A comparison of population-samples from various localities with the respective type specimens has shown that Sogatodes brazilensis (Muir) and S. orizicola (Muir) are conspecific. Under the provision of Becommendation 24A of the International Code of Zoological Nomenclature the name S. brazilensis is suppressed and becomes a synonym of S. orizicola.

Field observations indicated that a strain of Boophilus microplus (Can.) in the Rockhampton area in Queensland was resistant to an organophosphorus insecticide, dioxathion (applied as Delnav), which had previously controlled it successfully. Ticks of this strain were despatched to the Cooper Technical Bureau in England, where a culture was established and maintained under selective pressure from dioxathion.

The non-parasitic stages of the culture were maintained in an incubator under controlled conditions, and the parasitic stages on cattle in a tick rearing house designed for the purpose. Careful security precautions were taken to ensure that there was no dissemination of tick life outside the confines of the culture. The tick rearing house was provided with double doors, the inner ones of which were screened with copper gauze to prevent the possible transmission of anaplasmosis by biting flies.

The activities of 23 insecticides against larvae of this strain and larvae of a strain of B. microplus susceptible to organophosphorus poisoning were compared. The test method was an immersion technique, usually employing the insecticide in the form of an emulsion. Mortality was assessed 17 hours after treatment. These comparisons indicated that the strain was resistant to the organophosphorus insecticides carbophenothion (62 ×), dioxathion (25 ×), diazinon (15 ×), parathion (10×) and a carbamate, carbaryl (38×). It is suggested that these resistances may be due to a specific mechanism effective against these compounds.

Nine other organophosphorus insecticides, one carbamate and rotenone were subject to low-order resistance by this strain significant at P ≤0·05. This was considered to be non-specific resistance.

Dioxathion had been in use for tick control on the property concerned for four years before resistance was demonstrated. The property is situated in an area where the tick season lasts for ten months. In other parts of the world, dioxathion has been in use against Boophilus ticks for seven years without the development of resistance to it. The length of time resistance has taken to develop suggests that the resistance mechanism is not the expression of a single dominant gene, as has been demonstrated for dieldrin-resistance.

The results suggest that resistance to one or more organophosphorus insecticides will not necessarily prevent the use of other members of this wide and diverse group for tick control. Several of the insecticides shown here to be subject to low-order non-specific resistance are known to be effective tick dips, and one of them, ethion, has been used with success against the resistant strain.

An interesting corollary of the results was that the organophosphorus thions showed greater activity than their corresponding oxons against the susceptible strain.

Observations, largely based on regular catches along a fly-round, were made over the five years 1959–64 on a population of Glossina morsitans submorsitans Newst. in the Northern Guinea Savannah of Nigeria. The results showed that the largest numbers of flies were caught in the early dry season (November–January) and that, as the climate became progressively drier and more severe, fly numbers declined to reach their annual minimum at the end of the dry season or in the early rains (March–May). These results are tentatively interpreted in terms of the true density of the flies and their activity. Differences occurred between the various years, some of which could be explained by climatic differences.

Of the 7,412 flies caught over the five years, 1,128 (15·2%) were females; the percentage of females was highest in the dry season, rising to a peak of 24·1 per cent, in February, and was below 10 per cent. during the wet season. Many more females were caught on the bodies of the catching team than on vegetation or the ground near the team.

The flies rarely fed on civet cat (Civettictis civetta) or duiker (Cephalophus rufilatus, Sylvicapra grimmia), which were the potential hosts most frequently observed in the experimental area, but fed mainly on wart-hog (Phacochoerus aethiopicus) and man, the next most commonly observed potential hosts.

During the heavy rains, males of G. m. submorsitans were evenly distributed over the fly-round, but at all other seasons they were concentrated to some extent in areas of thicker vegetation. During the dry season, pupae were found in the dry soil of forest islands and riverine vegetation in the savannah; the wet season breeding sites were not discovered. Previously published data on the resting sites and trypanosome infection rate of G. m. submorsitans in the area are summarised.

The results are discussed and compared with the conclusions reached by other workers from earlier more extensive studies on G. m. morsitans Westw. in Tanganyika and on G. m. submorsitans in the Sudan Savannah vegetation zone of Northern Nigeria.

