The principal object of this work was to elucidate the swarming behaviour of Aëdes cantans and other mosquitos, Chaoborus crystallinus and certain Chironomids. In addition, ancillary studies were made of the general ecology of these species. The swarms were found to consist entirely of males and to bear no direct relationship to mating or to the search for food. There was no noticeable difference between the swarming habits of the different species of mosquitos and even the differences between the Culicids and Chironomids were very slight. Swarming was observed to take place at about sunset in the evening and sunrise in the morning. The evening swarms appeared to be formed in response to decreasing light intensity and to disperse at a light intensity of about 7 Lux. Low temperatures may delay the start of swarming. The morning swarms started at about the same threshold of light intensity and at this time also their duration was reduced by low temperatures. Below about 50°F. they were not formed at all. Atmospheric humidity appeared to be of minor importance.

The methods used to investigate the properties of dusts are described and, in a theoretical section, the relevance of the various physical properties to insecticidal action are considered.

In order to investigate the effect of toxic and non-toxic dusts on insects the experimental procedure was simplified to eliminate all difficulties associated with the formation of uniform dust clouds and deposits. In problems relating to the adherence of the dusts to insects the actual quantity of dust was measured either by weighing the insects before and after dusting or by dyeing the dust with Sudan III and determining the quantity colorimetrically. The experiments were all conducted under known conditions of temperature and humidity.

Non-toxic dusts killed insects by causing them to lose water. Not all non-toxic powders were equally effective when conditioned to the same relative humidity. All were without effect at saturated humidity and became progressively more rapid in action as the humidity at which the test was carried out was decreased (p. 32).

The non-toxic dusts caused the insects to lose water by abrading certain areas of the cuticle ; the more extensive the abrasion the more quickly the insects died (p. 27).

To be effective as an abrasive the dust must be hard and finely ground and, perhaps also, sharply angular. Thus materials which ranked high in Moh's scale of hardness were in general more effective than soft materials and hard materials became quite ineffective unless they contained material below about 10 μ, diameter. Presumably the coarser materials could not gain access to the articulations, etc., where abrasion usually occurred. This effect can be seen clearly with carborundum powders. Sharply angular glass was more effective than the same powder converted into rounded spheres (pp. 25–31).

A historical account of the Ox Warble Flies, Hypoderma lineatum and H. bovis, is given and the economic importance of the two species is discussed. Certain aspects of their biology and control have been studied over a period of several years in South Wales culminating in intensive investigations in the years 1945 and 1946.

The adults of H. lineatum emerge from their puparia in early May and those of H. bovis a month later. The adults of both species in their persistent endeavours to oviposit, worry cattle, whenever the weather is sunny and calm, from the end of May to early September but they tend to avoid the vicinity of water and shade.

The degree of incidence is high in Britain, at least in South Wales, judging by the larval populations present in the backs of cattle. Over 80 per cent. of the cattle included in the surveys conducted in 1945 and 1946 harboured infestations of varying intensity; in one case 152 larvae were recorded on a Shorthorn heifer.

The larvae spend, on an average basis, 42·5 days in the subcutaneous tissues of the backs of their hosts. They begin to depart in order to pupate much earlier than is normally assumed to be the case in Britain, more particularly in Worcestershire, where the proportion of larvae pupating prior to March 28th was as low as 1·4 per cent. compared with over 20 per cent. in the present studies.

The marked variations commonly observed in the seasonal incidence of the larvae in their final instars on cattle are correlated with the relative prevalence of the two species in different districts.

The incidence of both species, judging by the intensity of infestations, is associated with the topographical conditions of the locality. For instance, herds on upland pastures are generally less liable to severe attacks than those on lowland farms.

Cattle under three years old normally carry much heavier infestations of the larvae of both species than those exceeding this age when kept under identical conditions during the oviposition period.

Diphenyl-dichloro-trichlorethane and benzene hexachloride preparations are ineffective for destroying the larvae, at least in their final instars, but emulsions of the former insecticide when applied in the form of a spray to the legs and flanks of cattle during the period of oviposition proved highly promising and warrant further investigations.

