Population patterns of Anopheles albimanus and their significance to malaria abatement*

Bull. Org. mond. Sante 1974, 50, 307-315 Bull. Wld Hlth Org. Population patterns of Anopheles albimanus and their significance to malaria abatement* SAMUEL G. BREELAND 1 Data obtained between 1967 and 1972 on naturally occurring populations of A. albi- manus in El Salvador were used to construct 24-hour and seasonalpatterns ofactivityfor the species. The resulting patterns are discussed in terms of control implications, and specific examples are given to relate the importance of a knowledge ofpopulation dynamics to the success of conventional and alternative methods of anopheline control in malaria abatement programmes. A retrospective view ofa successful feasibility study of the sterile-male release technique illustrates the importance ofpopulation studies in the selection ofrelease sites, the formulation of release schedules, and the selection ofsurvey methodology for the evaluation of results. If anything has been learned about Anopheles and the transmission and control of malaria it is that the various vector species are dissimilar. A. gambiae is a complex of very different populations, and A. albi- manus is probably not simply A. albimanus through- out its range. Certainly, it varies considerably in its response to control measures. Mosquito populations must be considered as biological organisms in a complex environment and not merely as targets of insecticides, especially in cases where the method of application presupposes a " cooperative mosquito ". Antimalaria workers recognize that the develop- ment of insecticide resistance among vector species and the frequent destruction of their natural enemies have increased rather than reduced the need for developing alternative methods of control-methods that will not pollute the environment, will act selec- tively on the target species, and will be of more than temporary value. It is also recognized that the development of alternative and integrated measures will require considerably more attention to biological principles and entomological techniques than does the wall-spraying regimen of malaria eradication programmes. The importance of full knowledge of the population dynamics of the target species for recog- nition of the "Achilles heel" on which to base control measures is apparent. The vector population * Presented at the Symposium on Malaria Research, Rabat, Morocco, 1-5 April 1974. 1 Chief, Medical Entomology Branch, Vector Biology and Control Division, Bureau of Tropical Diseases, Center for Disease Control, Atlanta, Ga. 30333, USA. should be known sufficiently well to employ specific methodology against the target-i.e., " species sani- tation ". This, like integrated control, is hardly a new concept, but is certainly a neglected one. Population dynamics, in the simplest terms, relates the fluctuations in the numbers of individuals in the population to changes in time and in circumstances. Hence, population studies related to the clock and calendar are essential to a knowledgeable approach to planning, conducting, and evaluating control efforts. From 1967 to 1972, as a member of the staff of the Central America Research Station (CARS), the author had the opportunity of studying populations of A. albimanus and associated mosquito species in El Salvador with the objective of developing know- ledge necessary to the testing and evaluation of alternative methods of mosquito control that might be more effective than conventional malaria eradica- tion procedures. This emphasis resulted not only from the interest of CARS in alternative method- ology, but also from the realization that dwelling- oriented vector habits were adequately known in the region or that this aspect was receiving sufficient attention, as may be seen from the study of Rachou et al. (11) on the epidemiology of malaria in El Salvador, the report of Wright et al. (13) on OMS-33 (propoxur) as a residual spray, and the evaluation of OMS-33 by Lassen et al. (8). The research interests of CARS in alternative methodologies included such diverse measures as the aerial application of ultra-low-volume (ULV) insecticides, conventional 3193 - 307 S. G. BREELAND ground-level cold and thermal fogging, ground and aerial larviciding, source reduction projects, the development of genetic and biological control tech- niques, and combinations of these and conventional eradication procedures in integrated