Osmotic Regulation in Mosquito Larvae

The ability of various mosquito larvae to regulate the osmotic pressure of the haemolymph in the face of changes in the external medium has been clearly established by the work of Wigglesworth (1938) and of Beadle (1939). Particular interest attaches to two species of the genus Aedes: A. aegypti L., which is a fresh-water form incapable of living in sea water, and A. detritus Edw., which is normally found in saline, often in highly saline, waters.

Wigglesworth has shown that in A. aegypti the osmotic pressure of the haemolymph is equivalent to 0·80–0·89% NaCl with an average chloride content equivalent to 0-30% NaCl. More than half of the total osmotic pressure is thus exerted by non-chloride, presumably organic, solutes. When the larvae are placed in artificial sea water of various dilutions it is found that both the osmotic pressure and the chloride content of the haemolymph remain constant until the concentration of the external medium reaches 0·65–0·75% NaCl. In more concentrated media the chloride content of the haemolymph increases and the total osmotic pressure is thereby maintained slightly in excess of that of the medium up to an external concentration of 1·6% NaCl, at which the larvae die rapidly. In earlier papers Wigglesworth (1933 a-c) showed that whereas the general body surface was relatively impermeable to water and salts, these substances penetrated readily through the anal gills; and Koch (1938) demonstrated that the anal gills were responsible for the active uptake of chloride from very dilute external media.

The anatomy of the larva of A. aegypti is shown in Text-fig. 1, which is taken from Wigglesworth (19336). The mechanism of salt and water balance, as indicated by the work of Wigglesworth and of Koch, appears to be as follows. The larvae do not normally swallow the medium, except in so far as this is incidental to feeding. Water enters passively through the anal gills and salts are actively absorbed through these organs against the concentration gradient. Fluid—of unknown composition—is excreted by the Malpighian tubules and accumulates in the pyloric chamber; at intervals of a few minutes a drop of fluid passes down the intestine from the pyloric chamber to the rectum. In the rectum some of this fluid is resorbed and the rest is eliminated through the anus. The active uptake of chloride must involve osmotic work; but whether or not the activities of the Malpighian tubules and rectum involve osmotic work is unknown.

The larva of A. detritus shows essentially the same structure as that of A. aegypti except that the anal gills are reduced to small papillae. The osmotic relations are very different. In fresh water the total osmotic pressure and chloride content are of the same order as in A. aegypti, but in various dilutions and concentrations of sea water, up to the equivalent of 6% NaCl, the total osmotic pressure and chloride rise steadily to values of 1·4 and 0·9% NaCl respectively (Beadle, 1939). The body surface is almost completely impermeable to water and salts, exchange taking place through the gut. The gut wall appears to be freely permeable to water and to glycerol but not to sucrose; but whether the salt water of the external medium is passively absorbed or whether there is differential absorption with the performance of osmotic work is not known. Beadle also showed that after ligatures had been applied between segments v and vi there was some increase in osmotic pressure and chloride content, which suggests that organs in the posterior part of the animal play some part in the regulatory process.

The ability of these larvae to regulate the composition of the haemolymph is therefore fairly clearly defined and a good deal is known of the intake side of the mechanism. But on the side of elimination we have no quantitative data at all. The work to be described in the present paper was undertaken with the object of making good this deficiency.

Larvae of A. aegypti were bred from eggs and fed on a preparation of powdered dog-biscuit. In the third instar they were removed to clean media and thereafter starved. ‘Distilled water larvae ‘and ‘frog-ringer larvae ‘were reared in this way in the respective media. ‘Double frog-ringer larvae ‘were reared with food in normal frog ringer and were later transferred to clean frog ringer of double strength. Cultures of ‘frog-ringer larvae’ survived well for 6–8 weeks; in the same period 50% of the ‘distilled water larvae’ died (the distilled water was continually renewed); 50% of the ‘double frog-ringer larvae’ died in 3–4 days.

