1. The responses of adult mosquitos, consisting of several Canadian species of Aëdes, were investigated under field conditions by employing heated and clothed robots.

2. Moisture increased the attractiveness of a warm body two to four times when the air temperature exceeded 60°F. ; at cooler temperatures, surface moisture decreased the attractiveness.

3. Warmth increased the attractiveness of a body, so that a robot at 98°F. attracted three times as many mosquitos as one at 50–65°F.

4. It may therefore be concluded that moisture is the major attractant factor when the air temperature exceeds 60°F., and warmth when the temperature is less than 60°F.

5. Carbon dioxide vapour emitted from an artificial head almost doubles the attractiveness of a warm body.

6. The vapours of ether and gasoline were significantly attractive. A sweat-soaked jerkin was more attractive than a water-soaked one.

7. Light colours were less attractive than dark ones, and luminescent dyes decreased the attractiveness of cloths. Green is less attractive than red or blue.

8. Glossy satins were less attractive than matt broadcloths. Greenish-khaki nylon cloth was much less attractive than khaki drill or cotton.

9. There was no evidence of interspecific differences in the responses of the mosquitos investigated.

The relationship was studied between susceptibility of a number of different species of insects to HCN fumigation and the recovery of HCN from them immediately after fumigation.

The test insects used were Tribolium castaneum, seventh stage caterpillars of Corcyra cephalonica, first-and second-instar nymphs of Drosicha sp., third-and fourthinstar nymphs of Drosicha sp. and adult females of Drosicha sp.

The apparatus and methods used in the fumigation and in the recovery of HCN from the fumigated insects are fully described.

Preliminary expsriments showed that the processes of distillation and redistillation did not affect the recovery of HCN but that the result obtained for recovery from distillation could be affected if some volatile reducing substance were produced and carried over to the distillate. It was found that this did actually take place in the case of one of the test insects—T. castaneum—but that redistillation got rid of the impurity.

In the main experiments it was shown that, on the assumption that the concentration of HCN to which insects are exposed is the effective dosage, the susceptibility of the test insects varied in the following descending order : firstand second-stage nymphs of Drosicha sp. > third- and fourth-stage nymphs of Drosicha sp.>C. cephalonica> T. castaneum>the adult females of Drosicha sp.

When the same insects were arranged in descending order of the quantities of HCN recovered per 100 gm. of body weight, the order was identical except for the nymphs of Drosicha sp. which occupied a different relative position. The two categories of nymphs of Drosicha sp. were found to occupy a different relative position again with regard to the other three test insects when exposed to a superlethal concentration and assessed for recovery of HCN per 100 gr. body weight.

Parallel batches of T. castaneum and C. cephalonica were fumigated and the HCN was recovered from the dead and survivors. More HCN was recovered from the dead insects than from those that survived.

Both recovery and sorption of HCN were estimated separately in parallel batches of insects (adult females of Drosicha sp. and C. cephalonica). Recovery was found to be less than sorption showing that a part of the HCN absorbed is converted into a non-recoverable state. Further, that the weight of HCN sorbed per gram body weight of adult females of Drosicha sp. is much less than in the case of C. cephalonica under similar conditions of fumigation and that the amount of HCN converted into non-recoverable products is less in Drosicha adults than in C. cephalonica.

A comparison of the water content of T. castaneum, C. cephalonica and Drosicha sp. (adults) showed that there was a positive correlation between water content and higher susceptibility to HCN and greater recovery of HCN was also indicated. It is suggested that this may be a factor in the “ Surface Resistance ” of an insec to a fumigant.

The observations of previous workers that larger amounts are sorbed by or recovered (after fumigation) from more susceptible species than for those less susceptible was corroborated by the results obtained with C. cephalonica, T. castaneum and adult females of Drosicha sp. but not with those from nymphs of Drosicha sp.

When dosage-mortality graphs were prepared by taking the amount of HCN recovered per gram body weight as an index of internal dose, the order of resistance of different test insects based on this new criterion was found to be entirely different from that based on the usual criterion of the concentration of HCN in the fumatorium being the index of effective dosage.

