Larvae of the African midge, Chironomus brevibucca (Kieff.), are abundant in the littoral zone of Lake Bangweulu, Northern Rhodesia, where, at certain times of the year, imagines emerge in vast numbers shortly after the appearance of the full moon. Attracted by light, the imagines fly to houses and alight on the protective gauze screenings, which, at peak periods of emergence, become covered with a solid mass of midges. They have not been seen at other phases of the lunar cycle, so the adult life is certainly not more than ten days and is usually much less.

Imagines, which are inactive by day, tend to settle in different situations according to their sex. On the gauze screenings there was a marked preponderance of females (4·68:1) on the day following a night's swarming but, on several adjoining whitewashed walls, males predominated (2·92:1). On grasses near the lake margin females outnumbered the males.

When, the gauze screening of a house was diffusely illuminated, imagines were attracted to it during the early hours of darkness. The first flies to arrive were almost invariably males, but females followed on rapidly and finally constituted the bulk of the population. On one evening the percentage of males fell in three hours from 95·4 at 1844 hr. to 29·5 at 2145 hr. The percentage of males reaches a minimum between 1900 and 2000 hr. and thereafter shows some rise, probably due to a secondary influx of males which are thought to have been resting during the day and not to be newly emerged.

An indication of the adult age of males seems to be given by the condition of the antennae during the early part of their adult life.

Some miscellaneous observations on the behaviour of C. brevibucca are recorded.

Some aspects of the phenomena recorded are discussed. The recent hypothesis of Korringa, which seeks to explain the periodíc swarming of certain polychaetes just after the time of full moon as being the result of an additive effect of exposure to nocturnal illumination on the maturation of organisms that are not subjected to sunlight by day, is examined and shown to be applicable to C. brevibucca.

An insectary method for rearing the cacao Mirids, Distantiella theobroma (Dist.) and Sahlbergella singularis Hagl., is described. Adults are caged, for oviposition, on cacao seedlings grown in baskets. When the eggs hatch, the nymphs are transferred to, and reared on, unripe pods suspended in ventilated glass cages.

Laboratory and field tests were made to investigate the toxicity, mode of action and persistence of technical DDT and γ BHC when applied to cacao in Ghana to control the Mirids, Distantiella theobroma (Dist.) and Sahlbergella singularis Hagl. The insecticides were tested at concentrations previously used in field trials. D. theobroma, S. singularis and Tribolium castaneum (Hbst.) were used as test insects. Preliminary tests with aldrin, chlordane and dieldrin were also made.

Cacao pods were sprayed to drip point with the test insecticides and the deposit was allowed to dry. Mirid nymphs were placed on the pods and the rate of knockdown and the mortality after exposure for a test period was recorded. These tests showed that BHC had a higher toxicity and acted much more quickly than DDT. BHC greatly reduced feeding.

Fumigant action was tested by spraying cacao leaves to drip point with the test insecticides, allowing the deposit to dry, and then exposing Mirid nymphs and adults over discs cut from the leaves. BHC had a powerful but transient fumigant action; adults, fifth- and second-instar nymphs were killed after, respectively, ½, 4 and 3 hours’ exposure at 28–30°C. over leaves treated with 0·25 per cent, γ BHC. No fumigant effect was observed after prolonged exposure over leaves treated with 2·5 per cent, technical DDT.

Persistence was studied by treating cacao leaves and pods growing under natural conditions and then testing the residual deposit at intervals. Tests with T. castaneum showed that when leaves were sprayed to drip point the residual toxicity of 1·25 and 2·5 per cent, technical DDT fell steadily and was negligible after an interval of three and six weeks, respectively; that of 0·25–2 per cent, γ BHC was negligible after 2–3 days. Corresponding tests with Mirids on pods were less consistent but suggested that 2·5 per cent, technical DDT was still effective against first-instar nymphs after three weeks; 0·25 per cent, γ BHC had a residual toxic effect for at least several days and depressed feeding greatly for at least a fortnight. It is suggested that BHC may have a local systemic action which persists after the surface deposit becomes ineffective.

Of the other insecticides tested, aldrin had a high toxicity and powerful fumigant action and appeared to be the most promising alternative to BHC.

It was concluded that BHC would be more effective than DDT when treating mature cacao because, in addition to its higher toxicity and quicker action, its fumigant action would compensate for incomplete coverage by low-volume spraying.