A series of 24-hr. catches of Glossina from bait-oxen was carried out during 16 months in 1962–63 at Lugala, Uganda, where G. pallidipes Aust., G. palpalis fuscipes Newst. and G. brevipalpis Newst. occur. Female flies were dissected to determine their physiological age and the presence or absence of trypanosomes. Five age-categories were distinguished, depending on whether a fly had ovulated 0, 1, 2, 3 or more (4 + ) times. Flies remained in each of the first four categories for about 11 days and the fifth therefore comprised those over 40–50 days old. Trypanosome infections were classified as brucei-type, vivax-iype or congolense-type (i.e., attributable to trypanosomes of the groups of Trypanosoma brucei, T. vivax and T. congolense) according to the sites in which they were found.

In all three species of Glossina, vivax-tjpe infections were commonest and alone showed seasonal fluctuations in incidence. Infections of the brucei-type were rare. The total infection rate (all types) in G. pallidipes and G. palpalis fuscipes was highest in or immediately after months of greatest rainfall and relatively lower in dry months; the highest infection rates in G. brevipalpis occurred a month later than those of the other two species. Over 80 per cent, of infections in all three were found in category 4+ flies, the percentage of which in the catches varied in much the same way as the total infection rate, suggesting not only that the flies live longest during wet periods but also that fluctuations of infection rate are largely due to changes in mean age. The regression of total infection rate on percentage of category 4 + flies was significant for G. palpalis fuscipes over 14 months, and for G. pallidipes over 12 months, but insignificant for G. brevipalpis.

The age-composition of catches of G. pallidipes and G. brevipalpis but not G. palpalis fuscipes varied during the day. In G. pallidipes, the percentage of older flies was higher in the middle of the day than in the early morning and late evening, and these contrasts were reflected in the infection rate, which was highest in samples taken in the middle of the day. In G. brevipalpis, the percentage of oldier flies and also the infection rate were lower during the night than during the day.

Estimates were made of the mean number of bites by infected females that would be received by one ox in one day. The number varied from month to month, with peaks shortly after periods of high rainfall, mainly as a result of changes in fly density and relatively little as a result of changes in infection rate. G. pallidipes, the most numerous species, was responsible for most of the potentially infective bites.

An account is given of the disappearance in Kaduna, Northern Nigeria, of Culex nebulosus Theo. from the town streams and drains and other polluted collections of water, habitats in which until comparatively recently it appears to have been the dominant species, and its replacement by C. pipiens fatigans Wied., which was formerly rarely recorded in these sites. Similarly, the latter species has replaced the former as the commonest Culicine mosquito resting in houses. Increased urbanisation and the widespread use of residual organo-chlorine insecticides within recent years have both been advanced as possible explanations for this change.

A laboratory experiment with larvae of C. p. fatigans from Kaduna and of C. nebulosus from villages on its perimeter showed the latter species to have a slightly, but significantly, quicker rate of larval development than C. p. fatigans; this would possibly confer an advantage on it in circumstances where the two species are in competition for the same larval habitats. Field and laboratory observations showed that both species can tolerate high organic pollution in their larval habitats. There appeared to be little evidence in Kaduna, or in other towns in Nigeria, to support the theory that urbanisation has been responsible for the replacement of C. nebulosus by C. p. fatigans. Laboratory and field trials demonstrated that larvae of C. nebulosus were extremely sensitive to DDT, and although resistance was not found in C. p. fatigans, this species showed tolerance of the insecticide. It is believed that, in Kaduna and most probably in other towns, the application of DDT as a larvicide to streams and drains has been responsible for the eradication of C. nebulosus there, and in the absence of competition from C. nebulosus and because of the greater tolerance of DDT shown by C. p. fatigans the latter has been able to establish itself as the dominant species.

The persistence of DDT deposits on vegetation in the Northern Guinea Savannah zone of Nigeria was studied by chemical analysis of leaf samples and by bioassay tests with teneral females of Glossina palpalis (R.-D.). Two formulations of DDT were used, one (A) made by mixing a 75 per cent, wettable powder with water to give a 5 per cent, suspension, the other (B) made by adding a household detergent to A at the rate of 3·8 g. per gal. suspension. They were applied to foliage by knapsack sprayer in August 1964, about the middle of the rainy season.