An investigation of the seasonal incidence of I. ricinus on cattle in N.W. Cardigani shire showed that there are normally two peaks of activity, one occurring in spring and the other in autumn. The infestation curve varied considerably from farm to farm according to the husbandry methods practised. Delayed stocking of infested pastures caused an initial higher infestation of the cattle but, except on one farm, did not prolong the infestation beyond the normal period of tick activity. A lower infestation in spring than in autumn occurred on two farms. This resulted from either the lighter stocking of the infested grazings in spring than in autumn or the partial exhaustion of unfed tick population by sheep grazing with the cattle during the spring only.

There was no evidence for the occurrence of a unimodal curve of tick activity in the region studied.

Ixodes ricinus is the common tick infesting farmstock in Wales. Dermacentor reticulatus is recorded on cattle and sheep on two farms in N. Cardiganshire.

I. ricinus is widely distributed in the four major grassland zones of N.W. Cardiganshire. The most extensive areas of tick infestation are found on the marginal and hill farms. In lowland areas, except on extensive stretches of marshland, infested pastures are confined to isolated tracts of badly managed or poorly drained land. Infested grazings are invariably “rough” and contain one of the following plants as dominant: Festuca spp., Agrostis spp., Molinia caerulea, Nardus stricta, Pteris aquilina or Juncus spp. The degree of infestation of cattle depends on the relative area of the available grazing land colonised by the tick. Continuous stocking of infested pastures results in heavy infestations and vice versa.

Bovine piroplasmosis is the major tick-borne disease in N.W. Cardiganshire and predominates in the lowland districts. Outbreaks of the disease on marginal and hill farms are generally infrequent. Reasons are suggested for this phenomenon. Tick pyaemia is confined to lambs on hill grazings. The incidence of this disease is low.

The distribution of the tick in Wales and the border counties of Hereford and Salop shows a general relationship to areas of rough grazing. The tick is widespread in the western counties and is almost absent in the border counties. Radnorshire and Breconshire are interesting in that they have few infested farms in relation to the extensive nature of the “rough” grazing. A survey has also been conducted of the incidence of bovine piroplasmosis in Wales.

The variations in susceptibility of different varieties of peas to attack by the Pea Moth was investigated and an attempt made to determine and measure the factors concerned. Six varieties of peas differing widely in haulm length and earliness of maturation were used in each of two trials. In the first trial (sown 29th March) the early maturing varieties came into flower before the moths were recorded on the crops and suffered the lowest attacks. The later varieties were exposed to attack over a much longer period and suffered the heaviest infestations. In the second trial (sown 3rd May) the attack was more uniform over all varieties with the early varieties more heavily affected than in the first trial; they were exposed to attack from the beginning of flowering until harvesting.

An estimate of the changes in the active moth population during the flight period was obtained and the varieties were compared in relation to the proportion of this population to which each had been exposed. There was a strong positive correlation between the degree of exposure and the incidence of attack on the different varieties.

The infestation of the varieties was also found to be influenced by the amount of cover which each provided; those with the most dense cover suffered the heaviest attacks.

Statistical analyses showed that the two factors, exposure and plant cover, were closely associated and exerted a joint influence on subsequent attack.

Data from other trials corroborated these findings and showed that strains of peas bred to mature early suffered substantially lower pea moth attack than did the later maturing types from which these had been bred.

Experiments were carried out to investigate the persistence of toxicity to tsetse of hessian, impregnated with DDT, exposed to the normal weather conditions of the Inland Savanna Region of the Gold Coast; also, to establish a satisfactory method of estimating the loss or change in its toxicity, and to study the effects of impregnation with DDT on the efficiency of tsetse traps.

Hessian was impregnated by being soaked for ten minutes in a filtered saturated solution of commercial DDT in kerosene which gave 4·4 g. DDT per sq. ft. and tests were made at intervals with batches of Glossina tachinoides at 15- and 60-second contacts. The comparison of the survival times of these flies compared with those of controls showed an appreciable residual toxicity after 20 weeks  full exposure to the weather including 4·26 ins. of rain, a slight toxicity after 30 weeks, and a negligible toxicity after one year including 34·35 ins. of rain. The same preparation could still produce toxic symptoms in tsetse after 15-second contacts when exposed for one year to wind and daylight, but protected from rain and direct sunlight. There was no great difference on the whole between the results from 15-second and 60-second contacts.