Fluoroacetamide has been tested as a systemic insecticide against Aphis fabae Scop., Brevicoryne brassicae (L.), Myzus persicae (Sulz.) and the eggs and larvae of Pieris brassicae (L.). In the majority of tests, sodium fluoroacetate and schradan have been included for comparison.

In dipping tests against A. fabae and B. brassicae, fluoroacetamide was as effective as sodium fluoroaeetate and both were much superior to schradan. Fluoroacetamide was superior to sodium fluoroacetate in preventing eggs of P. brassicae from hatching and they were about equally effective against the newly hatched larvae. Schradan was already known to have little action on the eggs or larvae. Neither of the fluorine compounds was very effective against fourth-instar larvae.

When watered on to the soil, fluoroacetamide was more readily absorbed from sand than from soil—like other systemic insecticides. Against A. fabae it was as immediately effective and persistent as sodium fluoroacetate and both were superior to schradan. To B. brassicae and M. persicae, fluoroacetamide was more toxic than sodium fluoroacetate, M. persicae being somewhat more resistant to fluoroacetamide and much more resistant to sodium fluoroacetate. Slight phyto-toxic effects were observed only at the higher dosages.

When poured on to the soil around the roots of cabbage plants, neither compound prevented eggs of P. brassicae from hatching, but the young larvae were killed when they began feeding. Applied systemically in this way, fluoroacetamide was more toxic than sodium fluoroacetate to these young larvae and to fourth-instar larvae.

When John Innes potting compost was treated with solutions of the fluorine compounds and stored in the greenhouse it was found that both compounds were quickly decomposed at the dosages necessary for systemic action against aphids, but higher dosages were much more persistent.

In solution, given systemically via the roots to broad beans, fluoroacetamide and sodium fluoroacetate were about equally effective and both were more toxic than schradan to A. fabae. Fluoroacetamide was the more toxic of the fluorine compounds.

The technique of the cut tap-root showed that the approximate LD 100 for A. fabae on broad beans was: fluoroacetamide between 0·09 and 0·9 mg./kg.; sodium fluoroacetate 0·7 mg./kg.; schradan 50 mg./kg. When taken up by cabbage leaves through the cut petioles, 5 mg. of the fluorine compounds or 50 mg. of schradan per kg. of fresh plant tissue killed all B. brassicae in 2–3 days.

Leaves of cabbage treated in the same way with fluoroacetamide were fed to larvae of P. brassicae. It was found that the growth of the larvae feeding on the leaves was not prevented until a dose of about 20 mg./kg. was reached. The same value was found for sodium fluoroacetate. For schradan, more than 2,900 mg./kg. were required.

Finally, with applications made to the older leaves against all these species of aphids feeding on the younger leaves of their host-plants, fluoroacetamide was equally effective or slightly superior to sodium fluoroacetate as a systemic insecticide. Schradan was nearly as effective against B. brassicae but much less effective against the other two aphids.

Even after repeated applications none of the compounds acting systemically prevented eggs of P. brassicae from hatching or had a detectable toxic action on fourth-instar larvae.

This part of the study of the garden chafer, Phyllopertha horticola (L.), in the English Lake District covers the flight seasons 1948–1952. Annual sex ratios in the population of pupae (just before ecdysis to adult) ranged from 34·6 to 46·9 per cent. females.

The flight season begins with the emergence of some adult males upon the sward. If the weather continues favourable, the earliest females appear on the second or third day of the season. Unsuitable weather delayed the advent of females until the sixth day in two seasons. In a homogeneous population, primary emergence is complete in just under a fortnight, at the end of which time numbers of beetles seen on the grass sward are at their peak. Phase 1, swarming over the sward, continues, with dwindling numbers, for just over a week more (overlapping Phase 2, swarming on the bracken).

The female/male ratio rises throughout Phase 1 except for a tendency to fall in the last day or two. But generally for most and sometimes the whole of Phase 1, the proportion of females above ground is far below the true population ratio. This is partly because population emergence in the female is 1–2 days behind emergence in the male; but much more because after primary emergence the male is fully involved in above-ground activity every day until the beginning of Phase 2 or end of Phase 1 (depending on whether he emerges earlier or later), while the female is not.

In the same way, there are no females present on the bracken until after the first day of Phase 2, but the proportion of females rises more quickly and to a higher level than in Phase 1. By the middle of Phase 2, the proportion is close to the true population ratio. Since males are now confined to the bracken, this means that mass oviposition is more or less complete. During the last week of the season the female proportion rises steeply. This is because females live a few days longer than males besides being 1–2 days behind males in their original emergence.