It was noted that the particles deposited from B were finer and more evenly distributed over the leaf surface than were those from A. Chemical analysis of leaf samples showed that the deposit from B was initially the higher. Both deposits declined greatly during the first six weeks after application, coinciding with the period of heaviest rainfall. Those from A and B reached 49·8 and 46·2 per cent., respectively, of their initial levels at the end of the second week, and 9·2 and 5·2 per cent, at the end of the sixth. From then to the end of the sixteenth week, when the observations ended, no further decline was observed. In the tests with G. palpalis, deposits from A and B caused mortalities of 60.3 and 66.8 per cent., respectively, in the fourth week and 79·4 and 82·8 in the fifth, no mortalities below the latter being subsequently observed up to the tenth week.

It is considered that rainfall was the main factor responsible for removing the deposits from foliage. After the end of the rainy period, losses of the deposits became negligible. Deposits from B at first declined faster than those from A. It is considered that this was due to the faster removal by rain of the smaller particles.

Heavily weathered deposits were highly toxic to G. palpalis under the test conditions, and it seems likely that under natural conditions of contact they would have exerted still greater toxicity.

A series of forty 24-hour catches was carried out at seven levels (20-ft. intervals from 0 to 120 ft.) on a steel tower in Zika Forest by means of unbaited and unlighted suction traps that acted simply as air samplers. The object was to study the general flight activity of mosquitos in a tropical forest environment from samples that would not be subject to the selective effects exercised by bait-traps and light-traps.

Altogether, 4,151 mosquitos belonging to thirty-four species or species-groups were taken. Eight species, Aedes (Stegomyia) apicoargenteus (Theo.), Mansonia (Coquillettidia) aurites (Theo.), M. (C.) fuscopennata (Theo.), M. (C.) maculipennis (Theo.), M. (C.) metallica (Theo.), M. (C.) pseudoconopas (Theo.), M. (Mansonioides) africana (Theo.) and M. (M.) uniformis (Theo.), were taken in numbers sufficient for detailed analysis. By far the most abundant of these was M. aurites, comprising 55·7 per cent, of the total sample. The remaining twenty-six species together constituted only 7·6 per cent.

Only 144 or 3·5 per cent, of the mosquitos taken were males; and of the females, only 19 or 0.5 per cent. were blood-fed and only 11 or 0·3 per cent, were gravid. Thus, catches in the suction traps consisted almost entirely of unfed non-gravid females. It is tentatively concluded that male mosquitos in the forest are generally very inactive in the absence of stimuli such as lights or baits. The paucity of engorged and gravid females was expected; few of these would come within range of the unlighted unbaited suction traps.

Dissection of 3,072 females belonging to the eight most numerous species showed all to be inseminated. Age-composition was determined in species of Mansonia only; the majority of females in the subgenus Coquillettidia were nulliparous, whereas in the Mansonioides, nulliparous and parous females occurred in roughly equal proportions.

The unlighted, unbaited suction traps caught far fewer mosquitos than did light- and bait-traps under similar conditions.

In general, the vertical distributions and flight-activity cycles of the various species as indicated by the suction-trap catches differed distinctly from those indicated by bait- and light-trap catches reported by other workers. Very little activity was recorded above the forest. Although, in general, activity was irregular, three main patterns could be distinguished: largely nocturnal activity continuing up to about midday but almost completely suppressed in the afternoon, as in M. aurites, M. fuscopennata, M. maculipennis, M. metallica and M. pseudoconopas; activity spread throughout the entire 24-hour period, as in M. africana and M. uniformis; and mainly diurnal activity, as in A. apicoargenteus.

Measurements of oxygen consumption by a series of single pupae of Glossina morsitans orientalis Vanderplank maintained at 30°C. showed that the time course of respiration during this stage follows a U-shaped curve and thus resembles that of other holometabolous insects. The rate of pupal respiration was found to be directly proportional to wet weight at all temperatures investigated, and did not appear to be affected by the temperature to which females were subjected during pregnancy. The relation between temperature and the logarithm of metabolic rate was found to be curvilinear, tending to a relatively stable level between 35 and 40°C., the highest part of the range investigated.