These apparently anomalous observations may be explained by assuming that the resistance shown by an insect in an actual fumigation operation, i.e., to the concentration of HCN to which it is exposed (external dose) is what may be called the total “ Effective Resistance ” and that this “ Effective Resistance ” is the resultant of (a) “ Surface Resistance ” to the entry of fumigant and (b) “ Internal Resistance ” to the amount of HCN which actually gains entry into the body in some way or other. Thus the “Effective Resistance ” of an insect may be due to a combination either of low “ Surface Resistance ” and high “ Internal Resistance ”, giving a very low “ Effective Resistance ” as in the case of C. cephalonica, or vice versa giving the maximum “ Effective Resistance ” as in adult females of Drosicha sp. The lower recovery of HCN from the nymphs of Drosicha sp., although they were more susceptible to fumigation than C. cephalonica, is explained by their higher “ Surface Resistance ” combined with a very much lower “ Internal Resistance ”, leading to a lower “ Effective Resistance ”.

Information about the recorded distribution of the tick, Rhipicephalus sanguineus (Latr.) was accumulated in the same way as that gathered and mapped by Buxton (1941) for certain fleas. The search of the literature which was begun by Miss Benzian was continued by me to January, 1949. Besides the examination of original papers the following were consulted : Review of Applied Entomology, Series B, Tropical Diseases Bulletin and Zoological Record.

The movement of weevils in a mass of grain was investigated using metal pipes 4 ins. in diameter. These grain columns were broken down into samples. It was found that the predominant movement of weevils was towards the bottom.

The major factor affecting movement was tightness of packing but although tight packing restricts movement it does not prevent oviposition. Small insects move more easily in grain than bigger ones and also show a greater tendency to move downwards. Increase of moisture content which increases the amount of airspace also increases the ease of insect movement.

The results of experiments giving the weevils a choice of moisture content were somewhat contradictory. This is partly due to the choice of the weevil being affected by the humidity it experienced prior to the test and partly due to the moisture choice being overridden by the greater attraction of the end of the grain column. On the whole, however, damper grain appears to be preferred to dry grain.

In the field, owing to the tighter packing, movement is likely to be more restricted than in these experiments. The tight packing will prevent much of the downward penetration of weevils observed in these experiments but some such movement may occur on the walls and posts. Where insects are originally placed in the centre of a bulk of grain they will often be immobile but will lay eggs and be liable to cause heating. It is clear that the insects will tend to move toward the edges of the bulk and remain there.

The breeding-place of Simulium neavei Roub. has remained undisclosed, despite many attempts to discover it, ever since Dr. S. A. Neave first collected the females in 1911. In May of that year Neave made collections near the Yala River, Kenya, and in November of the same year caught more specimens in Western Ankole, Uganda. Roubaud described and named the species in 1915. Dry in 1921 captured some specimens feeding on man at Ngoina, near Kericho, Kenya, and made the important observation that the local natives connected the fly with the occurrence of a chronic skin disease in this district which has since been identified as onchocerciasis.

Simuliid flies are a well known feature of riverain land in the northern Sudan during months of low Nile. They have been discussed in detail by Lewis (1948). Two species occur, Simulium damnosum Theobald, and S. griseicolle Becker. The former is larger and darker in colour. The latter is smaller with characteristic grey shoulders and is usually much more numerous. S. damnosum bites humans readily, especially round the ankles. S. griseicolle is usually less effective in its attempts to bite humans but susceptible persons may be badly bitten. It commonly attacks the head and shoulders and may cause great irritation by invading the eyes, ears, nose and mouth. Temporary protection for up to four hours can be obtained by rubbing the skin with di-methyl phthalate which acts as a repellent against both species.

1. The reactions of Xenopsylla cheopis Roths, to uniform and alternative relative humidities have been investigated.

2. X. cheopis aggregates in the dry side of the alternative chamber. There is an indication that the intensity of aggregation increases with increase in percentage difference between the alternatives and that it is also high when the alternatives are high in the scale.