After primary emergence, the average female lives about a fortnight, the male rather less. Approximately the first half of adult life is spent in the grass sward area, the second half on the surrounding bracken (or hedges or trees).

As Phase 1 proceeds, the pairing rate of females decreases. This is partly because the proportion of males is falling and therefore fewer males are available to search for each female over the grass sward, and partly because of a change in mating behaviour of the female as she ages. On the other hand, pairing rate is at its highest (average 82%, maximum 95%, of available females) in Phase 2 when the proportion of males is at its minimum (actually the true proportion in the population). This is because on the bracken the beetles are confined to a much smaller area than on the grass and therefore males find females more readily.

In flying weather, the proportion of beetles in the air at any moment is at best no more than about 70 per cent.; usually it is less than 50 per cent. On the whole males are very much more given to flying than females.

During the winter months, when larval populations of shoot flies (Diptera, Acalypterae) are relatively static, pure clumps of grasses in southern England were sampled. The tillers were dissected and the shoot-fly larvae removed and identified to determine what species were present and how and in what numbers they were distributed within the species of grasses. Concurrently, a permanent pasture was sampled in which the grasses were closely intermixed, and it was found in the two cases that the same host preferences of the insects were broadly maintained. Those preferences are due to the selectivity of the ovipositing female rather than the dietary requirements of the larvae, which are facultative parasites of most grasses and cereals.

A Table is given of the relative abundance of the larvae of 17 species in 16 different grasses and indicates the degree of host specificity of each insect. For example, Oscinella frit (L.), the frit fly of oats, is found in 11 of the common grasses but not in Dactylis glomerata whereas O. albiseta, (Mg.) is found only in this grass.

Horizontal migration of larvae from tiller to tiller can take place; vertical migration may also take place from ploughed-in grasses to cereals sown above. It is shown that migration is mainly limited to autumn and spring when soil temperatures are above the thermal threshold of activity of the larvae.

This paper presents further observations on the continued persistence of larvicidal treatments applied to small plots (10 yd. × 10 yd.) on Soutra Hill, Midlothian, to control larvae of the biting midge, Culicoides impunctatus Goetgh. Before doing so, consideration is given to the nature of the observed fluctuations in larval population encountered in the untreated plots. Evidence is produced which indicates that this is not due either to delayed hatching of eggs (this, if it occurs, is of minor importance), to change in technique or to the application of insecticide to neighbouring plots, but is due to the larvae responding to an oscillating factor (or factors) in the breeding site. It is suggested that one such factor might be soil water level. Therefore, it follows that changes in the untreated larval population must be taken into account when assessing the degree of control achieved by treatments, especially when the larval population is declining.

As regards insecticidal persistence three years after application: —

(i) Nine of the eleven DDT treatments (wettable powder, water-miscible concentrate and dust at 12, 50 and 200 mg. p,p′DDT/sq. ft.) gave 100 per cent. control. The other two (both wettable powder at 12 mg./sq. ft.) produced 75 and 22 per cent. reduction.

(ii) Only one of seven γ BHC treatments (wettable powder at 50 mg. γ BHC/sq. ft.) gave complete control and one treatment (w.p. 12 mg./ sq. ft.) was completely ineffective.

(iii) DDT retained its superiority to γ BHC as a larvicide against Culicoides, achieving 93 per cent. control compared with 55 per cent. for γ BHC.

(iv) The lowest dosages of dieldrin (3 mg./sq. ft.) and chlordane (6 mg./ sq. ft.) had begun to lose their larvicidal power, control declining from 76 to 53 per cent. and 67 to 29 per cent., respectively. There was no change at 6 mg. dieldrin/sq. ft. but the other treatments (12 and 25 mg. dieldrin and 12, 25 and 50 mg. chlordane/sq. ft.) had improved. Dieldrin at 12 and 25 mg./sq. ft. and chlordane at 50 mg./sq. ft. attained complete control.

The spray-volume treatments, now in their second year, showed an increase in activity. This was most marked in the wettable-powder treatments at 25 mg. p.p′DDT/sq. ft. which a year ago were completely ineffective and now five out of six trials gave complete control and the other 97 per cent. Improvement in the treatments at 50 mg./sq. ft. was less (from 43 to 83%) due in the main to the complete failure of one treatment (10 gal./acre).