approaches. This report offers a retrospective view of the elucidation of basic population dynamics of A. albi- manus in El Salvador and alludes to the applicability of that knowledge to the activities of CARS between 1967 and 1972. DIEL POPULATION STUDIES Typically, A. albimanus showed a repetitive pattern of behaviour over a 24-hour period. During daylight hours both males and females were found resting in well-protected niches of natural sites, such as rock crevices, lava, tree cavities, and ground holes; and in artificial structures, such as culverts, outbuildings, and cattle corrals; and beneath bridges. In 181 collections made between 1968 and 1970, represent- ing all months and many localities, 7 670 A. albi- manus adults were collected, with an average of about 42 specimens per collection in a female-to-male ratio of 3: 1. The diel resting cycle of females showed a pattern of resting during daylight hours, mass migration out of resting places at dusk, followed by several early evening hours of activity with no resting. Blood-fed females began reinfiltrating resting sites a few hours after dark. Resting populations reached a peak shortly after midnight, indicating a relatively short activity period during the first half of the night, beginning about dusk. At dawn there was some " resettling" followed by stabilization in a given site after full daylight. Males of A. albimanus were active all night and occupied the same type of daytime resting places as females, but were found in large numbers only near breeding areas. Detailed accounts of diel habits have been published (1). SEASONAL POPULATION STUDIES Fig. I illustrates the seasonal changes in larval habitats observed in El Salvador between 1968 and 1971. Observations during those 4 consecutive years followed the same habitats as they developed, stabi- lized, declined, disappeared, and reappeared along with the dry and rainy seasons. Early rainy season inundations in late May or early June in coastal pastures and lowlands resulted in a brood of floodwater mosquitos whose eggs had been dormant in the soil until the flooding (line 9). This was followed by anopheline production in the same areas within a matter of days, with some overlap (line 1). As the rains continued, accumula- tions of water filled marsh areas, ditches, and other lowlands; furthermore, water basins overflowed to provide additional habitats in marginal areas (line 2). These habitats were augmented still more by inunda- tions in roadside ditches, stream margins, inland pastures, and fields (line 3). Meanwhile, wet-weather ponds, lakes, and sloughs filled and, in some cases, overflowed the associated margins (line 8). Towards the end of the rainy season, receding water lines re- exposed marginal vegetation that had been sub- merged, and reflooding from intermittent flash rains produced breeding of the type represented by line 4. As rains diminished in November or December a gradual transition from rainy-season to dry-season habitats occurred. Estuaries (Fig. 1, line 5) were important dry- season breeding places and essentially supported only A. albimanus populations. A typical breeding place of this kind was Estero San Diego, on the El Salvador coast, in which breeding began after rivers had ceased flowing, and continued unabated until the estuary was opened to the sea by rainy-season floodings or by mechanical means. Rainy-season breeding in estuaries was negligible, although asso- ciated marshes and flats sometimes provided anophe- line habitats. Rivers (Fig. 1, line 6) ceased to flow shortly after rains diminished and by January portions with pools and vegetation supported substantial anopheline breeding. Breeding in rivers reached a peak in March and April and generally ceased in June, when dry- season water reserves began to decrease. Irrigation channels and associated overflows (Fig. 1, line 7) offered other dry-season habitats; although not usually major breeding sources, they occasionally became locally important. Permanent bodies of water having favourable levels and light conditions during the dry season offered good anopheline breeding conditions, but not all lakes and ponds were suitable, even though water was present. Small ponds with abrupt margins tended to recede below the vegetation line. Other bodies of water with larger surface areas were sometimes so shallow that rapid recession and turbidity combined to inhibit anopheline breeding, or rank growth of