It was found that survival times lengthened in October when the humidity was high but were shortened in December when the conditions became hotter and drier.

A further experiment was carried out with hessian carrying 4·4 g. DDT per sq. ft. and super-impregnated hessian with 7·4 g. DDT per sq. ft. Samples of both were exposed to the weather and others kept as controls. All samples were tested by 30-second contacts with G. tachinoides.

It was shown that contact with the super-treated cloth was invariably fatal; the standard cloth had an 82 per cent, toxicity but the flies that received a lethal dose took longer to die on the whole. It was found to be the survival time and not the toxicity that varied with seasonal conditions.

A full analysis of the survival times from the exposed samples in relation to those from the control samples showed that exposure to weather during late rains followed by dry season conditions caused only a slight but progressive deterioration of the killing power of exposed DDT-impregnated hessian. This amounted to 10 per cent, in 3–4 months including 7·42 ins. of rain in the case of standard cloth, and 10 per cent, in six months with the same amount of rain in the case of the super cloth.

Hessian impregnated with 25 per cent., 50 per cent, and neat DDT emulsion was found to be highly toxic and to retain its efficiency satisfactorily for at least 12 weeks' exposure in the dry season.

Hessian freshly impregnated with DDT in kerosene solution caused flies to die more quickly as long as traces of kerosene persisted but moisture, on the other hand, interfered with the action of DDT unless kerosene was also present.

An “animal” tsetse trap covered with standard impregnated hessian showed a 40 per cent, superiority in the numbers of G. palpalis and G. tachinoides taken over a period of 20 weeks. Corresponding to the results of the laboratory tests this superiority was lost in October but reappeared in December.

Comparisons of catches made by traps covered with standard impregnated hessian and hessian sprayed with 25 per cent. ATSO emulsion showed that, although the emulsion might initially give much better results, its superiority relative to the standard impregnation tended to diminish, especially during the wet season.

The increased efficiency of an impregnated trap is more than the measurable amount of the increased catches since flies which have investigated the trap, but have not been caught, are likely to have received a lethal dose of DDT.

The application of DDT to traps represents an economical use of the insecticide which can conveniently be applied in sufficient concentration to ensure that brief contacts will be fatal to tsetse over a period of several months.

Mosquitos continue to enter occupied, untreated native huts throughout the night, with peak periods of entry at dusk and dawn. Early morning mosquito catches do not give a true picture of the numbers entering and leaving huts during the night.

In a series of catches 63 per cent. of 5,576 mosquitos and 79 per cent. of 506 Anopheles gambiae were caught resting on the underside of the thatch roof.

By using five traps inserted in apertures one foot below the top of the wall, the numbers of mosquitos attempting to leave a hut were determined. Of 1,014 mosquitos entering huts before 10 p.m., 63 per cent. remained inside until 6.30 a.m., that is for 8½ hours. Catches to estimate numbers entering and leaving at different times during the night were also made.

Treatment of huts with DDT wettable powder and DDT-kerosene solution did not interfere with the normal behaviour of mosquitos as far as entry was concerned. Biting occurred in the treated huts.

The DDT wettable powder appeared to be more effective than the DDT-kerosene solution.

Some mosquitos entered the treated huts, fed and then left before acquiring a lethal dose. After making contact with treated surfaces mosquitos became restless but, under the conditions existing in the huts during the experiments, activation did not result in more leaving the treated huts than the untreated one. Unfortunately there were few A. gambiae and the predominant species entering the huts was Taeniorhynchus fuscopennatus.

Some of the female A. gambiae released into unoccupied DDT-treated huts escaped into the traps before acquiring a lethal dose. Although there was a tendency for more to enter the traps of a DDT-treated hut than those of an untreated hut, the data are insufficient to show a significant difference.

The majority of mosquitos entering the traps did so within one hour of their release.

No mosquitos were still alive 12 hours after their release in huts treated 17 weeks previously with DDT wettable powder or DDT-kerosene solution, or in the hut treated 12 weeks previously with "“Gammexane” wettable powder.

The brackish water form of A. gambiae on the East African coast—and probably in Mauritius—is not the same as A. melas of West Africa.