It is confirmed that spray volume in itself is unimportant in larvicidal applications providing the coverage is adequate.

Previous work on strains of Amphorophora rubi (Kalt.) is discussed and evidence is presented to show that the form on European raspberry can be divided into three strains. These are designated European raspberry strain 1, 2 and 3. The distribution, abundance and importance of the strains in relation to plant breeding for resistance to A. rubi is discussed.

One exceptional instance of a mycosis, with epidemic character, in white grubs has been studied in Southern Rhodesia.

The hosts were the larvae of Anomala exitialis Pér. and Schizonycha profuga Pér., which are tobacco pests, the entomophagous fungi being Torrubiella sp. (most prevalent), Beauveria bassiana (Bals.) Vuill. and Metarrhizium anisopliae (Metsch.) Sorokin.

Soil with an abnormally high content of nitrogenous organic matter favours the breeding of the two insects but at the same time provides a suitable medium for the survival and development of a high degree of virulence in the cryptogamic pathogens.

In the British Solomon Islands Protectorate, the incidence of the immature nutfall of coconuts that is caused by species of Amblypelta depends upon the distribution of (1) the species that are concerned, which determines on what islands nutfall occurs, and (2) certain species of ants, which affects the local distribution of nutfall on those islands. A. cocophaga China is the most important member of the genus as regards responsibility for nutfall; A. cristobalensis Brown and to a less extent A. costalis Van Duzee also frequent the coconut palm and cause nutfall, but not on a serious scale; A. gallegonis Lever rarely, if ever, attacks coconut.

The current view that Oecophylla smaragdina (F.) protects coconuts against Amblypelta, and that Pheidole megacephala (F.) and Iridomyrmex myrmecodiae Emery give little or no protection, is accepted, but evidence is brought forward showing that Iridomyrmex appears to give some protection against A. cristobalensis on San Cristobal. Most attention has been given to a fourth species of ant, Anoplolepis longipes (Jerd.), the relation of which to nutfall has hitherto been in some doubt; evidence from various sources shows that, at least on Guadalcanal and Malaita, it gives nearly, if not quite, as much protection against Amblypelta as Oecophylla does.

A method based on crop estimation is used to give some indication of the losses due to nutfall in the European plantations that have been surveyed. The accuracy of the method is discussed in an Appendix. Varying numbers of nuts survive to maturity in nutfall areas, depending on the intensity of attack. In addition to the loss caused by the failure of nuts to survive to maturity, Amblypelta attack causes further losses by reducing the amount of copra obtainable from surviving nuts, and by rendering plantations where the attack is sufficiently severe unprofitable to work, so that they become, in effect, a total loss, even though a proportion of the nuts may survive attack.

An attempt is made to evaluate and compare the fraction of the crop lost on some of the principal islands, but it is impossible to give an even approximately accurate figure for the losses due to nutfall in the islands as a whole, because essential data, comprising the planted acreage of each plantation and island, and, except for a very few plantations, details of annual production figures over a period of years, are often lacking; also, limited time has made it impossible to survey many of the plantations in sufficient detail for an accurate assessment of nutfall.

A list is given of the species of ants recorded from coconut plantations during the present work.

In the second of this series of papers on the losses caused by the Coreid, Pseudotheraptus wayi Brown, to coconuts in Zanzibar, descriptions are given in summary form of a number of Government-owned plantations of coconuts on the island, together with the yields recorded over the 12 years, 1944–55. The variation from year to year in any group of plantations is not large, and over all those considered does not exceed 25·8 per cent, of the 12-year mean.

In the plantations on poorer soils, predominantly occupied by the ants Anoplolepis longipes (Jerd.) or A. custodiens (F. Sm.), the vidaka damage rate (V.D.R.) during 1955 varied from 60–80. The plantations on better soils are occupied predominantly by a complex of Oecophylla longinoda (Latr.) and Pheidole punctulata Mayr and the V.D.R. determined on a part of these in 1955–56 varied from approximately 41 to 55, with a mean of 47·1. The mean value of the V.D.R. calculated from 200 localities taken at random in Zanzibar during June–August 1955 was 60·6. Taking a value of 50 as a conservative estimate of the V.D.R. for the whole island, and inserting this in the regression equation Y = a + bX relating yield (Y) with V.D.R. (X), derived in the first of this series of papers, together with the average values of the constants a and b obtained therein, it is shown that about two-thirds of the potential crop of coconuts is lost as a result of damage by P. wayi.

