Experiments were made at three different sites in Malaya between 1952 and 1957 to compare the numbers of mosquitos attracted to man, calf or goat under the same conditions. The baits being compared (either two men and one calf or two men and two goats) were exposed simultaneously in two net traps placed about 50 yd. apart. This procedure is referred to as comparative trapping.

The total number of mosquitos caught with each bait showed calf to be much the most attractive, followed by goats and then men. However, there were three species, the dark-winged form of Anopheles barbirostris Wulp, Aëdes albo-pictus (Skuse) and Culex pipiens fatigans Wied. that were attracted in larger numbers to man than to calf, and six species attracted in larger numbers to man than to goat.

To compare the host preferences of the different species as between man and calf, the numbers of each caught with man and with calf have been expressed as a ratio (the man: calf attraction ratio). For most species the ratios from the different sites were much the same, showing that the method produced repeatable results. The data from the three sites were therefore combined.

On the combined data the ratios range from 3·4:1 in favour of man for the dark-winged form of Anopheles barbirostris to 1:82 in favour of calf for A. vagus Dön.

As might be expected, the ratios show that the vectors of malaria, filariasis and virus diseases are among the more anthropophilous species.

Attraction ratios have been calculated from suitable data published by various authors for ten different species in five countries. The ratios obtained place the Anophelines in an order that agrees well with their status as vectors of malaria.

It appears that if the results of comparative trapping are expressed as attraction ratios, this offers a method of comparing the host preferences of mosquitos of the same or different species, at different times and places in a country, or even in different countries.

The larvae of the Lymexylid beetle, Melittomma insulare Fairm., bore into the bases of coconut palms. Associated with them are micro-organisms which cause extensive rotting of the bole, particularly at ground-level, and this results in the palm falling.

During 1953–58, an attempt was made in Praslin Island, Seychelles, to obtain maximum control of this pest using paradichlorobenzene as a fumigant. At the time of this treatment, 77 per cent, of the 90,500 coconut palms in the island were infested. Of these, about five per cent, were felled as worthless mountainside palms and nine per cent, fell during or within a week of treatment.

In 1960 it was found that a further 15 per cent, of the palms had fallen and of those still standing 53 per cent, remained infested. The continuing high loss of palms after treatment was due to the closed, moist fumigation chambers within the trunks encouraging the spread of rot, coupled with an incomplete kill of larvae.

A new method of treatment is described in which the necrotic wood is excised and a formulation of coal tar creosote and coal tar liberally applied. Praslin Island was retreated, using this tar method, and a plan to treat all the coconut palms in Mahé, the main island of the group, was subsequently commenced. Losses of palms during and within a week of treatment have been reduced to less than one per cent., and results are quoted which show that a single treatment can reduce a serious infestation in which 80 per cent, of palms are more or less heavily attacked to one in which only 18 per cent, of the palms are attacked and, on average, contain only five larvae. The treatment ensures that the entrances to the tunnels of these are exposed to view, and accordingly the surviving larvae can easily be killed.

Following successful small-scale trials, an attempt was made, by aircraft application of insecticide, to eradicate Glossina morsitans Westw. and G. pallidipes Aust. from an isolated block of savannah woodland, approximately 11 sq. miles in extent, at Chungai in the Central Province of Tanganyika between July 1958 and January 1959. A single-engined aircraft, fitted with two belt-driven rotary-cage atomisers, was used to apply a 5 per cent, solution of γ BHC in power kerosene at a nominal dosage of 0·08 gal. per acre (0.04 lb. γ BHC per acre). Seven applications were made at approximately 28-day intervals, the time taken to complete an application varying from five to eight days. The operation failed to control the flies. Although each of the first two applications reduced the apparent density of G. morsitans by about 90 per cent, and that of G. pallidipes by a lesser, although still considerable, factor, later applications gave varying and often low mortalities, and the populations increased slowly for some time, the insecticide applications causing only temporary depressions in numbers. Numbers fell towards the end of the operation, but final reductions were only about 50 per cent, or less. Kills of female flies were low, and this undoubtedly led to the eventual failure.

The low volume-dosage, a drop spectrum that possibly contained too few droplets of the required size, meteorological conditions, and biological factors that apparently favoured the survival of female flies are suggested as contributory elements to the low mortalities.

