Seasonal Abundance of Plague Vector Xenopsylla brasiliensis from Rodents Captured in Three Habitat Types of Periurban Suburbs of Harare,Zimbabwe

Abstract

The study was aimed at determining the seasonal abundance of Xenopsylla brasiliensis, an important vector of plague in Zimbabwe, from rodent hosts captured in selected habitat types of two periurban suburbs of Harare, Zimbabwe. The removal-trapping method was used to capture the rodents, from which fleas were collected and identified. Percentage incidence index (PII) and specific flea index (SFI) were calculated for X. brasiliensis in relation to rodent species host. Mastomys natalensis, Rattus rattus, Tatera leucogaster, and Rhabdomys pumilio were the rodent species present in the study areas and all species were infested with X. brasiliensis. PII for T. leucogaster in relation to X. brasiliensis was significantly higher (p < 0.05) compared with that of the other rodent species and T. leucogaster also recorded the highest SFI, whereas R. pumilio recorded the lowest indices. In both formal and informal settlements, the highest PII of X. brasiliensis was attained for M. natalensis, followed by R. rattus. In the cultivated habitat, T. leucogaster recorded the highest indices and R. pumilio the lowest. X. brasiliensis was found to cohabitat with Dinopsyllus lypusus and Ctenophthalmus calceatus on M. natalensis, R. rattus, and T. leucogaster. No cohabitation was recorded for R. pumilio. For all the rodent species captured, both the PII and SFI of X. brasiliensis were highest during the hot-dry season, followed by the hot-wet season, with the cold-dry season recording the lowest indices. The overall cohabitation was highest during the hot-dry season and lowest during the hot-wet season. The findings of the present study fit the reported occurrence of plague outbreaks during the hot-dry season in periurban Zimbabwe.

Introduction

T he genera of fleas mostly associated with plague transmission in eastern and southern Africa are Xenopsylla, Nosopsylla, and Leptopsylla (Service 1986). Xenopsylla brasiliensis and X. cheopis are the most important plague vectors within this genus and X. brasiliensis has been reported as the principal plague vector in southern Africa (Gratz 1999). The highest percentage of human plague cases and mortalities have been reported in Africa and most outbreaks occurred in southern Africa, which includes Zimbabwe (WHO 1999).

The range of host specificity of fleas to certain genera of rodents (Mastomys, Tatera, Rattus, Rhabdomys, Arvicanthis, and Otomys) is varied, with some species being adapted to only one host and others to a wide variety of hosts (Haeselbarth et al. 1966, Bahmanyar and Cavanaugh 1976, Borchert et al. 2007). Xenopsylla cheopis and X. brasiliensis are capable of reproducing and maintaining populations on several species of the host genera mentioned above (Haeselbarth et al. 1966). In Africa and India, X. cheopis appears to have an affinity for rats living in burrows, whereas X. brasiliensis is attracted to wild rodents and their fleas living above ground in thatched roofs and walls, such as R. rattus (Haeselbarth et al. 1966, Bahmanyar and Cavanaugh 1976).

Plague occurs naturally in rodents, which become infected through the bite of infected rodent fleas such as X. brasiliensis and X. cheopis. Humans get infected with the causative agent of plague when they enter zones with infected wild rodents through activities such as cultivation and hunting. In southern Africa, the primary source of human infection is from domestic rodents such as R. rattus and peridomestic such as M. natalensis (Bahmanyar and Cavanaugh 1976, Service 1986, Stevenson 1987). Xenopsylla brasiliensis has been reported to be the most common flea species on rodents in southern and eastern Africa and an efficient vector of plague (Heisch et al. 1953, Haeselbarth et al. 1966, Lewis 1972, Kilonzo and Mhina 1982, 1983, Shephard and Leman 1983, Shephard et al. 1983, WHO 1983).

Knowledge of seasonal variations in infestations of rodents by X. brasiliensis is important for understanding the epidemiology of plague. The aim of this study was to investigate the seasonal abundance and distribution of X. brasiliensis on rodent species trapped in selected habitat types of two periurban suburbs of Harare, Zimbabwe, and its cohabitation with other flea species. This information will assist in elucidating the risk factors associated with outbreaks of plague in periurban Zimbabwe.

