The termites most injurious to crops and trees in West Africa are subterranean or mound-building species of the Termitidae, mainly Macrotermitinae with some Amitermitinae and Nasutitermitinae. Insecticides for their control may be applied generally to the soil, locally around the plant or directly to the colony.

In investigations in Northern Nigeria, dusts containing aldrin or dieldrin were mixed with the top six inches of soil of a type commonly cultivated (a slightly humic, brown, loamy sand), exposed to weathering in the field and tested for persistence of the insecticides by bioassay at intervals, using workers of Trinervitermes ebenerianus Sjöst., a locally common surface-foraging species. The concentration of the insecticide in the soil was measured in terms of the time in days taken for 50 per cent. of the insects to be killed (T50). After 33–34 months, between one-third and one-fifth of the insecticide remained in soil originally treated with 0·5, 2 and 5 lb. active ingredient (a.i.) per acre.

T. ebenerianus proved very sensitive to dieldrin; the T50 value was 1·48–4·10 days for single samples, from each of five colonies, exposed to filter paper containing 0·0018 parts per million, as compared with 10–24 days for the controls. It is suggested that general soil treatments should be used with caution until more is known of their effects on termite populations, which are important in facilitating aeration, and penetration of water, in tropical soils.

Application of dieldrin emulsion, at a dose equivalent to 1 lb. a.i. divided among the planting holes for one acre (1,225), during planting of one-year-old, root-pruned seedlings of Eucalyptus camalduensis, resulted in a mean survival after 2½ years of 60 per cent. of the young trees, as compared with 17 per cent. in untreated controls. Pot-grown seedlings of Eucalyptus spp. were similarly treated at 8 oz. a.i. dieldrin per 700 pots prior to setting out in the field, when they showed very low mortality due to termites over the next 1½ years, attack only occurring where too short a pot allowed access by Macrotermes natalensis (Hav.) to the tap root. Four hundred pot-grown cacao seedlings similarly treated with 4 oz. a.i. dieldrin showed only four deaths due to termites one year after planting out. Pre-treatment of potting soil for Eucalyptus seedlings at 5–10 oz. of 2 per cent, dieldrin dust per cubic yard (sufficient for 500 pots) has given promising preliminary results.

Colonies of M. natalensis, which constructs large mounds, were successfully poisoned with 2½ fl. oz. of aldrin 40 per cent, emulsifiable concentrate in six gallons of water applied through three auger holes made into the central ‘ hive ’, containing the queen cell and associated chambers. It is considered that this dose could safely be reduced.

Dieldrin, γ BHC, DDT, Telodrin and Baytex have been applied topically in oil solutions to G. swynnertoni Aust. caught as wild adults. The order of toxicity of these compounds was the same as that to laboratory-emerged G. morsitans Westw. and all were relatively more effective. Wild females were 2–3 times as tolerant of the chlorinated hydrocarbons as the males, and pregnant females required nine times as much DDT as the males. Greater tolerance to Baytex did not exceed the difference in body weights between the sexes. These results confirm those obtained with G. morsitans and are attributed to the increased tolerance to chlorinated-hydrocarbon insecticides shown by females as they age and become fertilised and pregnant.

There is no such increase in tolerance of females to Baytex and the importance of this in planning practical control measures is indicated.

A series of six substituted phenyl N-methylcarbamate insecticides were applied to laboratory-reared adults of G. morsitans Westw. as solutions in di-isobutyl ketone. Only the 2-isopropoxy and the 3-isopropyl compounds were toxic at 0·0068 µg. per fly and the former, being apparently more active, was evaluated more fully. It was found to be equitoxic with γ BHC to young flies, and not synergised to any significant degree by five parts of piperonyl butoxide. Pregnant flies were considerably more tolerant than young flies to low dosages but not to high ones. The suggested explanation postulates that the developing larva is a potential site of detoxification but that the rapid action of high dosages results in the death of the parent before the insecticide can reach the larva and be detoxified by it.

Many mature larvae of the khapra beetle, Trogoderma granarium Everts, enter a facultative diapause and leave stored food to cluster in crevices in the storage buildings over the whole range of temperature in which larval development occurs (20 to 40°C.). The effects of changing temperature and food conditions on the behaviour of these diapause larvae have been studied in the laboratory.