Operational costs were considerably lower than in previous work.

In a study of the effect of different concentrations of radioactive phosphorus (32P) in the larval medium on the growth and development of Culex pipiens molestus Forsk., 32P was found to have little noticeable effect on the growth of the larvae up to a concentration of 3·0 microcuries (μc.) of 32P per ml., but, above this concentration, larval growth was greatly retarded.

The period of larval development was increased at concentrations greater than 1·0 μc. of 32P per ml., and pupation occurred two weeks later than in the controls. In concentrations higher than 5·0 μc./ml., pupation was completely inhibited, larvae became sluggish, stopped feeding and finally died.

The effect of 32P in the larval medium on the emergence and radioactivity of the resulting adults was also studied. On the basis of this study, it is recommended that, for efficient utilisation of radio-phosphorus in large-scale field experiments, a concentration of 32P of 1·0 μc./ml. be employed so that emerging adult mosquitos may be sufficiently radioactive to be readily detectable.

The difficulty of deciding the systematic status of variations within and between populations of Tyrophagus is considered in relation to the wide distribution of the genus and the many materials it infests.

Variation within populations of what was thought to be one species infesting cheese is explained by the isolation of three closely similar forms which it is difficult, if not impossible, to separate by measurement alone. Frequency distributions overlap slightly for four linear measurement ratios on which separation is attempted, but differences between them are emphasised by differences in minute structural characters and they are therefore accepted as distinct species, identified as longior (Gervais, 1844), palmarum Oudemans, 1924, and putrescentiae (Schrank, 1781).

Variation between populations attacking plant materials and other stored products in addition to cheese is accounted for to some extent by the recognition of three further Tyrophagus species, oudemansi Robertson, 1959, brevicrinatus Robertson, 1959, and tropicus Robertson, 1959.

Within the species recognised, significant differences are recorded in dorsal hair characters and body proportions between populations on cheese produced in different geographical areas, and also between cheese populations and those infesting other materials, and these differences are thought to be at a racial level.

T. longior, T. palmarum and T. putrescentiae are found to be distributed around the world, but as cheese pests they occupy somewhat different, although overlapping, geographical zones, since longior is a temperate-to-cool form, palmarum a temperate form and putrescentiae a subtropical or tropical form. The three are also found to occupy different, but overlapping ecological zones, making possible their association in mixed populations.

Confusion in ecological data previously published on Tyrophagus is thought to be attributable in part to similarities in the morphology, geographical distribution and ecology of its species, and to their close association.

An account is given of a survey of the mosquitos of the Jos area of Northern Nigeria, during May–June 1960, particular attention being given to those species likely to be concerned with the transmission of arthropod-borne viruses. The area concerned is in the savannah belt at an altitude of about 4,000 ft., with numerous rocky outcrops. The dry season is fairly severe, and extends from about October to May; during this time, drinking water for the villages is obtained from deep wells. At the time of the survey, that is, at the beginning of the rainy season, the most abundant species were Culex univittatus Theo., Aëdes ingrami Edw., A. aegypti (L.), A. africanus (Theo.), A. luteocephalus (Newst.) and A. vittatus (Big.). The biting cycles of these species are illustrated.

A survey was carried out of the breeding habits of A. vittatus in rock holes; those most favoured were about a foot across and six to nine inches deep, and appeared to be of human origin. Larvae appeared rapidly in the holes after the first rains, and predators, particularly dragonfly nymphs and tadpoles, were common in the larger rock holes. Under natural conditions, the duration of the life-cycle was about five days. The survival of the early stages under adverse conditions was investigated in simple laboratory experiments. At normal temperatures, the pupae did not survive desiccation for six hours, and, though the resistance of larvae to drying was not tested fully, it is considered that they are probably no more resistant than are pupae. Eggs survived desiccation for 10 but not for 18 weeks. Water temperature in exposed rock holes was not a factor limiting development, but the temperature of the exposed surface of dry rock holes was such that it appeared unlikely that eggs in these exposed situations would be able to survive the dry season. It was suggested that there may be collections of water in sheltered tree holes or rock holes which support a small but more or less continuously breeding population; this, combined with the ability of the eggs of A. vittatus to withstand desiccation for at least ten weeks at normal temperatures would enable the species to survive the relatively prolonged dry season.

