THE MECHANISM OF FOOD DISTRIBUTION TO MIDGUT OR

DIVERTICULA IN THE MOSQUITO

By M. F. DAY\)

[Manuscript received July 2, ·19.54]

Summary
An examination has been made of factors involved in the despatch of
sugars to the diverticulum and blood to the midgut in the mosquito. It is
suggested that sense organs observed in the buccal cavity of the mosquito
detect the presence of some sugars and components of blood. The resulting
nervous impulse is probably transmitted via the stomatogastric nervous system
to cause the contraction of sphincter muscles of the diverticula or the proven-
triculus. The significance of this' switching mechanism in mosquito feeding
is discussed.
I. INTRODUCTION

A knowledge of how mosquitoes transmit virus diseases requires an under-

standing of their processes of ingestion and digestion. One of the peculiarities
of mosquito digestion that has been known for a long time is that, whereas whole
blood passes directly to the midgut, solutions of sugars generally pass to the
diverticula, rather than to the midgut. The problem of the distribution of
Hquids to these regions of the mosquito gut has recently been studied by
Trembley (1952). Her main conclusions, incorporating some of those of earlier
workers whom she quotes, may be summarized as follows:
(1) Blood, irrespective of the manner in which it is fed, goes primarily to
the midgut. In some species practically none reaches the diverticula,
in others a considerable amount of blood may go to them.
(2) Solutions of sugars (glucose or honey) go primarily to the diverticula,
mainly the ventral diverticulum. Traces may go to the midgut at once,
but normally the contents of the ventral diverticulum are later passed
to the midgut, a small amount at a time.
The establishment of these facts does not answer the question of how
various liquids reach the respective regions of the gut and, in fact, no com-
prehensive suggestion has been advanced that does provide an explanation of
what Fisk (1950) has called the "switching mechanism." Boissezon's (1930)
suggestion that the diameter of the ducts to the diverticula was not great
enough to permit the passage of erythrocytes does not account for all of the
observations, and Marshall and Staley's ( 1932) conclusions concerning the
action of sphincters at the openings of the diverticula offer no indication of
how these sphincters are activated. Fisk (1950) performed some experiments
to determine what properties of blood enabled it to reach the midgut, but
stated that his results did not solve the problem.
"Division of Entomology, C.S.I.R.O., Canberra, A.C.T.
516 M. F. DAY

Consideration of the conclusions of Trembley referred to above and of the

anatomical relationships of the gut of the Diptera suggested that sense organs
located in or near the buccal cavity may be stimulated by components of the
food, resulting in the transmission of motor impulses to the sphincters of the
diverticula or the proventriculus, or both.
Attempts to obtain data to evaluate this suggestion are reported in this
paper. The results permit the formulation of a hypothesis that appears to pro-
vide an explanation of the "switching mechanism."

II. MATERIALS AND METHODS

Of the species of mosquitoes examined by Trembley (1952), Aedes aegypti

( L.) appeared to exhibit the switching mechanism to a marked degree, and it
was therefore suitable for this work.
As the method of feeding does not influence the destination of liquids
<

ingested (Bishop and Gilchrist 1946; Trembley 1952), heparinized rabbit blood
and blood and sugar mixtures were fed to starved A. aegypti from cotton wool
or filter paper soaked in the fluid. Liquids other than blood were generally
coloured with carmine. Mosquitoes, kept overnight in small glass vials covered
with mosquito netting, could frequently be stimulated to ingest blood and sugar
mixtures by having them briefly probe the skin of the forearm and immediately
removing them to small dishes (containing the cotton wool and substances to
be tested) standing on a thermostatically controlled warm plate. Difficulty
was experienced in getting numbers of mosquitoes to accept some of the mix-
tures, particularly those lacking erythrocytes. Larger numbers of fed mosquitoes
of both sexes were obtained in several instances by collecting insects as they
fed from saturated filter papers in dishes containing the blood-sugar mixtures.
Dissections were performed in a drop of Drosophila Ringer.
Ten-micron serial sections of mosquitoes stained with Mallory's triple con-
nective stain or by Bodian's protargol method were used to study the sense
organs. The anatomy, innervation, and musculature of the gut were studied
also in whole mounts of the gut after staining with methylene blue by the method
of Kuwana (1935). Phase-contrast microscopy, using a Leitz Heine condenser,
was used to study the crop musculature and innervation.