Materials and Methods

Study areas

The study areas were Hatcliffe (17°39′S; 31°10′E) and Dzivarasekwa (17°47′S; 30°54′E), which are periurban suburbs of the city of Harare, Zimbabwe, and lie approximately 16 km northeast and 14 km west of Harare, respectively (Fig. 1). The two study areas are approximately 35 km from each other, sufficiently far enough to serve as two isolated sites. Three habitat types—informal settlement, formal settlement, and cultivated land—were selected for rodent trapping in each study area (Fig. 1). A formal settlement refers to a sanctioned or authorized urban residential area built on surveyed and serviced municipal land with standard water; refuse collection, road, and sewage system. The houses are built following laid-down standards of the Department of Urban Planning and Building Inspectorate of the city. An informal settlement refers to unsanctioned or unauthorized squatter or illegal settlement at the urban periphery without water, refuse collection, road, and sewage system. The houses are temporary units constructed from various kinds of materials such as cardboard, polythene, scrap wood, and metal. Formal settlements, lying approximately 3 km to the south of each informal settlement, were included in the study for comparison purposes. In selecting the study sites, consideration was given to similarities in physical and social characteristics of the sites.

FIG. 1.

Map of southern Africa showing the geographical position of Harare, Zimbabwe (A), Hatcliffe (B), and Dzivarasekwa (C) study areas with the three habitat types and site of trap lines.

Trapping procedure of rodents

Trapping was conducted in all the habitat types—formal and informal settlements and the cultivated land. Removal-trapping was conducted for three consecutive trapping nights in 2005 and 2006 during the months of February (hot-wet season), June (cold-dry season), and October (hot-dry season), using Sherman live traps baited with a mixture of peanut butter and maize bran. In the informal settlement and the cultivated land, traps were placed in two lines of 25 traps each and approximately 10 m apart, giving a total of 50 traps for each habitat type. Fifty residential houses were used as a control in the formal settlement, with a trap being placed behind each house. Each residential house was on a 200-m² stand and the distance between adjacent residential houses was approximately 10 m. In the formal and informal settlements, trap stations were established at suitable points just outside the dwelling structures following consent of householders. Traps were placed in the afternoon and trapped rodents were collected the following morning and taken to the laboratory for processing. Traps were left open in the same place during the day with productive traps being immediately replaced every morning over a 3-day trapping session.

Rodent and flea identification

The captured rodents were taken to the laboratory for processing and identification. They were euthanized in chloroform following the method described by Bahmanyar and Cavanaugh (1976). Fleas were collected from each rodent by carefully brushing the fur of the rodent using a toothbrush as described by Bahmanyar and Cavanaugh (1976). After collection of fleas, the rodents were processed and identified following the illustrations and descriptions by Smithers (1975). Each identified species was confirmed by the Natural History Museum of Zimbabwe. Fleas collected from each identified rodent species were transferred into labeled Eppendorf tubes containing 70% ethanol and were processed for identification following descriptions by Haeselbarth et al. (1966) and Bahmanyar and Cavanaugh (1976).

Data analysis

The effect of habitat type and season on the flea population density of X. brasiliensis on rodents was measured by using the flea percentage incidence index (PII) and specific flea index (SFI) as described by Bahmanyar and Cavanaugh (1976). The PII of each rodent species was calculated by dividing the number found infested with X. brasiliensis by the total number trapped and the result was expressed as a percentage, whereas the SFI was calculated as the number of fleas per rodent species. The X. brasiliensis indices were calculated for each rodent species, in each habitat type and season. Comparisons were made for each pair of variables. For the PII, the χ ² test was used to measure the differences between categories and values of p < 0.05 were considered as significant.

Results

The dominant rodent species trapped were M. natalensis, R. rattus, T. leucogaster, and R. pumilio in that order (Table 1), and all were found to be infested with X. brasiliensis. Population density of X. brasiliensis as measured by PII and SFI varied from 5.2% to 62.5% and 0.2 to 3.0, respectively. Both indices were highest for T. leucogaster and lowest for R. pumilio. The PII for T. leucogaster was significantly higher (Table 1) compared with M. natalensis (χ ² = 22.7, df = 1, p < 0.001), R. rattus (χ ² = 17.9, df = 1, p < 0.001), and R. pumilio (χ ² = 76.0, df = 1, p < 0.001). Mastomys natalensis (χ ² = 40.5, df = 1, p < 0.001) and R. rattus (χ ² = 42.6, df = 1, p < 0.001) also recorded a significantly higher PII compared with R. pumilio.

