Cerambyx cerdo (Cc) is a protected saproxylic beetle in Europe, although it is increasingly reported as an oak ‘pest’. Cc ecological features are relatively well known, but, its reproductive biology is still poorly understood. Hence, we investigated the reproductive traits of Cc under laboratory conditions. In females, body length was 44.1 ± 0.9 mm, 28–53 (mean ± SE, range); fecundity 143 ± 11 eggs, 33–347; fertility 78 ± 1%, 65–93; oviposition period 44 ± 3 days, 13–128 and longevity 59 ± 5 days, 16–157. Fecundity was positively correlated with female size, longevity and oviposition period. Daily fecundity was 3.5 ± 0.2 eggs/day, 0.9–6.5 showing a fluctuating synovigenic pattern with a slight decreasing trend over time. Egg length was 3.74 ± 0.01 mm, 2.3–6.0 and egg volume 5.45 ± 0.04 mm3, 2.4–9.6. Egg size was correlated with female size, but, the relative size of eggs was larger in smaller females. Incubation time was 13.5 ± 0.1 days, 7–28. Hatching was superior in larger eggs and neonate size was positively correlated to egg volume. Females were polyandrous (up to 19 matings), but, multiple mating did not enhance fecundity or fertility. In males, body length was 41.8 ± 0.8 mm, 29–53 and longevity 49 ± 3 days, 9–124. Male longevity was unrelated to body size. Males were polygynous (up to 16 matings) and mating number did not affect male longevity. Overall, females were larger and lived longer than males. Cc reproductive traits are compared with those other Cerambycidae, especially with the congeneric pest Cerambyx welensii. Our data may be valuable to improve the protection/management measures of Cc in dehesa woodlands and other oak forests.

This paper deals with a number of collections of African sandflies sent for identification by the Imperial Institute of Entomology and by Dr. P. A. Buxton of the London School of Hygiene and Tropical Medicine. I have to thank Prof. S. Adler for putting some of this material at my disposal.

Oryctes rhinoceros (L.) breeds in the rotting stumps of oil palms and rubber trees, and is a potentially devastating pest of young oil palm replants, the unopened leaves of which are attacked by the adults. In view of the planters' belief that the pest is controlled by heavy ground vegetation, the effects of cover (natural, legume or grass) and bare ground on beetle populations in rotting logs and incidence of beetle attack on young palms were compared in a series of four experiments.

Three lepidopteran cereal stemborers, Chilo partellus (Swinhoe) (Crambidae), Busseola fusca Fuller, and Sesamia calamistis Hampson (Noctuidae) were collected from maize and sorghum in Ethiopia. The noctuid stemborers are indigenous to Africa while C. partellus is an introduced species from Asia. In 1999, the Asian stemborer parasitoid, Cotesia flavipes Cameron (Braconidae) was found to be widespread in Ethiopia, even though it had never been released in the country. In addition to attacking Chilo partellus, Cotesia flavipes was reared from B. fusca and S. calamistis. The origin of C. flavipes in Ethiopia may have been Somalia where it was released in 1997 near the border with eastern Ethiopia. Percent parasitism of borers by C. flavipes was higher in eastern Ethiopia than other surveyed regions, and parasitism was higher in 2000 than 1999. Parasitism was higher when cereals were intercropped with other plants and when wild grass hosts of stemborers were present.

A polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) analysis using mitochondrial (A + T-rich region; mtDNA) and genomic (zen-region; nDNA) DNA was performed on 182 female individuals of Episyrphus balteatus (DeGeer), a widespread aphidophagous hoverfly with supposed migratory behaviour. Specimens originated from 13 sampling sites in six European countries. The analyses revealed 12 and 18 haplotypes, respectively, for the two DNA types, several of them with a wide distribution, although seven and eight haplotypes, respectively, occurred only in one location. In contrast to other studies on mobile insects, the genetic diversity was relatively high. However, lack of population subdivision, low genetic distances between populations, the very high gene flow rates, and the complete lack of isolation by distance suggest that E. balteatus populations are largely connected and that there is an absence of large-scale geographic structuring. These results support the hypothesis that E. balteatus is a migratory hoverfly species, capable of moving over large distances. These findings related to the seasonal migrations of this species are discussed.