In salt-water gambiae a variable proportion of the females have an additional dark band on the palps, resembling 4-banded melas, but the remainder are indistinguishable from typical gambiae.

Eggs and larvae of salt-water gambiae show no morphological differences from those of fresh-water gambiae, thereby differing from A. melas of West Africa.

Larvae of the two forms show a clear-cut difference in reaction to sudden changes in salinity, and a simple test has been worked out whereby wild-caught females can be accurately identified by the reactions of their progeny.

This physiological test has formed the basis of all work in comparing the incidence, habits, and infectivity of salt and fresh-water gambiae in Dar-es-Salaam.

Exposed to equal chances of infection in the same village during 1947 and 1948, fresh-water gambiae had a sporozoite rate of 9·4 per cent. while that of salt-water gambiae was 0·8 per cent.

About 4 per cent, of both forms were infected with filaria larvae, but monthly figures showed that infection rates in salt-water gambiae may rise to 22 per cent.

Fresh-water gambiae show little tendency to leave African houses at dawn after feeding, whereas in salt-water gambiae over one-third of freshly blood-fed females leave the house at dawn.

In fresh-water gambiae many half-gravid females leave the shelter of the house at dusk on the night after the blood feed. There is no marked difference in infectivity between those which leave the hut and those which remain indoors at this stage.

Blood-fed and gravid females of fresh-water gambiae, funestus, and salt-water gambiae have been found in outdoor resting places, gravid females predominating in the case of the first two.

Although larvae of salt-water gambiae can complete their development in pure sea water, in nature increasing salinity becomes a limiting factor before it reaches that of sea water, continuous breeding being no longer possible at salinities over 83 per cent. sea water.

Salinity as a limiting factor explains the rather restricted breeding of salt-water gambiae on the coast, and suggests that certain coastal fresh-water swamps at Dar-es-Salaam could be cleared of all Anopheline breeding by salinifying with sea water.

The paper describes field and laboratory investigations on the bionomics of Cephalonomia waterstoni, a Bethylid parasite of Laemophloeus spp. A table is given in which are listed all the Bethylids attacking insect pests of stored products to which reference could be found in the literature.

An infestation of Laemophloeus, associated with two “hot spots” in Manitoba wheat, which supported a large population of Cephalonomia is described.

A simple technique for the laboratory culture of Cephalonomia is described.

The life-cycle of C. waterstoni with Laemophloeus ferrugineus as host has been worked out.

The lengths of egg, larval and cocoon (prepupal and pupal) stages at combinations of 25°C, 30°C. and 60 per cent., 80 per cent. R.H. are given. The egg and larval stages are short, lasting for about six days at 25°C. and four days at 30°C.

Within the limits used, the relative humidity appears to have no effect on the duration of development at any stage. On the other hand, temperature exerts a considerable influence; the life-cycle at 30°C. is completed in two weeks but at 25°C. it takes three weeks.

Again within the limits used, the mortality appears to increase with decrease in saturation deficit. Mortality ranged from 9 per cent, at S.D. 12·7 mm. to 36·5 per cent, at S.D. 5·0 mm.

Without food or water at all combinations of 25–30°C. and 60–80 per cent. R.H. adults live for about four days, with a range of 0·5–9·5 days. There is no difference between the sexes. Unexplained contradictory results were obtained in two experiments.

With normal or paralysed host larvae available at 30°C. and 80 per cent. R.H., males live no longer than when no food or water is available but females live for about five weeks at 25°C. and 80 per cent. R.H. and for about four weeks at 30°C. and 80 per cent. R.H.

Males fed with sucrose solution at 30°C. and 80 per cent. R.H. live for several days longer than when starved : females live for the same length of time as when fed with host larvae.

The pre-oviposition period at 25°C. and 80 per cent. R.H. is about five days; at 30°C. and 80 per cent. R.H. about one and a half days.

Fecundity. At 25°C. and 80 per cent. R.H., Cephalonomia lays about 40 eggs on 30 host larvae : at 30°C. and 80 per cent. R.H., about 65 eggs on 40 larvae.

Cephalonomia females readily oviposit on larvae that have been paralysed some time previously, and can be induced to oviposit on larvae already bearing eggs.