A method of obtaining locust eggs which could be expected to be deficient in carotene is described. It involved rearing a parental generation on an artificial diet which, because of certain pigmentation abnormalities it induced, was probably itself deficient in carotene.

Using crowded hatchlings of Locusta migratoria (L.) from eggs thus modified, it was shown that they were sensitive to dietary carotene, whereas hatchlings from normal eggs were indifferent. The absence of carotene was marked by inferior growth, lessened activity and most notably by an extremely different coloration in both hoppers and adults. Without carotene, melanisation was absent or greatly reduced and in many cases the integument developed a greenish-blue colour. With carotene, heavy melanisation occurred in the hoppers, and the colour of the adults was characteristically gregarious.

It was concluded that in normal eggs the amount of carotene present is usually sufficient to mask the expression of a dietary deficiency during nymphal growth, but that with normal eggs having low amounts of carotene, a dietary deficiency might become apparent as it did with modified eggs.

It is suggested that the blue colour of modified Locusta hoppers reared without carotene is due to mesobiliverdin, a prochromogen of insectoverdin, the green pigment of the solitaria phase of locusts. Insectoverdin itself could not be formed because it contains carotenoids, and these were absent from the diet. The effect of extreme deprivation of carotene is therefore to induce solitaroid tendencies, notably in regard to colour (suppression of melanin and production of mesobiliverdin) but also in regard to activity.

Crowded hoppers of Schistocerca gregaria (Forsk.) reared on synthetic diets lacking carotene usually had turquoise-blue blood by the fifth instar. It was shown, by the electrophoresis of yellow, green and blue bloods in parallel, that the blue chromoprotein of blue blood is the same as the blue chromoprotein of green blood. Its prosthetic pigment must therefore be mesobiliverdin.

The abnormalities of coloration which arise in crowded Locusta and Schistocerca when deprived of carotene are normal in the solitary phase. Moreover, they resemble those abnormalities of coloration consequent upon the implantation of additional corpora allata into gregarious hoppers. The implications of this are discussed in relation to the validity of phase criteria, and an attempt is made to relate these findings to hypotheses on the humoral regulation of phase.

In the North Block at Chungai, comprising about 11 sq. miles of thorn savannah and thicket in Central Province, Tanganyika, an Auster J5G aircraft was used between July 1959 and March 1960 to apply a 2·5 per cent. solution of dieldrin in oil at the rate of 0·125 gal. per acre in an attempt to eradicate Glossina morsitans Westw. and G. pallidipes Aust. Eight applications were made at approximately four-weekly intervals. Swath width was 55 yd., and the aircraft emitted the insecticide as a coarse aerosol of volume median diameter 50–60μ as it flew in both directions over the block.

The operation suffered delays, and three applications were incomplete to varying degrees. Kills of G. morsitans per application appeared to be 85 per cent. or higher, but the final reduction of 99·5 per cent. could theoretically have been attained with consecutive mortalities of only 65 per cent. It is suggested that this discrepancy may be due to the higher lethal dose required by pregnant females. G. pallidipes was also reduced by 99·5 per cent. Numbers of fly were reduced sufficiently for large-scale settlement with cattle, which should complete the work of exterminating the fly.

The experiment is compared with that of the previous year in the same block, using γBHC, which reduced fly catches by less than 50 per cent. It is concluded that a combination of reduced swath width, greater volume dosage, more lethal insecticide and smaller lethal drop, together with improved flying technique and the more reliable performance of the disseminating equipment, was responsible for the improved result. It is thought that without delays, interruptions, incomplete applications and reinfestation, even better results would be obtained.

This was the cheapest and one of the most successful aerial operations carried out against savannah tsetse. Costs actually over the ground were £301 per sq. mile; incidental costs due to the locality of operations were £66 per sq. mile. There is little chance of reducing the costs of flying directly, but economies are possible by the use of other insecticides or, more probably, by more efficient dispensing equipment.