vegetation with excess shade hindered breeding. Much insight into the population dynamics of anophelines can be gained by observing adaptability to drastic changes in environment during the transi- 308 MOSQUITO POPULATION PATTERNS Floodwater species, Aedn a Psfbop Floodwater species(Precede peaks in lines 1, 2, & 3) (Broods precede peaks of line 4) _Wdt wothr pods lizs a sloughs, Ae ofSimous ParmnaneW bodies of viatw with stabl dryon -6k A ppvnc wpzn s (Ex. Play Grande) levels, predomi.nantly oablmanue (Ex. Lake Jocaton) I rrigation wotw predominantly A. alblhwnus , Rvers with pooled wtotr: Ope sunlit situations -A. offilmonq; UJght canopy - _ poeudopunctlponn s i_ _Estuaries, A. albman sus _ 5- Reflooding of receding shorelines -or A. 5llou4- pormanint shallowr margins rnexpossedt- Miscellaneous habitats frorn 3- ___~~......... peak Inundations, A. olblmonus - nland ma r hes, lak m rglns, ditch s, pastures, and Ilow- land werf Iows. A. offilmenus a A. psewsopunct1pennIs 2- C -- Z Coastal marshes and lowlands, brackish or freh, predominantly A. a/b/manes 1- -1 June . July , Aug. , Sept. , Oct. , Nov. Dec. Jan. , Feb., March , Aprll , May RAINY SEASON DRY SEASON Fig. 1. General schematic representation of mosquito breeding in El Salvador. Lines 1-8, breeding of Anopheles; line 9, breeding of floodwater species; Culex follows the pattern of Anopheles. Based on data from more than 2000 collections between January 1968 and June 1971. tion from one season to another, particularly the development of rainy-season habitats after the dry season. Such changes in coastal El Salvador were observed closely during every year from 1967 to 1972, but specific studies were made in 1970. During Table 1. Anopheline densities of rivers during transition from dry to rainy season Date No. of No. of No. of larvae No. of (1970) dips larvae per 100 dips a adults b 1 April 300 625 208 - 7 300 676 225 116 13 - - - 160 20 300 890 297 - 6 May 300 337 112 - 7 - - - 109 14 - - - 67 21 400 442 110 - 25 - - - 264 5 June - - - 73 16 400 23 6 - 22 - - - 44 24 400 0 0 - 6 July - - - 3 26 400 0 0 0 17 Sept. - - - 0 a Totals from Rios Aquisquillo, Huiza, Jute, and San Antonio. b Totals from daytime collections under bridges of the same rivers. that year, the rains began in the second week of May and, as flow increased, river-breeding areas were flushed out. Reflecting this seasonal change (Table 1), the density of larvae in rivers of the San Diego area of La Libertad, which had reached a peak in April, began to decline in May, and breeding was negligible after mid-June. The adult population remained somewhat longer, but began to decline in June. Some adults were found under bridges as late as 6 July, but not thereafter. Estuaries, illustrated by Estero San Diego (Table 2), showed a similar pattern. Between April and June this estuary was closed to the ocean by a sandbar, providing a body of quiet water ideal for anopheline breeding. Larval density reached a peak on 24 June before flushing opened the bar. With the exception of a small number of larvae on 20 July, no larvae were found after the estuary opened to the ocean on 26 June. Adults, sampled by means of a New Jersey light trap, decreased rapidly following the depletion of larvae, indicating that the trap was measuring estuary breeding. Since rivers and estuaries are the main coastal sources of Anopheles in the dry season, it is interest- ing to examine anopheline adaptation to the loss of these habitats. The data of Table 3 reflect changing larval densities in coastal marshes and associated basins, exemplified by Pantano Ticuiziapa, near La 309 S. G. BREELAND Table 2. Anopheline densities in Estero San Diego a during transition from dry to rainy season Date Average No. of Adults in (1970) larvae per 100 dips light trap 2 April 154 - 6 164 - 20 - 5 27 - 10 5 May 160 17 11 - 45 22 362 164 29 - 397 1 June 390 158 8 - 856 22 - 1 951 24 648 - 29 - 1 127 6 July - 65 20 32 1 1 3 Aug. 0 5 20 0 5 18 Sept. 0 16 a The estuary was opened to the ocean on 26 June 1970. Libertad. First this and later other types of depres- sion were prepared as breeding habitats by the same rains as had rendered the rivers and estuaries unsuit- able for anopheline breeding. Larvae were first detected in Pantano Ticuiziapa on 22 May, shortly before the virtual disappearance of river breeding, thus providing the necessary overlap. The productiv- APRIL MAY JUNE JULY AUGUST SEPTEMBER. Fig. 2. Seasonal changes