Distribution of Necator americanus and Ancylostoma duodenale in Liberia

Bull. Org. mond. Sante 1972, 47, 317-324 Bull. Wld Hith Org.f Distribution of Necator americanus and Ancylostoma duodenale in Liberia H. C. HSIEH,1 N. R. STOLL,2 E. W. REBER,3 E. R. CHEN,' B. T. KANG,4 & M. KUO 5 In the first country-wide survey ofhookworm infection in Liberia, diagnosis by test-tube filter-paper cultivation revealed a prevalence of more than 90 %. This sensitive technique also permitted differential species determination. Necator americanus was found to occur in all regions and in a high proportion of the persons examined. Ancylostoma duodenale was mainly confined to a narrow coastal belt, and usually parasitized individuals harbouring N. americanus as well. The average number ofeggs per gram offaeces (EPG) ofinfected individuals was lowest in the population of the hinterland. During the latter part of the studies a modified test-tube filter-paper cultivation technique was used in combination with dilution counting to obtain data on intensity of infection separately for the 2 species. For N. americanus there was little regional difference in the intensity of infection, but the EPG for A. duodenale was much higher in people living in the coastal area. The species of hookworm found most frequently in Liberia is Necator americanus (Stiles, 1902). During a survey conducted for the World Health Organization, Stoll & Hsieh " found Ancylostoma duodenale there in June 1961. In the investigation reported in this paper further information was sought on the extent of infection with A. duodenale and with N. amnericanus in all 9 counties of Liberia. The study extended from April 1964 to December 1968. Along the coast of Liberia there is a relatively level sandy strip, about 10-16 km wide, cut by lagoons, tidal creeks, and marshes. The interior consists of rolling countryside except for a belt of tablelands and a number of steep mountain ranges. There are 6 main rivers and numerous smaller streams, which flow almost entirely over rock and have hardly any well developed valleys or flood-plains. As reported by Gnielinski (1967) Liberia lies within the tropical rainforest belt and has a short dry season. It has an even and fairly high temperature 1 Present address: Kaohsiung Medical College, Kaoh- siung, Taiwan. The Rockefeller University, New York, N.Y., USA. Liberian Institute of the American Foundation for Tropical Medicine, Harbel, Liberia. ' Present address: International Institute of Tropical Agriculture of the Ford & Rockefeller Foundations, Ibadan, Nigeria. ' Present address: Taiwan Provincial Pingtung Institute of Agriculture, Ping Tung, Taiwan. ' Stoll, N. R. & Hsieh, H. C., unpublished data, 1961. and high humidity. Rainfall is higher along the coast than in the interior. Monthly mean temperatures vary from 26.7° to 32.2°C during the day, the maxi- mum being 37.8°C, and from 21.10 to 23.9°C during the night, the minimum being 7.2°C. The relative humidity is 85-95% during most of the year. Villages comprising from 2 to 600 or more huts are scattered throughout the forest and along highways and the coast. Most of them are clean but there are no general systems of sewage disposal, drainage, or water supply. Water for drinking is taken from wells, rivers, and streams. Very few latrines are to be found and the people may urinate and defecate in the streams or in the bushes. As a result, the secondary bush surrounding villages and towns is considered to be the most important source of infection with hookworm. Manures such as human nightsoil are not applied to the soil and farms are not regarded as an important source of infection. MATERIALS AND METHODS Field surveys were begun in the Charlesville Dis- trict of Marshall Territory (Monteserrado County) and were gradually extended inland and to other parts of the coast until all 9 countries had been in- cluded. Most faecal specimens were examined at the laboratory of the Liberian Institute of the American Foundation for Tropical Medicine in Charlesville; 2902 317 H. C. HSIEH ET AL. those from Lofa County were examined at the field station of the Liberian Institute in