An experiment was done in Kenya to test several formulations of lindane and malathion dusts as protectants of bagged maize, with which they were mixed, against reinfestation (post-fumigation infestation) by store pests. The experiment was in the form of a 9 x 9 Latin square employing 162 bags. Sampling initially and after 4 and 5¼ months gave data on bag weight-losses, dust present per bag, numbers of adults of Tribolium castaneum (Hbst.), Oryzaephilus surinamensis (L.), Cadra cautella (Wlk.) and Sitophilus oryzae (L.), and moisture-content changes.

The dust weights (which were expected to be the most reliable measure of treatment efficacy) showed that there was very little difference between malathion admixed at 8, 4 and 2 parts per million. The other six treatments were far less satisfactory and all were significantly different from one another, with the following order of values: lindane at 6, 3 and 1·5 p.p.m., diatomite, kaolin and control. All the other data generally supported these results.

The extreme importance in this experiment of cross-infestation from nearby stacks has been shown in the analysis and has been illustrated by plotting numbers of T. castaneum against dust-weight data. Interpretation of this curve has lent support to work with lindane and malathion published by E. A. Parkin in 1960.

The results emphasise the significance of the contribution of grain and mould metabolism to the total bag weight-losses. Some estimates are made of the savings relevant to the admixture of malathion.

Field and laboratory experiments were carried out in Ghana on the tsetse fly, Glossina medicorum Aust., to discover the senses used by the fly in finding its host. In laboratory experiments on vision a model locomotive with a black screen attached to it was used as a ‘host’. The flies were stimulated to move by the movement of the screen. When the screen was stopped and started for short periods the increase in activity became less each time the screen was put in motion but the fall off was less rapid than in experiments with continuous movement. The possible significance of this is discussed in the light of the normal habitat of the species in the field.

In the field, G. medicorum did not respond to a moving object more than 25 ft. away but G. morsitans Westw. showed a good response at distances up to 150 ft.

Tethered flies were attracted to the model screen and showed a preference for landing on the lower edge near the centre.

Laboratory experiments using a guineapig and a goat as sources of smell produced no response from the flies but a strong reaction was obtained with the vapour of dilute acetic acid. In field experiments, using cows which were not visible to the fly, a very marked response to the smell of the cows was obtained. Orientation to a source of smell probably involves orientation to air movement and G. medicorum showed a strong upwind orientation in a wind tunnel even when blinded.

It is concluded that G. medicorum, which lives in thicket or forest where visibility is poor, probably responds initially to the smell of a potential host. An upwind orientation would then bring it to the general vicinity of the animal and the final approach might be visual.

Solutions of six chlorinated-hydrocarbon insecticides in kerosene have been applied in drops of about 0·02 microlitre (µl.) to adults of Glossina morsitans Westw., 2–5 days old, one day after the first meal. This species is found to be unusually susceptible to this group of insecticides. In order of increasing toxicity they are: DDT and aldrin, γ BHC, dieldrin and endrin, Telodrin (1,3,4,5,6,7,8,8-octachloro-3a,4,7,7a-tetrahydro-4,7-methanophthalan). The LD50 of DDT was 0·0165 µg. (males) and 0·025 µg. (females), that of Telodrin was 0·00062 µg. (males). Susceptibility of the two sexes to DDT, γ BHC and dieldrin did not differ significantly. Two batches of flies tested with an interval of eight months differed by about two times in their response to dieldrin and γ BHC; but the response to DDT was unchanged. This difference was not seasonal.

For practical use, dieldrin is the best and cheapest available insecticide, a fact confirmed in the field. Only Telodrin might replace it.

The cyclic biting activities of East African mosquitos of the genus Aëdes, belonging to the subgenera Mucidus, Diceromyia, Finlaya and Stegomyia are discussed. The total number of 24-hour catches involved was 448, all in East Africa, of which 401 were made in forest in Uganda.

It is pointed out that many species in the subgenera of Aëdes dealt with show arboreal tendencies of varying degree and, while a species may show a preference for some particular level, many of them appear to make daily vertical migrations. Consequently, biting may reach its maximum intensity at different times at different levels. It is also pointed out that the habits of a species may vary from one area to another, and therefore it is often best to discuss a representative series of catches from a productive area, noting such differences as may have been observed elsewhere. It has been found further that, under certain circumstances, a species may fail to show a clear pattern of biting behaviour and the possible reasons for this are discussed.