The predators Macrolophus basicornis (Stal), Engytatus varians (Distant) and Campyloneuropsis infumatus (Carvalho) consume large numbers of tomato pests such as Bemisia tabaci (Gennadius) and Tuta absoluta (Meyrick). However, they are zoophytophagous and feed on plant parts as well. We evaluated the type and effect of injury caused by nymphs and adults of these mirids on tomato seedlings and fruit in the absence of prey. For each mirid species, seedlings were exposed to groups of 20 nymphs or adults for 72 h, and fruits were exposed for 48 h to groups of four nymphs or adults. Type and the number of injury on stems, petioles and leaflets of tomato seedlings and fruits were recorded after removal of insects. Nymphs and adults of these mirids caused necrotic rings on the leaflets, but no injury was observed on stem and petioles. The necrotic rings on leaflets consisted of blemishes, characterized by feeding punctures surrounded by a yellowish, bleached area. The number of necrotic rings did not exceed one per individual mirid and seedlings developed normally. Nymphs also caused feeding punctures on tomato fruit, but in even lower numbers than on leaflets. Two weeks after the start of the experiment the tomato fruit still looked fresh and feeding punctures had disappeared. Adults did not cause any injury to tomato fruit. The results indicate that nymphs and adults of these zoophytophagous mirids cause little injury to tomato seedlings and fruit, even when present in high densities and in the absence of prey, making them interesting candidates for biological control.

So far as could be determined from laboratory experiments the sequence of events relating to blood feeding and the laying of eggs in the life of Stegomyia fasciata was found to be as follows:—

On the second or third day after emergence from the pupa the mosquito takes her initial feed of blood. If she has previously been fertilised she lays her first batch of eggs on the sixth or seventh day, and is ready to feed again on blood a few hours later. Thereafter she lays eggs regularly every third or fourth day, and feeds on blood once only after each batch is laid in preparation for the next batch. The female continues to lay batches of eggs throughout her life, and in this way at least fifteen may be deposited. In order that eggs may be laid at all, both fertilisation and a meal of blood are required, and it is necessary that the fertilisation should precede the blood feed. Once fertilised, the female may continue to lay batches of fertile eggs for at any rate 37 days without being reimpregnated. If the eggs are laid in the early morning, she feeds during the day-time; if in the afternoon or evening, she feeds at night; but sometimes she refuses an offer to feed in daylight in favour of the next opportunity to feed in the dark.

Uscana lariophaga Steffan is an egg parasitoid of bruchid species in West Africa. The hosts are Callosobrucbus maculatus (Fabricius) and Bruchidius atrolineatus Pic, insect pests of stored cowpea. Experiments were carried out to study the selection between hosts of different ages and of different species in choice and no-choice situations. Results were obtained through parasitization studies and by direct behavioural observations. A description of the parasitization behaviour of Uscana lariophaga is given. At 30°C, C. maculatus eggs of 0–2 day old are readily accepted for parasitization, but in a choice situation the 0 day old eggs are preferred. Of the older eggs only those of 3 days old are parasitized but significantly less than those of 0–2 days. At 30°C, B. atrolineatus eggs older than 24 hours are significantly less parasitized than younger eggs also in the no-choice situation. When eggs are older than one day, the time needed for penetrating the egg increases with host age. Both bruchid species are suitable hosts for U. lariophaga. In a choice situation B. atrolineatus is the preferred host and especially so when U. lariophaga has been reared on this species. However, in a no-choice situation the highest rate of successful parasitization is obtained on hosts of the species from which they have been reared. The higher parasitization rate of C. maculatus in the field in Niger can be explained by the difference in the susceptible period between the two host species.

Quercetin is ubiquitous in terrestrial plants. The cotton bollworm Helicoverpa armigera as a highly polyphagous insect has caused severe crop losses. Until now, interactions between this pest and quercetin are poorly understood at the biochemical and molecular levels. In this study, we investigated the in vivo effects of quercetin on performance of cotton bollworm and on cytochrome P450 (P450) expression. Deleterious effects of quercetin on the performance of the cotton bollworm, including growth, survival, pupation and adult emergence were observed after oral administration of 3 and 10 mg g−1 quercetin to larvae since the third instar, whereas no significant toxic effect was found at 0.1 mg g−1 quercetin treatment. Piperonyl butoxide treatment enhanced the toxicity of quercetin. In vitro metabolism studies showed that quercetin was rapidly transformed by gut enzymes of fifth instar larvae of the cotton bollworm. qRT–PCR results revealed that the effect of quercetin on P450 expression was tissue- and dose-specific. Quercetin regulated P450 expression in a mild manner, and it could serve as P450 inducer (CYP337B1, CYP6B6) or repressor (CYP337B1, CYP6B7, CYP6B27, CYP9A14, CYP6AE11, and CYP4M7). These findings are important for advancing our understanding of the biochemical and molecular response of insects to plant toxins and have implications for a smart pest control.

