1. Work on the factors controlling the avidity of Aëdes aegypti females for a blood meal previously reported by Seaton & Lumsden (1941) was continued.

  2. Mosquitoes were kept after hatching for 96–120 hr. in darkness at about 25° C. and 80% relative humidity, and were then offered a blood meal on the human arm in darkness.

  3. Wide variations in the light intensity during the few minutes’ manipulation immediately previous to offering the host did not significantly affect the numbers feeding.

  4. The relative humidity of the environment at the time of biting appeared to be of little or no importance.

  5. The optimum environmental temperature for biting was near 35° C. At 35° C., out of batches of ten mosquitoes offered 5 min. opportunity to feed, means of 8·8 ± 0·44 and 9·2 ± 0·33 fed at high and low humidities respectively.

  6. Large proportions of A. aegypti females, confined within 8·5 cm. of a host and in darkness, were able to locate the host and to feed when environmental temperatures approximated closely to the skin temperature or even were above it. Some factor other than the presence of a temperature gradient rising to the skin must therefore be responsible for their orientation to, and probing of, the host.

  7. In the majority of fed mosquitoes blood was found in the stomach alone, but partly fed mosquitoes showed a higher proportion with blood in the diverticula than did those fully gorged. Such an effect is probably due to regurgitation of blood from the stomach when feeding is interrupted before its normal completion.

A summary of the results of earlier workers has already been given by Seaton & Lumsden (1941), of whose work the present is a further extension. In that paper were reported investigations to determine the effects of age, fertilization and light on the avidity for a blood meal of female Aëdes aegypti (L.). There remained to be worked out the effects of variations in temperature and humidity of the microclimate at the time of biting, with which the present work deals.

The A. aegypti used were originally derived from West Africa in 1926 and have been maintained since then in the laboratory, the adult females being fed on rabbits. Stock eggs were allowed to develop at weekly intervals. The larvae were fed on a mixture of powdered dog biscuit and ‘Bemax’, air being bubbled continuously through the cultures to prevent scum formation. As pupae appeared they were transferred to bowls of clean water and hatching was allowed to take place in cages in an insectary at about 28° C. and 65 % relative humidity (means of seven observations 27·7° C. and 66% relative humidity). No attempt was made to segregate female pupae, as it had been found previously by Seaton & Lumsden (1941) that fertilization did not affect subsequent avidity for a blood meal. The bowls were transferred daily to clean cages, so that all the adults in a given cage were known to have ecloded in the previous 24 hr. Material for experimentation was usually derived from three of these successive daily cages or lots each week. Females were removed in batches of 12–15 to net-ended glass cylinders 7 cm. long and 3·5 cm. in internal diameter. Usually six of these batches made up from each lot. The cylinders were placed end to end in glass tubes 47 cm. long and 4·5 cm. in calibre, whose ends were closed by rubber bungs carrying short pieces of glass tubing serving as inlets and outlets for a stream of air. The tubes were connected in series; an air current, humidified by bubbling through a capillary immersed in glycerolwater solution of specific gravity 1·106, was passed through the whole series at a rate of about 8 l./hr. The glycerol-water solution and the tubes were maintained in a darkened incubator. The tubes were covered with opaque black paper so that the mosquitoes were in constant darkness, even when the incubator was opened. Records of temperature were kept by means of a thermograph, checked daily against a chemical thermometer, and humidity determinations of the outflowing air were made from time to time by means of a dewpoint hygrometer. The extreme temperatures recorded were 23·4 and 26·3° C. The aggregate degree-hours for each lot was recorded, covering an average period of 12 hr. in the insectary, for which the mean temperature was taken as 28° C., and 4 days, reckoned from noon to noon, under the accurately controlled and recorded conditions in the incubator. Temperatures were reckoned from 8° C., as this appears from the work of Lewis (1933) on longevity to be a fairly accurate estimate of the temperature threshold of A. aegypti adults. Humidity records varied between 75 and 87 % relative humidity, the mean of ten random observations being 81·6%. Batches of mosquitoes were starved under these conditions in the incubator for 4 days; they were used for experiments when 96–120 hr. old, i.e. in the age group 5 of Seaton & Lumsden (1941). Seaton & Lumsden estimated the percentage mortality in that age group at about 12. The mortality in the present work was recorded and was generally slightly less, probably owing to better humidity control. Lots in which the mortality exceeded 10% were excluded, as it had been found previously that higher mortalities were associated with heterogeneous behaviour of batches at the time of feeding. Such higher mortality rates were met with on a few occasions, as, for example, when defects in the incubator control allowed the temperature to rise for periods above the range mentioned above.