After an increase from one constant temperature to another, fewer diapause larvae move from the food to the crevices, and those already dormant in crevices tend to emerge, feed and pupate. Some larvae emerge but do not pupate and later return to the crevices. These activities increase with temperature: they are slight at 20 to 25°C., intermediate at 30°C. and great at 35 to 40°C. However, if the temperature is held at 30°C. without change from the hatching of the egg (controls), there is little activity during diapause, although early larval development is fast at 30°C. Large jumps in temperature, e.g., 20 to 30°C., are more effective in breaking diapause than small ones, e.g., 30 to 32·5°C., and changes in the upper part of the range are effective, e.g., 35 to 37·5°C. There is much variation between individual larvae in the time elapsing between the increase in temperature and pupation. Pupation is probably delayed slightly by heavy accumulation of faecal pellets in the food. Sex and the length of the diapause period do not influence the breaking of diapause, which can take place within a month, and probably much less, of the onset of diapause.

Diapause is broken by gradual as well as abrupt temperature increases, but regular daily fluctuations, such as between 25 and 40°C., are ineffective unless preceded by a period of cooling (e.g., 1 month at 20°C.). The renewal of food badly fouled with faecal pellets at 30°C. breaks the diapause of some larvae, but changes in temperature regarded as moderate are more effective.

These results provide additional evidence about the degree of diapause in T. granarium. Some features, including the breaking of diapause by cooling followed by a return to a favourable temperature, are typical of classical diapause, but the breaking of diapause soon after it begins and the effect of a jump in temperature in the upper half of the developmental range without prior cooling are not typical, suggesting that diapause is broken by a trigger effect starting or stopping some physiological process. It can be concluded that mature, dormant larvae of T. granarium undergo a fairly weak facultative diapause with some unusual features before becoming prepupae. This diapause is compared with that of other insects.

The diapause appears to be particularly suited to the environment in dry food stores. It helps larvae to survive cool periods and food shortage, until warming combined with loading of new food breaks the diapause and allows breeding at the beginning of favourable periods. It appears likely that in grain stores in the Punjab (Pakistan) the life-cycle of T. granarium is largely governed by the climatic seasons. In the very erratic, non-seasonal environment in British malt stores, the ease of breaking diapause at almost any time after it has begun, with or without cooling, enables T. granarium to breed as soon as conditions are favourable, without seasonal control. Some other features of the diapause are well suited to this environment.

Five insecticides (Thiodah, Bromodan, dimethoate, Dipterex (trichlorphon) and Sumithion) were applied topically in solution to young adults of the tsetse fly Glossina morsitans Westw. The toxicity of Thiodan (LD50 for males, 0·0012 µg.; for females, 0·0014 µg.) lay between that of Telodrin and dieldrin, and that of Bromodan (LD50 for both sexes combined, 0·0032 µg.) between that of dieldrin and γ BHC. Dimethoate, Dipterex and Sumithion were all less toxic than Baytex, but were not fully evaluated.

Results recorded in this paper and earlier ones in this series, and those of workers using Musca domestica L., are discussed in an attempt to rationalise the search for new toxicants for tsetse flies. Chlorinated hydrocarbons fall into two classes, in both of which there is an approximate positive correlation between toxicity to Musca and to Glossina. One class, including Telodrin, Thiodan and Bromodan, is exceptionally toxic to tsetse and more than 40 times as toxic to Glossina as to Musca. These have a molecule with a particular type of nucleus (hexachloro-l,4-methano-cyclohexene) with a single bridged unsaturated ring. The more common chlorinated hydrocarbons (γ BHC, DDT and aldrin) are two to five times, and dieldrin 12 times, as toxic to G. morsitans as to M. domestica, and all are very toxic to both species.

Among organophosphorus compounds there is very little correlation between toxicity to M. domestica and that to G. morsitans, but it appears probable that no compound of low toxicity to Musca (i.e., LD50 of the order of 0·5 µg.) would be likely to compete with those compounds already known to be most toxic to Glossina. Replacement of the nitro group on the benzene ring of Sumithion by a thiomethyl group, to form Baytex, increases toxicity to Glossina fivefold, but that to Musca by only one-fourth. There is less information on organocarbamates, and it may be a coincidence that 2-isopropoxyphenyl N-methylcarbamate, the most toxic of those tested on G. morsitans, was also most toxic to Musca.