The effect of three green gram cultivars (PDM 54, PUSA BAISAKHI and SAMRAT) on the biology of Spilosoma obliqua Walker (Lepidoptera: Arctiidae) was studied using age-stage, two-sex life table. We also studied food utilization efficiency measures of larvae on green gram cultivars. The nutritional and antinutritional factors of leaves of green gram cultivars were determined. The preadult development time of S. obliqua was shortest on PDM 54 (35.54 days) and longest on SAMRAT (39.29 days). The fecundity was highest on PDM 54 (318.32) and lowest on SAMRAT (250.20). The net reproductive rate (R0) ranged from 37.53 on SAMRAT to 79.58 on PDM 54. The intrinsic rate of increase (r) was higher on PDM 54 (0.1148 day−1) and PUSA BAISAKHI (0.1018 day−1) than SAMRAT (0.0875 day−1). The finite rate of increase (λ) was lowest on SAMRAT (1.0915 day−1). Mean generation time (T) was shortest on PDM 54 (38.12 days) and longest on SAMRAT (41.42 days). Population projection revealed that the population growth was slowest on SAMRAT. The growth rate of sixth instar larvae was highest on PDM 54 and lowest on SAMRAT. The lower level of nutritional factors such as total carbohydrates, proteins, lipids, amino acids and nitrogen content, and a higher level of antinutritional factors such as total phenols, flavonols and tannins influenced higher development time and lower fecundity of S. obliqua on SAMRAT than other cultivars. These findings suggested that SAMRAT is a less suitable cultivar to S. obliqua than other cultivars, and this cultivar can be promoted for cultivation.

The development of all stages of Spodoptera littoralis (Boisd.) was studied in controlled-environment cabinets under a range of constant and cycling temperatures. Larvae were tested on three food plants; development was faster on lucerne than on two cultivars of chrysanthemum, Taffeta and Fred Shoesmith. The durations of each instar were similar in relation to larval life span, but at low temperatures and on chrysanthemum many larvae went through an extra instar. Late-instar larvae were nocturnal but larval development was little disturbed by a night-break lighting regime as used in the production of all-the-year-round (AYR) chrysanthemums, and only slightly more so by a reversed temperature cycle. The latter changed the daily pattern of moth emergence which otherwise occurred mostly when temperatures were highest or in the dark. The development results were used to estimate duration of development by summing (a) daily increments of development (100/duration of development), (b) day-degrees (for eggs 28·6 day°C above 14·8°C; for larvae on lucerne 266·1 day°C above 11·5°C, on chrysanthemum 257·1 day°C above 13°C, and above 13°C 177·1 day°C for male pupae and 153·5 day°C for female pupae). The estimates were compared with observed development times for all stages in an environment where temperatures fluctuated naturally. Both methods gave satisfactory estimates when development times were short, but temperature summation was less accurate when maximum temperatures fell close to base temperatures. It is concluded that summation of development increments would give a more consistently accurate account of the life-cycle of S. littoralis on AYR chrysanthemum under glass to enable development to be predicted or past events in the life-cycle to be dated. Tables of development increments against daily maximum and minimum temperatures are given for each stage. To estimate duration of development from these Tables, the daily percentage development corresponding to each day's maximum and minimum temperature is summed up to 100%. This gives an estimate of the median point of hatching or pupation.

The account of the entomological work carried out by the writer in Southern Nigeria falls naturally into three sections, October-December 1911, January-June 1912, and June-October 1912. During the first period, work, principally on cotton pests, was carried on at the Agricultural Headquarters, Ibadan, particular attention being devoted to the Red Cotton Stainer Bug. The second period was given to travel (fig. 1), when the entomological problems of the different districts were studied on the spot, and lectures were given en route to meetings of native chiefs, native agricultural societies, schools and school teachers. The third period was spent in dealing with the insects of nursery cocoa, cotton, maize and yam. Lectures and demonstrations were given to agricultural pupils. A small amount of work with insecticides was attempted. In conjunction with the Mycologist, rules were formulated for conducting the three first Fumigatoria of the Colony.

