Observations on the relative attractiveness to Glossina pallidipes of different animal baits, a tsetse trap, and a fly-round patrol

Bull. Org. mond. Sante 1 1972, 47, 789-793Bull. Wid Hithi Org. Observations on the relative attractiveness to Glossina pallidipes of different animal baits, a tsetse trap, and a fly-round patrol E. C. ENGLAND 1 & D. A. T. BALDRY 2 Studies conducted in the Lambwe Valley, Kenya, have shown that in the absence of wild hosts, Glossina pallidipes was more attracted to a calf than to a sheep, a goat, a man, or a tsetse trap, although the latter attracted more flies than the nonbovid baits. Other investigations have shown that a newly developed tsetse trap was much more efficient at catching G. pallidipes than a standard fly-round patrol technique. An added advantage of the trap was that the sex ratio of flies caught by it was more representative of that existing in the fly population. Fly-round patrols always caught many more male than female flies. In the mid-1960s Langridge (1972) developed a tsetse trap for collecting Glossinapallidipes, which is economically and medically the most important species of tsetse in Kenya. The trap, which is now commonly known as the " Langridge " trap, is basically a modification of the trap developed by R. H. T. P. Harris in the 1920s for collecting G. palli- dipes in Zululand (see Buxton, 1955). During a variety of entomological investigations conducted in the Lambwe Valley, Kenya, from 1968 to 1971, the Langridge trap was used extensively to sample G. pallidipes populations. This account is therefore largely concerned with appraisals of the merits of the trap in comparison with the standard fly-round patrol technique of sampling G. pallidipes. Some observations on the relative attractiveness of the trap, man, and domestic animals to G. pallidipes are also described since this is relevant to the epidemiology of the trypanosomiases in the study area. MATERIALS AND METHODS All investigations were conducted in the Lambwe Valley, South Nyanza District, Kenya, a description of which has been given by Allsopp & Baldry (1972). 1 Project Technical Officer (Entomology). Present address, WHO, P.O. Box M 190, Accra, Ghana. ' Project Scientist-Entomologist. Present address, WHO Regional Office for Africa, boite postale 6, Brazzaville, People's Republic of the Congo. The results of studies on the hosts and trypanosome infection rates of G. pallidipes in this area have been described by England & Baldry (1972), while Allsopp et al. (1972) have discussed the relationships between G. pallidipes and the game animals resident in the area. The studies reported here on the relative attrac- tiveness to G. pallidipes of domestic animals, man, and the Langridge trap were conducted during February and March 1969 in three localities-the Obaluanda area, the new Ponge Dam at the eastern end of the Roo Valley, and near a small dam towards the western end of the Roo Valley. In each of the three areas a natural clearing was found inside a G. pallidipes thicket habitat, which was large enough easily to accommodate 5 baits around the periphery. The baits used were a calf, a sheep, and a goat, all of which were caged, a Langridge tsetse trap, and a man. Each cage was newly constructed and a modified Langridge trap8 was attached to the top. Each bait animal remained in its original cage throughout the investigation. Baits were arranged equidistantly around the clear- ing and their positions were rotated clockwise each day. To minimize the chance of tsetse being carried into the clearings by operators or vehicles, no vehi- cles were allowed within 200 m of a clearing, and 'Designed by Mr G. R. Jewell, formerly Project Tech- nical Officer (Entomology). 2967 -789- E. C. ENGLAND & D. A. T. BALDRY all persons approaching on foot were examined for following flies. The working day consisted of 4 1-hour periods. At the start of each day's study, the bait animals were placed in their cage, the Langridge trap was erected, and the man acting as bait took up his position in company with an attendant whose job it was to catch, with a hand net, flies alighting on him. At the end of each hour, the animals and the man acting as bait were allowed to rest for 15 min, the former being released from their cages. The cages of the Langridge trap and of the caged ani- mals were removed and the tsetse flies counted, classified, and recorded. The flies were not released until the end of the working day, some flies being kept for dissection in order to determine the trypano- some infection rates. Observations were made for 9 days at the Oba- luanda and Ponge Dam sites and for 13 days at the other dam site. For studies on the relative merits of Langridge tsetse traps and fly-round patrols, traps were made and used in the manner described by Langridge (1972). In using these traps three important points were always kept in mind-namely, that they must be carefully sited in a partly open and partly shaded position in order to realize their greatest catching potential; that