III. FEEDING EXPERIMENTS

Although several earlier authors have reported the destination of mixtures

of blood and sugars in the mosquito gut, the data are inadequate for the present
purposes. Experiments were therefore performed to determine what com-
ponents of the blood or blood: sugar mixtures were stimulatory and what were
the approximate thresholds for stimulation. The results (Table 1) demonstrate
that, in agreement with the work of earlier authors, blood reaches the midgut
whereas glucose solutions go mainly to the diverticula. '0 The addition of a
.. Frequently liquids were found only in the ventral diverticulum, whereas the dorsal
diverticula contained gas bubbles. Only the ventral diverticulum was considered in the
results presented.
 FOOD DISTRIBUTION IN THE MOSQUITO ,517

small amount of blood to 10 per cent. glucose solution results in the mixture
going partly to the midgut and partly to the diverticulum. Thus 1 part of blood
to 20 of glucose was sufficient to be detected and to activate the switching
mechanism. Over a wide range of concentrations of blood in mixtures of
glucose solutions the mixture went both to the midgut and to the diverticulum.

TABLE 1
DESTINATION OF MIXTURES OF 10 PER CENT, GLUCOSE AND WHOLE BLOOD WHEN INGESTED BY
AEDES AECYPTi

I Destination (%)
Molar Number -
Composition of Liquid Concn, of Engorged
Glucose Diverticulum Midgut
I I
Blood - 50 6* 100
Blood 3 ml; glucose 1 ml 0·08 25 16 92
Blood 2 ml; glucose I ml 0·13 20 40 70
Blood 1 ml; glucose 1 ml 0·25 30 73 80
Blood 0·5 ml; glucose I ml 0·30 30 83 86
Blood 0·25 ml; glucose I ml - 20 80 90
Blood 0·1 ml; glucose I ml 0·46 30 100 7
Blood 0·05 ml; glucose I ml 0·50 25 100 8
Glucose I ml - 50 100 10

* Trace only. (Note that particularly at intermediate concentrations the ingested liquid frequently
goes to both midgut and diverticulum in the same insect.)

If glucose is added to heparinized blood in sufficient concentration, the

mixture again goes both to the midgut and diverticulum. The conclusion is
inescapable that both whole blood and glucose are detected, and the relative
concentration of the components of a mixture determines its destination. These
results suggested the search for separate receptors to detect glucose and blood
(see Section IV below).
The experiment was repeated replacing glucose by other sugars. Equal
parts of 5 per cent. sucrose and heparinized blood went mainly to the diverti-
culum, indicating that sucrose was detected (Table 2). On the contrary, a
mixture of 2 mIlO per cent. lactose in 1 ml heparinized blood went entirely
to the midgut, suggesting that lactose was unable to stimulate the receptors.
Insufficient numbers of mosquitoes for a study of thresholds could be induced
to accept blood mixed with arabinose, mannose, and raffinose, but all three
could apparently be detected, for when these sugars were mixed with blood
at a concentration of 0,2 glml the mixture went both to the diverticulum and
to the midgut.
Experiments were next designed to determine what components of whole
blood were responsible for the switching mechanism. Similar experiments
performed by Fisk (1950) did not permit conclusions to bB drawn from them,
518 M. F. DAY

but Bishop and Gilchrist (1946) found that plasma, washed erythrocytes, and
haemolysed blood when ingested went to the midgut and not to the diverti-
culum. These data and the results presented in Table 2 suggest that both
plasma and erythrocytes are detected but the reaction to plasma is less marked
than that to washed erythrocytes. Solutions of haemoglobin and albumin with
glucose went mainly to the diverticulum, and were thus presumably not de-
tected; however, these substances were ingested by too few mosquitoes for
the results to be unequivocal.