3. Locomotor activity and speed of movement increase with rise in relative humidity.

4. Males kept at 90 per cent. relative humidity for five days aggregate intensely in the drier of the alternatives 65 and 85 per cent. R.H. Similar fleas kept five days at 70 per cent. relative humidity react weakly.

5. There is greater activity after five days' exposure to 90 per cent. relative humidity than after exposure to 70 per cent. R.H. for fhe same duration.

6. One concludes that high activity in wet air contributes towards aggregation in the dry side and that the intensity of aggregation is affected by the difference in activity which occurs in the fleas at different relative humidities.

7. The intensity of aggregation to the alternatives 80 and 90 per cent. relative humidity is unaffected by sex, or by the remains of a blood meal.

8. A high klino-kinesis may occur in X. cheopis.

9. Nosopsyllus fasciatus reacts in the same way as X. cheopis to uniform and alternative humidities, but there is no evidence of a high klino-kinesis.

10. When a roll of lint is dropped into a vessel infested with X. cheopis the number which jump on to it increases with relative humidity. The number which jumped on to a month-old rat did not show a clear relationship to humidity. Relatively more females than males jumped on to the roll of lint and rat.

11. A colony of 25 buck rats was infested with X. cheopis. In relation to size of flea population the flea index increases with rise in relative humidity. The proportion of each sex on a rat, to its respective number in the bedding material, does not appear to bear a simple relationship to humidity.

12. Humidity receptors are not confined to the distal three segments of the maxillary palps if indeed there are any there.

A combination of historical, geographical, and epidemiological studies has given sufficient insight into the ecology of sleeping sickness to enable the main factors influencing the development and spread of an epidemic to be traced.

The evidence shows that in West Africa sleeping sickness is not primarily a disease of the forest, where tsetse flies are most abundant, but belongs essentially to the dry country in the north of the savanna woodland zone, where the earliest occurrences and severest outbreaks have been located.

The first mention of sleeping sickness comes from the upper Niger and dates back to the 14th century. By the beginning of the present century intense though localised epidemics were devastating parts of the Mossi, Grounsi and Lobi country of the upper Volta rivers. At this time the disease was unknown on the coast and of sporadic occurrence only in the forest. A severe trans-Volta epidemic covering 60,000 square miles, developed between 1924 and 1940, but was confined to the north of the inland savanna zone with nothing comparable in the forest.

The epidemic spread in three principal ways : (1) Outwards from original foci of infection because of the dynamic nature of the disease. This produced a concentration of infection around headwaters, a feature characteristic of advanced epidemics. (2) Through the agency of travellers, originally from north to south but subsequently in both directions : a rapid method of spread producing linear distribution of infection along trade routes. The tempo was greatly increased on the pacification and development of West Africa after 1900. (3) A gradual southward shift in the main epidemic zone appears to be resulting from a long-term change in the African climate which is combining with man's activities to produce a southerly extension of xerophytic vegetation types and a regression of forest.

The most important spread was that caused by the trading caravans, more especially the cola traders, who have been coming down to the cola-nut areas in the Ashanti forest from the big markets on the Niger and Upper Volta since the 11th century. The caravans were formerly very large, up to one or two thousand strong, and were frequently made up of Mossi and Grounsi from the territory that was so heavily infected by the beginning of the present century. It is certain that a continuous introduction of infection would have been taking place into the forest ever since trypanosomiasis was prevalent in the north, that is for 100 years at least. And infection has been known in the forest for about that period, yet always to a mild degree, never reaching epidemic form. It has been sought for, because conditions in the forest, with the vector Glossina palpalis in contact with every village and path, appeared to be ideal for the transmission of infection and this drew the particular attention of the early workers from 1908 onwards. But the most that could be found was a threatened epidemic in north-west Ashanti, very significantly centering on the big cola markets which formed the termini for the northern traders.