Voinjama, Lofa County. Most specimens were collected by a tech- nician early in the morning and promptly brought to the laboratory by car. Those from Maryland, Sino, and Grand Gedeh Counties were transported by air to Roberts International Airport and from there to the Liberian Institute 8 km away by car. Specimens for examination were sought primarily from schoolchildren, but also from people of the vil- lages. After the purpose of the work had been explained to the head teachers and village chiefs and to the people directly concerned and after lists of prospective donors had been drawn up a waxed card- board carton with a tight-fitting lid was numbered for each person both on the lid and on the carton. Prospective donors were asked to furnish not less than 10 g of faeces. For school surveys the con- tainers were handed to students in the morning and collected within a few hours. For village surveys the containers were usually distributed in the evening and collected early the next morning. In the laboratory, the cartons were arranged in numerical order and the consistency of each sample was recorded according to the following classifica- tion: F (formed), SF (soft-formed), M (mushy), MD (mushy diarrhoeal), and D (diarrhoeal). Each sample was examined by test-tube filter- paper cultivation (Harada & Mori, 1951), as de- scribed by Hsieh (CCTA/WHO African Conference on Ancylostomiasis 1963), and by dilution counting (Stoll, 1962). When a sample was too small to be examined by the dilution technique, diagnosis and differentiation of the two hookworm species (as well as detection of Strongyloides stercoralis, if present) was possible by means of test-tube filter-paper cul- tivation alone. Stools that were shown to be positive for hook- worm solely by culture were assigned an arbitrary number of eggs per gram (EPG) of 100. After a series of specimens had been examined the counts were listed with appropriate notation of the consistency of the stools. The counts were then brought to " basis F ", by multiplying by 1.5 for SF specimens, by 2 for M specimens, by 3 for MD spe- cimens, and by 4 for D specimens. In the second half of the 1964-68 study period, a modified test-tube filter-paper method (MTFC) was introduced (Hsieh, 1971). This made possible, after a culture period of about a week, the examination not only of the filariform larvae that had reached the water at the bottom of the test tube during the cultivation period, but also of slower larvae that still remained on the faecal smear on the filter paper. Thus the MTFC test allowed reliable determinations of the ratio of the 2 species present per culture. The ratio of Necator to Ancylostoma obtained by the MTFC method when applied to the total dilution egg count gave the number of eggs to be recorded for each species. RESULTS Proportion ofpersons infected In the period from April 1964 to December 1968 faecal specimens were examined from nearly 6 000 in- dividuals in 75 localities in all 9 counties of Liberia (Fig. 1). Altogether 4 390 schoolchildren were examined in 62 schools, with from 14 to 214 pupils each. In 54 of these schools 90% or more of those examined were positive; in none of the other 8 were fewer than 55% positive. In each county 100-1 468 schoolchildren were examined, and the proportion that were positive for hookworm ranged from 90% to 97 %. In addition it was possible to examine 1 573 spe- cimens from a representative sample of the general population of the western part of Liberia. Of 901 re- sidents of different ages in 7 villages in Montserrado county, 91 % were positive, and of 507 from 3 vil- lages in Lofa county, 80% were positive. In one village in Grand Bassa County 87% of those ex- amined (54 persons) were positive, and in Nimba County (59 examined) 63% were positive. While in 12 villages the proportion of those examined that were positive for hookworm was 63- 100%, Gbatala Camp in Bong County was an un- usual exception. More than 100 American Negroes emigrated from the Chicago area of the USA in the summer of 1967 and started a new community in this locality, living in tents or temporary huts. Our survey of 52 