TABLE 2
DESTINATION OF MIXTURES OF SUGARS AND BLOOD AND BLOOD CONSTITUENTS WHEN INGESTED BY
AEDES AErjrPTI

Destination (%)

Composition of Liquid Number Fed

Diverticulum Midgut

Blood 5·0 ml; 3·5% sucrose 1·0 ml 20 0 100

Blood 0·3-0·2 ml; 3·5% sucrose 1·0 ml 20 100 100
Blood 1·0 ml; 5% sucrose 1·0 ml 20 100 10
Erythrocytes* 0·3 ml; 5% glucose 1·0 ml 20 501' 100
Erythrocytes* 0·1 ml; 5% glucose 1·0 ml 20 100 20t
Plasma§ 1·0 ml; 5% glucose 1·0 ml 20 100 45
Plasma§ I·d ml; 5% sucrose 0·5 ml 20 80 30
Plasma 20 100 20

* Prepared by twice centrifuging heparinized rabbit blood and resuspending the sedimented cells
in physiological saline to the original volume.
t In three of the 10 insects of the group, solution, but no erythrocytes were present in the
diverticulum.
:\: Erythrocytes present in midgut only, none in diverticulum.
§ Supernatant from first centrifugation of * above.

vVhen mixtures of washed erythrocytes and glucose were fed to mosquitoes,

the mixtures went to the midgut, the diverticulum, or to both, depending upon
the concentrations of the components. When the concentration of washed
erythrocytes was low, however, a number of mosquitoes were found to have
diverticula filled with a clear liquid, whereas the midguts contained liquid to-
gether with a few erythrocytes. The straining mechanism mentioned by Bois-
sezon (1930) appears to be effective in these instances. The occurrence of
erythrocytes in the diverticulum when blood: sugar mixtures in other concen-
trations are fed demonstrates, however, that this straining apparatus is not
normally able to control the switching mechanism.
A group of spines was observed in the neck of the ventral diverticulum
(Plate 1, Fig. 5). These do not seem to have been noticed by previous workers;
they differ in shape and arrangement from those described in the pyloric region
of the same species by Trembley (1951). Their presence would assist in keep-
ing erythrocytes out of the diverticulum when the circular muscles at the neck
 FOOD DISTRIBUTION IN THE MOSQUITO 519

were contracted, and they probably function in straining out the erythrocytes
as described in the previous paragraph.
The response to washed erythrocytes suggested that the stimulus might be
tactile rather than gustatory. An attempt to examine this possibility was made
by feeding mosquitoes on washed red cell ghosts. Rabbit erythrocytes were
haemolysed in distilled water, centrifuged, resuspended in distilled water, and
centrifuged again. The resulting ghosts, pinkish in colour, were resuspended
in a minimum of water and fed to 7-day-old, previously unfed mosquitoes. Five
ingested these ghosts and in all the suspension went mainly to the midgut. Thus
washed red cell membranes, though less readily accepted than erythrocytes,
were able to activate the switching mechanism.
It was possible that residual haemoglobin may have been responsible for
causing the membranes to be despatched to the midgut; it was therefore con-
sidered desirable to determine the destination of other particulate material
after ingestion by the mosquito. Several such materials (water-insoluble dyes,
starch grains) were not ingested. Finally, starved mosquitoes did ingest washed
sarcosomes separated from the thoracic muscles of the blowfly Lucilia cuprina
(Wied.). When suspended in distilled water, these approximated the diameter
of mammalian erythrocytes. In 20 A. aegypti females that ingested this suspen-
sion, the material was found both in the midgut and the diverticulum. The
supernatant from a centrifuged preparation went only to the diverticulum. It
seems likely,therefore, that it was the particulate nature of the sarcosomes
that caused some of the suspension to be despatched to the midgut.
In some experiments males were used in addition to females. The two
sexes reacted similarly; blood went to the midgut and sugar to the diverticulum
in the male in spite of the fact that males do not ingest blood under natural
conditions.
In some early experiments it seemed possible that th~ age of the mos-
quitoes or their previous feeding were responsible for yariations in their
behaviour with respect to the destination of blood : sugar mixtures. A series
of tests was therefore set up to determine whether these factors were contri-
buting to the variability of behaviour. In one test comparisons were made
between 2-day-old females previously unfed and 7-day-old females that had
had access to sugar water for the first 6 days after eme~gence. There was no
difference in the destination of blood : glucose mixtures in the two groups. In
a second test no difference in the destination of blood : glucose mixtures was
detected in a comparison between females that had had no previous feed and
females that had had a single blood meal 7 days previously. Variability in
destination occurred equally in all groups, and no method was developed which
resulted in completely uniform behaviour of any groups of insects towards
blood: glucose mixtures. There was much more marked uniformity, however,
in response to sugar or to whole blood offered separately.