Some effects on the adult of environmental conditions during the larval and pupal stages of Spodoptera littoralis (Boisd.) were investigated in the laboratory in Israel. The conditions studied were temperature during larval life, temperature and humidity during pupal life and the density at which larvae were kept, and their effects were found by varying one of them at a time. Except for such variations, larvae were reared in batches of 30–50 at 22–24°C. and 60–70 per cent, relative humidity, pupae were kept at 23–24°C. and 70–80 per cent, relative humidity, and adults were kept in pairs at 20–21 °C. and 75–85 per cent, relative humidity. Larvae were fed on clover leaves and adults on sugar solution.

High temperature during the larval stage reduced fecundity; whereas at 22°C. the resultant females laid about 1,200 eggs each and 12·5 per cent, produced non-viable ones only, at 30°C. they averaged about 700 eggs each and 50 per cent, produced non-viable ones only. Temperature during the pupal stage affected fecundity similarly and life-span to a lesser extent; at 20 and 34°C., 40 and 5 per cent, of the resultant adults lived 15 days or more, the percentages of females laying only non-viable eggs were nil and 70 and the mean numbers of eggs laid were about 1650 and 700, of which about 60 and 2 per cent, were viable, respectively. High temperatures were directly inimical to pupae, less than half of which survived at 26°C. or over.

The effects of humidity during the pupal stage were in general less than those of temperature. The mean life-span was longest in adults from pupae kept at an intermediate humidity of 76 per cent., as also were the proportion of barren females and the mean number of eggs laid; the percentage of viable eggs was reduced only amongst those from females kept as pupae at over 95 per cent, relative humidity. Pupal mortality was affected only by low relative humidities, which raised it.

Larvae reared singly tended to be brighter in colour than those reared in batches of 30–50; fewer died and the subsequent pupae were heavier and longer. The mean life-span of the resultant adults was longer, and they laid more eggs.

The results are compared with those of other workers on S. littoralis and other insects.

Leptoconops becquaerti (Kieff.) is one of the most troublesome species of biting midge in Jamaica. The gonotrophic cycle in this insect was studied and is described. In this species a single blood-meal is sufficient for the development of the ovaries and the maturation of eggs.

The stages in follicle development and digestion of the blood-meal adopted for other blood-sucking Diptera have been applied to L. becquaerti with slight modifications. A high degree of gonotrophic harmony is present in this insect.

Fully fed females were kept at five temperatures, 68, 77, 85, 91 and 98°F., to determine the time taken for the ovarian follicles to reach Stage V of development (egg mature, chorion present). The times were approximately 98, 64, 36, 30 and 27 hours, respectively.

It was demonstrated that the number of ovarioles in L. becquaerti was proportional to the size of the individual (as measured by wing length). The average number of ovarioles in an insect of wing length 0.83 mm. (the mean wing length of the population sampled) was about 103.

The number of eggs matured was also proportional to wing length, and was also dependent upon the temperature at which the flies were kept. Thus at the standard wing length of 0·83 mm., the average numbers of eggs matured at the five chosen temperatures (68 to 98°F.) were 78, 85, 65, 61 and 60, respectively. Females that had matured eggs were induced to oviposit by the decapitation technique.

Females provided with a source of carbohydrate (honey) during egg maturation produced slightly more eggs on the average than those fed on blood alone. The comparable figures at 91°F. in the standard individual were 65 and 61, respectively. The probable rôle of supplementary carbohydrate feeding in nature is discussed.

At the five experimental temperatures the greatest number of eggs hatched on days 11, 7, 5, 4 and 4, respectively, after laying. Egg-batches used in these experiments were deposited by decapitated females, and may therefore have been laid somewhat before the usual time in nature. The percentage fertility varied between individual batches at all temperatures, but the over-all fertility was consistently high (70–75%) at the lower four temperatures, and very low (4%) at 98°F. This difference almost certainly resulted from high temperature in the latter instance. Fertility in egg-batches deposited naturally (without decapitation) was of the same order as that in batches deposited by headless females.