The experimental chamber was of the same basic type as that used in previous work but differed in the following details. It was of slightly different dimensions; length 18 cm., internal diameter 4·5 cm., and the net diaphragm 8·5 cm. from the lower end. The thermometer was 1·5 cm. from the skin surface. A disk of asbestos sheet 3 mm. thick, with a central hole through which the mosquitoes had access to the skin, was added to the lower end of the chamber to reduce heat interchange between the chamber interior and the applied arm. The size of the biting aperture was increased to 500 sq.mm. The hygrometer used was of the dewpoint type described by Buxton (1931) and was immersed in the water-bath, the air current passing through it immediately before entering the chamber.

As before, glycerol-water solutions in flasks immersed in the water-bath were used to obtain high humidities; for low humidity work the flasks were removed, and air which had been passed through two towers of silica-gel was admitted direct to the copper coil and filter in the water-bath. The rate of the air current was about 28 l./hr., corresponding to a renewal of the total volume of air in the experimental chamber i-6 times/min. When much of the work had been completed, an imperfect joint was discovered, allowing some leakage of air from the upper end of the experimental chamber and therefore upsetting the precise determination of the atmospheric humidity therein. Further series of experiments were run at both high and low humidities, and where the earlier results are quoted, a note is added to make this clear.

The experimental procedure was similar to that already described, except that in the case of low humidities it was impossible to determine the humidity of the air stream using ether evaporation as the sole method of cooling. The thimble of the dewpoint apparatus was therefore cooled by means of a freezing mixture. This method was cumbersome, and it was considered sufficient to carry it out twice, before and after the experimental series run at low humidities, the conditions of the drying towers remaining unaltered in the meantime. Maintenance of the dewpoint hygrometer at—7·25° C. for several minutes resulted in no visible deposition of dew. This indicated a maximum possible vapour pressure of 27 mm, Hg (Landolt-Börnstein, 1923). The values plotted on the graph (Fig. 1) for the low-humidity series, are therefore maximum values only. Temperature and humidities recorded for experiments all refer to the conditions at the level of the chamber thermometer, 1·5 cm. from the skin surface. They are closely representative also of the conditions in the part of the chamber above this level, but below it the skin surface inevitably causes modifications dependent upon the temperature difference between it and the chamber air.

The temperature gradients close to the skin surface in the apparatus under normal working conditions are shown in Fig. 2. They were worked out by means of the resistance-type skin thermometer described by Bourne (1946). This instrument operates on resistance changes, due to temperature, in a minute bead enclosed at the tip of a glass probe. The instrument is designed primarily for the estimation of surface temperatures, i.e. with the bead in contact with a medium of fairly high conductivity. Calibration was carried out against a standard chemical thermometer in a mercury-bath enclosed in a water-filled thermos flask. If the instrument is used, as in the present work, for the estimation of air temperature, possible errors due to slowness of heat interchange between the bead and the surrounding air must be considered. The period required for stabilization in air was found to be less than 3 min., and as readings were taken at 3–4 min. intervals and only when the apparatus gave a constant reading, errors from this source were eliminated. The current necessarily passed through the bead to determine its resistance might be expected to have some heating effect resulting in too high values in media of low conductivity. The effect was, however, found to be negligible; two pairs of observations, in mercury and air, at 17 and 37° C. respectively, differed by under 0·15° C., and in both cases the value for air was actually below that for mercury. Eight temperature gradients near the skin were worked out, two at each combination of high and low temperature and high and low humidity. There was no consistent difference due to humidity, and the curves given in Fig. 2 are representative.

Mosquitoes were allowed to remain in the chamber usually for 3·0–5·5 min. before the arm was applied, thus allowing for replacement of the air in the chamber 5–8 times. Only on four occasions was the preliminary exposure longer, 7–11 min. Three of these were in the lowest temperature group of the high-humidity series when the small amount of dew deposited made reading of the hygrometer difficult. The other was in the 35° C. group of the same series. The mosquitoes were allowed a period of 5 min. access to the pronator aspect of the forearm, always of the same individual, in each experiment. In all experiments the chamber interior was in darkness.