Larvae of A. detritus were collected from a salt marsh near Walton-on-the-Naze, Essex. Third and fourth instar larvae were kept in the laboratory in clean sea water and survived unfed for 6 weeks. All the work on A. detritus described in this paper was carried out on this single collection. Unfortunately, when a second collection was called for it was found that prolonged drought and an effective sea wall had temporarily eliminated the breeding pools. The work on this species was thus unavoidably curtailed.

Collections were made of the haemolymph and of the fluids from different regions of the gut. Very often only relatively small quantities, of the order of 0·01 cu.mm., were available. Such quantities were more than adequate for freezing-point measurements to be made by a method previously described (Ramsay, 1949), but it was not possible to determine chloride. The results here reported are restricted to measurements of total osmotic pressure, expressed in terms of the percentage concentration of an NaCl solution having the same freezing-point.

The regions of the gut studied were the rectum, the intestine, the midgut and the caeca; and a variety of methods of collection were adopted. In all cases it was necessary to restrain the movements of the larva, but at the same time it was desirable to avoid unnecessary interference with the workings of the viscera. For this reason anaesthetics were not used. The larva was secured by fine silk ligatures, one around the base of the respiratory siphon and the other around the neck ; this second ligature, if desired, could be left sufficiently loose to permit swallowing. The larva was placed in a shallow layer of well-aerated medium and, although the tracheal system was occluded, activity was maintained up to 36 hr. or more. In well-aerated water sufficient gas exchange takes place through the general body surface to maintain life and activity. This was shown by Macfie (1917) for A. aegypti, and Beadle has shown that after ligation of the respiratory siphon in A. detritus the salt and water balance is undisturbed.

• Rectal fluid (R.F.). This was collected by inserting a cannula through the anal canal with the aid of a manipulator; in this way the fluid momentarily present in the rectum could be drawn up. By adding trypan blue to the external medium just before collection was made it could be shown that the external medium did not enter the cannula. Larger quantities could be collected over a longer period by tying a flared cannula* into the anal canal.

• Intestinal fluid (I.F.). It is possible to insert a moderately fine (100/x diameter) cannula through the anus and rectum into the intestine. The circular muscle of the intestine is normally contracted and prevents rectal fluid from reaching the cannula. When a droplet of fluid is carried down by peristalsis from the pyloric chamber nearly all of it passes into the cannula.

• Midgut fluid (M.F.). Owing to the loop in the intestine it is difficult to pass a cannula into the midgut from the anus. It is easier to reach the midgut by passing the cannula through the mouth, but the trick of doing this without injuring the larva was not discovered until a late stage in this work. In all the experiments of which the results are reported, collection from the midgut was made by a method involving dissection of the larva as described under (d) below.

• Caecal fluid (C.F.). It has not so far proved possible to insert a cannula into the caeca by way of any of the natural openings of the body, and it was therefore necessary to dissect out the gut and penetrate its wall. The larva was dried on cigarette paper and placed on a dry slide under a layer of liquid paraffin. A ligature was tied around the body between thorax and abdomen. The body wall was punctured and the haemolymph flowed out on the slide under the oil (at this stage the collection of the haemolymph was made). The tissues of the thorax were dissected away from the anterior region of the gut which was then impaled upon a fine silica ‘freezing pipette’ (Ramsay, 1949), connected with a screw plunger and filled with oil backed with mercury. When possible the point of the ‘freezing pipette’ was worked into the extremity of one of the caeca and a sample of fluid was drawn up. In other cases where this latter part of the operation did not succeed, the sample was taken from the lumen of the midgut in the region of the caeca. To collect from the extremity of a caecum is perhaps an unnecessary refinement in view of the periodic contractions of the caeca which cause exchange of fluid between the caeca and the neighbouring region of the midgut. The abdomen of the larva was then dissected, the midgut was impaled upon a second ‘freezing pipette’ and the sample was drawn up.