in larval densities in four coastal rivers, a coastal marsh, and miscellaneous inundations, 1970. ity of the coastal marsh was relatively short-lived, being limited by increasing pollution of the standing water. Meanwhile, other inundations (roadside ditches, irrigation overflows, puddles, etc.) developed in an overlapping pattern and compensated for the loss of the marsh habitat (Fig. 2). APPLICABILITY OF POPULATION PATTERNS During the period covered by this report, there were many opportunities to apply results of both diel and seasonal population studies. The 24-hour activ- Table 3. Larval densities in coastal situations during transition from dry to rainy season t Anopheline larvae/i 00 dips Date Pantano Miscellaneous Comments Ticuiziapa inundations April 0 0 dry conditions 22 May 196 0 marsh recently flooded; no miscellaneous inundations 2 June 200 0 marsh recently flooded; no miscellaneous inundations 10 131 0 marsh recently flooded; no miscellaneous inundations 24 210 0 marsh recently flooded; no miscellaneous inundations 2 July 130 50 roadside ditch; first miscellaneous breeding detected 14 10 marsh polluted and with broken surface film; habitat becoming unsuitable 22 0 92 marsh with solid surface film, unsuitable; breeding in flooded puddles 3 Aug. 0 65 marsh unsuitable; breeding in pools and puddles 4 Sept. 0 23 marsh unsuitable; breeding in drainage ditch 310 MOSQUITO POPULATION PATTERNS ity pattern influenced the selection of control tech- niques and evaluation of methodology, and it is obvious that seasonal changes exerted a profound effect on larval habitats and associated adult mos- quito populations. It is significant and important that no major anopheline breeding habitat, including permanent bodies of water, supported breeding with- out seasonal interruption. It is precisely this fact and the elucidation of specific patterns that hold out hopes for the success of well-timed malaria abate- ment efforts. A retrospective view of the El Salvador experience illustrates the applicability of population data to the development of control methodology. Diel population patterns and control implications The results of field trials involving the aerial application of ultra low-volume (ULV) malathion against A. albimanus in El Salvador during 1968 and 1969 were disappointing. The discovery of malathion resistance in the field population in 1969 was a factor, but probably more important to the failure of the method were the resting habits of the species in the target area. The local A. albimanus rested in well- sheltered sites, mainly deep crevices in rocks, and large numbers were inaccessible and protected from falling droplets during the time of day when treat- ments were made. The resting curve suggests that treatments during the early evening hours might be more effective. By contrast, the same technique against the same species in Haiti was more success- ful, possibly owing at least in part-to different resting habits of the species, i.e., more exposed in ground vegetation during the time of day when treatments were made. The reciprocal of the resting curve for A. albimanus in El Salvador represents the period of activity, showing a peak during the early hours of darkness. This can be validated by light trap, cattle corral, and human-bait collections (2). Since A. albimanus is more exophilic and exophagic and people tend to be out of doors during the early evening hours, outdoor malaria transmission is an important consideration, and the impact of indoor residual wall treatments is obviously lessened. By contrast, A. gambiae and A. funestus in the Kisumu area of Kenya are strongly endophilic and endophagic and people tend to stay indoors during the hours of biting activity (R. E. Fontaine, unpublished observations). Furthermore, these vectors spend longer periods of time resting indoors than does A. albimanus. Furthermore, A. albimanus is not particularly anthropophilic, and blood-engorged specimens collected from inside dwellings during the early morning hours have not always fed on man (3). The implications of such differing behaviour pat- terns can be profound. For example, WHO is now carrying out a stage VII large-scale trial of fenitro- thion in Kenya, and this compound is showing good results against local anophelines; however, the behaviour pattern of A. albimanus in Central Amer- ica suggests that fenitrothion would be much less effective there. The potential problems from