Virgin females produce only male offspring (arrhenotoky).

Eggs are laid in groups of one, two or three (rarely four) per larva. Single eggs produce mainly females ; pairs produce mainly one male and one female ; trios produce mainly one male and two females. At 25°C. and 80 per cent. R.H. more single eggs are laid than pairs ; at 30°C. and 80 per cent. R.H. more pairs are laid than singles. This results in the production of a higher proportion of females at 25°C. than at 30°C. The incidence of trios at both temperatures is low.

C. waterstoni is equally viable on L. minutus, L. ferrugineus and L. turcicus but shows a marked preference for L. ferrugineus.

An account is given of the life-history of Auchmeromyia luteola in a laboratory culture maintained in London for two-and-a-half years. Attention was centred on the bloodsucking larva, known as the Congo Floor Maggot, which is an intermittent ectoparasite specific to man. It was reared chiefly on the natural host, but a strain has been maintained on shorn guinea-pigs through several generations. It was also found possible to rear the larva on free blood.

The known distribution of A. luteola is reviewed on the basis of published records, museum collections and information from scientists in Africa. The species is highly successful in both the wettest and the driest parts of the Ethiopian region, but does not seem to extend south of Durban. It can flourish only where man occupies permanent settlements and makes his bed on the floor within reach of the maggot. Strains originating in Nyasaland, in the western part of the Belgian Congo, and in the Anglo-Egyptian Sudan were successfully cross-mated and produced fertile eggs. The second generation from these crosses, however, was not always fully fertile.

The method of cultivation of the material is described.

Batches of eggs hatched 36–60 hours after oviposition when kept at 26–28°C. and 50–50 per cent. R.H. In drier atmosphere development was delayed and took from 3 to 7 days at 23°C. and 10 per cent. R.H.

The habits of the larvae are discussed and it is shown that larvae took 20 minutes to gorge (although newly hatched specimens often stopped feeding after 10 minutes). Given the opportunity, a meal was taken daily except for a day missed before each moult. No larva could be induced to bite twice in one day unless the first feed had been interrupted.

The rate of growth of the larvae was found to be strongly influenced by the feeding schedule, those receiving. 4 meals a week having a much higher rate than those receiving only 2. There was also a correlation between rate of growth and temperature but the correlation with relative humidity was less well marked. After fasts of several days, meals of the order of two and a half times their own weight were taken by larvae of all three instars.

Larvae which were fed four times a week and kept at 23°C., besides growing faster than others fed at less frequent intervals, reached higher maximum and pupal weights and started to pupate on the 26th day. Larvae fed at less frequent intervals took larger meals but lighter pupae and flies were produced.

The time of moulting appeared to be directly related to weight and only through this to the environment. The first moult occurred after the larva had gorged to a weight of 1·5–2·1 mg. (usually at the 2nd or 3rd meal), the second moult after it had reached 12–19 mg. (at the 4th to 7th meal).

There are wide limits of weight between which the onset of pupation can occur, the upper limit being higher for female than male larvae. If the feeding schedule (in relation to climate) is favourable, the larvae reach a maximum weight which induces the onset of pupation regardless of other factors. If the meals are scarce, growth is retarded more than metabolism and time becomes the limiting factor; then, the scarcer the meals, the smaller, not the later, the pupae. If the meal schedule is so adverse that the larvae cannot reach the minimum weight for pupation (about 97·5 mg.) in the time set by their metabolic rate in the given climate, death ensues without pupation.

No larva has been known to complete an instar on a single meal. The minimum for complete development in any climate is probably six meals, two in each stage. Larvae reared at 28·5°C. in 60 per cent. R.H. on five meals a week, moulted after the rd and 7th meals and pupated after the 16th or 17th.

Failure to moult, followed by death, occurred in all strains in the laboratory. It is believed that this is sometimes due to overcrowding and sometimes to the larva being disturbed when ready to moult. Inability to moult was commonest in the dry atmospheres and among larvae fed only twice a week.

Saturated air was fatal to the larva but not to the pupa. The species tolerates a wider range of atmospheric humidity than most insects, and was even reared successfully in an atmosphere of 10 per cent. R.H.