individuals in April 1968 revealed no hookworm infections. N. americanus was found in a high proportion of those examined all over the country. Considered county by county, 88-96% of the schoolchildren and 63-87% of the general population of the villages we examined had this parasite. In contrast A. duodenale was found principally in the coastal counties, infections being especially pro- minent in the sandy lowland belts and near the mouths of rivers. The proportions of people examined that were positive for A. duodenale in the different counties 318 DISTRIBUTION OF NECATOR AMERICANUS AND ANCYLOSTOMA DUODENALE IN LIBERIA Fig. 1. The 75 examination centres in the 9 counties are close to the laboratory of the Liberian Institute. were as follows: Maryland 10%, Sino 48%, Grand Bassa 27%, Montserrado 33%, and Grand Cape Mount 9%. In the inland counties strikingly small proportions of the people examined were positive for this species, the averages being only 2-4% for Grand Gedeh, Nimba, Bong, and Lofa. In these four counties all individuals found positive for A. duo- denale were also infected with N. americanus. Many of the persons living inland who were positive for A. duodenale were later interviewed. They proved to be nonindigenous inhabitants or Liberians who had lived in the sandy coastal areas for different periods in recent years. Wherever the proportions positive for A. duo- denale were small in the coastal areas it was where high hilly land was close to the coast-e.g., near the promontories of Grand Cape Mount and Maryland counties, where the dominant soil is lateritic Parasitism with A. duodenale usually occurred as a double infection with N. americanus. However, in 20 of the 75 localities studied, infections with A. duo- denale alone were found in 1-22% of the individuals. of Liberia. Stations 5 and 6 in Montserrado County Average intensity of infection with hookworm When dealing with hookworm infections the num- ber of eggs per gram of faeces should always be in- terpreted with caution; although the EPG reflects the number of worms harboured, it is not necessarily exact and differs for the two species. A. duodenale females are more prolific egg layers, by a factor of at least 2, than are those of N. americanus. Data on intensity of infection with hookworm, sensu latu, are available from 1 564 children 5-14 years of age, and are given in Table 1 in relation to topographical region. For convenience we have designated 3 regions. The "6 coastland " is the sandy belt of coastal lowland about 10 km wide in the counties of Montserrado, Grand Bassa, and Sino and extending partially into Maryland and Grand Cape Mount (Fig. 1). The " midland " is somewhat arbitrarily limited to the region 11-50 km from the coast, comprising lowland interspersed with sandy areas and seasonal swamps along the lower parts of rivers. The term " hinter- 319 H. C. HSIEH ET AL. Table 1. Proportion of children with hookworm (without regard to species) and average EPG in three different regions No. of childiren ProportionRegion examined positive for Average EPGhookworm(% coastland 292 99 5 490 midland 628 94 5 290 hinterland 644 94 3 380 land" refers to the rolling and hilly countryside 51 km or more from the coast and near the pro- montories of Grand Cape Mount County and Maryland County. Throughout the three regions there was a striking similarity in the proportions of children with hook- worm infections, but there were differences in the intensity of infection, as shown in Table 1. Average intensity of infection with N. americanus and A. duodenale The 1 564 children for whom data are presented in the tables were those from whom specimens were available for both the dilution counting and the MTFC tests (Hsieh, 1971). Using these two tech- niques it was possible to obtain separate quantitative information on each species (Table 2). The highest EPG for children with N. americanus was found in the midland, the figures being almost one-third smaller in the coastland and the hinterland. Table 2 also shows that there was wide variation in the pro- portion of children with N. americanus infections only, the figure being highest