excessive dependence on insecticides and generalizations on their probable usefulness are apparent. Seasonal population patterns and control implications The advantage of knowing seasonal population patterns in different types of breeding area is well illustrated by the experience gained in an estuary study area-Estero San Diego, El Salvador. This estuary showed a repetitive pattern of seasonal breeding year after year. There was a high correla- tion between the larval population of the estuary and the adult mosquito population on the adjacent beach as measured by light traps. The plotting of such seasonal data (see Fig. 3), is useful in planning malaria reduction through mosquito control. The curve shows that the application of an adulticide, for example, during the rainy season, would serve no useful purpose in this particular area, whereas a more timely application based on mosquito densities and distribution might be effective for an entire year. Also, opening of the estuary, larviciding, or com- binations of these methods could be effective and might be indicated (Fig. 4). These data exemplify the importance of timing to the planning of control efforts. The failure to eradicate malaria in Central Amer- ica is often attributed to resistance to DDT, discon- tinuous walls, and outdoor transmission. The results of population studies have suggested certain modifi- cations in existing methodology and the development of alternatives. One such alternative is the space spray approach, exemplified by the aerial application ofULV malathion, which presumably would circum- vent the technical problems mentioned above. An- other alternative is the sterile-male release technique. A successful feasibility study of this technique has been completed in El Salvador in cooperative tests between the Insects Affecting Man Research Labora- tory (IAMRL) of the US Department of Agriculture, Gainesville, Fla., and the Central America Research Station (CARS). Detailed results of the study are about to be published in a series of papers by 311 S. G. BREELAND La-ae per 100 dips Adults per trap night E400X z 1 969 1970 Fig. 3. Average number of A. albimanus larvae per 100 dips, San Diego Estuary; and number of A. albi- manus per light trap night, San Diego Beach, January 1 969-June 1970. 2000 1900 1800 1400 1300 , 1200 1100 E ,z 1000 o 900 o 800 K 700 E 600 z 500 400 300 200 100 A (1951) Is I I . --- Adults in light trap II I Larvoe /100 dips I I II I I II I Ill II I I' ESTUARY *II, OPENE ./~~~ % ---- Fig. 4. Larval and adult population of A. albimanus in and near San Diego Estuary, 1970. Breeland et al. (4), Dame et al. (5), Lofgren et al. (9), and Weidhaas et al. (12). This study is an excellent example of the importance of doing the right thing, with the right mosquito, in the right place, at the right time and-since it illustrates the application of seasonal population knowledge to the successful application of control methodology-it is the basis for the remainder of this report. While CARS was elucidating the biology of A. al- bimanus in El Salvador, IAMRL demonstrated that field populations of Culex pipiens quinquefasciatus could be reduced by the release of laboratory-reared chemosterilized males into a natural population (10). This success, and the concurrent development at Gainesville of techniques for the mass-rearing of A. albimanus (6), renewed interest in the possible use of the sterile-male method against anophelines. It was recognized early that A. albimanus possessed certain behavioural characteristics that might make it adaptable. Its most important feature was that it could be readily colonized, indicating a similarity in the mating behaviour of laboratory-reared males and wild females. Prior to this test, CARS had given much attention to the study of diel and seasonal activities of A. albi- manus in 10 areas of El Salvador, including Lake Apastepeque, near San Vicente. This area had also been included in a year-long study, during 1968, by the national antimalaria programme, CNAP (Cam- pania Nacional Anti Paludica), which made its data available to us. The CNAP data, confirmed by our own observations in the same area and elsewhere in El Salvador, showed a seasonal pattern of breeding for A. albimanus, likely to be repetitive from year to year. Kumm and Zuniga (7) also reported a seasonal prevalence of A. albimanus during a study carried out in 1941-42 at Lake Ilopongo, near San Salvador, similar to the