This historical evidence and the reasoning from epidemiology lead to the conclusion that conditions in the forest are not conducive to the development of epidemic sleeping sickness and that the low state of endemicity found there is maintained by the constant introduction of infection from the true epidemic areas in northern savanna.

From this conclusion arises a practical point of the greatest importance. If the sources from which infection is introduced into the forest could be eliminated the disease there should eventually die out and the tsetse, from the human point of view, would be harmless. Tsetse control in the forest may prove difficult and expensive, and if it is attempted by clearing this might end in the literal destruction of the forest. Such measures would be hard to justify, so many other factors of possibly greater importance than trypanosomiasis are involved, both the intrinsic value of a forest for its products and the wider value through its influence on climate, soil and water.

In formulating a plan for the control of sleeping sickness, the habits of both vectors, human as well as insect, should be considered. The tsetse plays a major role in the development of the high infection rates characterising the epidemic outbreaks in northern savanna ; the human vector distributes infection from these sources along trade routes and into the forest. The elimination of the disease at its source, in true epidemic centres, which can be most effectively accomplished by eradication of the tsetse, will check the distribution of infections to the secondary areas of lighter infection which could then be cleared up by quite minor control measures or might even disappear spontaneously.

This plan is now in operation in the Gold Coast. The validity of the arguments on which it was based is being shown by the results that are already apparent : the high rates of reduction in the epidemic areas and the pronounced lowering of infection in neighbouring, uncontrolled areas, more particularly in the forest region of north-west Ashanti where it is entered by a trade route coming from the previously heavily infected country.

Experiments have been carried out on the insecticidal action of p-, m- and o-iodo-nitrobenzene and their dichlorides. The real toxicity of these substances was estimated with the micro-injection method of Dresden, the contact action with the petri-dish method of Krijgsman & Berger. The real toxicity of the p- compounds appeared to be somewhat higher than that of their isomers. The dichlorides showed a somewhat higher toxicity than the other compounds. The contact action of the p- compounds is much higher than that of their isomers. This is caused by considerable differences in permeation velocity. The position of the iodine atom affects the permeation velocity very considerably but has much less influence on the real toxicity. It is suggested that the enhanced effect of the p- position on the permeation velocity is a general rule.

P-iodo-nitrobenzenedichloride shows such a high permeation velocity that it must be considered as a promising contact insecticide despite the fact that its real toxicity is low as compared with several other modern insecticides.

Attention is again drawn to the presence of mating scars in G. palpalis. They vary in degree and are caused by the superior claspers. The function of the inferior claspers is described and a function for the vermiform appendices is suggested. Flies mated under certain laboratory conditions may fail to develop scars, for reasons given. There are no mating scars in the two other tsetse-flies found in Sierra Leone, viz., G. longipalpis and G. fusca owing to the character of the claspers in those species. For this reason it is likely that the scars are confined to flies of the palpalis group. A suggestion is made concerning the physiology of the production of the scars and attention is called to their bionomic importance. The results of spermatheca dissections of unscarred and scarred females are tabulated. From these it is concluded that the absence of mating scars provides a fairly reliable index of the rate of emergence at any particular time.

Perfusion of the isolated heart of Periplaneta americana with caffeine, digitalin, acetylcholine, nicotine and lobeline shows that the pacemaker of this heart is different from the “ myogenic centre ” of the vertebrate heart.

The action of strychnine, morphine and apomorphine on this heart preparation affords evidence of the existence of a neurogenic pacemaker.

This pacemaker is stimulated by suitable concentrations of acetylcholine, nicotine, lobeline and pilocarpine, while it is inhibited by atropine. Acetylcholine and tetraethyl pyrophosphate show a synergistic action. Atropine and tetraethyl pyrophosphate, like acetylcholine and curare, show an antagonistic action. These results prove that the neurogenic pacemaker possesses cholinergic properties.

Adrenaline stimulates the insect heart and ergotamine inhibits it, thus suggesting that the insect heart probably also has adrenergic properties.