The development of Ptinus fur (L.) and P. sexpunctatus Panzer at 70 per cent. R.H. and several temperatures has been studied.

The optimum temperature for rapid development of P. fur is about 23°C., at which temperature it completes its development in a mean period of 132·1 days on fishmeal. For P. sexpunctatus the optimum temperature for rapid development is slightly below 30°C., development being completed in approximately 111 days on wheatfeed at this temperature.

Eggs of P. sexpunctatus, but not those of P. fur, hatch at 30°C., which is near the maximum temperature for development of P. fur.

Larvae of P. fur normally moult three times at 23°C. on fishmeal, but some have an extra moult.

Well defined diapause as mature larvae in cocoons occurs at low temperatures in some individuals of P. fur ; at 23°C. this lasts about 220 days after normal larvae have pupated ; the period at 20°C. is about 280 days. Diapause at this stage also occurs in P. sexpunctatus. In experiments, the development periods of the diapausing and non-diapausing larvae overlapped to form a continuous series.

Some P. sexpunctatus pass through a period of quiescence as mature adults in cocoons ; at 23°C., this lasted 43·5 days on fishmeal and 103·0 days on wheatfeed (mean periods). In P. fur the pre-emergence period is 30 to 60 days, considerably longer than that of other warehouse Ptinids studied.

Of the foods, wheatfeed allows the quickest development for both species, but fishmeal produces the heaviest beetles.

The adult beetles live for several months. They require food and drinking water. Egg output is small, the mean number of eggs laid per female at 25°C. being 38·6 for P. fur and 21·3 for P. sexpunctatus.

P. sexpunctatus was able to overwinter in an unheated outbuilding and could possibly become a pest on some foods.

P. fur is unlikely to occur in large numbers, owing to its long development period, small egg output and tendency to larval diapause.

Comparative studies were made on three presumed sibling species of the genus Eubazus, parasitoids of European Pissodes spp. weevils, to clarify their taxonomy and define diagnostic characters. Several populations of E. semirugosus (Nees), E. robustus (Ratzeburg) and Eubazus sp. were compared with respect to their morphology (mainly through morphometric analyses), fecundity, isoenzyme patterns and host preference. Crosses were made to assess the genetic and behavioural compatibility of the populations. In addition, the North American E. crassigaster (Provancher), a parasitoid of Pissodes strobi (Peck), was compared to E. semirugosus, a species selected for introduction against P. strobi in Canada. The ratio of the length of the ovipositor sheath to the fore wing length was the most discriminating morphometric variable, but discriminant analyses including several measurements were needed to completely separate European species. A canonical discriminant function provided a total separation between males of E. crassigaster and E. semirugosus, but not between females. Eubazus crassigaster and E. semirugosus were totally separated by the banding pattern of the enzyme phosphogluconate dehydrogenase whereas hexokinase and esterase provided a diagnostic separation between Eubazus sp. and E. robustus. Eubazus sp. differed from all the other species by having a greater number of ovarioles and, consequently, a higher potential fecundity. In a two-choice oviposition test, E. semirugosus and Eubazus sp. showed a significant preference for their natural host, P. castaneus De Geer and P. piceae (Illiger), respectively. A similar test made with their progenies reared under standard conditions showed that the difference in host preference was genetically fixed. Males and females of different species did not mate readily, in contrast to individuals from the same species. All attempts to interbreed E. robustus and Eubazus sp. failed, but a few crosses between E. semirugosus and the two other European species produced fertile offspring. These observations strongly suggest that the complex of Eubazus spp. parasitoids attacking Pissodes spp. in Europe is composed of at least three sibling species, two of which appear to have specialized on distinct host species that occupy exclusive microhabitats.