IV. ANATOMICAL CONSIDERATIONS

The feeding experiments reported in Section III suggested that two types
cf sense organs might be necessary to detect both sugars and the components
520 M. F. DAY

of blood. Although pharyngeal sense organs have been recorded in several

insects (e.g. in larval Drosophila (Hertweck 1931), Glossina (Jobling 1933), and
in Panorpa (Grell 1938», no record of their occurrence in mosquitoes could
be found. Examination of the stylets showed the absence of appropriate re-
ceptors in these chitinous structures (see also Robinson 1939). Serial sections
of the head of Aedes aegypti revealed the presence of the dorsal and palatal
papillar sense organs in the buccal cavity. These were first described by Sinton
and Covell (1927), and by Barrand and Covell (1927), who showed that their
form and arrangement had considerable taxonomic value. The location and
arrangement of these sense organs in Aedes aegypti is shown in Figures 1 and
2. Sections showed that the sense organs were of four types, as illustrated in
Plate 1, Figures 1-4; the three pairs of palatal organs were spines, whereas the
three pairs of dorsal organs were hair-like receptors and there was a single pair
of campaniform papillae. Similar structures were found in male A. aegypti.
The cibarial sense organs described in adult Drosophila by Hertweck (1931)
and Miller (1950) are probably homologous with these papillar sense organs
of mosquitoes.

Fig. I.-Diagrammatic median longitudinal section of anterior part of a

mosquito, showing foregut and midgut and the stomodeal nervous system.

The precise function of the papillar sense organs cannot, of course, be cer-
tainly adduced from their structure; but the form of the palatal papillar organs
is not inconsistent with the suggestion that they are chemoreceptors; the struc-
ture and position of the dorsal papillar receptors suggest that they may be
capable of detecting the flow of erythrocytes and of the sarcosomes mentioned
in the previous section. One of the four types of sense organ is apparently
able to detect the presence of blood plasma.
 FOOD DISTRIBUTION IN THE MOSQUITO 521

The innervation of the seven pairs of sense organs could not be traced
with certainty, but all received nerves from the frontal ganglion. (Miller (1950)
believed also that the pharyngeal nerves that innervate the cibarial sense organ
of Drosophila are homologous with the frontal ganglion connectives, and Wenk
(1953) has shown that the labral nerves of the flea Ctenocephalus innervate
cibarial sense organs.) It therefore seems likely that nerve impulses from the
mosquito sense organs reach the stomatogastric nervous system. In all insects
the stomatogastric nervous system innervates the foregut as well as the incre-
tory organs of the head. This has been well shown in Calliphora (Graham-
Smith 1934), and a similar arrangement has been revealed in favourable methy-
lene blue preparations of Aedes aegypti. Orlov (1924) has shown that the
frontal ganglion, of Oryctes larvae at least, contains sensory as well as motor
nerve cells. The origin of these cells has never been elucidated. The experi-
ments of Bolwig (1952) suggest that the frontal ganglion is not concerned with
peristaltic movements in the fly gut and its function is not completely understood.

ANTERIOR

Fig. 2.-Diagram of number and arrangement of sense

organs in the dorsal wall of the buccal cavity of Aedes
aegypti. Palatal papillar sense organs ( PP); campani-
form sensillum (CS); dorsal papillar sense organ (DP),
and hair-like sensilla (HS). The parallel lines indicate
the plane of the section of Plate 1, Figure 4.