Table 1.

Indices of Xenopsylla brasiliensis According to Rodent Host Species Trapped During the Period 2005–2006

Rodent species	No. caught	No. infested	Total fleas	PII	SFI
Mastomys natalensis	489	188	702	38.4^a	1.4
Rattus rattus	350	141	491	40.3^a	1.4
Tatera leucogaster	120	75	359	62.5^b	3.0
Rhabdomys pumilio	97	5	23	5.2^c	0.2

Values in the same column with a different superscript letter are significantly different at p < 0.05

PII, percentage incidence index; SFI, specific flea index.

The epidemiological indices of X. brasiliensis on the four dominant rodent species captured according to habitat type are shown in Table 2. Mastomys natalensis was captured in all the three habitats, R. rattus in two habitats (formal and informal), whereas T. leucogaster and R. pumilio were captured in the cultivated habitat only. There was no significant difference in the PII of X. brasiliensis for M. natalensis and R. rattus among the habitat types studied, and similarly, the SFI of X. brasiliensis for M. natalensis and R. rattus were comparable among the habitat types studied (Table 2). In the cultivated habitat type, T. leucogaster recorded the highest indices and R. pumilio the lowest (Table 2).

Table 2.

Indices of Xenopsylla brasiliensis According to Habitat and Rodent Host Species Trapped During the Period 2005–2006

	PII			SFI
Rodent species	Informal	Cultivated	Formal	Informal	Cultivated	Formal
Mastomys natalensis	44.1^a	34.8^a	46.5^a	1.6	1.4	1.3
Rattus rattus	39.4^a	—¹	41.4^a	1.4	—	1.3
Tatera leucogaster	—	62.5	—	—	3.0	—
Rhabdomys pumilio	—	5.20	—	—	0.2	—

Values in the same row with a different superscript letter are significantly different at p < 0.05.

No rodent species trapped in the habitat.

Xenopsylla brasiliensis was found to cohabitat with Dinopsyllus lypusus and Ctenophthalmus calceatus on M. natalensis, R. rattus, and T. leucogaster. No cohabitation was recorded for R. pumilio (Table 3). Tatera leucogaster recorded the highest multiple infestation (40%) and R. rattus the lowest (13.4%). Xenopsylla brasiliensis cohabitation with both D. lypusus and C. calceatus was highest for T. leucogaster (20.8%), whereas with D. lypusus only it was highest for R. rattus (72.3%) and with C. calceatus only it was highest for M. natalensis (27.1%). The highest X. brasiliensis cohabitation was recorded with D. lypusus for all the three rodent species in which cohabitation existed.

Table 3.

Xenopsylla brasiliensis Cohabitation with Dinopsyllus lypusus and Ctenophthalmus calceatus on Rodent Host Species Trapped During the Period 2005–2006

		Double and triple infestation	Double infestation only		Triple infestation only
Rodent species	No. caught	Total (% of total caught)	XC (% of total infested)	XD (% of total infested)	XDC (% of total infested)
Mastomys natalensis	489	85 (17.4)	23 (27.1)	53 (62.4)	9 (10.6)
Rattus rattus	350	47 (13.4)	9 (19.1)	34 (72.3)	4 (8.5)
Tatera leucogaster	120	48 (40.0)	7 (14.6)	31 (64.6)	10 (20.8)
Rhabdomys pumilio	97	0	0	0	0

X, Xenopsylla brasiliensis; D, Dinopsyllus lypusus; C, Ctenophthalmus calceatus.

For all the rodent species trapped, both the PII and SFI were highest during the hot-dry season, followed by the hot-wet season, with the cold-dry season recording the lowest indices (Table 4). The hot-wet season recorded a significantly higher PII for M. natalensis (χ ² = 15.9, df = 1, p < 0.001), R. rattus (χ ² = 18.3, df = 1, p < 0.001), and T. leucogaster (χ ² = 58.3, df = 1, p < 0.001) compared with the cold-dry season (Table 4). Similarly, the hot-dry season also recorded a significantly higher PII for M. natalensis (χ ² = 28.6, df = 1, p < 0.001), R. rattus (χ ² = 37.8, df = 1, p < 0.001), and T. leucogaster (χ ² = 73.2, df = 1, p < 0.001) compared with the cold-dry season (Table 4). Tatera leucogaster recorded the highest indices for all the seasons and R. pumilio recorded the lowest.