Alighting of horse flies (Diptera: Tabanidae) and non-biting muscids (Diptera: Muscidae) was studied at Khartoum, Sudan, using black cylindrical models mimicking a goat, calf and cow. Flies were intercepted by attaching electrocution grids or clear adhesive film to models. Alighting sites and defensive behaviour were also documented on hosts through observation. Female Tabanus sufis (Jennicke), T. taeniola (Palisot) and Atylotus agrestis (Wiedemann) were the main tabanids captured. Muscids landed in equal numbers on all sizes of models. They had a strong preference for the upper portions of both models and hosts. Landings of T. taeniola and A. agrestis increased with model size, but not so for T. sufis. T. taeniola and A. agrestis scarcely alighted on the legs of models whereas 60–78% of T. sufis did so. Landings of T. sufis on artificial legs did not vary with model size. Landings of all tabanids on the lower and upper portions of a model increased with model size. For both hosts and models, most tabanids (88–98%) alighted on the lower half and legs. Most muscids (63–89%) alighted on the upper half. Landing of tabanids on the cow was 34.9 and 69.3 times greater than that on the calf and goat, respectively. These results are discussed in relation to strategies for the control of blood-sucking flies associated with farm animals using either insecticide-treated live baits or their mimics.

The greenhouse whitefly Trialeurodes vaporariorum Westwood, 1856 (Hemiptera: Aleyrodidae) is an invasive and highly polyphagous phloem-feeding pest of vegetables and ornamentals. Trialeurodes vaporariorum causes serious damage due to direct feeding and transmits several important plant viruses. Excessive use of insecticides has resulted in significantly reduced levels of susceptibility of various T. vaporariorum populations. To determine the genetic variability within and among populations of T. vaporariorum from Serbia and to explore their genetic relatedness with other T. vaporariorum populations, we analysed the mitochondrial cytochrome c oxidase I (COI) sequences of 16 populations from Serbia and six neighbouring countries: Montenegro (three populations), Macedonia (one population) and Croatia (two populations), for a total of 198 analysed specimens. A low overall level of sequence divergence and only five variable nucleotides and six haplotypes were found. The most frequent haplotype, H1, was identified in all Serbian populations and in all specimens from distant localities in Croatia and Macedonia. The COI sequence data that was retrieved from GenBank and the data from our study indicated that H1 is the most globally widespread T. vaporariorum haplotype. A lack of spatial genetic structure among the studied T. vaporariorum populations, as well as two demographic tests that we performed (Tajima's D value and Fu's Fs statistics), indicate a recent colonisation event and population growth. Phylogenetic analyses of the COI haplotypes in this study and other T. vaporariorum haplotypes that were retrieved from GenBank were performed using Bayesian inference and median-joining (MJ) network analysis. Two major haplogroups with only a single unique nucleotide difference were found: haplogroup 1 (containing the five Serbian haplotypes and those previously identified in India, China, the Netherlands, the United Kingdom, Morocco, Reunion and the USA) and haplogroup 3 (containing the single Serbian haplotype H3 and haplotypes from Costa Rica, the USA and Spanish Canary Islands). Collectively, our data indicate a rather limited value of COI as a genetic marker for discrimination between different T. vaporariorum populations in the investigated area. Possible explanations for the observed lack of COI sequence variability, such as specific genetics of biological invasion and/or the influence of bacterial symbionts that manipulate insect reproduction, are discussed.

The biology and development of immature stages of summer (SR) and winter rainfall (WR) strains of Onitis caffer Boheman were studied in South Africa in a summer rainfall area (Pretoria). A pair of adult beetles buried up to 650 ml of dung from one dung pad, and females laid an average of 1 egg for every 62 ml of dung buried. Broods were buried at an average depth of 50 cm in watered soil and 24 cm in unwatered soil. Eggs laid in June (winter) hatched 10 weeks later. The third instar was reached by late October. WR larvae suffered higher mortality (82%) with the onset of summer rains than did SR larvae (53%). By the following June, 6% of surviving WR individuals had emerged as adults, 12% were adults in faecal shells and 82% were diapausing third-instar larvae. Of the SR individuals, 67% emerged by June and 33% were diapausing third-instar larvae. In a second experiment, two groups of broods of the SR strain of O. caffer of different ages were set out in the field in June. One group, equivalent to broods produced early in the season (March), entered diapause as third-instar larvae, and adults emerged the following autumn. The second group represented broods produced later in the season (May); the majority did not diapause, but adults also emerged the following autumn. A field population of SR adults of O. caffer was sampled for two years. Adults were first recorded at the end of February (late summer), and newly-emerged specimens were caught up until the end of April. Maximum numbers were recorded in late March to early April (autumn), and activity ceased in June-July (winter).