Statistical methods used follow those given by Simpson & Roe (1939). Standard errors are used throughout.

Seaton & Lumsden (1941) have shown that direct illumination at 0·5 m.c. of the skin area available to mosquitoes reduces almost by half the numbers biting. Tate & Vincent (1935) found that up to about 20° C. prolonged illumination induced about 70% of females of the English anautogenous race of Culexpipiens L. to gorge, whereas only 6% or less of those kept in the dark did so. This effect occurred also at 25° C. but was not so well marked. Light did not significantly influence three autogenous races in this respect. In view of these findings it was decided first to see whether, after 4 days in the dark at about 25° C., variations in light intensity during the period of transference from the dark of the tubes in the incubator to the dark of the experimental chamber exerted any influence on the numbers of females biting. For this purpose two parallel series of experiments were run; in one series transference took place in the ordinary room conditions of diffuse daylight, and in the other the mosquitoes were manipulated in the lowest light intensity at which the operation could be satisfactorily carried out. This was in fact very low; it was found that with partially dark-adapted eyes the mosquitoes could be quite easily dealt with when seen only as vague silhouettes against a pale background. Eight experiments were carried out after manipulation in ordinary diffuse daylight, six with mosquitoes of lot 70 and two with mosquitoes of lot 71. Seven experiments were carried out after manipulation in dim light, four with mosquitoes of lot 70 and three with mosquitoes of lot 71. The aggregate degree-hours for these lots were both 1795, and their mortalities were respectively 7·3 and 9·8% during the 4 days in the incubator. The means and standard errors of-the temperature and humidity records of the two series of experiments, based on both the initial and end readings, were:

Diffuse light series: 25·01 ±0·16° C. and 77·2±0·92% relative humidity.

Dim light series: 25·05 ±0·20°C. and 78·75 ± 1·1 % relative humidity.

The humidity records are, however, inaccurate owing to the chamber defect noted above, but this affected both series, so that as far as the effect of preliminary manipulation is concerned the series are considered comparable.

In the series manipulated in diffuse daylight five mosquitoes fed in one batch, six in one, seven in four and eight in two, a total of eight batches. In the corresponding series manipulated in dim light four mosquitoes fed in one batch, six in three and seven in three, a total of seven batches. The mean numbers feeding for the two series were:

Diffuse light series: 6·14 ± 0·37.

Dim light series: 6·875 ±0·33.

These means are not significantly different, t=1·38 and P>0·1. It is therefore concluded that as such a wide variation in light intensity during the period of transference to the apparatus does not significantly affect the results the much smaller normal variations can also be neglected.

All experiments were carried out with the chamber interior in darkness after transference of the mosquitoes in diffuse daylight. Two successive series of experiments were run, one at high and one at low humidity. The distribution of lots to experimental batches is shown in Table 1. To counteract possible variations in the behaviour of lots, batches from a given lot were distributed as much as possible throughout the temperature values tested. It was not, however, practicable to distribute batches from one lot to both humidity series, as the time involved in changing the humidity controls frequently and waiting for the apparatus to stabilize would have made working impossibly slow. There was, however, no difference in the treatment of the mosquitoes used in these different series with the exception that owing to warm weather, which occasionally resulted in room temperatures slightly above the chosen incubator temperature, there was a slight increase in the aggregate degree-hours for some of the lots in the high-humidity series. The values for all the lots are shown in Table 1, which also shows the percentage mortalities.

Table 2 and Fig. 1 show the results of both series of experiments. The mean temperatures for each group and their standard errors are calculated from the initial and end temperatures of each experiment. The means of corresponding temperature groups in the two series are close together with the exception of the lowest pair. The temperature and humidity values plotted in Fig. 1 are the mean values for each experiment in the case of temperature, but, as noted above, the humidity values in the low-humidity series are mean maxima. Initial and end relative humidity estimations were almost always close together; the greatest divergence was 13% and most were less than 5% apart.