On the basis of the present work and the results obtained by other investigators, a theory is put forward that the heart mechanism of most arthropods consists of a neurogenic pacemaker with adrenergic properties, controlled by a cholinergic accelerating nerve. This mechanism bears some resemblance to the sympathetic nerve system of vertebrates.

Rotenone strongly counteracts the action of acetylcholine, tetraethyl pyrophosphate, nicotine, lobeline, pilocarpine and digitalin on the insect heart. Its point of action, however, remains obscure at the moment.

The opposing action of rotenone and tetraethyl pyrophosphate is an indication that a combination of these insecticides for pest control is not to be recommended.

An account is given of an attempt to estimate changes in the rate of egg laying in a warehouse and in an unheated outbuilding at the laboratory. Small samples of food were placed on pieces of newspaper on the floor, and were renewed regularly, weekly or fortnightly in the warehouse and daily at the laboratory. Larvae were bred from these food samples and counted.

The rate of oviposition was found to depend on temperature, especially in winter, and on relative humidity, especially in summer. It was seldom cold enough indoors, in unheated premises, to prevent oviposition entirely for more than a few days at a time, 5°C. being the approximate minimum temperature for oviposition. It was never so dry at the laboratory that egg laying stopped but this did happen at the warehouse in July, 1946. Free drinking water is necessary for continued oviposition and this was available from rain in both buildings.

Darkness causes an increase in activity and in oviposition and most of the eggs in the foods supplied were laid at night. Foods placed near the daytime hiding places of adults were most used for oviposition.

The garden chafer, P. horticola (L.) is widely distributed in England and Wales but is more abundant in the western half of the country. Its natural habitat is poor quality permanent grassland characterised by a diversity of species of flowering plants and a high proportion of weeds. The soil of the infested areas is invariably light, the land is usually hilly and the rainfall relatively high.

Adults emerge suddenly in May or June according to the weather. Males emerge some days before females. Daily activity increases during the flight period which lasts 3–4 weeks. Beetles only fly in warm sunny weather.

The females lay an average of 14 eggs each. Fecundity is correlated with weight and is related to the food supply of the larvae. In nature, the weight of the adults may be limited through the exhaustion of the food supply of the larvae. Females will mate and lay fertile eggs without feeding. In general, feeding by the adult does not affect fecundity but in one experiment beetles fed on salad burnet produced and laid more eggs than beetles fed on bracken, blackberry, grass or kept without food. Over 90 per cent. of the eggs are fertile ; fertility decreases with the age of the adult. The incubation period of the eggs is about 22 days in the laboratory but about 28 days in the field.

The spatial distribution and seasonal changes in population of P. horticola have been studied within an infested field on the Tickenham Ridge, Somerset. The eggs are laid in groups from which the larvae disperse. At the end of the feeding period the larvae may be distributed at random but there is some evidence that regrouping at feeding sites occurs. Most of the eggs are laid in the top 3 ins. of soil. The larvae feed in the root zone but go deeper in the soil in winter to hibernate. The population fell in successive years within the area studied. Initially there was a higher infestation in part of the field where the turf was damaged by the larvae but the relative intensity of infestation changed and subsequently more eggs and larvae were found in part of the field where the turf was undamaged. There was a higher mortality among larvae in the damaged part of the field. These population changes have been related to a number of environmental factors which may account for them, in particular to dispersal of the adults, density of the vegetation, consolidation, soil moisture and predators. In a field at Alton Pancras, Dorset, predation by birds reduced the larval population by 50 per cent. between 1st September and 1st November.

A study of the effect of infestation upon the vegetation showed that damage was largely due to the destruction of the grasses. Weeds were unaffected and there was no evidence of an increase of weed area by colonisation of bare patches resulting from damage. Brachypodium pinnatum resisted attack owing to its tough rootstock. Dactylis glomerata and Lolium perenne were relatively resistant, Poa spp. and Festuca spp. were damaged but regenerated when feeding stopped and the population declined. Salad burnet, a favoured food plant, was more abundant in damaged areas and its distribution may influence the distribution of P. horticola.

When infestation is severe and the grass roots are destroyed, the soil aggregates are broken down. The soil structure was restored when the grasses regenerated, and by consolidation.