Comparison of the yields from palms occupied by Oecophylla in two of the plantations on better soil with those of the whole of those plantations suggests that about one-half of the potential crop is being lost.

The annual value of the total exports of unconverted nuts and coconut products (copra and coconut oil) varied from £636,254 to £1,008,194 during 1949–56, and the annual value of the total production of nuts, calculated from the estimated total acreage (104,000), the number of palms per acre (50), the 12-year average yield per palm on Government-owned plantations (22·5 nuts per annum) and the market price in 1948–56 (Shgs. (E.A.) 90–150 per thousand nuts), is £526,500 to £877,500. Taking into consideration the lower and upper estimates of the amount of crop lost and of the value of what remains, the minimum and maximum estimates of the financial loss caused by P. wayi in Zanzibar are of the order of £500,000 and £2 million, respectively.

An undesirable effect of using a dust containing 0·4 per cent, lindane for the control of Pseudotheraptus wayi Brown on coconut palms is the eradication of Oecophylla longinoda (Latr.), a species of ant inimical to this Coreid, and in the search for a more suitable preparation, formulations were tested in Zanzibar that contained DDT, with synthetic resins to prevent its absorption through the treated surface of the palm. One such formulation, prepared by dissolving 23·5 lb. of a proprietary coumarone indene resin and then 235 lb. 80 per cent. p,p′DDT in a mixture of 100 gal. power kerosene and 100 gal. diesoline, heated to 40–45°C., is here termed 10 DRDK and its properties described.

When 10 DRDK was applied by pneumatic hand-sprayers to the centres of palm crowns, small species of ants (Pheidole spp., Crematogaster spp., Paratrechina longicornis (Latr.)) and Anoplolepis longipes (Jerd.) were eradicated from such sites, whereas O. longinoda increased in numbers and other large species (Camponotus spp., Polyrhachis spp., etc.) were unaffected. Although 10 DRDK was lethal to all species when applied as an aerosol in the laboratory, the residual deposit from it killed only the smaller species, apparently because their bodies make contact with the DDT crystals whereas those of the larger species (except A. longipes) are held clear. Under Zanzibar conditions, this deposit, which is at first gummy, gives rise within 2–3 weeks, or following mechanical disturbance, to needle-like crystals of DDT, 10–120 μ long × 2–10 μ wide projecting from the sprayed surface and forming clumps up to 1·5 mm. high.

In field tests, paper bands dipped in 10 DRDK and attached round palm trunks prevented the passage of A. longipes for six months, and of Pheidole for nine months, despite heavy rain; that of the larger species of ants was unaffected. Determinations of the deposit on a sample band selected each week at random showed great variability but fell from 392–478 μg. DDT per sq. cm. initially to 92–176 μg. after 31 weeks. In concurrent tests of similar bands exposed in the laboratory, the deposit fell only to 282 μg. per sq. cm. over approximately the same period.

In tests of 10 DRDK applied topically to the dorsum of Pseudotheraptus wayi, complete mortality of adults was obtained within 48 hours with droplets 85 μ in diameter (containing 0·31 μg. DDT) and the size required diminished progressively, the earlier the stage treated. When individuals of P. wayi were allowed to walk for one minute on glass plates bearing known deposits from 10 DRDK, complete mortality of all stages, from 2nd-instar nymphs to adults, followed within 24 hours of exposure to 2·7 mg. DDT per sq. ft., but not to 1·6 mg., except in the case of recently moulted adults, 5th- or 4th-instar nymphs.

Two experimental areas totalling some 1,000 palms, of which 10 per cent, had Oecophylla nesting in them in October 1952, were sprayed with a commercial formulation containing DDT from November 1952 to June 1953, and the percentage fell to 2·5; the areas were then treated with 10 DRDK by hand-sprayer until January 1954 and by machine for the rest of 1954. The percentage of palms having Oecophylla, in October 1953, October 1954 and February 1955 increased to 7, 15 and 38, respectively. In a control area one mile away, the percentage rose from 23·5 to 36 over the same period. In a third area, 12 applications of 10 DRDK totalling 10 gal. per acre were made by machine during 1953, 12 totalling 33 gal. in 1954 and 9 totalling 30 in 1955, and the percentage of palms colonised by Oecophylla rose from nil to 21·7 in 1953, to 26·1 by March 1955 and to 30 by April 1956, while the percentage colonised by A. longipes fell over the same period from 90 to 1.5. In the control areas, A. longipes remained dominant and no Oecophylla became established.