The main activity zone for feeding lies, in both series, from 25 to 35° C., with an optimum near 35° C. Humidity variations of wide amplitude appear to exert little influence, as only at 20° C. does the t test show any statistically significant difference between corresponding groups worked at high and low humidities. However, in view of the close correspondence of all the other groups, this difference may be regarded as fortuitous until it receives further confirmation. Lumsden & Bertram (1940) considered that the application of saliva to the skin of the fowls used for feeding Aëdes aegypti females induced gorging, and at that time the apparent effect was thought probably to be due to a resultant rise in atmospheric humidity.

It was therefore unexpected to find in the present work that such wide variations in atmospheric humidity were without pronounced effect. Bishop & Gilchrist, similarly, in early work (1944) gained the impression that saliva increased the rate of gorging in mosquitoes offered a blood meal through animal membranes, but later (1946) proved that it had little effect. On the other hand, Lewis (1933) found a reduction in the proportion of A. aegypti feeding with lowered atmospheric humidity at both 25 and 30°C. His method was similar to that used in the present work, though there were several small differences in the treatment of the material. Lewis used females 6–7 days old kept in ‘damp air’ in the dark at 23°C.; his preliminary exposure was longer—10 min. ; the area of skin per individual female was greater—about 240 sq.mm, instead of 50 sq.mm. ; the chamber atmosphere was changed more slowly—only about 0·2 time/min.; sulphuric acid was used as the drying agent for low-humidity experiments. These differences are probably insignificant, but it is noteworthy also that his experiments were carried out in light, while in the present work the chamber interior was always in darkness. The present work, and almost certainly that of Lewis also, is concerned with the first blood meal of A. aegypti females. It is generally agreed that in nature the first meal is taken in daylight though Marchoux, Salimbeni & Simond (1903) in South America consider that for later meals the mosquitoes tend rather to bite at night. Edwards (1941) quotes Flu in Surinam, and Cardamitis in Greece, as having come to the same conclusion. It is possible that the difference between the present results and those of Lewis may be due to different light conditions; perhaps low humidity as a factor inhibiting biting activity is only operative in the presence of light, since only during the day would the mosquito encounter humidities low enough seriously to endanger life. It is significant in this connexion that when Lumsden & Bertram (1940) obtained the impression that saliva encouraged gorging, light was not completely excluded. A second possible explanation is that in mosquitoes starved for long periods hunger may become a master factor causing activity in search of food despite unfavourable environmental conditions. Larsen (1943) observed several Noctuid moths on the wing by day following a period in which bad weather, particularly at night, had inhibited normal nocturnal activity.

The temperature gradients close to the skin in the present apparatus under normal working conditions are shown in Fig. 2. The lower curve, obtained with the apparatus at room temperature, shows that the temperature gradient is limited to the region within 10 mm. of the skin. The more gradual gradient of the upper curve must be attributed to slight cooling of the air during its passage down the chamber when working well above room temperature. The linear velocity of the air current in the chamber was about 0·5 cm./sec., which is well below the limit—30 cm./sec.—of the Beaufort Scale number o (‘Calm’) (Meteorological Observer’s Handbook, 1926), so that a gradient closely similar to, or even steeper than, the lower curve must obtain under natural conditions. Both temperature and olfactory perception are generally accepted as being of importance in the location of the host by bloodsucking insects, and the available information has been summarized by Wiggles-worth (1939). As far as mosquitoes are concerned there seems to be no information relating to the radius of action of the olfactory or other perception which must be responsible for the location of a host from a distance. It is fairly clear that perception of the host by temperature effects is limited to a few centimetres. One may cite in this connexion the work of Howlett (1910) with several mosquito species and Rivnay (1931) with Cimex lectularius. Wigglesworth & Gillett (1934), working with Rhodnius prolixus, considered that the effect was due to air temperature and not to radiant heat. The present work on the extent of the temperature gradient close to the skin corroborates the conclusion that temperature sense is a short-range one which must be responsible only for the final orientation of the insect and the eliciting of a probing response. Bishop & Gilchrist (1946), working with Aëdes aegypti feeding through animal membranes, have adduced evidence that a temperature gradient steeply rising to the membrane was an important factor in attracting mosquitoes to feed. A higher rate of gorging was obtained when a gradient of from 24 or 28 to 42° C. was present than when both membrane and environment were at 37°C. Their work, as also the present work, was carried out with the mosquitoes confined artificially close to the blood source where temperature would be expected to be the predominant factor. That A. aegypti females are able to locate their host at short range in the absence of a temperature gradient rising to the skin is, however, clear from the present experiments, as a considerable proportion fed when the environment was warmer than the skin surface, and the highest proportion fed when environment and skin temperatures were close together. The discrepancy between the present results and those of Bishop & Gilchrist (1946) may perhaps be explained by the alcohol-washed membranes used by these workers lacking some of the olfactory characters normal to skin. Rivnay (1931) has recorded that mouse and rabbit skin and human stratum corneum lost their attractiveness to Cimex lectularius after washing with soap or with ether.