The ecology of four tsetse species of the group of Glossina fusca (Wlk.) was studied near the Ugbobigha field station of the West African Institute for Trypanosomiasis Research in southern Nigeria. G. tabaniformis Westw. and G. nigrofusca Newst. occurred in Lowland Rain Forest and G. medicorum Aust. in forest outliers in Derived Savannah; G. fusca was present in both habitats. Meteorological records showed that the climate in the rain-forest was cooler and more humid than in the forest outliers where, in turn, it was more equable than in the savannah. The distribution of the four species, each with its own range of tolerance, was thought to be largely determined by the climatic conditions within the various habitats.

A fly-round, using a white Zebu ox as bait, had been established in the rain-forest in August 1956; 2,549 fusca-group flies were caught in just over three years, 83 per cent, were G. tabaniformis, 16 per cent. G. fusca and 1 per cent. G. nigrofusca. The number of flies caught on the fly-round declined during the catching period; it is suggested that this was caused by the necessity to kill all flies caught in order to identify them. This disadvantage was overcome in further experiments, in which fusca-group flies were caught from an ox tethered in the rain-forest, and in which all flies caught were marked and released; the proportion of each species in the total catch was known, although the identities of individual flies were unknown. A total of 14,577 fusca-group flies were caught once per day during 22 days in July 1957 (wet season) and 7,694 during 22 days in January 1958 (dry season). G. fusca and G. medicorum, which can be readily distinguished on external characteristics, were caught on an ox fly-round laid out in an area of riverine forest in the savannah and further studies on all species were made in a primitive fly-breeding room at Ugbobigha.

G. tabaniformis was most plentiful in the middle of the rains and scarcest during the dry season. Estimation of the numbers of G. tabaniformis and G. fusca combined, during July 1957 and January 1958, suggested that flies were approximately three times more abundant in the rains. The dry-season decrease was considered to have been caused by the adverse effects of high temperatures and low humidities. Largest numbers of G. tabaniformis were caught in those areas of the rain-forest where the host animals, especially the red river hog (Potamochoerus porcus), were least disturbed by human activity. In both wet and dry seasons there were early-morning and late-afternoon and early-evening peaks in activity of G. tabaniformis. It is incorrect to describe the species as crepuscular in habit; during July 1957 flies of a mixed population composed of approximately three-quarters of G. tabaniformis and one-quarter of G. fusca were caught at a rate of only just less than one a minute even when least active. Activity at night was negligible. Analysis of the sex ratio throughout the day showed that rather more than half of the catch of G. tabaniformis before 0700 hr. and after 1800 hr. were males. During the middle of the day females greatly outnumbered males. On the ox fly-round, 58 per cent, of the catch of G. tabaniformis were females; from the tethered ox, 65 per cent, (wet season) and 70 per cent, (dry season) of the mixed catch of G. tabaniformis and G. fusca were females. The percentage of females amongst flies that emerged from puparia of G. tabaniformis in the fly-breeding room was 57.

The numbers of G. fusca caught on the ox fly-round in the rain-forest were insufficient to provide definite evidence for seasonal fluctuations in population density, although such fluctuations probably did occur. In experiments in riverine forest, 2,547 individuals of G. fusca were caught once per day in July–August 1958 (wet season) and 304 in January–February 1959 (dry season). It is suggested that forest outliers in the savannah at Ugbobigha were a suitable habitat during the wet season but that the high temperatures and saturation deficits in the dry season approached the limit of endurance of the species and prevented larger numbers occupying the habitat; mean length of life was probably much reduced in the dry season. G. fusca in the wet season was nearly equally active throughout the day, with a suggestion of peak activity just before dark. In the severe conditions in the dry season in the riverine forest, activity was negligible between about 1200 and 1700 hr. Analysis of the sex ratio of G. fusca in the wet season showed that the proportion of males increased during the day. In the rain-forest, 44 per cent, of the examples of G. fusca caught on the fly-round were females; in the riverine forest 47 per cent, (wet season) and 40 per cent, (dry season) were females. The sexes emerged from the puparia in equal proportions.

Only 29 individuals of G. nigrofusca were caught on the ox fly-round in the rain-forest, of which only three were females.

Only 120 individuals of G. medicorum were caught in the wet-season experiment in the riverine forest and 36 in the dry season; 54 per cent, were females in the wet season and 42 per cent, in the dry season. The flies were most active early in the morning. The emergence rate (82%) from puparia of G. medicorum in the fly-breeding room was higher than for G. tabaniformis (56·5%) and G. fusca (54%).