• Haemolymph (B.F.). This was collected directly into a ‘freezing pipette’ at the end of each experiment by drying the larva and puncturing it on a slide under oil, as described under (d) above.

It was possible to make all five collections from one and the same larva, but most experiments were conducted on a less ambitious basis. In every case a collection of haemolymph was made and this served as the standard of reference with which the other fluids were compared.

For purposes of histological examination material was fixed in Camoy. Paraffin sections were cut at 10 p and stained in iron haematoxylin.

The types of larvae studied may be placed in categories as follows: (1) Aedes aegypti, distilled water larvae; (2) A. aegypti, frog-ringer larvae; (3) A. aegypti, double frog-ringer larvae; (4) A. detritus, sea-water larvae; (5) A. detritus, distilled water larvae.

The larvae were kept for at least 48 hr. in the specified external medium before being used for experiment.

As described in the previous section, collections were made from the rectum, the intestine, the midgut and the caeca. At least six collections (each from a different larva) were made from each position within each of the first four categories. A sample of haemolymph was collected from each larva. Results of the freezing-point determinations made on these collections are given in Tables 1–4. The differences in osmotic pressure between haemolymph and gut fluid have been subjected to the ‘t’ test for significance, and the corresponding values of ‘P’, taken from Fisher & Yates (1938), are given. P<0·01 is considered to be significant and P<0·01 is considered to be highly significant. Owing to the shortage of supply of A. detritus (for reasons already given) results of measurements in the fifth category, detailed in Table 5, amount to only two collections from each region and are insufficient for statistical treatment.

The data of Tables 1–5 are also shown graphically in Text-fig. 2.

The most important points which emerge are the following : (a) the caeca fluid is always hypertonic (to the haemolymph); (b) the midgut fluid is either isotonic or hypertonic; (c) the intestinal fluid is isotonic, except in the case of A. detritus in sea water, where it is hypertonic; (d) the rectal fluid is hypotonic in both A. aegypti and A. detritus when the larvae are in distilled water. In A. aegypti, when the larvae are in frog ringer or double frog ringer, the rectal fluid is still slightly hypotonic. In A. detritus, when the larvae are in sea water, it is strongly hypertonic.

From these results certain main conclusions may be drawn at once.

1. In A. aegypti, in fresh water, osmotic work is carried out in the rectum with the effective dilution of the fluid which enters from the intestine.

2. In A. detritus, in sea water, osmotic work is carried out in the rectum with the effective concentration of the fluid which enters from the intestine.

3. In A. aegypti there is no evidence that the rectum can effect concentration of the fluid which enters from the intestine.

4. In A. detritus there is evidence, from two examples only, that the rectum can effect dilution of the fluid which enters from the intestine.

These main conclusions, and others of a more tentative nature, are further discussed in section IV.

In view of the conclusions reached in the preceding subsection about the functions of the rectum, it is obviously of interest to know whether there are any differences in the structure of the rectum in A. detritus as compared with A. aegypti, to which the observed differences in function may be related.

The histology of the rectal epithelium of Aedes larvae does not appear to have been studied in detail. Wigglesworth (1933 b)—see also Text-fig. 1—gives a small-scale cross-section and refers to the ‘relatively large epithelial cells’. The investigation reported here is admittedly incomplete, but seems to indicate a possible correlation between histological structure and physiological function.

The rectal epithelium of A. aegypti is fairly deep and folded, with large nuclei but without distinguishable cell boundaries. The cytoplasm shows evidence of striation and appears to be more dense along the margin adjoining the lumen of the rectum. This dense border appears to be continuous and interposes itself between the nuclei and the lumen. Very often the nuclei appear as it were to be forced out of the striated cytoplasm altogether and to lie within its folds ; that is to say, the nuclei lie on the margin of the epithelium which adjoins the body cavity—see Pl. 1 A. Six specimens of A. aegypti have been examined in serial sections, and in all of them the epithelium appears to be uniformly of this type throughout the rectum.