All members of the subgenus Mucidus so far adequately studied are arboreal and nocturnal. In the subgenus Finlaya, A. ingrami Edw. is a mosquito of the forest understorey which bites most freely in the hour before sunset. It makes vertical migrations, however, and the biting cycle at ground-level differs from that in the trees. A. longipalpis (Grünb.) is a diurnal species of the forest canopy. A view formerly put forward by the writer, that the form of the biting cycle in this species could be attributed to different physiological groups biting at different times, can no longer be supported.

In the subgenus Stegomyia, A. dendrophilus Edw. and A. apicoargenteus (Theo.) are diurnal species with rather irregular biting cycles, A. dendrophilus

Only a few individuals in the first generation of adults of Pieris brassicae (L.) reared in captivity will feed on artificially prepared flowers. By breeding from the eggs laid by these individuals and from a few others which have been helped to feed it is possible to maintain the stock. In each successive generation a few more individuals feed, and after about four generations most of the insects feed voluntarily. Finally, after many generations, all the insects use the flowers, and the experiments described are concerned with this adapted stock.

Many factors influence the attraction of the insects to the artificial flowers and the volume of food taken. Colour, blue or yellow, is responsible for attracting the adults to the flowers, and it is only when they are very close to the flowers and under the stimulus of colour that they show any detectable reaction to the odour of the honey solution in the flowers. Larger blue flowers are more attractive than smaller blue flowers.

The insects find the artificial flowers more easily, at first, if they are high in the cage. More honey solution is taken at 30 than at 20°C. and more in a large cage than in a small cage. When given a choice, the insects take more 10 per cent, (v/v) than 1 per cent, (v/v) honey solution and more 20 per cent, (v/v) solution than 10 per cent, (v/v) solution. When offered two of these solutions in separate cages the insects at first take less of the 1 per cent, solution than of the 10 per cent, solution. Later, presumably in an effort to obtain sufficient nutriment, they take more of the weaker solution than of the stronger solution. The insects live for a much shorter time on the 1 per cent, solution than on the 10 per cent, solution. Fresh honey solution is preferred to a solution 4–5 days old, and older honey solution is quite unsuitable as food. When the survival of adults fed on honey and sucrose solution is compared it is found that females, especially, survive longer when feeding on honey.

About 50 per cent, of the adults fed on 10 per cent, honey solution in the normal stock cage under glasshouse conditions live for 18 days. Very few survive for longer than 36 days. Starved insects all die by the llth day under similar conditions. When starved at 12·5°C. and 60 per cent, relative humidity, about half the males live for 17 to 19 days, but all are dead by about the 23rd day, whereas half the females live for about 29 days and some survive up to 40 days.

A study has been made of the development of Cryptolestes pusillus (Schönh.) at various combinations of high temperatures and humidities. Larval development is much quicker at 90 than at 70 per cent. R.H., and the optimal temperature is probably about 37°C. Under optimal conditions the developmental period is less than has so far been recorded for other species of Cryptolestes, the mean duration of the larval stage being as little as 8½ days at 33°C. and 90 per cent. R.H. The oviposition rate is similarly much increased at both high temperatures and high humidities. There is evidence that middle-aged females lay more rapidly than either younger or older individuals. Under optimal conditions the average oviposition rate can exceed 7 eggs/ ♀ /day, about 17 per cent. of the weight of the adult.

In cultures of adults on flour, there was no clear indication of a reduction of the oviposition rate or of a juvenile mortality at densities ranging from 1·5 g. per adult down to 0·03 per adult, but there were signs that the latter density was near a critical value.

Eggs of C. pusillus were cultured on whole English grain with the germs in poor condition, at humidities ranging from 50 to 90 per cent. The endosperm rather than the germ was attacked in almost all cases. At 50 per cent. R.H., only one individual out of 21 completed its development whereas at 90 per cent. R.H. few failed to do so. Humidity had a considerable effect on the rate of development in the larval stage. At 90 per cent. R.H. it was only slightly slower on grain than on flour.