A simple apparatus incorporating a photographic light meter is described, which is used for rapid measurement of the amount consumed by insect larvae of black filter-paper impregnated with plant extracts.

Many crops in southern Tanganyika are attacked by termites, especially those grown in the miombo woodland (Brachystegia-Pseudoberlinia), which covers much of the southern and western parts of the country. The termites chiefly concerned are Microtermes albopartitus (Sjöst.), M. redenianus (Sjöst.), Ancistrotermes latinotus (Hlmgr.) and Allodontermes tenax (Silv.). A pit dug to investigate their nest systems showed that M. albopartitus was the most numerous, 425 occupied cells being recovered from an area 50 × 24 ft. excavated to a depth of 6 ft. Cells of the other species included 166 of M. redenianus, 54 of Ancistrotermes latinotus and 49 of Allodontermes tenax. The cells of Ancistrotermes latinotus were aggregated into small clusters connected by long runways; those of the other species were more or less randomly scattered. The greatest concentration of cells of all species was within the top 2 ft. of the soil, but a number of cells were found at greater depths and it is certain that 6 ft. is not the maximum nesting depth. Descriptions of the nest systems and fungus combs are given.

Observations at Nachingwea in 1959–63 showed that loss of maize stands from termite attack averaged 27 per cent, over the five years, and that of soya 33 per cent. The damage was often great on individual plots, as much as 90 per cent, being recorded from one plot of maize. Plant-population counts over three growing seasons showed that maize is lost mainly towards the end of the season, whereas soya may be attacked any time after planting, increased losses occurring during dry spells and at the end of the season.

Control was achieved by the use of aldrin or dieldrin applied as dusts to the soil before planting or used in seed dressings. It was found in a series of field trials that good control of termites was afforded by aldrin worked into the soil along the planting ridges at a rate of 1 lb. active ingredient per acre, but that in practice ½ lb. per acre gave nearly as good results when measured by yields. In spite of the considerable losses, it was considered uneconomic to use aldrin to protect soya from termite attack, but the increased yields of 200–400 lb. maize per acre justified the expenditure on insecticide. The aldrin dust protected both crops for at least one further year after application, the effects on yield being of the same order as before.

A seed dressing of 75 per cent, dieldrin gave excellent control and was considerably cheaper than soil application of aldrin. The cost was only a few shillings an acre when the dressing was used at the rate of 0·1 oz. per 1 lb. seed, and the effect was to raise yields of maize by 330–500 lb. and of soya by 110–180 lb. per acre.

Trenches were dug to investigate the effects of soil treatment with aldrin on the nest systems of the termites. There was little effect on cell density eight months after application, but after 20 months there were signs of a reduction, especially in the top 9 in. of the soil.

In tests of insecticides applied in a spray tower to the ventral surface of nymphs of Distantiella theobroma (Dist.) collected from lindane-resistant and lindanesusceptible populations in cocoa plantations in Ghana, the levels of resistance of the former (ratios of the LC50's of the two populations) were 20·1, 74·4 and 73·3 for lindane, dieldrin and aldrin, respectively. There was no cross-resistance to DDT. Out of 18 other compounds tested, carbaryl (the most toxic), two other carbamates and two organophosphates were more toxic to the resistant population than was lindane to the susceptible one.

Details are given of a trap for Phlebotomine sandflies attracted to a small mammal placed in a cage incorporated in the trap. It was developed in British Honduras for the purpose of studying the sandflies attracted to proven or possible hosts of Leishmania, other than man. The results of initial tests carried out at Central Farm from November 1964 to February 1965 are reported. These provided evidence that the sandflies caught in the trap were attracted to the rodents used as bait, and did not enter by chance; unbaited traps caught virtually no flies. In traps baited with rodents known to be reservoir-hosts of Leishmania, the species caught in greatest numbers was Phlebotomus apicalis Floch & Abonnenc, and amongst those caught in smaller numbers were P. cruciatus Coq. (including P. diabolicus Hall) and P. deleoni Fairchild & Hertig; it is suggested that all three are possible vectors of leishmaniasis in British Honduras. The relative numbers of the sandfly species caught in the traps were very different from those caught on man in British Honduras and by direct catching on animals in Panama.