Female larvae in the third instar take more blood than males and lose more by excretion ; they also grow larger and produce heavier pupae and adults.

The temperature lethal in one hour to larvae having completed the second moult is denned within about one degree (42·5–43·5°C.) and does not appear to vary according to atmospheric humidity.

The Floor Maggot can survive fasts perhaps longer than any other Dipterous larva. At 28·5°C. and 90 per cent. R.H. survival of first-instar larvae, unfed, after one meal and after two meals, was 9–20, 8–21 and 4–21 days respectively ; third-instar larvae after moult (i.e. at 12·19 mg.), and at about 90 mg. survived 17–18 and 28–47 days respectively. At the same temperature but 10 per cent. R.H. much more weight was lost and the survival time was much less. At 23°C. newly hatched unfed larvae survived 5–37 days at 60 per cent. R.H. and 9–22 days at 10 per cent. R.H., while third-stage larvae survived 25–48 days at 90 per cent R.H. and 9–19 days at 30 per cent. R.H. Newly hatched larvae lived for 3–8 days at 35°C. in 60 per cent. R.H.

The pupal stage lasts about 9 days at 34°C., 11 days at 28·5°C. and 15–16 days at 23°C. and is unrelated to atmospheric humidity. Losses in weight at different atmospheric humidities were studied; the proportion lost was about 16 per cent, at 90 per cent. R.H. and about 25 per cent, at 10 per cent. R.H.

The habits of the adult flies are discussed ; human faeces appear to be the staple diet. The male seeks the female persistently and mating is protracted and occurs repeatedly. One male can fertilise several females. Oviposition and development continue all the year round without diapause. In the laboratory at 22–23°C. a female normally laid about 54 eggs at her first oviposition, and in one case a female laid as many as 6 batches of fertile eggs. In warm weather, the first batch was laid about 16 days after emergence (or after 20–23 days at 23°C), smaller batches being laid subsequently at intervals of 5–8 days. The female lived in the laboratory up to 93 days and the male up to 85 days.

The life-cycle under natural conditions is roughly estimated as 10 weeks, so that five generations a year might be expected.

Comparative tests have been carried out in the laboratory to assess the relative toxicities of Chlordane, BHC and DDT to the leather beetle, Dermestes maculatus. The method used was simple enough for routine industrial purposes when comparisons of new, untried contact insecticides are required, and consists in a measurement of the toxicities of solid films produced by a blood-smear technique on glass photographic plates from standard solutions of insecticide.

Among factors influencing the application of the method, the most important were size of insect and size of crystal clusters as influenced by the volatility of the solvent used to form the film. From LD50 measurements, BHC and Chlordane proved, respectively, 140 and 600 times as toxic as DDT to the leather beetle. DDT had comparatively little effect on eggs, larvae or adult beetles.

A controlled technique is described, whereby the comparative protective values of various dips and sprays used in sheep maggot fly control can be accurately assessed.

In order to determine the exact mode of action of the agents under test, each is examined from the point of view of its power to inhibit oviposition by the adult fly and also for its larvicidal properties.

The “anti-adult” test involves the weekly exposure of treated groups of sheep to an active and dense fly population, maintained in a closed chamber under controlled conditions of temperature and light.

The larvicide test is a modification of a former method of larval implantation, and is applied to the sheep when the “anti-adult” protection, indicated by the previous test, is found to have failed.

In the second part of the paper, a number of results are presented which serve to illustrate the working of the technique, and demonstrate the comparative value of several methods of maggot fly control which are in current use.

A dip containing 0·5 per cent. DDT in water suspension was found to give the most satisfactory protection, and to achieve this result purely by its action against the adult fly. A 0·5 per cent. DDT benzene emulsion dip was less satisfactory, as alsowas spraying as a method of applying the insecticide. In general, spraying has given about half the protection of dipping, whilst on sheep in full fleece, the application of half a gallon of spray was found to be inadequate. Protection was again found to be poor when the fleece was carefully dressed with half a pound of a dry powder containing 5 per cent. DDT.

In contrast to DDT, BHC and arsenic were shown to possess virtually no “antiadult” properties at the concentrations tested. BHC and arsenic both acted as efficient larvicides and provided a valuable and similar degree of protection against the sheep maggot fly, as an incidental effect, when they were tested at the concentrations normally used for their primary function of sheep scab control.