A. detritus is a larger species than A. aegypti, and in addition the rectum of A. detritus appears relatively somewhat larger and better supplied with tracheae. As displayed in dissection there is nothing to suggest any division of the rectum into two regions. But when sections are examined it is clear that two kinds of epithelium are present. The posterior portion, extending from the anal canal to about the middle of the rectum, is lined with an epithelium which, although deeper than the epithelium of A. aegypti, appears to have the same character, i.e. a layer of dense cytoplasm on the side next to the lumen and the nuclei on the side next to the body cavity. The anterior portion, extending from the middle of the rectum to the intestine, is lined with an epithelium which is likewise deep and folded, and shows evidence of striation; but on the whole its cytoplasm is less dense, and if anything the density is greater on the side away from the lumen. The nuclei may be found in any position : on the side of the lumen, away from the lumen or in the middle. Further, the transition between these two types of epithelium is quite sharp ; the anterior portion is overlapped and enclosed by the posterior portion for a short distance over which the anterior portion thins rapidly and projects inwards in the form of a small flap—see Pl. 1B. Four specimens of A. detritus have been sectioned and all show the two types of epithelium as described above.

The marked difference in size and thickness of the epithelium in the two species makes it difficult to claim with certainty that the epithelium of A. aegypti belongs to one or other of the two types of epithelium in A. detritus. It is clear, however, that the epithelium of A. aegypti bears much more resemblance to the posterior than to the anterior epithelium of A. detritus, and the interpretation placed upon these results is that A. detritus possesses in the anterior portion of the rectum a special type of epithelium which is not represented in A. aegypti.

The conception of the mechanism of salt and water balance in A. aegypti, as indicated by the work of Wigglesworth and of Koch, has been outlined in section I. We are now in a position to add that the fluid passing down the intestine is isotonic with the haemolymph, and that the rectal epithelium carries out osmotic work in rendering this fluid hypotonic before it is eliminated. Whether this is brought about by resorption of salts or by secretion of water is not yet demonstrated ; but Wiggles-worth’s (1933b) observations suggesting a resorption of fluid in the rectum do not support the second alternative. Attempts made during the course of the present work to demonstrate changes of volume in the rectum were inconclusive.

Mainly on the basis of experiments in which larvae were fed on solid trypan blue, Wigglesworth suggested that fluid is secreted into the lumen of the gut by the cells of the posterior midgut region, and that this fluid passes forward to the caeca where it is resorbed. These experiments have been repeated, using trypan blue and also phenol red which is even more instructive.

After a larva has been allowed to swallow a saturated solution of phenol red the gut has a characteristic appearance. The caeca appear red (pH 7·6–8·0), the midgut is crimson (pH > 8·4), the region of the pyloric chamber is orange-red (pH 7·4–7·8) and the rectum is yellow (pH < 7·2).* When a droplet passes into the rectum from the intestine the colour change is complete in about a minute.

About a day later most of the dye has disappeared from the gut; what remains is to be seen in the caeca and extending some little distance down the midgut, a distribution which seems to indicate a general forward movement of the fluid in the midgut.

The orange-red fluid of the pyloric chamber is presumably midgut fluid mixed with fluid excreted by the Malpighian tubules. The excretion by the Malpighian tubules could of itself make available a supply of fluid to support the forward movement in the midgut, but the change of colour in the middle region of the midgut indicates some activity on the part of the cells of this region. The osmotic pressure measurements show that the midgut contents are somewhat hypertonic to the haemolymph. It is possible that osmotically active substances are secreted into the lumen and that water follows passively. This implies that the gut wall is readily permeable to water, and a simple experiment demonstrates that this is, in fact, the case. A larva is placed in a 2M-sucrose solution saturated with phenol red. The larva swallows this medium, and after a few minutes the dye is seen in the anterior part of the midgut. The neck and anal canal are then ligatured to prevent any further exchanges with the external medium. In 20 min. the gut is seen to have become enormously distended, almost filling the body, and it is obvious that the fluid in the gut has increased at the expense of the haemolymph.