Table 4.

Seasonal Indices of Xenopsylla brasiliensis According to Rodent Host Species Trapped During the Period 2005–2006

	PII			SFI
Rodent species	Hot-wet	Cold-dry	Hot-dry	Hot-wet	Cold-dry	Hot-dry
Mastomys natalensis	46.5^a	22.6^b	53.3^a	1.7	0.5	2.5
Rattus rattus	45.3^a	18.6^b	57.4^a	1.4	0.4	2.4
Tatera leucogaster	73.5^a	38.8^b	83.8^a	4.0	1.0	4.7
Rhabdomys pumilio	3.9^a	0	14.3^a	0.04	0	0.8

Values in the same row with a different superscript letter are significantly different at p < 0.05

The overall cohabitation was highest during the hot-dry season and lowest during the hot-wet season, and no cohabitation with C. calceatus was recorded during the hot-wet season (Table 5).

Table 5.

Xenopsylla brasiliensis Seasonal Cohabitation with Dinopsyllus lypusus and Ctenophthalmus calceatus on the Rodent Host Species Trapped During the Period 2005–2006

	Season
	Hot-wet				Cold-dry				Hot-dry
Rodent species	XDC	XD	XC	Total (%)	XDC	XD	XC	Total (%)	XDC	XD	XC	Total (%)	Grand total
Mastomys natalensis	0	13	0	13 (15.3)	6	16	10	32 (37.6)	3	24	13	40 (47.1)	85
Rattus rattus	0	11	0	11 (23.4)	2	6	6	14 (29.8)	2	17	3	22 (46.8)	47
Tatera leucogaster	0	13	0	13 (27.1)	8	7	2	17 (35.4)	2	11	5	18 (37.5)	48
Rhabdomys pumilio	0	0	0	0	0	0	0	0	0	0	0	0	0

X, Xenopsylla brasiliensis; D, Dinopsyllus lypusus; C, Ctenophthalmus calceatus.

Discussion

Results from our study agrees with earlier reports by Taylor et al. (1981), which showed that M. natalensis, R. rattus, T. leucogaster, and R. pumilio were the dominant rodent species in Zimbabwe and are hosts of X. brasiliensis. The presence of the four major rodent species infested with X. brasiliensis, which has been implicated as the principal plague vector in Zimbabwe (Taylor et al. 1981), is a potential risk factor for plague outbreaks in the study areas.

Flea indices give useful information for the surveillance of plague and serve as indicators of potential plague transmission (Bahmanyar and Cavanaugh 1976, WHO 1976). The overall SFI range of 0.2–3 for Xenopsylla brasiliensis recorded in this study is comparable to that reported from Kenya (Heisch et al. 1953) and Tanzania (Kilonzo and Mhina 1982). An SFI of at least 0.5–1 is considered sufficient to maintain plague in a locality (Hirst 1927) and an index ≥1 is reported to represent a potentially dangerous situation with respect to plague risk (WHO 1999). Thus, the X. brasliensis SFIs recorded for T. leucogaster, M. natalensis, and R. rattus in our study represent a potential risk for plague outbreak in the study areas.

As reported in other studies (Hirst 1927, Heisch et al. 1953, Kilonzo and Mhina 1982, Makundi et al. 2003), variations of flea densities according to host species were also observed in our study. The observed variations of flea densities on different rodent species may be attributed to factors such as the immune status and grooming behavior of the rodent host, numbers of fleas lost over distances traveled by the rodent host species, and the type and degree of protection from external conditions offered by the shelter of the rodent host (Haeselbarth et al. 1966, Nelson et al. 1975, 1977, Bahmanyar and Cavanaugh 1976, Kim 1985).

In southern Africa, T. leucogaster is reported to be the principal host for X. brasiliensis and plays a major role in sylvatic plague transmission (Davies 1948, Smithers 1975). Xenopsylla brasiliensis is also reported to be attracted to rodents living above the ground in thatched roofs and walls such as R. rattus (Haeselbarth et al. 1966, Bahmanyar and Cavanaugh 1976). In contrast, R. pumilio is reported not to be often infested with X. brasiliensis (Haeselbarth et al. 1966). The results of the present study support these observations as the population density of X. brasiliensis was demonstrated to be highest on T. leucogaster, followed by R. rattus and M. natalensis, with R. pumilio recording the lowest.