pattern in Lake Apastepeque. Thus, when Lake Apastepeque was chosen for our studies early in 1971, there was sufficient background know- ledge to allow the confident prediction of a seasonal pattern-so necessary for the planning of a sterile- male release programme. Prerelease studies con- ducted between January 1971 and April 1972 con- firmed earlier observations and formed the basis for the selection of release sites, the formulation of release schedules, and the selection of survey meth- odology for evaluating the results. Fig. 5 shows the seasonal densities of A. albimanus females for 1968 based on captures from Establo Apastepeque, and for 1971-72 based on larval collec- tions from Lake Apastepeque and adult captures 312 I7 II i 8 15 2229 6 132027 4 11 18 25 2 9 16 23 30 6 13 20 27 2 9 16 23 APRIL MAY JUNE JULY AUGUST SEPTEMBER L MOSQUITO POPULATION PATTERNS A % '*-Ift / % I/ " II Iv/. \ I I \ilII I I I ,1A.o/bimanus Females/ .: -- A.a/bimonus larvae/dil 'Capture +1 p .....No collection J F M A M J J A S O N D 1968 Fig. 5. Seasonal density of A. albimanus females from from Lake Apastepeque (1971-72). from Establo Apastepeque. Regular releases began in mid-April 1972 to coincide with a low natural population, just prior to the beginning of the ex- pected seasonal build-up predictable from the curve. The cumulative pressure from periodic releases of sterile males was to be continued sufficiently long to overwhelm the natural population and prevent the seasonal peak expected during the last quarter of the year. The monthly mosquito population levels for 1968 and 1971 at Establo Apastepeque (Table 4) show that during normal cycles high densities of adults occur- red in the fall (September-November) and very low densities in late winter and early spring (Febru- ary-April). The 1972 data also show that the objec- tive of preventing the large increases in A. albimanus in the fall was achieved. Measurements by other collecting techniques were equally impressive. Not only did seasonal density curves contribute to the success of the sterile-male release programme, but data obtained in the project were used by F M A M J J A S 0 N DI J F M A 1971 1972 Establo Apastepeque (1968 and 1971-72), and larvae Table 4. A comparison of the number of A. albimanus females collected per night at Establo Apastepeque in 1968, 1971, and 1972 Average No. of females per collection Month 1968 1971 a 1972 January 197 161 February 10 109 March 1 10 April 0 21 May 10 26 34 June 181 42 22 July 114 67 22 August 170 122 7 September 1 062 593 0 October 1 403 1 311 2 November 912 403 24 December 68 181 15 a In January and April, no collection was made; in February and March, collections were made in the morning. 1500- 1000 - 100 5 - ; 10 - 313 n i 314 S. G. BREELAND Weidhaas et al. (12) to determine the density of mosquitos in the release area, the rate of growth of the population, the competitiveness of released males, the survival of the various stages, and the ability of the mosquito to transmit malaria in the area. Basic population parameters were used to construct a model of population dynamics and vector capacity. It was concluded from data by Weidhaas et al. (12) that, in the Apastepeque experiment, A. albimanus sterile males competed well with wild males in the field, their competitiveness ranging between 25% and 100% of that of their rivals. The latter figure is probably more accurate, but even the former would have enabled the release of the sterile males in that experiment to be qualified as successful. It was further calculated that survival rates of immature stages of the mosquito varied from 0.02 to 0.15, whereas the estimated average daily survival of adult females varied with the season of the year from 0.65 to 0.91, the higher value being associated with the wet season. With these estimates, it was possible to construct models of population dynamics and vector capacity for a specific mosquito population in a specific place at a specific time. Estimates of absolute densities and the survival of the adult females in the test area allowed the age distribution of the popula- tion to be calculated. This, together with knowledge of the biting behaviour of the females and epidem- iology of malaria in the area, allowed Dr Weidhaas and his colleagues to construct a population model to determine the number of adult females capable of transmitting malaria at different times. With such estimates it should be possible to simulate