The types of habitats occupied by the Nigerian species of the fusca group are described.

Females of Glossina palpalis (R.-D.) were dissected and examined for pregnancy at the Ugbobigha field station of the West African Institute for Trypanosomiasis Research in southern Nigeria between January 1958 and October 1959. Of 2,490 females dissected, 46 (2%) were virgin; 764 (31%) had an egg in utero and 221 (9%) a larva in utero. Highest pregnancy rates were found in months with heaviest rainfall and lowest mean temperature, saturation deficit and evaporation rate.

The pregnancy rate in G. palpalis near Kaduna, in Northern Nigeria, was nearly twice that found at Ugbobigha, the major difference being in the number of females with a larva in utero. It is suggested that in the Ugbobigha area an adverse factor operates in the late egg or early larval stage, causing abortion; this is probably not a direct effect of climate or of insufficient food but may partly be related to the possible inability of females to obtain a blood-meal at just the right stage in gestation. Low pregnancy rates may, at least in part, explain the low densities of G. palpalis commonly found in the forest belt of West Africa, despite the wide distribution of the species in such areas.

Resistance to γ BHC and dieldrin in various insects usually gives a characteristic resistance spectrum, probably indicating a common defence mechanism. Resistant house-flies (Musca domestica L.) are slightly anomalous in showing greater tolerance of γ BHC than of endrin and isodrin, unlike other resistant strains. A possible explanation is that they can develop an additional defence, specific towards γ BHC, as well as the usual group resistance. This might well be enhanced enzymatic degradation of BHC, which is known to occur in flies, but was not found in resistant strains of Anopheles gambiae Giles or Cimex lectularius L.

To confirm this hypothesis the rates of BHC elimination were measured in normal and two resistant strains of flies and also in normal and resistant strains of Lucilia cuprina (Wied.) (which showed the more usual resistance spectrum). To avoid toxic effects, the non-insecticidal α BHC was used and its elimination after six hours measured in extracts by gas chromatography. The two resistant fly strains showed significantly increased degradation of BHC, correlated with their greater γ BHC resistance, whereas rates were about the same in normal flies and the two strains of L. cuprina.

The effects of piperonyl butoxide on the toxicity to adult mosquitos (Anopheles stephensi List, and Aedes aegypti (L.)) and house-flies (Musca domestica L.) of pyrethrins, two synthetic pyrethroids, three organophosphorus compounds and their oxygen analogues, and three carbamates (Sevin, and both 3-isopropylphenyl and 3,5-di-isopropylphenyl N-methylcarbamate) were determined quantitatively. Solutions, in di-isobutyl ketone, containing insecticide alone and insecticide/ piperonyl butoxide mixtures in the proportion of 1:10 were applied topically to individual insects.

A high degree of synergism was obtained only with natural pyrethrins and carbamates and house-flies, and the effect on the toxicity of these insecticides to adult mosquitos was much less pronounced. Insecticidal potency of the synthetic pyrethroids allethrin and dimethrin was influenced less than that of natural pyrethrins.

Piperonyl butoxide antagonised the action of malathion in M. domestica, and to a slight extent in A. stephensi, but increased the toxicity of malathion and malaoxon to Ae. aegypti. It had little effect on the potency of diazinon and parathion and their oxygen analogues (diazinoxon and paraoxon) with the exception that it increased the toxicity of diazinon and diazinoxon to Ae. aegypti.

The results are discussed in relation to the hypothesis that piperonyl butoxide inhibits detoxification in insects. From a practical point of view, they suggest that no great advantages would be derived from the addition of piperonyl butoxide to formulations of organophosphorus compounds and carbamates used for residual contact action against adult mosquitos.

Dissections of over 93,000 grass stems obtained by random core-sampling of natural grasslands at a site in Berkshire largely confirm earlier conclusions regarding the associations of different species of dipterous stem borers and grasses. Allowing for the number of stems of each grass species per unit area, the densities of larvae were greatest in Anthoxanum odoratum, Poa and Lolium and lowest in Dactylis glomerata. Oscinella frit (L.) was less abundant than Cetema spp. and Opomyza germinationis (L.); it was not found in D. glomerata but was most abundant in Lolium and Festuca. Sampling showed that overwintering larvae of O. frit suffer at least 50 per cent, mortality between September–December and January–March.