The addition of lindane to 10 DRDK to give concentrations of 0·1, 0·5 or 1·0 per cent, adversely affected the selective properties of the formulation, and although the highest strength gave a good kill of P. wayi and other insects, apart from Coccids, which increased considerably, it reduced Oecophylla to a greater extent than it did other species of ants. No advantages are obtained by the addition of dieldrin, endrin, pyrethrins or rotenone.

The importance of certain species of Stomoxys in Uganda, the possibility that these species may be responsible for the mechanical transmission of trypanosomiasis, and the effects of severe biting of Stomoxys on cattle causing loss in milk yields and weight in beef animals, is pointed out.

A satisfactory method is described for rearing large numbers of Stomoxys calcitrans (L.) under the conditions that prevail in East Africa.

A description of special cages for keeping the adults and for egg production is given, together with a technique for supplying the adults with 0·2 per cent, citrated bullocks' blood.

A satisfactory egg-laying and breeding medium contained 9 oz. of dried cow dung, 5 oz. of dried blood granules and 1 oz. of sugar.

A detailed description of the construction and working of a constant-temperature and humidity room for rearing larvae and pupae is given, the temperature and relative humidity of the room being maintained at 90°F. and 90 per cent., respectively.

The maximum number of eggs that can be introduced into 1-lb. and 2-lb. Kilner jars without causing overcrowding and subsequent under-nourishment in the later larval instars was found to be 200 and 400, respectively.

Cashew, Anacardium occidentale, an increasingly important cash crop in the coastal region of Tanganyika, is damaged by two widely distributed species of Helopeltis, H. anacardii Miller and Helopeltis sp.

At Bagamoyo, in June 1955, the period from egg to adult for H. anacardii occupied 35·9 days; that for Helopeltis sp. occupied 39·2 days.

Damage is caused to the leaves, shoots, buds, young fruits and to the cashew apple. Damage in a young tree is more marked than that in an old tree and may result in its being malformed or stunted.

The build-up of Helopeltis populations on cashew follows the main growth period of the crop which begins shortly after the end of the main rains in May. The insects are readily found from July to October-November, when the crop matures.

Application of a 0·5 per cent, γ BHC dust and a 5 per cent. DDT dust to two-year-old trees increased the amount of crop but the increase was not economic owing to the fact that the trees were coming into bearing for the first time. Cross-infestation from untreated plots may also have reduced the effectiveness of the treatments.

Three or four applications by a knapsack-type power duster of 0·5 per cent. γ BHC dust at 15-day intervals previous to flowering of the main crop gave promise of successful control in the field.

Observations were made at Dorimon in the extreme north-west of Ghana, during the hot months of March and April 1954, of the infectibility of Glossina palpalis (R.-D.) and G. tachinoides Westw. with Trypanosoma vivax.

Pupae were obtained from the Black Volta where shade temperatures ranged from 70 to 98°F. and the flies that emerged from them were kept singly in small glass tubes in the laboratory at a temperature of 85 to 95°F. and a relative humidity of 60 to 80 per cent. They were given several feeds on infected goats. The following percentage rates of infection were observed.

Labrum and (or) hypopharynx: General rate, 56·1; males of G. palpalis, 72·2, females, 34·6; males of G. tachinoides, 68·4, females, 56·8.

Labrum only and labrum & hypopharynx: General rate, 37·4; males of G. palpalis, 27·8, females, 15·4; males of G. tachinoides, 68·4, females, 40·9.

Early infection is the rule but late infections occurred. Some flies appear to be uninfectible. Thus, out of 107 flies, 47 remained uninfected in spite of having had on the average 10·6 infective meals per fly. A high percentage of flies lost their labrum infection.

The following percentage loss of labrum infection was recorded: General rate, 48.8; males of G. palpalis, 66·7, females, 62·5; males of G. tachinoides, nil, females, 43·8.

G. tachinoides appears to be more tenacious of infection than G. palpalis. Feeding infected flies on goats in the negative phase and (or) on spurwing goose (Plectropterus gambensis) did not appear to have any adverse effect on the colonies of trypanosomes in the labrum.