It was found in preliminary work that mosquitoes which had taken only small quantities of blood could easily be overlooked on macroscopic examination. All mosquitoes, therefore, were dissected, and the stomach and diverticula examined under the binocular dissecting microscope. It should be restated here that mosquitoes were excluded from the skin surface at the end of the feeding time by the interposition of a card, chloroformed for removal from the chamber, and dissected, usually immediately but sometimes as long as 90 min. subsequently. By visual examination the organs were placed in one of four classes; the results of this examination are shown in Table 3.

Of the total number of 590 mosquitoes, 200 showed no blood in any organ. Only 42, or 10·8%, of the remaining 390 mosquitoes which had taken blood showed it in any of the diverticula. It is apparent from the large preponderance of mosquitoes with blood in the stomach only that this is the usual destination of the whole blood meal in A. aegypti.

Marshall & Staley (1932) have summarized some of the conflicting results of previous workers regarding the function of the oesophageal diverticula in relation to blood meals, and have further carried out careful examination of this matter in British species of Anopheles, Aëdes, Theobaldia and Culex. After normal feeding they found small quantities of blood in the diverticula in 16, or 9·9%, of 161 mosquitoes. Bishop & Gilchrist (1946) examined the destination of the meal in considerable detail, varying both the character of the meal and the means by which it was offered, by open drops or through membranes. They concluded that the nature of the food rather than the method of feeding determined its destination, whether this was the stomach or diverticula of the mosquito. Only on very few occasions were blood constituents (haemoglobin in plasma or in water) found in the diverticula. The present work agrees, therefore, with the general conclusion of Marshall & Staley (1932) and of Bishop & Gilchrist (1946). Philip (1930) stated that interruption of feeding resulted in regurgitation of blood into the diverticula, and Marshall & Staley (1932) have further examined this point by chloroforming twenty-five mosquitoes in the act of biting. Four of these mosquitoes, dissected subsequently, had blood in the diverticula perceptible with low-power magnificatinn, and the total was raised to seventeen by high-power examination. The occurrence of blood in the diverticula in the present work has been examined in relation to the temperature and humidity at which the mosquitoes were fed and the percentage mortality and degree-hour aggregates of lots. No significant correlation was found. If, however, the mosquitoes are divided into two groups depending on whether they were fully, or only partly, distended with blood, and these groups are associated with the presence or absence of blood in the diverticula, the arrangement given by Table 4 results. X2 for this arrangement is 11·04 and with one degree of freedom, P<0·01; the difference is significant. The conclusion of Philip (1930) and of Marshall & Staley (1932) that interruption of feeding results in regurgitation of blood to the diverticula is therefore confirmed. It is remarkable, however, that in two cases (see Table 3) a large quantity of blood was found in the diverticula, while the stomach appeared quite free from blood under the binocular dissecting microscope. It seems improbable that regurgitation from the stomach could be so complete as to leave no trace under such examination, and it is probable that the blood in these instances passed direct to the diverticula. The rarity of such an occurrence is, probably, an index of its abnormality,

It is of value to summarize here the results of the present work and those of Seaton & Lumsden (1941) as far as they are important in deciding a standard technique for work in which it is desired to obtain regularly a high proportion of gorged A. aegypti after a single opportunity of feeding. The mosquitoes should be starved in the dark at about 24° C. and 80% relative humidity after hatching. They should be used between 72 and 120 hr. old. They should be applied to the host in a small container whose interior is in darkness and maintained at a temperature of 30–35°C. Precise control of the humidity of the air in the container seems unnecessary. An exposure time of 10 min. is ample.