The innervation of the diverticula of Aedes is shown semi-diagrammatically

in Figure 1. The arrangement of the musculature in an expanded ventral diver-
ticulum and the relative size of the nerve and its method of branching are
shown in Plate 1, Figure 6.
522 M. F. DAY

The origin of the nerves innervating the muscles of the diverticula is still
not definitely known, but it seems most likely that they form part of the stomato-
gastric nervous system. It is not essential for the present purposes to know the
precise neural pathway between the sensory receptors and the effector system.
It is entirely possible that some central connections exist. However, the path-
way through the frontal ganglion and the stomatogastric nervous system appears
possible and is the simplest way of accomplishing the connections.

V. DISCUSSION
The following hypothesis to explain the switching mechanism in the mos-
quito is suggested by the data presented above. Certain sugars stimulate the
pit organs in the buccal cavity of the mosquito, causing impulses that result
in the relaxation of the sphincters of the diverticula. Certain components of
the blood of vertebrates stimulate the papillar sense organs, resulting in relaxa-
tion of the cardiac sphincters. Simultaneous stimulation of both groups of
sense organs may result in relaxation of both series of sphincters. Both male
and female mosquitoes react in the same manner to these ingested materials,
even though the males never obtain a blood meal under natural conditions.
The thresholds of the pit organs to sucrose and glucose are somewhat high
in comparison with tarsal and palpal receptors (see Dethier 1953) but these
receptors would respond to the concentration of sugars normally found in
nectars.
The function of the diverticula as containers of ingested sugar solutions
does not, of course, preclude them from having additional functions. In some
of the present experiments they were found to function also as "air separators"
as previously shown by MacGregor (1930).
The question may be raised of the survival value of the switching mechan-
ism in the mosquito and of the absence of such a mechanism in, for example,
the blowfly, in which the necessary anatomical requirements would appear to
be well developed. It is suggested that the answer is to be found in the differ-
ences in the feeding behaviour of the two Diptera. In the majority of species
female mosquitoes require a blood meal in order to mature their eggs. The
opportunity to bite a suitable vertebrate to provide the blood meal will come
at irregular intervals and the insect must be capable of utilizing the opportunity
when it is presented. Nectar is more generally available and is sufficient to
keep the insect alive. But the ability to take a blood feed in spite of a recent
nectar meal would seem to be of survival value, and this is what is accom-
plished by the "switching mechanism"; a crop containing some nectar does not
preclude the taking of a blood meal should this become available.
Trembley (1952) has shown that the switching mechanism is not as well
developed in some species of mosquitoes as it is in A. aegypti. It may be that
the species of Anopheles and Culex that she studied, in which erythrocytes
were often found in the diverticula as well as the midgut, have either less
sensitive receptors in the buccal cavity, or a less well-developed sphincter
musculature in the necks of the diverticula.
DAY PLATE 1
FOOD DISTRIBUTION IN THE MOSQUITO

Aust. J. Bioi. Sci., Vol. 7, No.4

 FOOD DISTRIBUTION IN THE MOSQUITO 523

VI. ACKNOWLEDGMENTS
Thanks are due to Mrs. M. J. Bennetts, who performed many of the earlier
experiments, and to Mrs. L. Abbot for assistanee in the later stages of the work.
The colony of mosquitoes used originated from Rockhampton, Qld., and was
obtained through the kindness of Mr. P. R. Wilkinson, of the Division of Ento-
mology. Figures 1 and 2 are the work of Mr. L. A. Marshall, and the photo-
micrographs were taken with a Leitz "Panphot" by Mr. D. Wilson.