An account is given of marking and release experiments with Anopheles gambiae Giles in a coastal area of Tanganyika. Laboratory-reared mosquitos were used, labelled either by the topical application of paint or by the introduction of radioisotopes into the larval breeding pans. Two different isotopes were used, 32P and 35S, and recaptures were recognised by autoradiography.

Routine catching stations were established within a circle of radius of 1¼ miles. Releases were made either in the centre or near the periphery of the experimental area, so that recaptures were possible up to a maximum of 2¼ miles.

The following results were obtained:

1. Of 132,000 mosquitos released, 1,019 were recaptured.

2. The mean flight range of females released in the centre was estimated to be 0·64 mile, and of males 0·52 mile. Of females released on the periphery, the mean range of dispersion was estimated to be 0·98 mile. Individuals of both sexes were caught at the maximum range of 2¼ miles.

3. The dispersion of recaptured mosquitos was shown to be non-random and to be related primarily to the distribution of human settlements.

4. In certain series of releases the direction of the prevailing wind had a definite effect on the dispersal of mosquitos. But in general this was a minor factor.

5. Dispersion during the first day or, in many instances, during the first two days after release was more restricted, compared with that of older mosquitos. But no difference in the distribution of catches was detected between those aged three-nine days and those more than nine days old.

6. Marked females were recaptured up to 23 days after release. Apart from the first two days, the regression of density on age amounted to a daily loss of 16 per cent, of mosquitos from the experimental area. The effect of emigration could not be assessed quantitatively, but it was held to be a minor component of the total daily loss. The relatively high level of mortality suggested by these figures is attributed to the use of laboratory-reared mosquitos.

7. The corrected sporozoite rate in marked females at the time of recapture was 0·8 per cent.

8. The survival of males was only slightly lower than that of females.

9. It is concluded from the survivorship curve that the mortality rate remained constant throughout the period in which marked females were recovered.

Seven strains of Aëdes aegypti (L.), three DDT-resistant (from Trinidad (2) and Haiti) and four susceptible (from West Africa (2), Arabia and U.S.A.) have been investigated with respect to their attraction to oviposition sites illuminated at different levels. The resistant strains have tended to choose the darker sites (at 0·02 lumen per sq. ft. and below) and susceptible strains those illuminated at between 3·5 and 6·5 lumens per sq. ft.

Genetical crosses between a ‘dark-laying’ (resistant) strain and a ‘light-laying’ (susceptible) strain indicated that ‘light laying’ was fully dominant over ‘dark laying’ and it is suggested that the mechanism involved is probably simple. No linkage with the gene for physiological resistance to DDT could be demonstrated.

The expressivity of this character varied in certain circumstances with a tendency among most strains to lay at darker sites after the second and subsequent blood-meals. In spite of these variations, behaviour at any one time remained reasonably predictable.

The reflectance of the oviposition site did not appear to be important in determining the difference in behaviour between ‘dark-laying’ and ‘light-laying’ strains. On the contrary, it seemed probable that ‘dark-laying’ strains were attracted more to the dark situation in which the dimly-lit site was found rather than to the site itself. There was some evidence of exploratory activity in the initial stages of oviposition by a ‘dark-laying’ strain.

Resting habits proved very similar in two strains differing markedly in oviposition habit. For both strains, 0·1 lumen per sq. ft. was apparently too low to be attractive and 12 lumens per sq. ft. too high. Besting behaviour did not appear to be under such critical control by light as was oviposition.

Possible explanations for the differences in ‘oviposition light preference’ are discussed.

If adults of Cryptolestes ferrugineus (Steph.), of which the females have already been fertilised, are cultured on flour at low densities, the rate of oviposition is independent of the density of adults and juvenile mortality is negligible. At high densities there is a heavy mortality amongst eggs and young larvae, and there is evidence that, providing no great degree of conditioning of the medium has occurred, an equilibrium is reached with the number of eggs plus larvae present at any time being independent of the density of adults. Cannibalism is undoubtedly a major cause of the juvenile mortality under these conditions. Where crowding of adults is so great that conditioning of the medium becomes important, the number of eggs plus larvae present at a given time after the cultures have been set up decreases with increasing density of adults, there being some indication of an inverse linear relation.