Given the interaction of domestic rodents, R. rattus, with peridomestic, M. natalensis, and wild, T. leucogaster, rodents in the cultivated, formal, and informal sectors (Zimba 2008), the presence of X. brasiliensis on these rodents poses serious potential plague outbreaks in the studied areas. However, the results of the present study seem to suggest that the informal settlement could be at a higher risk of plague outbreaks as higher SFIs were recorded from this sector. Habitat variations of flea density have been also reported in Sri Lanka, where X. cheopis index varied from 1.7 in customs grain storage facilities, 0.8 in principal town grain markets, to less than 0.02 in domestic premises (Hirst 1927). The possible reason for this variation could be probably that the environmental conditions in the informal settlement were more favorable for the breeding and growth of the rodents (Zimba 2008). In addition, cracks, crevices, and debris on floors of dwelling structures in the informal settlement could have created favorable conditions for the larval stages of the fleas as observed earlier by Service (1986).

Observed cohabitation of X. brasiliensis with other flea species on the same host agrees with earlier findings from other countries such as Sri Lanka, Kenya, and Tanzania, respectively (Hirst 1927, Heisch et al. 1953, Kilonzo and Mhina 1983). The highest observed cohabitation was with D. lypusus and this flea has been reported as an important vector of plague among wild rodents in scattered foci of the disease in equatorial Africa (Haezelbath et al. 1966). The epidemiological significance of cohabitation is not clear; however, it may be possible that cohabitation of efficient plague vectors such as X. brasiliensis and D. lypusus may have a more severe impact on the outcome of an outbreak.

Dinopsyllus lypusus has been reported to be common on rodents in East Africa, where it is associated with plague transmission (Kilonzo and Mhina 1982, 1983, Makundi et al. 2003). The flea has been also reported to be an important vector of plague among wild rodents in scattered foci of the disease in equatorial Africa (Haeselbarth et al. 1966). In southern African countries such as Botswana, the Democratic Republic of Congo, Malawi, Mozambique, and Zambia, X. brasiliensis and D. lypusus have been reported to be the most important plague vectors (WHO 1999). The role of C. calceatus in the transmission of plague is currently unknown, but in East Africa, a related subspecies (C. calceatus cabirus) has been reported to be a vector of the disease and has been collected from rodents involved in plague epizootics (Heisch et al. 1953). The role of D. lypusus and C. calceatus in plague transmission in Zimbabwe has not been studied. In the present study, it is assumed that should the two species be involved in transmission, then the severity of a plague outbreak would be amplified.

Seasonal variations in rodent flea densities have been reported in India, Sri Lanka, Kenya, and Tanzania, respectively (Hirst 1927, Heisch et al. 1953, Kilonzo and Mhina 1982). During the present study, the hot-dry season recorded the highest flea indices and incidence of cohabitation. Temperature ranges between 10°C and 30°C and a high humidity have been reported to favor the viability of fleas, the infectivity, and the survival of the causative organisms in the fleas (Hirst 1927, Mason-Bahr and Bell 1987). In addition, among the rodent fleas, X. brasiliensis and X. cheopis have been reported to have the greatest tolerance for dry conditions (Haeselbarth et al. 1966). The findings of the present study support the reported occurrence of plague outbreaks during the hot-dry season in Zimbabwe (Pugh and Parker 1975, Epstein 1997, WHO 1999). Hence, rodent control efforts should be targeted during the cold-dry season when the population density is high (Zimba et al. 2008). Control measures should take the form of an integrated approach involving environmental management and chemical control, particularly in the informal settlements where risk of plague outbreaks appears to be high, and periurban cultivation should be discouraged to prevent wild, peridomestic, and domestic rodent interactions.

Footnotes

Acknowledgments

The authors acknowledge financial support from the city of Harare, Department of Health, and the National History Museum of Zimbabwe for confirming rodent identifications.

Disclosure Statement

No competing financial interests exist.

References

Bahmanyar

, Cavanaugh

. Plague Manual. Geneva: World Health Organization, 1976; 76.

Borchert

, Mach

, Linder

, Ogen-Odoi

, Angualia

. Invasive rats and bubonic plague in northwest Uganda. USDA National Wildlife Research Center Symposia. Managing Vertebrate Invasive Species. University of Nebraska–Lincoln, 2007; 281–293

Davies

DHS

. Sylvatic plague in South Africa; history of plague in man 1919–1943. Ann Trop Med Parasit, 1948; 42:207–217.