in the hinterland; the figures for all N. americanus infections, however, showed great similarity. The EPGs of those who were positive for A. duodenale differed greatly in the three regions and were somewhat parallel to the pro- portions of infections found in the same areas. There were few infections with A. duodenale alone, and none were found in 1 021 samples from the hinter- land counties of Bong, Nimba, and Grand Gedeh. The intensity of infection with hookworm, sensu latu, in the hinterland was almost identical with that for N. americanus, and the figures for the midland did not differ greatly. In the coastland, however, there was a considerable difference, accounted for by A. duodenale infections. Subgroups of intensity of infection with hookworm Additional insight into parasitism is furnished by dividing the subjects into subgroups on the basis of their intensity of infection, as shown in Table 3. In- fections that were diagnosed only by culture and that contained too few eggs (<400) to be detected by routine dilution tests are shown separately in Table 3 and Fig. 2; they were arbitrarily assigned an EPG value of 100. When the infections were considered without regard to species over half the children showed an Table 2. Proportions of the population with pure and mixed infections of N. americanus and A. duodenale, and average EPG in relation to region Coastland (292) b Midland (628) b Hinterland (644) b Infections a Average Proportion Average Proportion Average Proportion EPG positive EPG positive EPG positive (%) (%) (%) Na only Na when with Ad Na total Ad only Ad when with Na Ad total Total Na plus total Ad 5 010 24 4 750 65 3 290 91 2 800 68 4 460 29 3 030 3 3 380 92 4 660 94 3 280 94 4 730 7 4 290 0 0 1 840 68 500 29 100 3 2110 75 630 30 100 3 5 490 99 5 290 94 3 380 94 a Na = Necator americanus; Ad = Ancylostoma duodenale. b The numbers of children examined in each region are shown in parentheses. 320 1 DISTRIBUTION OF NECATOR AMERICANUS AND ANCYLOSTOMA DUODENALE IN LIBERIA 321 Table 3. Subgroups of intensity of infection with N. americanus and A. duodenale Percentages in subgroups of EPG subgroups Naa in Ada in Hookworm sensu latu a Boys Girls Boys Girls very high (10 000) b 9 10 5 1 1 high (5 000-9 900) 16 20 8 3 2 moderate (3 000-4 900) 16 17 16 2 2 low (400-2 900) 48 44 55 16 14 very low (<400) 5 4 7 6 6 negative 6 5 9 72 74 total 100 100 100 100 100 number examined 1 564 1 038 526 1 038 526 a Na, N. americanus; Ad, A. duodenale. b The highest EPGs were: 58 480 for N. americanus and 22 800 for A. duodenale. 80 60- A. duodenole in 1064 M -in 526 F 40 0. 20- 0 0 60C7 _ 0) IV. omericonus 40 - in 1064 M - in 526 F 20 Neg L M H VH Neg L M H VH Fig. 2. Percentage distribution of hookworm infections according to EPG in 1 064 male and 526 female children in Liberia. The letters at the bottom of the figure indicate the intensity of infection (L, low; M, moderate; H, high; VH, very high). The stippled part of the L column indicates the percentage of infections of very low intensity. EPG lower than 3 000, but only 6% were negative. However, 25% were in the two highest EPG sub- groups. The data for the different species show that the greater part of the intensity of infection was ac- counted for by N. americanus, and the subgroups for the boys for this species were similar to those for hookworm sensu latu. The proportions of girls in the two highest subgroups were only half (or less than half) of those shown by the boys, while in the lowest subgroups the proportions of girls exceeded those of boys. For A. duodenale infection, the sub- groups were similar for both boys and girls. DISCUSSION The above results are of interest in relation to data collected several decades ago, in neighbouring Sierra Leone, where presumably the Necator-Ancylostoma ratio was similar to that in Liberia. Gordon (1925), using the dilution counting technique, examined 137 male natives of Freetown. Our divisions at 3 000, 5 000, and 10 000 EPG can be used to classify the 114 positive cases (83 %) reported by Gordon as 53% low (including very low), 10% intermediate, 9% high, and 11 % very high. The overall average EPG for the positive cases was 4 840, basis F.1 These figures differ but little from the results of our study. 