Two strains of a virus disease, “A” and “B,” which possibly may be two distinct viruses, occur on cacao in Trinidad. These are compared with the more virulent “swollen-shoot” and related viruses that are widespread in West Africa.

The most important resemblance between the Trinidad and the West African viruses is that both are carried exclusively by mealybugs of the family Pseudococcidae. There are also points of similarity in the symptoms, which in the Trinidad virus consist mainly of a transient red vein-banding, with or without a more or less discontinuous yellow vein-flecking (which does not disappear when the leaf matures) and, on certain varieties of cacao, red-mottle on the pods. Swellings on the shoots, a conspicuous symptom of most of the strains of the West African viruses, have not been observed in Trinidad.

Four species of mealybugs are definitely known to be vectors: Pseudococcus citri, which is the commonest and undoubtedly responsible for most of the natural spread of the disease; P. brevipes; a species near P. brevipes but almost certainly distinct; and Ferrisia virgata. Certain other mealybugs have been found on cacao in Trinidad but the virus has not yet been transmitted by them.

Most transmission experiments have been made with mealybugs bred up on potato tubers and with Posnette's technique of feeding potential vectors, after infection-feeding, on cacao beans from which one of the cotyledons has been removed, so that the insects can feed on the convoluted surface of the remaining cotyledon or on the radicle. Notwithstanding some drawbacks to this method it has many advantages over using young growing plants as tests.

The “latent period” of the virus in the test plants grown from dissected beans has varied between 20 and 123 days, though the symptoms most commonly appear from 30 to 50 days after infection of the bean. The symptom of red vein-banding has appeared much more often than that of yellow vein-flecking, irrespective of the symptoms on the virus source plant. When both symptoms appear the red veinbanding almost always shows up first.

All three immature instars, and young adults, of all four vector species are probably almost equally efficient as vectors, except that there is some indication that P. brevipes may be slightly more efficient than the other three species.

Starvation of the mealybugs before infection-feeding does not increase their capacity to transmit, though it usually helps to make them settle and feed more readily on the source plant.

The time for which the mealybugs feed on the virus source plant has little if any effect on their capacity to transmit the virus. Mealybugs have become infective in just over one half-hour infection-feeding, though the proportion of transmissions obtained when the infection-feeding time has been between one-half and two hours is rather less than that for over two hours, probably because in the former tests some of the insects had not actually fed.

The duration of feeding on the test bean is probably also immaterial, provided of course that the mealybugs do actually feed. The shortest test-feeding time that has resulted in transmission is 90 minutes, but this included a considerable “settling” time.

Transmission may be effected whether a mealybug feeds on the cotyledon or the radicle of a test bean.

Mealybugs can still transmit the virus if they are starved after infection-feeding for a period up to 22½ hours, but no transmission has yet been obtained when the post-infection starvation has been over 23 hours. It is not yet known whether a short period of feeding after infection-feeding renders the mealybugs incapable of transmitting, but post-infection feeding for 66 hours does so.

A transmission rate of just under 14 per cent, has been obtained with single mealybugs to each test bean but with larger numbers the rate has not risen in accordance with the mathematical expectation. A possible explanation for this might be that there are “active” and “inactive” races of P. citri, but so far practically no evidence for this has been obtained.

Mealybugs can become infective by feeding on the symptom-free parts of flush leaves showing symptoms in other parts, on entirely symptom-free flush leaves from infected trees, on the stem of a young infected plant, and on the leaves of a young infected plant after the disappearance of the transient symptoms of red vein-banding. It appears, however, that they pick up the virus more readily from flush leaves actually showing symptoms, and these have been used as the source of the virus in most of the experiments.

Few experiments have yet been made with strain “B” of the virus, but this strain has been transmitted with P. citri and P. brevipes.

All stages of P. citri, but especially young adults, often wander about of their own accord and are thus capable of spreading the virus, particularly to trees actually in contact with an already infected tree. The transport within a plantation of cacao pods during harvesting is thought to be a likely cause of isolated new infections.