Records are given of the species of mosquitos captured at Mombasa in the years 1943–44 and 1949–52. They were caught by four different methods, namely: (1) day-time net catches in the bush, (2) window-trap catches, (3) diurnal (bait) biting catches in the bush and (4) twenty-four hour (bait) biting catches in three different situations. Sixty-six species were taken in the day-time net catches and nine others which did not figure in the net catches were taken in one or other of the remaining three methods.

The habits of the commonest species, as indicated by these catches, are discussed in some detail.

Culex pipiens fatigans Wied. occurred in large numbers in biting catches, net catches and window-trap catches. In 24-hour catches most biting occurred after midnight, but, in the window traps, activity was greatest between 18 and 21 hr.

Aëdes simpsoni (Theo.) was not taken in window traps or houses. Seasonal occurrence, as indicated by larvae in banana axils, net and biting catches in the bush, is discussed. Twenty-four hour catches confirmed earlier findings that this mosquito bites almost solely during daylight.

Aëdes pembaensis (Theo.) occurred throughout the year with little seasonal variation; this is thought to be connected with its breeding in crab-holes. It is often found in fair numbers in houses when these are near the breeding grounds. Window-trap catches show that most activity takes place at night though some also occurs during the day. The biting rhythm as shown by 24-hour catches in houses, bush and on tree platforms is discussed.

Aëdes albicosta (Edw.). Previously this mosquito may have been confused with Aëdes albothorax (Theo.) and Aë. lineatopennis (Ludl.). It can occur at times in great numbers when it may bite viciously both indoors and outside.

Anopheles gambiae Giles and A. funestus Giles were not evident in large numbers. In window traps most activity took place at night although some also occurred during the day. In 24-hour catches, A. gambiae fed mostly just before dawn though biting occurred also from midnight onwards. A. funestus appeared suddenly and in fair numbers in window traps but was absent for much of the year.

Aëdes aegypti (L.), though fairly common, is recorded only in Tables and in one figure as its bionomics have already been discussed elsewhere.

Notes are given on other species, which include Culex duttoni Theo., C. nebulosus Theo., C. tigripes Grp., Aëdes metallicus (Edw.), Aë. albocephalus (Theo.), Eretmapodites subsimplicipes Edw. and E. quinquevittatus Theo.

Preliminary experiments having shown that dimefox (bis(dimethylamino) fluorophosphine oxide) placed in holes drilled at ground-level in cacao trees gave encouraging results in controlling the most important of the mealybugs that transmit swollen-shoot virus (Pseudococcus njalensis (Laing), Planococcus citri (Risso) and Ferrisiana virgata (Ckll.)), two further experiments to investigate the technique and determine the optimum dosage were performed in Ghana in April-June and September-November 1953, respectively. A randomised-block layout was used, with four replicates (each comprising a plot of seven trees) of each treatment at Tafo and, in the first experiment, another two at Bunso. The first experiment disclosed no significant difference in effect when the implantation holes were spaced four inches as against two inches between centres and the former spacing alone was used in the second experiment. The dosage levels of dimefox, based on the weight of the aerial part of the tree estimated from the girth—weight ratio, were 75, 150, 225 and 300 parts per million w/w in each experiment, the lowest representing about one-twelfth of the soil-application rate, and also 375 p.p.m. in the second experiment. The results were assessed in terms of the total number of mealybugs, of all three species, counted on the various parts of the trees, which were felled and examined six weeks after treatment; counts made on a duplicate set of trees in the second experiment that were felled at nine weeks were significantly greater.

Statistical analyses of the logarithmic transformations of the population figures obtained at six weeks showed the regression of population on dosage to be a curve of quadratic form. There was thus a limit to the response obtained, the minimum populations of total mealybugs (adults + nymphs) being achieved at a dose rate of 281 p.p.m. in the first experiment and 317 p.p.m. in the second. In the former, the minimum value of the average population of total mealybugs per tree was below that required for efficient control (one per tree), but the variability of the response was such that the population on 35 per cent, of trees so treated would have been above that level; in the latter, the minimum average value attainable was only slightly less than one, and consequently the population on nearly half the trees would have exceeded the level for effective control. Analysis of the combined results from those replications of each experiment that were sited at Tafo showed that a given dosage rate appeared only about half as effective, in terms of the residual population (expressed logarithmically), in October-November as in April-May, but concurrent records from untreated control trees suggest that this may partly have been because populations were falling naturally at the earlier date. Slight leaf-shedding occurred in some of the treated trees.