Field populations were reduced by ploughing and by reseeding. Dusting with 3·5 per cent. benzene hexachloride at 70 lb. per acre during the flight period gave promising results when applied in favourable weather.

Losses from Pink Bollworm (Platyedra gossypiella) and from Spiny Bollworm (Earias insulana) were studied in the American Upland cotton crop grown on Zeidab pump irrigation scheme in the Northern Province of the Sudan. In 1948, a year of normal attack, 12·9 per cent. of the estimated potential yield was destroyed by bollworms. Pink Bollworm was responsible for 10·7 per cent. or a loss of 135 pounds of seed cotton per acre. Spiny Bollworm damaged 2·2 per cent. or 28 pounds per acre.

Four-fifths of the total yield is derived from the bottom crop which is harvested before the end of September and consists of high grade seed cotton. The other fifth is produced as a top crop and consists of low grade seed cotton ripening in October and November. Only 3·5 per cent. of the bottom crop is lost on account of Pink Bollworm as against 33·3 per cent. of the top crop. The corresponding figures for Spiny Bollworm are 1·6 per cent. and 4·1 per cent.

Some factors influencing the extent of pink bollworm attack are outlined. The Zeidab crop escapes serious damage through rapid maturity and even growth. Correct agricultural management is of great importance. Attack by Laphygma exigua in the seedling stage may result in increased damage by Pink Bollworm in the ripening crop.

The origin of the initial pink bollworm infestation is discussed. Available evidence, which is not conclusive, suggests that local carry over of the dead season may not be sufficient to account for the initial infestations observed in the crop.

Atmospheric water vapour saturation deficit appears to influence the duration of larval diapause in Pink Bollworm in the northern Sudan. Maximum emergence of moths from resting larvae in three areas occurs at or near the lowest saturation deficit of the year. The possibility is discussed that the degree of fixity of the diapause may be related to the gradient of the saturation deficit curve during the development of the resting larvae in the crop, a rising curve being associated with a well fixed diapause which lasts until the saturation deficit falls, and a falling curve being associated with a poorly fixed diapause which does not last but begins to break at once.

Cordia macrostachya (Jacq.) R. & S., is a Boraginaceous shrub indigenous to the West Indies, Central America, and northern South America. It was accidentally introduced into Mauritius about 1890, probably from British Guiana, and became a weed of major importance.

A comparative ecological study of Cordia in Mauritius and in Trinidad showed that its invading power in the former island rested upon a profuse production of viable fruits, whereas in Trinidad, where the plant is of no agricultural concern, fruit production was controlled by the combined action of a number of insects. It was accordingly decided to introduce suitable species of these insects into Mauritius.

The first to be introduced was a leaf-eating Cassidid beetle, Physonota alutacea, but as described elsewhere, this introduction, which took place in 1947–48, was not a success.

The introduction of the second species, Schematiza cordiae Barber, and its result, is described in the present paper. This Galerucid is also a leaf-feeder, but it will sometimes feed upon the inflorescences.

Through the agency of the Commonwealth Bureau of Biological Control, air consignments of Schematiza were received between June and September, 1947. Breeding was started, and after feeding tests in both Mauritius and Trinidad had established its specificity to C. macrostachya, the release of laboratory reared adults was commenced in March, 1948. In February, 1949, after about 18,000 had been released, breeding was no longer necessary for large populations had developed in the field and were inflicting severe injury to Cordia scrub. Dispersion of the insect over the island followed quickly, and was assisted by the collection and distribution of about 60,000 adults. Schematiza became extremely abundant so that by the middle of 1950 most expanses of scrub in the island had either been stripped of leaves and inflorescences, or had generally been severely defoliated.

As a result, much scrub is dead owing to a prolonged period without foliage, much is being subjugated by other vegetation, while partial defoliation and direct inflorescence attack have reduced the reproduction of other bushes.

This rapid control of C. macrostachya is attributed to the high fecundity of the Schematiza combined with its freedom from parasites and important predators, and to its suitability to the climates of all Cordia infested regions.