1 See p. 318. H. C. HSIEH ET AL. Four other studies carried out in different coun- tries offer interesting comparisons, although the data, including the age and EPG groupings, are not all parallel. Hill (1927) reported that a community at San Lorenzo, Puerto Rico, had an average EPG of 8 330, basis F, the highest that had been found in that country. According to the author, the area was a " self-limited" long valley that had never been influenced by sanitation. Of the 261 males and 250 females aged from 3 to more than 50 years examined by dilution counts 98% were positive and 28% showed on EPG of 11 100 or more. Among the 207 children tested, aged 4-14 years, 206 were posi- tive with an average EPG of 6 690. A high infection rate was found by Fulleborn et al. (1928) and Fulleborn (1929) in Corrientes, Argentina, where dilution counting showed 96% of 396 persons of both sexes aged from 3 to more than 50 years to be positive. The average EPG was 5 200, basis mushy stools, 25% of the persons had an EPG exceed- ing 5 000 and 11 % one exceeding 10 000). The average EPG for 257 boys and girls aged 6-15 years was 3 750, basis mushy stools, and 11 others had EPG counts in excess of 20 000. By contrast, a lower infection rate was found by Sweet (1929) in Mysore State, India. Examination of 11 972 people of both sexes and all ages by the Willis salt flotation and dilution counting technique showed 59% to have hookworm. The average EPCC (con- sidered to be equivalent to EPG) was 1 600, basis F; 46% of 2 825 children 5-14 years of age were positive and had an average EPCC of 1 340. In the three areas mentioned above-Puerto Rico, Argentina, and India-the predominant parasite was N. americanus; in addition, an estimated 5-10% of the persons examined had A. duodenale infections. In the south of the USA N. americanus was identified as a new species and the sole species by Stiles (1902). The prevalence and intensity of hookworm infection was reexamined in 8 of the southern states in 1930- 38. Keller et al. (1940) reported that 19.3 % of the 424 511 persons of the rural white population examined by dilution counting were positive, and that hookworm was present in all but 13 of the 441 counties investigated. Of 132 218 white children 5-14 years of age living in rural areas, 18.7% were positive with an average EPCC of 3 430, basis F. In the same period a comparative study of rural white and black populations was undertaken in Florida, Mississippi, North Carolina, and South Carolina. The results showed that 21.5% of the former (59 208 persons, average EPCC 3 400) and 5.3% of the latter (17 458 persons, average EPCC 1 600) were infected. Maass (1927) reported that 96% of the 276 people he examined were positive for " Ancylostomum " at the Mission Hospital at Bolahun, Lofa County, Liberia. This lies within the area in which N. ameri- canus is dominant in the northwestern hinterland but Maass specifically mentioned finding both species. Despite the high prevalence of hookworm he saw only 3 cases of severe anaemia resulting from this infection. Schwab (1947) described hookworm as very common in Liberia, with perhaps 50% of the population in the hinterland harbouring this parasite and suffering from chronic anaemia and decreased vitality. Poindexter (1949) found, on the basis of a single smear, only 59 positive cases among 4000 dispensary patients examined in Monrovia over a period of 19 months in 1947-48, and at the Ganta Mission dispensary in Nimba County, near the Guinea border, only 6% of 550 individuals were positive. However, at the Ganta Leprosy Station 48% of 150 patients showed hookworm ova. Poin- dexter (1953) later reported on a study of 3 000 mem- bers of the Gola tribe in Lofa County, but noted the results of examinations for only 100 adolescent boys and 100 girls, of whom 53 and 36, respectively, showed hookworm. Bray (1958), in a brief survey for intestinal protozoa by zinc sulfate concentration, found hookworm ova in 94% of 100 persons of all ages living in the Charlesville-Harbel district. Vogel (1958) noted a hookworm prevalence of 86% in 22 localities in the Central Province of Liberia during a schistosomiasis survey using the Teleman concentration method. He examined 1 136 persons of 5 years of age and over, and in half the villages over 90% were positive. The first demonstration of the presence of A. duo- denale, sensu strictu, in Liberia appears to be that by Stoll & Hsieh (1961), who examined stools from 193 persons in the Charlesville-Harbel area by test- tube filter-paper cultivation. The findings resembled those made in 243 individuals near Brazzaville: in the two groups combined 34% of the 436 persons were positive for A. duodenale. In contrast, 127 persons from villages near Ibadan, Nigeria, examined, in the same way were negative for A. duodenale. Of the total of 563 individuals about one-third were under 15 years of age, and the overall prevalence ofN. ame- ricanus was 87 %. The presence of A. duodenale in Liberia was verified by Franz et al. (1964) in a study of 524 patients at the Firestone Hospital, Harbel. Faecal 322 DISTRIBUTION OF NECATOR AMERICANUS AND ANCYLOSTOMA DUODENALE IN LIBERIA 323 specimens from 263 hospital patiens and 261 out- patients of both sexes, aged 4 months to 60 years, were selected at random for test-tube cultivation after hookworm eggs had been found by direct smear in routine hospital laboratory examinations. Of those specimens already known to contain hook- worm eggs, 16% were found to be positive for A. duodenale, about two-thirds of them showing mixed infections with N. americanus. Davis & Vardy-Cohen (1962), in a health survey of 16 schools in Monrovia, found 13.2% of 912 pupils positive for hookworm by the thiomersal-iodine- formaldehyde method. This is higher than the 1.5% reported from Monrovia by Poindexter (1949) for 4 000 dispensary patients examined by means ofsingle smears. Both figures, however, are much lower than those found in our examinations of schoolchildren in West Point, a slum suburb of Monrovia near the harbour, where more than 90% of those examined were positive with a high frequency of A. duodenale accompanying the N. americanus. Although our survey showed N. americanus to be present in a large proportion of those examined throughout Liberia, high endemicity of A. duodenale was found only on the coastal belt within approxi- mately 50 km of the Atlantic Ocean, in the counties of Montserrado, Gran Bassa, and Sino. The highest proportion of A. duodenale infections (94%) was found in Fawen, 10 km from the coast. The total A. duodenale rate exceeded 50% in the following 9 localities (the distances in kilometres from the ocean are shown in parentheses): Lower Careysburg (20), Upper Careysburg (25), Snafu (3), Zopehtown (0.5), West Point Morning and West Point After- noon (0.5), Duazon (0.5), and Gemeh (3) in Mont- serrado County, and Greenville (1) in Sino County. Many migrants from the Liberian hinterland who live on the coast show A. duodenale. It has already been noted that A. duodenale may be considered either absent from or nonendemic in the hinterland. The climate of the hinterland differs from that of the coast, but is within the range in which A. duodenale is endemic in other parts of the world. Furthermore, the diet and the customs of the rural tribes are on the whole similar throughout Liberia. Further in- vestigations are necessary to elucidate the reasons for the unusual distribution of A. duodenale in the hinterland and in the coastal region of Liberia.' 1 For a study of soil characteristics in relation to the distribution of the two species in Liberia, see Hsieh et al. (1971). ACKNOWLEDGEMENTS The authors are grateful to Mrs E. W. Reber, Dr W. A. Sodeman, Mr and Mrs J. Baum, Mr N. G. N. Williams, Mr G. Tamba, Mr J. K. Davis, Mr D. D. Jallah, Mr S. K. Logan, Mr H. Zusu, and other colleagues of the Liberian Institute of the American Foundation for Tropical Medicine for their help during the survey. The authors also thank the town chiefs and schoolteachers for their cooperation in facilitating the collection of faecal samples. The assistance of the following persons is also acknowledged: Mr H. W. Hermerling, Manager, and