Cultures of C. ferrugineus on a mixture at 70 per cent. R.H. of sound grain and grain already slightly damaged by Crypiolestes are capable of initiating dense infestations in the laboratory, at least at moderate and high temperatures. Under optimal conditions the infestations develop within 15 days. Values deduced for the pre-adult period at various temperatures were similar to those obtained by Rilett (1949). At 17°C., development was extremely slow but the low temperature did not increase mortality. Cannibalism and competition by adults and fully developed larvae appear to play an important part in limiting the production and survival of the juvenile stages.

Sound grain at 70 per cent. R.H. is unsuitable as a medium for culturing Cryptolestes in the laboratory but it becomes suitable after relatively small additions of flour and damaged grains. At 33°C. at least, a point is suddenly reached where an infestation is initiated, the critical feature being increased survival of larvae rather than an increase in the number of eggs laid. An increase in humidity has a similar effect to adding food.

Estimates of the production of CO2 at various temperatures gave values higher than previously reported. From 21 to 33°C. the rate of increase of the respiration rate and the increase in the speed of development are proportional to rise in temperature. This is in accord with other indications that C. ferrugineus has a high optimum temperature.

Solutions of six organophosphorus compounds, Sevin, and pyrethrins (alone or synergised with piperonyl butoxide) were applied by microburette in drops of constant volume (0·0216 μl.) to the dorsum of the thorax of young adults of Glossina morsitans Westw., 2–5 days old that had taken their first blood-meal the previous day. The solvents used were decalin (decahydronaphthalene), toluene and lighting kerosene, respectively.

Malathion, methyl-parathion, DDVP (dichlorvos) and Sevin were eliminated in preliminary tests as insufficiently toxic. Muscatox (coumaphos) was reasonably toxic but not readily soluble and was therefore not considered further. Diazinon and Baytex (fenthion) were fully evaluated; the former was about as lethal as DDT, the latter as γ BHC (LD50 about 0·004 μg.). The LD50 of diazinon for males (0·0115 μg.) was significantly smaller than that for females (0·016 μg.) but the LD95 was much the same for both sexes.

Pyrethrins were about equitoxic with dieldrin (LD50, 0·002 μLg.); when synergised with 15 parts of piperonyl butoxide to one of pyrethrins the LD50 for males was reduced to less than half this value, and there was a significant difference in susceptibility of the sexes, the LD50 for females being 1·4 times that for males. However, the slopes of the regression lines were such that at LD95 the difference between synergised and plain pyrethrins was too small to be of any practical use.

These results show that, judged by innate toxicity to young flies, none of these insecticides can compete with dieldrin or Telodrin for practical control, although Baytex is a useful reserve should Glossina acquire resistance to chlorinated hydrocarbons.

Sorghum midge, Contarinia sorghicola (Coq.), was discovered in Nigeria in 1953, and a survey which indicated the widespread occurrence of the midge in the country was followed by the investigations reported in this paper.

Food-plants of the midge in Nigeria are guineacorn (Sorghum vulgare, sensu lato), which annually provides about two million tons of grain for human consumption, and the wild grasses, Andropogon gayanus and Sorghum arundinaceum.

At Samaru, Zaria, Northern Nigeria, midges emerge from infested guineacorn heads in the early morning with maximum emergence between 7.45 and 8.15 a.m. After mating, females fly to recently flowered heads where they lay eggs within the spikelets, laying twice as many eggs on the glumes as on the pales. Each female may lay about 50 eggs and both males and females usually die within ten hours of emergence. The egg hatches within four days and, after ten days' feeding, during which the ovary shrivels up, the larvae pupate within the spikelet. The cycle from egg to adult is completed in 19 to 22 days during the growing season but towards the end of the rains larvae spin cocoons and enter diapause.

Large numbers of diapause larvae are carried through the dry season in late-flowering heads which, because they are severely attacked by midge, are not harvested and remain on the stems which are kept in stacks and used for building and fencing or as fuel. Smaller numbers of larvae are present in threshing trash. The emergence of adults from the diapause population was observed from 1955 to 1959. In cages, the first adults were observed to emerge half way through the rains, about 5 to 7 weeks after the weekly mean R.H. had exceeded 60 per cent, and emergence continued for 9 to 12 weeks. Field observations confirmed experimental observations and showed that the build-up of the midge population before the main crop comes into flower is largely dependent on the presence of early-flowering varieties. So long as sufficient guineacorn is in flower and the weather is favourable the midge population builds up steadily to a peak in October and early November.

