyS,

S..::~5:*:.

132

'SM BRA/V,,S/

ECOLOGY AND VECTOR CONTROL 133

RtSUMI

L'une des fonctions majeures des etudes ecologiquesen Afrique du Sud a ete d'entreprendre des enquetes surles especes d'invertebres et de vert6bres directement ouindirectement interesses dans la transmission de maladieshumaines. On a eu pour but d'etablir une nomenclaturefixe a l'6chelon de l'espece, de determiner la place desesp&ees, et de tenter de degager les principaux elementsdu milieu susceptibles de limiter les especes. Lorsque lesesp&es ont ete classees de facon relativement suire, onles etudie sous l'angle de la repartition et de la frequencede telle ou telle maladie particuliere. En meme temps,lorsque l'occasion s'en pr6sente, on entreprend desenquetes plus detaillees sur le terrain en vue d'unemeilleure compr6hension des cycles de transmission natu-rels, d'une part, et d'une base pour la lutte contre lesvecteurs, d'autre part.

Ceci n'est peut-etre pas de l'ecologie au sens strict duterme. Cela n'en est pas moins essentiel pour projeteret mener a bien un travail ecologique sur les groupementset populations d'animaux jouant un role dans le main-tien d'un agent pathogene (qu'il s'agisse de virus, debacterie, de protozoaire ou d'helminthe).La conception biogeographique adoptee s'est form6e

lorsqu'on a compare la distribution geographique dela peste humaine transmise par des rongeurs sauvagesa la distribution des petits mammiferes et de leurs pucesrespectives. Ceci a largement eclaire l'etude des facteursde limitation ou de maintien de la peste. Ces etudescomparatives ont 6te facilitees par l'idee d'utiliser descartes quadrillees (par degres carres) du Service trigono-

metrique et d'y coucher les observations sur la repar-tition des especes. Le systeme de division permet declasser facilement les renseignements et d'en dresser laliste de fagon que les notions essentielles sur la frequencedes maladies, les especes d'h6tes et de vecteurs, l'habitat,etc. soient retenues ou ajout&es en termes d'unites geogra-phiques. Si l'on prend en meme temps comme referencesles cartes de distribution, il est possible de retrouver rapi-dement les notes originales prises sur le terrain, lesspecimens se trouvant dans les collections des musees ettous autres renseignements se rapportant au sujet etcatalogues dans le systeme.

Cette unification est un preliminaire essentiel a touttravail plus fondamental. C'est en un sens un service6cologique permettant d'unir etudes sur le terrain etrecherches de laboratoire. I1 joue un r6le en contribuantit la recherche fondamentale et en fournissant une basea la lutte contre les vecteurs. Des qu'un programmeordinaire de lutte a etd mis au point, des progres devraientse realiser a partir de donnees ecologiques. Bien souventil n'en est pas ainsi. Parmi les resultats apparemmentsurprenants des campagnes d'eradication certains ontete prevus, certains autres eussent pu etre evites si l'onavait pris garde aux cons6quences possibles d'une inter-ference avec l'equilibre dynamique des populationsanimales. Ceci est au cceur de 1'ecologie et l'impasseactuelle ou se trouve la lutte contre les vecteurs au moyend'insecticides residuels doit inciter l'ecologiste a trouverdes solutions pratiques.

ACKNOWLEDGEMENT

This paper is published with the kind permission of the Secretary for Health, Republic of South Africa.

REFERENCES

Davis, D. H. S. (1939) S. Afr. J. Sci., 36, 438Davis, D. H. S. (1948) Ann. trop. Med. HIyg., 48,

207Davis, D. H. S. (1953) J. Hyg. (Lond.), 51, 427Davis, D. H. S. (1962) Ann. Cape Prov. Mus., 2, 56

Meillon, B. de, Davis, D. H. S. & Hardy, F. (1961)Plague in southern Africa. L The Siphonaptera(excluding the Ischopsyllidae) of southern Africa,Pretoria, Government Printer

Steyn, J., De Beer, J. J., Greeff, J. G. & Coetzee, W. L.(1960) S. Afr. med. J., 34, 90