In the great majority of organisms, the reproducing individual does not engender all of its progeny simultaneously. Between the beginning and the end of the reproductive period there is usually a certain interval of time, during which the production of offspring continues in a manner depending, on the one hand, upon the specific characteristics of the organism, and on the other, upon the conditions under which it lives. From this simple fact follow some interesting consequences.

If the successive generations are separated by a long interval of time and conditions during this interval are of such a nature that the organism is inactive and in the condition of obligatory dormancy, known as diapause, then, in spite of the increase of the population, the generations will remain separate and distinct, and a comparison of the total populations in the successive reproductive periods will enable us to determine the rate of increase.

On the other hand, if the interval between generations is short and favourable to activity and the organism does not pass through a period of obligatory dormancy at this time, then, as the species increases, the successive generations will come to overlap, and the amount of overlapping will increase progressively as time goes on.

This is primarily due to the fact that when there is reproduction over a period of time-intervals, in a typical individual of the species, and conditions remain constant then the number of time-intervals during which births occur will increase from generation to generation.

The taxonomy of African species of Antestiopsis, some of which are important pests of coffee, has hitherto been in confusion owing to the difficulty of finding decisive morphological characters. In the present study an attempt is made to clarify the relationships of the coffee pest species and some of their close allies by the use of both biological and morphological characters.

The species of Antestia and Antestiopsis occurring on coffee in Africa are differentiated in a key, and descriptions are given of the following: Antestiopsis orbitalis (Westw.), with three clearly distinguishable subspecies, orbitalis, bech-uana (Kirkaldy) and ghesquierei Carayon, several local colour forms of these, and a few variable populations that cannot be referred definitely to any one of the subspecies; A. intricata (Ghesq. & Carayon); A. facetoides sp.n. for the species hitherto known as “ faceta ” in eastern Kenya and Tanzania; A. crypta, sp.n. confined to Katanga; A. falsa (Schout.), a rare and imperfectly known species; A. lepelleyi sp.n., which does not feed on coffee; and A. littoralis sp.n., from the Kenya coast, known to the writer only from museum specimens.

Evidence for the arrangement adopted is discussed, and the taxonomic value of certain morphological characters is assessed. The distributions presented on maps are discussed, and it is concluded that the activities of man are unlikely to have obscured the natural patterns. A list is given of known host-plants, which, except for those of A. orbitalis orbitalis and form facetus, are with only a single other exception members of the Eubiaceae. The distribution on a local scale and recent changes of abundance of species in competition are discussed, and it is shown that A. orbitalis ghesquierei and A. o. bechuana appear to be spreading at the expense of A. intricata in Uganda and A. facetoides in Tanzania, respectively. Kesults of crossing East African species and forms show that those here regarded as species will not cross but that the subspecies of A. orbitalis will do so, producing intermediates that, when allowed to interbreed, are not as viable as pure strains. Finally an attempt to find morphometrical characters is discussed, and the results are given of a calculation of discriminant functions based on two characters. These, while, not conclusive, support the conclusions drawn from other evidence.

I remained in the proclaimed area till 6th August 1915, then returning to the vicinity of Monkey Bay for the purpose of endeavouring to establish artificial breeding places on a large scale.

While in the proclaimed area I took the opportunity of completing my survey of the distribution of Glossina morsitans, especially in the neighbourhood of Rifu and Kuti described by Dr. Shircore (Bull. Ent. Res., v, p. 87) as “primary centres 1 and 2,” which I had been unable to examine last season before the advent of the rains. As in the case of “ centres 3 and 4,” at Nyansato and Lingadzi respectively, I have not been able to find that the fly is sufficiently localised, even when the dry season is far advanced, as to render feasible any attempt to control it by prophylactic clearing of the bush. In the Rifu district there is a range of rocky hills and high ground running more or less parallel to the lake, with corresponding modification of the soil, so that a zone of scrub has sprung up, from half a mile to two miles in width, consisting very largely of thorn bush, among which are a few big trees. Towards the north this gradually dwindles, to be replaced by the borassus palms usually growing in the sandy ground along the Lake shore, and towards the south it gradually widens out and becomes continuous with the Kuti bush some five miles distant. Throughout its whole extent the fly was plentiful.