Epstein

. Climate, ecology and human health. Consequences, 1997; 3:1–23.

Gratz

. Emerging and resurging vector-borne diseases. Annu Rev Entomol, 1999; 44:51–75.

Haeselbarth

, Segerman

, Zumpt

. The arthropod parasites of vertebrates in Africa (Ethiopian region) Zumpt

. Publications of the South African Institute for Medical Research: Johannesburg, 1966; 13:117–250.

Heisch

, Grainger

, D'Souza

JSAM

. Results of a plague investigation in Kenya. Trans Roy Soc Trop Med Hyg, 1953; 47:503–521.

Hirst

. Rat-flea surveys and their use as a guide to plague prevention measures. Trans Roy Soc Trop Med Hyg, 1927; 21:87–108.

Kilonzo

, Mhina

JIK

. Observations on the current status of plague endemicity in the western Usambara mountains, north-east Tanzania. Acta Trop, 1983; 40:365–372.

10.

Kilonzo

, Mhina

JIK

. The first outbreak of human plague in Lushoto District, north-east Tanzania. Trans Roy Soc Trop Med Hyg, 1982; 76:172–177.

11.

Kim

. Evolutionary relationships of parasitic arthropods and mammals. Kim

. Co-evolution of Parasitic Arthropods and Mammals. A Will-Inter-science Publication, John Wiley and Sons: New York, 1985; 3–80.

12.

Lewis

. Notes on the geographical distribution and host preferences in the order Siphonaptera. Part 1: Pulicidae. J Med Entomol, 1972; 9:511–520.

13.

Makundi

, Kilonzo

, Massawe

. Interaction between rodent species in agro-forestry habitats in the western Usambara mountains, north-eastern Tanzania and its potential for plague transmission to human. Singleton

, Hinds

, Krebs

, Spratt

. Rats, Mice and People: Rodent Biology and Management. Australian Centre for Agricultural Research: Canberra (ACIAR), 2003; 124–130.

14.

Manson-Bahr

PEC

, Bell

. Plague and melioidosis. Manson-Bahr

PEC

, Bell

. Manson's Tropical Diseases. London: Bailliere Tindal, 1987; 586–602.

15.

Nelson

, Bell

, Clifford

, Keiran

. Interaction of ectoparasites and their hosts. J Med Entomol, 1977; 13:389–428.

16.

Nelson

, Kierans

, Bell

, Clifford

. Host-parasite relationship. J Med Entomol, 1975; 12:143–166.

17.

Pugh

, Parker

. Plague: Rhodesia's first recorded outbreak. Cent Afr J Med, 1975; 21:92–96.

18.

Service

. Lecture Notes on Medical Entomology. London: Blackwell Scientific Publications, 1986; 265.

19.

Shephard

, Hummitzsch

, Leman

, Hartwig

. Studies on plague in the eastern Cape Province of South Africa. Trans Roy Soc Trop Med Hyg, 1983; 77:800–808.

20.

Shephard

, Leman

. Plague in South African rodents 1972–81. Trans R Soc Trop Med Hyg, 1983; 77:208–211.

21.

Smithers

RHN

. Guide to the Rats and Mice of Rhodesia. Publication of the Trustees of the National Museum: Rhodesia, 1975; 50.

22.

Stevenson

. Arthropod-borne infections. Stevenson

. Davey and Lightbodys' The Control of Disease in the Tropics. London: Lewis, H.K. and Co. Ltd, 1987; 265–273.

23.

Taylor

, Gordon

, Isaacson

. The status of plague in Zimbabwe. Ann Trop Med Parasit, 1981; 75:165–173.

24.

WHO. Plague Manual. Geneva: WHO, 1999; 171.

25.

WHO. Weekly Epidemiological Record. Geneva: WHO, 1976; 51:344.

26.

WHO. Weekly Epidemiological Record. Geneva: WHO, 1983; 35:265–266.

27.

Zimba

. Species composition, abundance, seasonal variations and reproductive patterns of rodents and fleas captured in three habitat types of Hatcliffe and Dzivarasekwa suburbs of Harare. ZimbabweD.Phil. thesis2008; 1–220.