I am deeply indebted to Prof. P. A. Buxton, F.R.S., and to Dr Kenneth Mellanby for their interest, help and encouragement and to Messrs S. A. Smith and D. Sheffield for the preparation of apparatus and maintenance of larval cultures.

Bishop
,
A.
&
Gilchrist
,
B. M.
(
1944
).
A method for collecting sporozoites of Plasmodium gallinaceum by feeding infected Aëdes aegypti through animal membranes
.
Nature, Lond
.,
153
,
713
.
Bishop
,
A.
&
Gilchrist
,
B. M.
(
1946
).
Experiments upon the feeding of Aëdes aegypti through animal membranes with a view to applying this method to the chemotherapy of malaria
.
Parasitology
,
37
,
86
.
Bourne
,
W. A.
(
1946
).
A resistance-type skin thermometer
.
Brit. Med. J
.
2
,
332
.
Buxton
,
P. A.
(
1931
).
The measurement and control of atmospheric humidity in relation to entomological problems
.
Bull. Ent. Res
.
22
,
431
.
Edwards
,
F. W.
(
1941
).
Mosquitoes of the Ethiopian Region. III
.
Culicine Adults and Pupae
.
London
:
British Museum
.
Howlett
,
F. M.
(
1910
).
The influence of temperature on the biting of mosquitoes
.
Parasitology
,
3
,
479
.
LANDOLT-BöRNSTEIN
(
1923
).
Physikalisch-Chemische Tabellen
, 5th ed. edited by
Roth
,
W. A.
&
Scheel
,
K.
Berlin
:
Julius Springer
.
Larsen
,
E. B.
(
1943
).
The importance of master factors for the activity of Noctuids. Studies on the activity of insects I
.
Ent. medd
.
23
,
353
.
Lewis
,
D. J.
(
1933
).
Observations on Aëdes aegypti L. (Dipt. Culic.) under controlled atmospheric conditions
.
Bull. Ent. Res
.
24
,
363
.
Lumsden
,
W. H. R.
&
Bertram
,
D. S.
(
1940
).
Observations on the biology of Plasmodium gallinaceum Brumpt, 1935, in the domestic fowl, with special reference to the production of gametocytes and their development in Aëdes aegypti (L.)
.
Ann. Trop. Med. Parasit
.
34
,
135
.
Marchoux
,
E.
,
Salimbeni
,
A.
&
Simond
,
P. L.
(
1903
).
La fièvre jaune
.
Ann. Inst. Pasteur
,
17
,
665
.
Marshall
,
J. F.
&
Staley
,
J.
(
1932
).
On the distribution of air in the oesophageal diverticula and intestine of mosquitoes. Its relation to emergence, feeding and hypopygial rotation
.
Parasitology
,
24
,
368
. Meteorological Observer’s Handbook (1926).
London
:
H.M.S.O
.
Philip
,
C. B.
(
1930
).
Possibility of mechanical transmission by insects in experimental yellow fever
.
Ann. Trop. Med. Parasit
.
24
,
493
.
Rivnay
,
E.
(
1931
).
Studies in tropisms of the bed bug Cimex lectularius L
.
Parasitology-
,
24
,
121
.
Seaton
,
D. R.
&
Lumsden
,
W. H. R.
(
1941
).
Observations on the effects of age, fertilization and light on biting by Aëdes aegypti (L.) in a controlled microclimate
.
Ann. Trop. Med. Parasit
.
35
,
23
.
Simpson
,
G. G.
&
Roe
,
A.
(
1939
).
Quantitative Zoology
.
New York
:
McGraw-Hill Book Co
.
Tate
,
P.
&
Vincent
,
M.
(
1935
).
The biology of autogenous and anautogenous races of Culex pipiens L. (Diptera-Culicidae)
.
Parasitology
,
28
,
115
.
Wigglesworth
,
V. B.
(
1939
).
The Principles of Insect Physiology
.
London
:
Methuen
.
Wigglesworth
,
V. B.
&
Gillett
,
J. D.
(
1934
).
The function of the antennae in Rhodnius prolixus (Hemiptera) and the mechanism of orientation to the host
.
J. Exp. Biol
,
11
,
120
.