their efficiency might decline if they were not repositioned at least 150 m from their previous position, at least once a month; and that their efficiency might fall with age, unless they were reconditioned monthly. In practice it was found convenient to replace traps with reconditioned ones every month and to change their positions at the same time. Traps were usually used for two 24-hour catching periods per week. During the intervening periods the traps were left open, i.e., with the cage removed, so that any flies that entered could easily escape. Patrols involved a minimum of four men. Two men carried between them a black cloth screen suspended from a wooden pole; with the ends of the pole placed on the shoulders, the hands of the men were free to catch, by means of hand nets, flies alighting on either side of the screen. A third man determined the sex and condition of the flies that were caught and placed them in a cage for subsequent release on completion of the patrol. A fourth man recorded on a special data sheet the condition of each fly in relation to the position of the patrol team. Patrols operated along defined paths (fly-rounds) of a convenient and known length, so that apparent densities (i.e., the number of nonteneral males per 10 000 m) could be calculated if desired. Fly-round patrols were conducted at the same time each morn- ing, but in opposite directions on alternate days in order to compensate for any irregularities caused by flies following the catching party. Studies on the comparative merits of traps and fly-round patrols, and on the decline in the efficiency of traps kept in one position for a long period, were made in the Otuok thicket area, while the effects of ageing on trap efficiency were studied in the Ruma thicket area. All investigations were made between 1968 and 1970. RESULTS The relative attractiveness to G. pallidipes ofdomestic animals, man, and the Langridge tsetse trap During the first 23 days of the investigation the baits available to G. pallidipes were a calf, a sheep, a goat, a man, and a Langridge trap. The results (Table 1) clearly show that the calf was the most attractive bait, more than half the total catch of ffies being obtained from this animal. During the last 8 days of the investigation the calf was removed from the experimental area, with the result that almost equal proportions of flies were attracted to the remaining baits (Table 1). Table 1. Numbers of G. pallidipes collected from bait animals, a Langridge tsetse trap, and man, in the presence and absence of a bait calf Numbers and percentages of flies attracted to different hosts and to a tsetse trap Bait With calf present a Without calf b No. % No. % calf 1 101 54.1 not applicable not applicable sheep 172 8.4 55 25.2 goat 132 6.4 47 21.5 man 260 12.7 63 28.9 trap 369 18.1 53 24.3 totals 2 034 99.7 218 99.9 a 23 4-h working days. b 8 4-h working days. 790 ATIRACTIVENESS OF VARIOUS BAITS .TO GLOSSINA PALLIDIPES When the calf was available, the proportions of G. pallidipes feeding on the three species of domestic animal were as follows: calf, 40.7%; sheep, 1.2%; goat, 0. When the calf was removed, 9.1 % of the flies attracted to the sheep fed on it while none of the flies attracted to the goat attempted to feed. Comparisons between fly-roundpatrols and Langridge trap catches Between 15 December 1969 and 12 January 1970, the G. pallidipes population of part of the Otuok continuous thicket was concurrently sampled by fly-round patrols and by trapping. Five traps were positioned equidistantly along the fly-round (1 460 m), each trap being allowed to catch flies for two 24-hour periods per week. The twice-weekly fly-round patrol, conducted by a standard patrol team, took approxi- mately 3 hours to complete on each occasion. During the study period, the 5 traps were emptied 6 times and caught a total of 7 749 flies. This re- presents a mean catch of 1 291 flies for the 5 traps per round, and a mean individual trap catch of 258 flies. The proportion of females was 63.9%. At the same time, 7 screen patrols collected 837 flies, i.e., 119 flies per patrol. The proportion of females was 18.9%. Under the conditions of the patrol route it is unlikely that the traps had an attractive influence on the fly population farther than 150 m on either side and it can therefore be assumed that the 5 traps were probably effective along the whole 1 460-m length of the round. On this basis, the efficiency of the traps at catching G. pallidipes was over 10 times that of the patrols. At the end of November 1970, both sampling techniques were employed along a transect in another part of the Otuok thicket. A Langridge trap located at the thicket edge produced a mean trap catch of 902 G. pallidipes (3 trappings); 70.8% of these flies were female. The mean patrol catch along a 30-m sector traversing the thicket edge was 7.3 G. palli- dipes (3 patrols) during the same period; 27.8% of these flies were female. A second trap, located 150 m inside the thicket, produced a mean catch of 594 