VII. REFERENCES
BARRAND, P. J., and COVELL, G. (1927)~-The morphology of the buccal cavity in anopheline
and culicine mosquitoes. Indian J. Med. Res. 15: 671-80. .
BISHOP, A., and Gn.CHRIST, B.M. (1946).-Experiments upon the feeding of Aedes aegypU
through animal membranes with a view to applying this method to the chemo-
therapy of malaria. Parasitology 37: 85-100.
BOISSEZON, P. DE (1930).-Contribution a l'etude de la biologie et de l'histophysiologie de
Culex plpiens L. Arch. Zool. Exp. Gen. 70: 281-481.
BOLWIG, N. (1952).-Hunger-reaction of Hies (Musca) and the function of their stomato-
gastric system. Nature 169: 197.
DETHIER, V. G. (1953) .-Chemoreceptiori in "Insect Physiology." Ed. K. D. Roeder. (John
Wiley & Sons Inc.: New York.)
FISK, F. W. (1950) .-Studies on proteolytic digestion in adult Aedes aegypti mosquitoes.
Ann. Ent. Soc. Amer. 43: 555-72.
GRAHAM-S:MlTH, G. S. (1984).-The alimentary canal of Calliphora erythrocephala L. with
special reference to its musculature and to the proventriculus, rectal valve and rectal
papillae. Parasitology 26: 176-248.
GRELL, K. G. (1988) .-Der Darmtraktus von Panorpa communis L. und seine Anhiinge bei
Larve und Imago. Zool. lb. (Anat.) 64: 1-86.
HERTWECK, H. (1981).-Anatomie und Variabilitat des Nervensystems und der Sinnesorgane
von Drosophila melanogaster (Meigen). Z. wiss. Zool. 139: 559-668.
JOBLING, B. (1938).-A revision of the structure of the head, mouthparts and salivary glands
of Glossina palpalis Rob. Desv. Parasitology 24: 449-90.
KUWANA, Z. (1985).-The innervation of the alimentary canal of the silkworm larva. Annot.
Zool. lap. Tokyo 15: 247-60.
MACGREGOR, M. E. (1980).-The artificial feeding of mosquitoes by a new method which
demonstrates certain functions of the diverticula. Trans. R. Soc. Trop. Med. Hyg.
23: 829-81.
MARSHALL, J. F., and STALEY, J. (1982).-On the distribution of air in the oesophageal
diverticula and intestine of mosquitoes. Its relation to emergence, feeding, and
hypopygial rotation. Parasitology 24: 868;81.
Mn.LER, A. (1950).-The internal anatomy and histology of the imago of Drosophila melano-
gaster, in "Biology of Drosophila." Ed. M. Demerec. (John Wiley & Sons: New
York.)
ORLOV, J. (1924).-Uber den histologischen Bau der Ganglion der Mandmagennervensys-
terns der Insekten. Z. mikr.-anat. Forsch. A 2: 89-110.
ROBINSON, G. G. (1989).-The mouthparts and their function in the female mosquito Ano-
pheles maculipen.nis. Parasitology 31: 212-42.
SINTON, J. A., and COVELL, G. (1927).-The relation of the morphology of the buccal cavity
to the classification of anopheline mosquitoes. Indian J. Med. Res. 15: 301-8.
'TREMBLEY, HELEN L. (1951).-Pyloric spines in mosquitoes. J. Nat. Malar. Soc. 10: 213-15.
TREMBLEY, HELEN L. (1952).-The distribution of certain liquids in the esophageal diverticula
and stomach of mosquitoes. Amer. I. Trop. Med. Hyg. 1 (4): 693-710.
WENK, P. (1983).-Der Kopf von Ctenocephalus canis (Curt.) (Aphaniptera). Zool. lb.
(Anat.) 73 (1): 108-64.
524 M. F. DAY

EXPLANATION OF PLATE 1
Photomicrographs of Aedes aegypti.
Fig. I.-Longitudinal section of dorsal wall of buccal cavity showing structure of the palatal
papillar sense organ projecting into the lumen. The spine is directed posteriorly.
Fig. 2.-The same, showing one of the six hair-like sensillae.
Fig. 3.-The same, showing structure of a dorsal papillar sense organ.
Fig. 4.-Longitudinal section of the buccal cavity, showing two palatal papillar sense organs
(PP), a campaniform sensillum (CS), and a dorsal papillar sense organ (DP) •
. The salivary pump (SP) is also shown. The anterior end of the insect is to the
right of the illustration.
Fig. 5.-Whole mount of neck of ventral diverticulum, showing spines. These vary in length
from about 15 to 30 14. Phase contrast.
Fig. 6.-Whole mount of expanded crop showing arrangement of circular muscles (CM)
and irregular network of fine muscle fibres (MF) connecting them and the crop
nerve (CN), and its method of branching. The dark oval bodies are nuclei of
the epithelium. Phase contrast.