The evidence that the fluid of the midgut is resorbed in the caeca, combined with the evidence of higher osmotic pressure in the caeca, must be taken to indicate that the cells of the caeca are capable of osmotic work. But it does not appear that this internal circulation of water from haemolymph → midgut → caeca → haemolymph plays any direct part in the osmotic relations between the larva and the external medium.

Since in A. aegypti the fluid passing down the intestine is isotonic with the haemolymph, it does not appear likely that in this species the Malpighian tubules have more than a passive role in osmotic (as distinct from ionic) regulation. In A. detritus, on the other hand, the fluid collected from the intestine is definitely hypertonic to the haemolymph and to the midgut fluid. Although this observation suggests osmotic activity on the part of the Malpighian tubules it is to be accepted with caution for the following reason. It was observed in A. detritus that when a small droplet passing down the intestine reached the rectum there was sometimes an incipient retro-peristalsis, the posterior portion of the intestine becoming momentarily distended with fluid from the rectum. (Such retro-peristalsis was never observed in A. aegypti unless the anus was obstructed and an excess of fluid had accumulated in the rectum.) Although the cannula was thrust as far up the intestine as possible, it may have happened that some rectal fluid found its way into the cannula in this way. Beadle was inclined to attribute an active excretion of salts to the Malpighian tubules on the basis of his experiments with ligatures, but since the ligatures he applied must also have eliminated the activity of the rectum, the part played by the Malpighian tubules remains uncertain.

When this investigation was nearing completion my attention was drawn to the work of Boné & Koch (1942) on the larva of Limnophilus flavicornis. Boné & Koch collected fluids from the anus and from the intestine and determined their chloride content. They assumed that the intestinal fluid was derived solely from the excretion of the Malpighian tubules; this may well be true, but it is not to be forgotten that the intestine is in free communication with the midgut and may receive fluid from the midgut as well as from the Malpighian tubules. It was shown that when the larva was placed in an external medium of 0·1 % NaCl the chloride content of the intestinal fluid was greater than that of the haemolymph; when the larva was placed in 0·001 % NaCl the chloride content of the intestinal fluid was less than that of the haemolymph, and further reduction of the chloride content was shown to occur in the rectum. In so far as they indicate a reduction in the concentration of osmotically active substances in the rectum the results obtained with Aedes aegypti are in agree-ment with the results of Boné & Koch on Limnophilus. At first sight it would appear that there is disagreement as to the part played by the Malpighian tubules. But there is no reason to suppose that the differences in chloride content as between the intestinal fluid and haemolymph in Limnophilus are incompatible with equality in total osmotic pressure ; it may be that the chloride/non-chloride ratio of osmotically active substances is capable of being varied in the Malpighian tubules, while the total osmotic pressure remains the same as that of the haemolymph. On the other hand, in view of the anatomical and histological differences known to exist in the Malpighian tubules of different orders of insects a radical difference in the physiological capability of the Malpighian tubules in Limnophilus as compared with Aedes need not be surprising. This uncertainty serves to emphasize the desirability of determining both chloride content and total osmotic pressure in the case of animals whose body fluids contain a high proportion of ‘organic’ solutes, and it is unfortunate that there is no method suitable for chloride determination on the small quantities of fluid obtainable from the gut of the mosquito larva.

As far as differences in total osmotic pressure (as distinct from possible differences in chemical composition) are concerned the rectum is the organ in the mosquito larva which plays the most important part in regulation. It remains for us to consider whether the activity of the rectum—and of other organs such as the anal gills—is sufficient to account for the observed ability of the larvae to survive in various media. In A. detritus the reduced anal gills are believed to be without function; the larva swallows the external medium and therefore it must be able, broadly speaking, to eliminate a fluid of the same osmotic pressure if it is to maintain the constancy of the haemolymph. Reference to Table 4 shows that in three cases out of seven the rectal fluid was substantially more concentrated than the external medium, showing that in normal sea water the rectum operates with some margin of safety. How it fares in more concentrated sea water, equivalent to 6 % NaCl, has not been studied.