Previous studies have largely concentrated on populations in different species of grasses, but it is advantageous to consider the production of O. frit in terms of short (grazed or regularly mown) grasslands and long grasslands (hay fields, uncut commons and rough grazing); comparisons of figures obtained by core-sampling for larvae and emergence traps for adults indicate a production from short grassland and long grassland of about 11 and under 1 overwintered adult per sq. yd., respectively. Having regard to the greater productivity of common leys, dominated by Lolium and Poa, a national average for the first figure would be 15. There is little increase from generation to generation in grassland populations during the year, in contrast with the situation in oats.

From tentative estimates of the acreages and productivities of the different sources of O. frit in England and Wales as a whole, it is clear that the majority of the adults of O. frit are produced by grasslands in the spring and summer and by the oat panicles in the autumn, the latter population being much the greater. As O. frit is a migratory insect, such a conclusion is especially significant and indicates that during the autumn the ‘ population pressure ’ in the grasslands will be very different to that during other seasons.

Previous work in Britain has shown that insecticides, even systemic ones, fail to protect brassica plants in the seed-bed from virus diseases transmitted by aphids, and an explanation was sought through experiments carried out at Rothamsted. Observations on the effect on apterae and alates of Myzus persicae (Sulz.) of contact with leaves detached from cauliflower plants previously treated with a 0·2 per cent. DDT emulsion spray showed that exposure for up to 30 min. was insufficient to kill the majority of those tested, although many were temporarily incapacitated; alatae of Brevicoryne brassicae (L.) were even less affected.

In cage experiments using infected and healthy cauliflower seedlings, apterae of M. persicae were able to transmit cauliflower mosaic virus (ClMV) and cabbage black ring spot virus (CBRSV) from infected to healthy seedlings when no spray was used, but failed to do so when the infected plants were sprayed with DDT; when alates were used in similar tests using only CBRSV, only low rates of transmission were obtained. In similar experiments using turnip seedlings and ClMV, transmission by apterae of M. persicae was high on untreated plants when the infector plant was itself untreated, but distinctly lower when it was sprayed with DDT; alates of M. persicae were found to transmit ClMV from a DDT-treated infector to untreated plants but not to a treated plant. Virus spread was thus reduced but not prevented by the use of DDT sprays. It is considered that these results lend support to the hypothesis that increase of virus within a seed-bed is preponderantly the result of the arrival, in succession, of many individual viruliferous alates from outside.

The age-composition of mosquitos biting at different times and levels was investigated in Zika Forest, Uganda. The four species studied were Mansonia aurites (Theo.) group (material mainly of actual M. aurites (Theo.) but probably including some examples of M. microannulata (Theo.)), Aedes africanus (Theo.), A. ingrami Edw. and Culex annulioris Theo. In each case the biting cycles of nulliparous and parous females were computed separately.

The material examined came from eleven 24-hour catches made by Dr. A. J. Haddow. Samples were taken at seven levels, at 20-ft. intervals, from the ground to a height of 120 ft. The top of the canopy was at 70–80 ft.

In each species the biting cycles of nullipars and pars were closely similar, both at the preferred level and at all levels combined. In only one case (that of the group of M. aurites) was there a significant and consistent difference in the average age of females at different levels, there being a relative excess of older females above the canopy, as found previously in M. fuscopennata (Theo.). This difference persisted throughout the period of biting activity.

These and other recent findings are briefly discussed. It is concluded that the hypothesis which relates the form of the mosquito biting cycle to the age-composition of the biting population and to the larval habitat of the species concerned cannot be accepted as a valid generalisation.

Oviposition was investigated in five strains of Aedes aegypti (L.), two being DDT-resistant (from Trinidad (STR) and Haiti) and three susceptible (two from West Africa and one (AS) from U.S.A.). Strains were compared with respect to their oviposition at sites illuminated at two levels (0·02 and 6·5 lumens per sq. ft.), each site being composed of a strip of moist white paper and an approximately equal area of open water set against a dark background.

Oviposition on the water surface was consistently less at the lower level of illumination. Generally the moist white paper was preferred by all strains, but oviposition on the water surface was practised more readily by strain AS than by the others, and differences between certain strains (including AS and STR) were significant.