Female ticks are almost entirely confined to head, axillary and inguinal regions of cattle. Larval and nymphal stages attach on the head and below the hock joints on the fore- and hind-legs.

The percentage of females on the forequarters of cattle increases during the course of the season, whether in spring or autumn; during the early stages of a season's activity the hindquarters carry a greater number than the forequarters. Later in the season the position is reversed. Suggestions are put forward to account for this phenomenon.

The head, axillary and inguinal regions are the major attachment sites of the female tick on sheep. In ewes the head region carries the higher percentage, but on lambs the axillary region is the most heavily infested. This difference may be due, to some extent, to the fleece of the ewes forming a barrier to the movement of unfed ticks towards the axillary and inguinal regions.

Lambs carry a heavier infestation between the 21st April and 6th June than ewes on the same grazing. This phenomenon has been discussed in the light of previous work in Northern England.

There is no reliable method of estimating the total infestation on cattle. A good picture of the seasonal incidence of the female tick on the host can be obtained from the hindquarters count. The standard count (the forebody) on sheep may be used for estimating total infestations.

Under the conditions prevailing at Kilifi G. pallidipes is evenly distributed throughout all vegetation types with the exception of cultivation from which it is at all times almost entirely absent. There is a preference for the denser vegetation in the dry season from which the tsetse spread out into the lighter vegetation in the wet and humid seasons. There is never at any time any concentration of G. pallidipes along the edges of extensive thicket or forest.

The distribution of G. austeni is similar to that of G. pallidipes except that there is no tendency to spread into lighter vegetation in the wet season.

There is some reason to think that G. brevipalpis prefers thicket of medium density and does not share with G. austeni a preference for tall heavy thicket.

G. pallidipes was the only species staged for hunger, except for austeni in the Sokoke forest. Males are replete at all seasons. G. austeni showed a remarkably large proportion of recently fed males.

Pupae and pupal cases of G. pallidipes were found on the floor of the coral rag thicket and forest. G. austeni and G. brevipalpis showed a preference for the undersides of large logs.

The entomology of swollen shoot of cacao is a complex and unique problem involving the inter-relations in the field of over 120 insect species of four insect and two arachnoid orders. Briefly, there are 17 species of pseudococcids, 75 species of ants, 16 species of hymenopterous parasites, three predatory beetles, one predatory dipteran, and three arachnid species involved in vector relationships directly, with a further 18 Coccid species involved indirectly (it is possible, of course, that further work will show that some of these 18 species are directly concerned as vectors).

In the present paper an attempt has been made to reduce this assemblage of insect material to some semblance of order. The Coccid species are named and a series of preliminary observations on their biology and field behaviour detailed. The ant species, some of which are obligatorily associated with certain vector species, have been sorted into groups where specific determination has proved impossible or unnecessary and information has been included on their field habits and relative abundance.

There are three distinct but complementary ecological niches involved in the problem. The first, and most important, is the association between the mealybug virus vectors and the Myrmicine coccid-tending ants. The second is the association between mealybugs of the genera Paraputo and Formicococcus and the wild forest tree hosts of swollen shoot virus, and the third is the negative association between the mealybug tending Myrmicine ants and Oecophylla and Macromischoides, the latter species acting in certain circumstances as barriers to the spread of the mealybugtending species and hence to the spread of mealybugs and virus. These problems will be dealt with on a quantitative basis in a further paper.

Some of the structures illustrated here have already been noticed in two previous papers (Jackson, 1948a & b). All the present drawings were made by one of us (E.B.) from living material of Glossina swynnertoni Austen. The eggs and the first- and second-instar larvae were dissected from wild female flies and examined in saline under the microscope; the third-instar larvae were obtained by keeping females in captivity until larviposition occurred, normally in the late afternoon.

Coccids on cacao have received much attention in recent years in connection with the transmission of swollen-shoot disease in the West African cacao belt and many papers on them have been published by British entomologists.

♂ (Type). Antennae somewhat longer than the head and pronotum together. Frontal ridge flat, punctured, gradually narrowed to the apex, its margins obtuse. Fastigium of the vertex slightly concave, oval, with rounded apex and almost smoothed lateral carinulae. Occiput with a scarcely distinguishable median carinula.