The case of A. aegypti is rather more difficult, because of the parenteral absorption of water and salts through the anal gills. But when the animal is kept in a current of distilled water the amount of salt absorbed by the anal gills must surely be insignificant. Under these conditions the fluid collected from the rectum has an osmotic pressure equivalent to 0·07% NaCl (Table 1). No doubt some part of this is attributable to nitrogenous excretory matter, but the possibility of some salts being present can hardly be denied. Wigglesworth (1933 b) estimates that the volume of fluid eliminated by a larva living in tap water ‘certainly does not exceed the cubic capacity of two anal gills per hour ‘; this means that the larva will eliminate something like 20% of its own volume in 24 hr. Wigglesworth has also shown (1938) that when the larva is placed in distilled water the chloride content of the haemolymphs falls steadily to 0·05% NaCl in 8 days and thereafter remains constant. Even if the chloride content of the rectal fluid is of the order of 0·005% NaCl, as found for Limnophilus by Boné & Koch, the whole chloride content of the larva might be lost in 4–5 weeks. There is, however, some evidence that during prolonged starvation in a current of distilled water there is reduction in the volume of fluid eliminated by Aedes aegypti. Animals taken for examination often show a more or less continuous column of faeces extending through the intestine to the rectum. When observed on a slide, especially if the larva’s movements are restrained by the pressure of a cover-slip or by ligatures, the elimination of faeces begins almost at once, and after tfie solid contents of the hindgut have been evacuated the characteristic regular passage of drops down the intestine can be seen. It is difficult to believe that the solid column of faeces could have been established at a time when droplets of fluid were regularly passing down the intestine. This observation suggests that when the larva is undisturbed the elimination of faeces and fluid from the anus is in abeyance, that if fluid continues to be excreted by the Malpighian tubules it passes forwards to the caeca and that the intake of water through the anal gills is reduced.

Wigglesworth (1933b) considered the possibility that the passive entry of water through the anal gills might be actively resisted by a mechanism analogous to that which is known to exist in frogs’ skin. But his experiments showed that when elimination of fluid was prevented by a ligature between segments vi and vii the posterior part of the animal swelled up and might even burst, and he therefore recognized the lack of any evidence for active control of water intake. Nevertheless, the idea that there may be an active resistance to the entry of water at the anal gills and that this resistance readily breaks down when the larva is harassed is one which should not be discarded prematurely.

The histological observations reported in section IIIB are not in sufficient detail to merit extensive discussion. It is interesting, however, to have some clue as to the possibility of identifying certain physiological processes of osmotic regulation with discernible histological characters. Pagast (1936) has observed differences in the appearance of the Malpighian tubules of A. aegypti according to the medium in which the larvae were kept. Hamisch (1934) has also applied histological methods to the problem of osmotic regulation in Chironomus thummi, but reaches conclusions which are difficult to accept. Progress with the same general problem is reported by Pettengill & Copeland (1948) and by Copeland (1948), who have identified the chloride-secreting cells in the gills of the fish Fundulus. Mosquito larvae would appear to provide promising material for an extension of these studies.

I am greatly indebted to Dr V. B. Wigglesworth for reading the manuscript of this paper and for most valuable discussion. I wish to thank Sir Rickard Christophers for providing me with the eggs of A. aegypti and for advising me on the method of rearing this species; and I also wish to thank him and Mr P. G. Shute for going to great trouble to help me obtain supplies of A. detritus. Nearly all the freezing-point measurements and preparation of sections were carried out by my assistant, Miss J. Gukenbiehl.