Mass crosses and back-crosses were performed between strains AS and STR. The results were in close agreement with the hypothesis that laying on open-water surfaces is controlled by a genetic system in which there is no dominance.

The frequency of oviposition on an open-water surface was not correlated with two other characteristics (oviposition time and oviposition light preference) that have been investigated in these five strains, nor with DDT resistance.

The work described in this paper forms the final part of an investigation into the biology and breeding of Pieris brassicae (L.) in captivity and concerns the larvae and the pupae.

The larvae of the Cambridge stock used in this investigation were found to pass through five instars in the course of their development at temperatures between 12·5 and 30°C. At the lower temperature, development was completed in 46·5 days and at the higher temperature in 11 days.

The average width of the head capsules in each instar was not affected by the temperature at which the larvae were reared, it showed little variation, and it never overlapped with that of the preceding or ensuing instar and, therefore, provides a certain way of determining the instar of any larva.

At 20°C., isolated larvae and larvae kept in crowded cultures completed their development in approximately the same time—19·6 and 18·8 days, respectively.

The average consumption of food during the whole larval period was determined in two experiments, in which it was found to be 1·42 and 1·29 g. of fresh leaves per g. of larva per day, respectively.

The duration of the pupal period ranged from 7·5 days at 30°C. to about 40 days at 12·5°C.

The adults showed a definite diel rhythm of emergence. When kept at a constant temperature, with a photoperiod from 6 a.m. to 10 p.m., nearly all the insects emerged during the dark period and that immediately following it—actually between the hours of 1 a.m. and 9 a.m. If the photoperiod is displaced 12 hours, the emergence is also displaced by the same amount, to correspond with the new dark period. If, instead of keeping the temperature constant, with the photoperiod 6 a.m. to 10 p.m., it is allowed to fluctuate, as it does naturally in June, the emergence is delayed and instead of occurring in darkness and the early hours of the morning as it does at a constant temperature, it takes place mainly during the morning and the afternoon. When insects, which have been reared at a constant temperature and a photoperiod from 6 a.m. to 10 p.m., are allowed to emerge at a constant temperature, in continuous light, there is very little evidence of a diel rhythm of eclosion but if the insects are kept in continuous darkness they show a definite rhythm of emergence. If the pupae are kept in constant light but the temperature is allowed to fluctuate, most of the adults emerge during the warmer period of the cycle.

Diapause in the pupa of P. brassicae is mainly determined by the photoperiod and the temperature during the larval stages. At 20°C., larvae reared in continuous darkness do not form diapause pupae; as the daily photoperiod increases, the percentage of diapause pupae formed also increases until, at a photoperiod of 12 hours, only diapause pupae are formed. Beyond this point the percentage of diapause pupae again declines until, with a photoperiod of about 18 hours, only non-diapause pupae are formed. At higher temperatures similar trends are observed but lower percentages of diapause pupae are formed at all photoperiods.

In P. brassicae there is no evidence that a short, sharply defined period of a day or two exists in the course of the life of the larvae during which the photoperiod operates to influence diapause.

Non-diapause pupae produced from larvae reared in continuous darkness and from larvae reared in long days (over 15 hours' light) appear to contain a growth-promoting hormone capable of causing the emergence of diapause pupae.

Features of the development of Stomoxys calcitrans (L.) are described, based on observations in Uganda on stocks maintained in laboratory culture at 80°F. and 80 per cent, relative humidity. Eggs hatch within 24 hours of deposition, and the mean viability of 10 random batches was 76·5 per cent. The larvae are about 1 mm. long when newly hatched, reach maximum length (about 11 mm.) in the third (final) instar on the seventh day, and pupate one day later, the puparium being barrel-shaped and about 5 mm. long and 11 mg. in weight. The pupal period is two days. Females mate when six days old and oviposit 1–2 days later. The mean values, for five newly emerged, once-mated females, of total eggs laid, number of batches and duration of egg-laying (in days) were, respectively, 376, 10·6 and 12·2.

To be suitable for breeding of Stomoxys in the field in East Africa, sites require the presence of rotted cattle dung, rotted straw or foliage, and shade. In the dry season, such conditions occur only within cattle enclosures and particularly in the small, grass-thatched structures, provided with straw bedding, that are used for housing calves. In the wet season, suitable breeding conditions occur in open park-land and banana plantations, wherever rotted cattle manure and vegetation occur together.