the effects of single or integrated methods of control on popula- tion densities and to predict vector potential. Thus mosquito population data are of considerable im- portance for malaria abatement through mosquito control. R1ESUMt LES CARACTtRISTIQUES DE POPULATION D'ANOPHELES ALBIMANUS ET LEUR IMPORTANCE AU REGARD DE LA LUTTE ANTIPALUDIQUE Des donnees recueillies de 1967 a 1972 concernant les populations d'Anopheles albimanus existant en El Salva- dor ont servi a etablir les caracteristiques nycthemerales et saisonnieres de l'activite de cette espece. Ces caracteris- tiques sont examin&es sous l'angle de leurs consequences pour la lutte antipaludique. Des exemples sp&ifiques montrent l'importance des connaissances relatives a la dynamique de population pour assurer l'application effi- cace des methodes de lutte contre les anopheles - qu'il s'agisse des methodes classiques ou de methodes de rem- placement - dans les programmes d'eradication du palu- disme. On expose les resultats obtenus lors d'une etude de faisabilite concernant la technique des males steriles, illustrant la necessite de proceder A des etudes de popula- tion avant de choisir les endroits propices au lacher, les schemas de liberation et la methodologie la mieux adaptee a l'evaluation des resultats. REFERENCES 1. BREELAND, S. G. Studies on the diurnal resting habits of Anopheles albimanus and A. pseudopunctipennis in El Salvador. Mosquito news, 32 (1): 99-106 (1972). 2. BREELAND, S. G. Methods for measuring anopheline densities in El Salvador. Mosquito news, 32 (1): 62- 72 (1972). 3. BREELAND, S. G. Studies on the ecology of Anopheles albimanus. American journal of tropical medicine and Hygiene, 21 (5): 751-754 (1972). 4. BREELAND, S. G. ET AL. Release of chemosterilized males for the control of Anopheles albimanus in El Salvador: I. Characteristics of the test site and the natural population. American journal of tropical medicine and hygiene, 23 (2): 274-281 (1974). 5. DAME, D. A. ET AL. Release of chemosterilized males for the control of Anopheles albimanus in El Salva- dor: II. Methods of rearing, sterilization and distri- bution. American journal of tropical medicine and hygiene, 23 (2): 282-287 (1974). 6. FoRD, H. R. & GREEN, E. Laboratory rearing of Anopheles albimanus Wiedemann. Mosquito news, 32 (4): 509-513 (1972). 7. KUMM, H. W. & ZUNIGA, H. Seasonal variations in the numbers of Anopheles albimanus and A. pseudo- punctipennis caught in stable traps in Central America. American journal of hygiene, 39 (1); 8-15 (1944). 8. LASSEN, K. ET AL. Preliminary report on the effect of selective application of propoxur on indoor surfaces MOSQUITO POPULATION PATTERNS 315 in El Salvador. American journal of tropical medi- cine and hygiene, 21 (5): 813-818 (1972). 9. LOFGREN, C. S. ET AL. Release of chemosterilized males for the control of Anopheles albimanus in El Salvador: III. Field methods and population control. American journal of tropical medicine and hygiene, 23 (2): 288-297 (1974). 10. PATTERSON, R. S. ET AL. Suppression and elimination of an island population of Culex pipiens quinque- fasciatus with sterile males. Science, 168 (3937): 1368- 1370 (1970). 11. RACHOU, R. G. ET AL. Synoptic epidemiological studies of malaria in El Salvador. American journal of tropical medicine and hygiene, 14: 1-62 (1965). 12. WELDHAAS, D. E. ET AL. Release of chemosterilized males for the control of Anopheles albimanus in El Salvador: IV. Dynamics of the test population. American journal of tropical medicine and hygiene, 23 (2): 298-308 (1974). 13. WRIGHT, J. W. ET AL. Orthoisopropoxyphenyl methylcarbamate (OMS-33) as a residual spray for control of anopheline mosquitos. Bulletin of the World Health Organization, 40 (1): 67-90 (1969). DISCUSSION NAJERA: It is necessary to correlate the data on the dynamics of outdoor resting behaviour to night biting activity and the data on variations in general mosquito densities to malaria incidence before judg- ing the relevance of such variations to malaria transmission. WHITE: Age-grading of vector populations is par- ticularly likely to be biased by inadequate sampling methods. GRAMICCIA: I agree. The improvement of en- tomological sampling methods is necessary in order to secure representative data for statistical evaluation. PARISI: It is important that sample sizes should be adequate, and care should be taken in generalizing from limited local observations to the conditions in a large area. 12