Dr C. E. Smale, Firestone Cavalla Plantation Co., Maryland County; Mr G. Braams, General Manager, and Dr W. Bucken, African Fruit Co., Sino County; the Reverend and Mrs J. Lynn, Grand Gedeh Mission; the Reverend and Mrs R. Prunty, Bendu Mission, Grand Cape Mount County; and the physicians at the Ganta Hospital, the Methodist Mission in Nimba County. Acknowledgement is made of the invaluable support and administrative help kindly given by the National Public Health Service of the Republic of Liberia, the Liberian Medical Association, the Firestone Plantation Co. at Harbel, and the airport authorities at Robertsfield. Gratitude is also expressed to Miss R. Mandlebaum, Graphic Services, The Rockefeller University, and to Miss H. E. Hesse, Mrs E. G. Gamache, and Mr J. F. Salisbury, American Foundation for Tropical Medicine. This study was supported by US Public Health Service Research Grant No. AI-05775 from the National Insti- tutes of Health. RISUM1 REPARTITION DE NECATOR AMERICANUS ET D'ANCYLOSTOMA DUODENALE AU LIBtRIA De 1964 'a 1968, on a recherch6 la prevalence de I'ankylostomiase dans un groupe de pres de 6000 sujets de tous 'ages, dont 4390 6coliers, habitant dans tous les comt6s du Lib6ria. L'examen des feces a et6 effectu6 par la m6thode de culture sur papier filtre en tube ai essai et par numeration des aeufs selon la m6thode de Stoll. La grande sensibilit6 de ces techniques a facilit6 le diagnostic, a permis de differencier dans chaque cas 324 H. C. HSIEH ET AL. Necator americanus et Ancylostoma duodenale et a fourni des donnees numeriques concernant l'intensite relative de l'infection par l'une et l'autre espece. Dans presque toutes les localit6s prospectees, la pr6- valence de l'ankylostomiase etait de l'ordre de 90%, aussi bien chez les enfants que chez les adultes. On rencontrait N. americanus chez une proportion elev6e des sujets examines dans toutes les regions du pays. Par contre, A. duodenale infectait principalement les habitants des zones c6tieres. L'interrogatoire des sujets vivant a l'int6rieur du pays et infectes par A. duodenale a r6vele qu'il s'agissait d'etrangers ou de Lib6riens ayant sejourne dans les regions cotieres pendant un temps plus ou moins long. Les num6rations d'aeufs ont montr6 une intensit6 de l'infection par N. americanus tres semblable dans toutes les regions, alors que le parasitisme a A. duodenale, chez les habitants de la cote, etait particulierement marque. REFERENCES Bray, R. S. (1958) Trans. roy. Soc. trop. Med. Hyg., 52, 186 CCTA/WHO African Conference on Ancylostomiasis (1963) Wld Hith Org. techn. Rep. Ser. No. 255 Davies, A. M. & Vardy-Cohen, D. (1962) W. Afr. med. J., 11, 207-214 Franz, K. H. et al. (1964) W. Afr. med. J., 13, 110-113 Fulleborn, F. et al. (1928) Arch. Schiffs- u. Tropenhyg., 32, 441-481 Filleborn, F. (1929) Brit. med. J., 1, 755-759 Gnielinski, S. V. (1967) University of Liberia Journal, 7, 9-19 Gordon, R. M. (1925) Ann. trop. Med. Parasit., 19, 429-460 Harada, Y. & Mori, 0. (1951) Igaku to Seibutsugaku, 20, 65-67 (In Japanese) Hill, R. B. (1927) J. preventive Med., 1, 537-545 Hsieh, H. C. (1971) Chin. J. Microbiol., 4, 25-39 Hsieh, H. C. et al. (1971) Chinese J. Microbiol., 4, 132-156 Keller, A. E. et al. (1940) Amer. J. trop. Med., 20, 493-509 Maass, E. (1927) Abh. Geb. Auslandsk., Bernard Nocht Festschrift, pp. 268-273 Poindexter, H. A. (1949) Amer. J. trop. Med., 29,435-442 Poindexter, H. A. (1953) Amer. J. trop. Med. Hyg., 2, 30-38 Schwab, G. (1947) Tribes of the Liberian hinterland, Cambridge, Mass., USA, Harvard University (Peabody Museum of American Archaeology and Ethnology, Paper No. 31 Stiles, C. W. (1902) Amer. Med., 3, 777-778 Stoll, N. R. (1962) Ann. N. Y. Acad. Sci., 98, 712-724 Sweet, W. C. (1929) In: A Survey of Mysore State for enlarged spleens andfor hookworm and other helminthic infections, Bangalore, Mysore Department of Health, pp. 29, 37 (Mysore Department of Health Bulletin No. 7) Vogel, H. (1958) Bilharziasis survey in the Central Pro- vince In: Annual Report of the Research Activities of the Liberian Institute of the American Foundation for Tropical Medicine, Harbel, Liberia, pp. 40-47