The species commonly known in the West Indies as the pepper flower-bud moth has been shown not to be Gnorimoschema gudmannella (Wlsm.), as hitherto supposed, but an undescribed species, for which the name Gnorimoschema capsica is proposed. The species is described and differentiated from G. gudmannella and other species in the Grnorimoschemini.

The distribution on cotton plants of the eggs, nymphs and to a lesser extent the adults of Empoasca lybica (de Berg.) was studied on the varieties Bar XL1 and BAJ 7/57 in the Sudan during 1961–64. Counts of jassid nymphs in the process of hatching from eggs buried in the leaf veins showed that hatchlings were most numerous on the third to fifth expanded leaves from the main-stem apex and that most eggs had been laid either close to the point of insertion of the petiole into the leaf lamina or about half-way along the length of the veins on the underside of the leaves. The greatest number of hatchlings was found in the central leaf vein. It was concluded that eggs are most frequently laid in this position on newly expanded leaves at the apices of the main stems or side branches.

The distribution of the early instars on the main-stem leaves was peaked, with maximum numbers occuring on or about the fifth expanded leaf from the apex; a similar distribution with a less prominent peak nearer to the apex leaf was found on side branches. The later instars, being more mobile, had a more even distribution. In the over-all distribution of nymphs on a typical cotton plant, a combination of the patterns found on main-stem and side-branch leaves, the greatest numbers of jassids occurred in the upper half or upper third of the plant. Co-variance analyses showed that the relation between numbers of jassids per zone and number of leaves per zone was not close and that the manner in which the plants were divided into zones for analysis of the relation greatly affected the correlation coefficients obtained.

Regular counts of nymphs and adults on the upper surface of leaves throughout a period of two complete days showed that both nymphs and adults moved to the upper surface of leaves each day between 1400 and 1800 hr. local time and returned to the under surface at about 0300–0400 hr. Neither the environmental factor that stimulated this movement nor the way in which it benefited the insect was clear.

Pupae of Glossina morsitans Westw. collected in the Zambezi valley near Kariba, Rhodesia, were brought to the laboratory, and the effects of the chemosterilant tepa were investigated by observations on the resulting adults. In the standard test, 25 pairs of adults were caged together for 28 days at 79°F. and 70 per cent, relative humidity and fed on guinea-pigs, and the survival of both sexes, the numbers of pupae and adults produced and (in some tests) the insemination rate of the parent females were recorded. With untreated flies, the insemination rate was 94 per cent, or more, averages of 28 pupae and 26 adult progeny per cage were produced, and 72 and 77 per cent, of males and females, respectively, survived for the 28 days. These results were compared with those obtained when the flies of one sex in the cage had been treated with tepa either as pupae or as adults.

Injection of 1 μg. tepa into the thorax of young males produced complete sterility. Dipping the pupae in 5 per cent, tepa solution for one minute caused complete sterility of adults emerging during the first two post-treatment weeks and partial sterility of males emerging during the third week. Adults emerging from dipped pupae washed one day after treatment were fertile. Dipping pupae in solutions containing 1 or 0·5 per cent, tepa did not completely sterilise the resulting adults. Contact exposure of adult males and females of various known ages usually resulted in complete sterility (99·7%) after exposures ranging from 15 to 240 min. to deposits on glass of 10 or 50 mg./ft.2 tepa. Males exposed for 240 min. to 10 mg./ft.2 tepa retained their sterility throughout a 42-day test period and, in special tests, competed well with untreated males for the females, but their life length was reduced by 25 per cent, when unmated and by 33 per cent, when mated. Males exposed for 60 min. or less survived as well over a 42-day test period as untreated males, but those exposed for 15 min. recovered an undesirable degree of fertility. No treatment affected the ability of the male to inseminate the females. Sperm from treated males were motile and appeared normal in their behaviour in the female spermathecae. Dominant lethality induced by tepa was usually expressed during the embroyonic stages, but occasionally was delayed until the pupal stage.