G. pallidipes (3 trap- pings); 71.1 % of these flies were female. The mean patrol catch along a 90-m section of the round adjacent to the trap, was 14.3 flies, of which 23.3 % were female (3 patrols). A third trap was located 90 m farther inside the thicket. The mean trap catch was 895 G. pallidipes of which 72.3% were female. The mean patrol catch along 90 m of fly-round adjacent to the trap was 23 flies, of which 33.4% were female. None of the three traps was positioned directly on the patrol round. They were concealed a few metres inside the thicket and thus not readily visible to flies moving along the round. Considering the density of the thicket, it also seems improbable that they could be seen through the thicket vegetation by flies as far away as 100 m. Therefore, if it is assumed that the traps were effective across a dis- tance of only 90 m in any one direction, it can be concluded that in all cases the number of G. palli- dipes caught by the traps was approximately 20 times that caught by the corresponding patrols. The decline in the efficiency of traps kept in one position for a prolonged period As stated earlier Langridge-type traps were be- lieved to decline in efficiency if they remained in any one position for more than about a month. This was amply confirmed by data collected in the Otuok area between April and November 1970. Five traps were located in isolated thicket clumps adjacent to the Otuok thicket in mid-April 1970. Although, the traps were replaced by new or recon- ditioned ones monthly, their positions were not changed until September 1970. The performance of these traps declined until they were repositioned, whereupon their performance increased remarkably (Table 2). The decline in the efficiency of traps as a result of ageing In the course of aerial spraying trials it seemed likely that the weathering of traps might be causing Table 2. The effect of non-repositioning of Langridge tsetse traps on the catches of G. pallidipes in Otuok thicket, Lambwe Valley; April-November 1970 Mean trap Month catch Remarks (G. pallidipes) April/May 66.5 June 53.7 traps replaced only July 41.4 August 11.0 September 24.8 October 84.8 traps replaced and repositioned November 81.2 791 E. C. ENGLAND & D. A. T. BALDRY them to become less attractive to G. pallidipes. The situation was investigated during December 1968. Altogether, 10 Langridge traps were placed in the Ruma thicket; 5 had been in constant use since late September 1968 and thus were over 2 months old, while 5 were new traps. During 6 trapping periods the mean catches of G. pallidipes were 0.24 flies for old traps and 5.64 flies for new traps. The new traps were thus catching 23.5 times more G. pallidipes than the old traps. DISCUSSION The results of observations made on the attrac- tiveness of different types of animal bait and the Langridge type tsetse trap to G. pallidipes show that in the area studied a calf not only attracted the greatest number of flies but, of the animals available, it was the bait most frequently fed on. The proportion of flies attracted to the ox, sheep, and goat (54.1%, 8.4%, and 6.4%, respectively) seems to go a long way towards explaining the incidence of animal trypanosomiasis in these ani- mals in the Lambwe Valley (17.0%, 5.0%, and 2.1 %, respectively) as reported by Robson & Ashkar (1972). Although they do not agree with the limited data available from blood meal analysis on the feeding preferences of G. pallidipes in the area, as described by England & Baldry (1972), they agree fairly closely with the findings of Weitz (1963) for the species over a larger area of East Africa (1.6% feeds from ox, 0.1 % from sheep/goat). The proportion of G. pallidipes caught from the man acting as bait during the same experiment (12.7%) is misleading, and undoubtedly gives a false impression of the attractiveness of man to G. pallidipes. Of 923 blood meal squashes analysed by England & Baldry (1972) only 2 (less than 0.2%) were derived from man; this is compatible with Watson's (1972) findings that the annual incidence of sleeping sickness varied between 0.18% and 0.48 %. There are three possible explanations to account for the apparently greater attractiveness ofman observed during the experiment. Firstly, although only one man was catching flies from the man acting as bait, both men (the bait and the catcher) were probably exerting an attrac- tive influence on G. pallidipes in their vicinity. This could have increased the fly catch from the human bait by anything up to 100%. Secondly, whereas those flies that visited the other baits and the trap were caught by passive methods, i.e., they had to find their own way into the trap and thus had some chance to escape, those which went to the man acting as bait had little chance of escaping since they were actively caught by the catcher as soon as they alighted. The proportion of flies that are attracted to a trapping device and then escape is not known, although personal experience leads the authors to believe that, on occasions, the