A. gayanus appears to be of little importance as an alternative food-plant, and preliminary evidence of the existence of a distinct biological race of C. sorghicola on A. gayanus is presented. Wild sorghum (S. arundinaceum) may be an important alternative food-plant in the south, where it is most abundant, but is of little importance in the main guineacorn-growing areas of the north where it is uncommon. Eupelmus popa Gir., Eupelmus sp., Aprostocetus sp. and two species of Tetrastichus parasitise C. sorghicola, and examples of the three genera are generally present at Samaru in the ratio 3 of Eupelmus: 1 of Aprostocetus: 1 of Tetrastichus. They are only of importance late in the season. Two spiders, a species of Thomisus and a species tentatively referred to as Diaea, prey on ovipositing midges but are apparently of little importance.

Experiments show a significant negative linear regression of yield on the proportion of spikelets attacked. There is no evidence of compensation and estimation of the proportion of spikelets attacked by midge gives a direct measure of the loss of yield. In 1957 and in 1958, random samples of guineacorn spikelets were taken from farmers' crops throughout Northern Nigeria. The proportion of spikelets containing midge larvae and pupae was estimated by dissecting 50-spikelet sub-samples taken at random from each sample, and the maturity of the sample was measured in each sub-sample by counting the number of spikelets which had not flowered. Samples should have been taken between 7 and 21 days after 50 per cent, of the crop had flowered, but many of the samples were immature and were discarded. In 52 mature sub-samples obtained in 1957, 10·5 per cent, of all spikelets contained midge and, in 167 obtained in 1958, 4·1 per cent, contained midge. In 1957, the results did not permit estimation of over-all crop losses but in 1958 it was estimated that at least 91,100 tons of grain, valued at £1,822,000 and representing the produce of 218,200 acres, were lost to sorghum midge.

The pattern of intensity of midge attack in the survey samples suggests that in the main guineacorn-growing areas, which lie above 9°N. at an altitude exceeding 1,000 ft., midge damage is less severe than in lower latitudes and at lower altitudes where heavier attacks probably result from the longer growing seasons, higher humidities and more abundant wild sorghum.

Nigerian farmers recognise the empty heads caused by midge but are unaware of the midge itself and, until they have learned to recognise cause and effect, control measures depending on their co-operation may be unsuccessful. Cultural control by disposal of crop residues and the growing of a uniformly flowering crop would be the most natural and effective method in the main guineacorn-growing areas. There is little possibility of achieving chemical control at present and, though the field resistance of the Nunaba group of varieties (Sorghum membranaceum) from the Gold Coast has been confirmed in Nigeria, laboratory experiments suggest that when the midge is not free to choose between resistant and non-resistant varieties it is able to adapt its behaviour and will then oviposit on resistant varieties.

This paper is intended to serve as a basis for further studies of sorghum midge in Nigeria and elsewhere in West Africa.

The situation following different degrees of discriminative clearing against G. morsitans Westw. was examined in eastern Ankole District, Uganda. In one area, where the treatment had been more intensive, the fly density was low, and it seemed that the population might be maintained by immigration from an adjacent area to the west, where treatment had been less intensive and in parts of which fly density was relatively high. The application of more intensive discriminative clearing of the fly's concentration sites to the western area, together with completion of the treatment of the eastern, failed to result in any marked effect on the tsetse population except as regards its distribution. Before this intensive clearing, concentrations of G. morsitans occurred in a double-storey vegetation type, the chief component of the upper storey of which consisted of Acacia gerrardii. After the upper-storey trees in this were felled, concentrations were found in groups of tall Acacia hockii, 12–18 ft. high, which had appeared to be of little importance to the fly before clearing. A year after the treatment of the western area there was no apparent reduction in fly density. The feeding habits and rate of utilisation of reserves of the flies showed no significant change as a result of clearing.

It seems that the failure of the discriminative clearing régime applied in this area, compared with the effect produced by less intensive clearing elsewhere in Ankole, is due to the presence in this area alone of the groups of tall A. hockii.