Taxonomic characters of the adult, pupa and larva of the coconut flat moth, Agonoxena argaula Meyr., are compared with those of the three other known species comprising the family Agonoxenidae, and keys to the species are provided. One new species, A. phoenicia, is described from northern Queensland. The systematic position of the family and the generic status of the four species are discussed. Although differences of possible taxonomic importance have been noted, the species are considered sufficiently homogeneous to allow assignment to the single genus Agonoxena Meyr. Haemolytis miniana Meyr. is transferred to Agonoxena, and Haemolytis Meyr. is suppressed as a junior synonym.

Three-day-old virgin females of Glossina austeni Newst. that had emerged in the laboratory in Britain from wild puparia collected in Africa or from subsequent laboratory cultures were caged with males over nine days old for 24 hours and dissected after death. Of the 5,474 flies so treated, 38 per cent. died within 14 days after mating day, whereas of the 593 found to be uninseminated, 73 per cent. died during that period, mostly within the first six days. A possible explanation of the discrepancy is that females with a poor expectation of life are unattractive to males, and this could account for the very small proportion of uninseminated flies in field collections, since such flies would tend to die before they could be captured.

Polovodova's technique for determining the physiological age of mosquitos was used in a study in 1962–64 of the age-composition of populations of Anopheles gambiae Giles and A. funestus Giles resting in houses in two areas of Tanzania. One area was around Muheza, 25 miles from the coast, where the climate is humid and equable, and the other was around Gonja, 80 miles inland, where hardly any rain falls for five months of the year.

It was found that the age-composition was almost identical in populations of A. gambiae and A. funestus at Muheza, about 20 and 23 per cent., respectively, being 3-parous and older and 1 per cent. 7-parous and older in both species. At Gonja, the population of A. gambiae was much younger, 14 per cent, being 3-parous and older and only 0·3 per cent. 7-parous and older. The oldest mosquitos found at Muheza included one 12-parous female of A. gambiae and one female of A. funestus believed to have laid eggs 14 times. No examples of A. gambiae older than 8-parous were found at Gonja.

Dissections to determine the condition of the ovariolar sacs in A. gambiae at Gonja showed that in 87 per cent, of freshly fed parous females an interval of at least 24 hours had occurred since oviposition. At Muheza, 72 per cent, of individuals of this species in the cool season and 52 per cent, in the hot season were in a similar condition.

Marking and recapturing experiments were carried out with females of A. gambiae in order to be able to correlate calendar age with physiological age. The oldest recaptured was 34 days old and was found to have laid eggs 10 times. From data on 60 recaptures, it was concluded that, although there was some irregularity, the first gonotrophic cycle lasted 3–4 days and later cycles 3 days.

Age-specific sporozoite rates in A. gambiae rose from 4·1 per cent, for 3-parous to 32 per cent, for 7-parous and older females, and in A. funestus from 3·2 per cent, for 3-parous to 30 per cent, for 7-parous and older females. Most of the infected 3-parous females were gravid, indicating that few were infective at the beginning of the fourth cycle. On this account it was concluded that some 80 per cent, of malaria infections were transmitted in the fifth, sixth and seventh cycles.

Analysis of the distribution of age-groups indicated that both A. gambiae and A. funestus showed a deficiency of nulliparous females, presumably because greater numbers of this group rested outside houses. From the second to seventh cycles the proportions of successive age-groups in both species at Muheza declined regularly at a rate corresponding to a mortality of 37·8 per cent, per cycle for A. gambiae and 38·6 per cent, for A. funestus, or 14·6 and 15·0 per cent, per day, respectively. Beyond this age the mortality was considerably higher. At Gonja, the population of A. gambiae declined at a rate corresponding to a mortality of 51·5 per cent, per cycle for the second to sixth cycles, or 20·9 per cent, per day. Above this age, the mortality was estimated to be higher still.

From the regression of infectivity on age it was estimated that 6·8 and 6·1 per cent, of A. gambiae and A. funestus, respectively, became infected at each blood-meal.

These findings are discussed in the light of current epidemiological theory.