proportion could be appreciable. Thirdly, although G. pallidipes is selective in its choice of hosts under natural conditions, it is suf- ficiently adaptable to overcome its dislike for some animals and man, when its preferred hosts, i.e., bushbuck, bushpig, and buffalo, are not available. The data presented in Table 1 suggest that under such circumstances man may be relatively less dis- liked than either sheep or goat. The results of comparative studies of trapping and fly-round patrol techniques clearly show that Langridge traps are much more efficient at sampling a G. pallidipes population than the other techniques, provided that they are reconditioned and reposi- tioned at intervals of approximately 1 month. These observations confirm the findings of Tarimo et al. (1970) who, after making similar investigations, stated " It is concluded that the fly-round may give a false assessment of reduction of G. pallidipes following insecticide application. Consequently, the high reduction recorded by fly-rounds in an earlier experiment with pyrethrum (Irving et al., 1969) should be interpreted with caution". Tarimo et al. (op. cit.) arrived at this conclusion after simulta- neously using Langridge traps and fly-round patrols to assess the percentage reductions in a fly popula- tion after three aerial applications of a pyrethrum/ DDT mixture. Percentage reductions according to patrols were 64, 36, and 29, and according to traps, 11, 7, and 0. An additional advantage of Langridge traps, par- ticularly when G. pallidipes populations are being sampled for ecological and epidemiological pur- poses, is that they catch more females than males (proportion of females during the present studies, 63.9-72.3%). Since female flies are believed to live longer than male flies, and therefore constitute the larger proportion of the fly population as a whole (see Jackson, 1937, 1940), the Langridge trap probably attracts and catches a representative cross- section of the fly population. Conversely, male flies predominate in fly-round patrol catches (pro- portion of females during the present studies, 18.9- 33.4 %). 792 ATTRACIlVENESS OF VARIOUS BAITS TO GLOSSINA PALLIDIPES 793 ACKNOWLEDGEMENTS The writers are grateful to the Director of Veterinary Services and to the Chief Zoologist, Ministry of Agricul- ture, Kenya, for providing facilities in the Lambwe Valley. They also thank Mr C. Rodrigues, Field Zoo- logist, Ministry of Agriculture, Kenya, for technical assistance in the field, and Mr G. R. Jewell, for develop- ing the bait traps used during the first studies. RtSUMt OBSERVATIONS SUR L'ATTRAIT RELATIF EXERCE SUR GLOSSINA PALLIDIPES PAR DIFFERENTS APPATS ANIMAUX ET UN PIEGE A GLOSSINES; EFFICACITIE COMPAREE DU PItGE ET D'UNE EQUIPE DE CAPTURE Des etudes ont et6 faites dans la vall&e de la Lambwe sur les merites respectifs de differents appats animaux et d'un piege a glossines de type # Langridge * en tant que source d'attrait pour Glossina pallidipes. Les glossines etaient beaucoup plus attir&es par un bovin (54,1 Y) que par un mouton (8,4%), une chevre (6,4%), un homme (12,7Y) ou le piege (18,1 %). En l'absence de bovin, les autres appats, y compris l'homme, et le piege les ont attir6es en nombres sensiblement equi- valents (21,5 i 28,9 %). Ces observations corroborent les donn6es relatives a la pr6valence de la trypanosomiase animale chez les especes etudi&es: 17,0% d'infections chez les bovins, 5,0% chez les moutons et 2,1 % chez les chevres. Le nombre de glossines captur6es sur l'appat humain donne probablement une ide fausse de I'attrait reel exerce par l'homme sur l'insecte. Trois hypotheses sont formulees a cet egard. On a aussi compar6 1'efficacite du piege et des captures effectuees par une 6quipe d'environ quatre hommes pour l'echantillonnage des populations de G. pallidipes. Le piege a permis de capturer un nombre beaucoup plus 6lev6 de mouches. En outre, le rapport des sexes parmi les specimens obtenus par cette methode (63,9 A 72,3% de femelles) 6tait proche du rapport considere comme normal au sein des populations naturelles, alors que parmi les insectes recueillis par les captureurs on ne comptait qu'une faible proportion (18,9 a 33,4%.) de femelles. REFERENCES Allsopp, R. & Baldry, D. A. T. (1972) Bull. WldHith Org., 47, 691 Allsopp, R. et al. (1972) Bull. Wld Hlth Org., 47, 795 Buxton, P. A. (1955) The natural history of tsetse flies, London, Lewis, p. 816 England, E. C. & Baldry, D. A. T. (1972) Bull. Wid Hlth Org., 47, 785 Irving, N. S. et al. (1969) Bull. ent. Res., 59, 299 Jackson, C. H. N. (1937) Proc. zool. Soc. Lond., 1936, 811 Jackson, C. H. N. (1940) Ann. Eugen., 10, 332 Langridge, W. P. (1972) In: Proceedings ofthe 12th meet- ing of the International Scientific Council for Trypano- somiasis Research, Bangui, Lagos, Organization of African Unity (in press) Robson, J. & Ashkar, T. S. (1972) Bull. WldHlth Org., 47, 727 Tarimo, C. S. et al. (1970) Bull. ent. Res., 60, 221 Watson, H. J. C. (1972) Bull. Wld Hith Org., 47, 719 Weitz, B. (1963) Bull. WId Hith Org., 28, 711