Arising out of a routine mosquito survey, carried out in a coastal district of the Gambia in 1946, it was discovered that specimens of typical A. gambiae showed a variation in maxlllary indices, suggesting the possible differentiation of the species into two races distinct from the known variety “ melas ”.

Material from catching stations, considered beyond flight range from possible breeding places of “ melas ”, was subjected to statistical analysis as were also collections of adults emerging from different categories of fresh-water breeding places.

The analysis showed that the maxillary indices of the samples obtained from control house catches were distributed in an abnormal manner and were therefore unlikely to have been drawn from the same population. At the same time, the samples collected by window-traps and spray-catches were found to differ significantly in their means, and this supported the view that a mixed population was being sampled. The samples obtained by trapping emerging adults from different categories of breeding places, where any adulteration by “ melas “ was a physical impossibility, showed significantly different mean maxillary indices. It was therefore very unlikely that they could have been drawn, by random sampling, from a pure population. When the samples, obtained by trapping over-breeding places, were divided according to their mean indices, it was found that those falling into the group with a mean index of 13·107 were taken from collections of water which could be classified as casual and those falling into the group with a mean index of 15·33 were collected from permanent pools.

Preliminary research on the micro-flora of fresh-water pools has shown that certain distinct phases, marked by dominant plants, can be distinguished. The distinction between casual water and permanent water is capable of more accurate definition by reference to the dominant species. It has been observed that, within the limits of statistical differentiation, the race having the larger mean maxillary index is never found in water where photosynthesis, and hence chlorophyll, is absent whereas the race with the smaller mean maxillary index is never found in water where photosynthesis has been established amongst the micro-flora.

The apparent difference in behaviour between the two races, as shown by the predominance of the 13·107 index group in the window-traps, and the predominance of the 15·33 index group in the spray-catches, cannot it is thought, be accepted as evidence unless the existence of a fatigue factor, accounting for house-resting habits in members of the latter group which have travelled far and exhausted their energy reserves, can be excluded.

It is considered that the results of the work described in this paper may be of value in permitting control house catches to be analysed. From such analysis it should be possible to assess the proportion of adults which had originated in casual water and the proportion coming from more permanent breeding places. If, of course, it should be proved that the apparent differences in adult behaviour are, in reality, linked with differences in mean maxillary indices, it will be necessary to take this into consideration in assessing the value of residual insecticides and mass spraying. It is further argued that, even if the differences in behaviour can be explained on the fatigue factor hypothesis, it will still be necessary to take them into account in attempting to assess the value of mass spraying technique.

It is considered that the method of statistical analysis could be more widely utilised in exploring many problems arising in the field, and that other variations, apart from those of maxillary dentition, might prove even more fruitful as a basis for analysis, although they would call for more time and labour than can usually be spared.

Thanks to the courtesy of Mr. G. S. Dun, Government Entomologist to the Territory of Papua and New Guinea, we have been able to study three interesting new species of Mirids found damaging cacao plants in New Britain. The first was discovered by Mr. Dun early in 1950 in abundance at Kabeira Plantation about six miles from the Lowland Experimental Station at Keravat where considerable damage had been done to the crop. This species was at first thought to be the West African Bryocoropsis laticollis Schumacher, 1917, which might have been accidentally introduced into New Britain. Closer investigation showed that it belonged to a distinct genus and species (Parabryocoropsis typicus, gen. et sp. n.). In September 1950, Mr. Dun found a second species attacking an isolated block of Cniollo cacao at the Keravat Experimental Station. On examination this proved to belong to still another genus and species (Pseudodoniella pacifica gen. et sp. n.).

Four species of Haemagogus have been found in Brazil, the commonest and most widely distributed being H. spegazzinii.

Keys and diagrams are given to facilitate the identification of the Haemagogus of this country.

Two of the proven vectors of sylvan yellow fever, H. capricornii and H. spegazzinii, can be distinguished one from another only by their male terminalia.