ABSTRACT
Control of body volume has been studied in closely related freshwater and salt-water mosquito larvae (Aëdes aegypti and A. detritus).
In A. aegypti osmotic inflow of water is equal to the rate of extrusion in urine and is less than the rate of production of Malpighian tubule fluid. Some of this fluid is conveyed to the rectum by intestinal peristalsis dependent on adequate hydrostatic pressure in the pyloric chamber, the rest is conveyed to the midgut by retroperistalses and is resorbed into the haemolymph.
Retroperistalses are normally partly inhibited by nervous control probably involving the stomatogastric system and stretch receptors in the body wall. When inhibition is stopped (by interruption of the anterior part of the nerve cord) pressure in the pyloric chamber stays low, intestinal peristalses and hence urine production are curtailed, and the larvae swell.
In A. detritus urine production is apparently independent of the slow osmosis and is balanced normally by drinking. Larvae with the mouth sealed shrink in consequence of urine output. Toleration of shrinkage is probably advantageous when the larvae are in concentrated brine.
INTRODUCTION
The larva of Aëdes aegypti gains water from the dilute external medium by osmosis. The osmotic inflow amounts, at room temperatures, to about 33% of the body weight/d in larvae with anal papillae of normal size, and it occurs principally through these papillae (Wigglesworth, 1933 a; Shaw & Stobbart, 1963). In order to stay in water balance the larva must extrude water at an equivalent rate and it is very likely that this extrusion is performed entirely by the excretory system. The fluid produced by the Malpighian tubules is discharged into the pyloric chamber of the gut (Text-fig. 1) and is then either passed forwards into the midgut from which it is absorbed into the haemolymph, or backwards to the rectum as discrete droplets by means of peristaltic action of the intestine. In the rectum some water and most of the ions are resorbed from the fluid before it is discharged to the exterior (Wigglesworth, 1933 a, b; Ramsay, 1950, 1953). Ramsay (1953) was able to show, by stretching the body of the larva, or by distending it with additional haemolymph, that the rate at which droplets of fluid were passed back to the rectum was increased when the body wall was under tension, and he supposed that the peristaltic activity of the intestine was under nervous control. No information was available, however, about the way in which the flow of fluid from the Malpighian tubules was divided between the midgut and hind gut.
During an investigation into the control of sodium uptake by fed sodium-deficient larvae (Stobbart, 1970) it became apparent that larvae swelled greatly when (i) they were ligatured at the neck or between thorax and abdomen (ii) the thoracic ganglia or the nerve cord in the first abdominal segment were pinched or frozen. These unexpected results led me to make some semiquantitative observations on urine production in unoperated larvae and larvae ligatured in various ways, and these results I now wish to report together with the data on swelling as they allow us to take the question of regulation of body volume a stage further, and also provide an interesting comparison with the control of urine production in terrestrial insects (Maddrell, 1963; Berridge, 1966).
Some observations on larvae of Aëdes detritus are also reported. This species lives in salt marshes and will tolerate waters more concentrated or more dilute than sea water. The ionic and osmotic compositions of the haemolymph are regulated, despite wide variations in the external medium, at values similar to those of A. aegypti (Beadle, 1939). The anal papillae are very small and it appears that regulation is achieved by the larvae drinking the medium and excreting appropriate volumes of dilute or concentrated rectal fluid depending upon the current situation (Beadle, 1939; Ramsay, 1950; Shaw & Stobbart, 1963). The cuticle of the general body surface seems to be very impermeable (Beadle, 1939; Ramsay, 1950) in common with that of A. aegypti (Wigglesworth, 1933 a; Treherne, 1954).
To avoid ambiguity we define certain terms as follows: tubular fluid, the unmodified secretion of the Malpighian tubules: pyloric fluid, the probably unmodified tubular fluid during its stay in the pyloric chamber: rectal fluid, tubular fluid in process of modification due to resorption of ions and water by the rectal epithelium: urine, the completed excretory fluid as discharged to the exterior.
MATERIALS AND METHODS
Fourth-instar A. aegypti larvae of stock L (Stobbart, 1967) were used in this work. They were reared at 28° C in three different ways: (i) under sodium-deficient conditions (Stobbart, 1960) in which case the anal papillae were greatly hypertrophied (Text-fig. 2) and the sodium content was low; (ii) in a medium containing 2 mm/1 NaCl (Stobbart, 1959, 1965, 1967); (iii) in Newcastle tap water, of sodium concentration ∼ 0·3 mm/1. The last two methods yield larvae with normal anal papillae and normal sodium concentrations. Larvae were used either in the ‘fed’ condition (Stobbart, 1970) or starved for 60–72 h.
Larvae of A. detritus were collected from Seaton Sluice in Northumberland. They were kept in sea water and fed, like A. aegypti, on dog biscuit and ‘Bemax’. After reaching the fourth instar they were starved in sea water for 60–72 h before use.
Operations on larvae (Aëdes aegypti)
The techniques described earlier (Stobbart, 1970) were used for ligaturing the larvae at various places and for destroying parts of the nervous system by pinching and freezing.
Sealing of mouth and anus ( both species)
This was achieved with resin-beeswax mixture (Krogh & Weis-Fogh, 1951). The seals were tested by placing larvae in sea water coloured with amaranth or phenol red; those with faulty seals were discarded.
Measurement of swelling and shrinkage (Aëdes aegypti)
Swelling (or shrinkage) was recorded visually or photographically, or as a change in wet weight, the larvae being weighed to an accuracy of better than 1% either individually or in groups of ten on torsion balances (Stobbart, 1970); any changes in weight due to uptake of ions must have been quite negligible. Swelling was allowed to occur in the very dilute rearing solution used for producing sodium-deficient larvae, in medium containing 2 mM/1 NaCl, in 2 mm/1 NaCl, or in Newcastle tap water. As the osmotic pressure of the haemolymph is equivalent to 142 mm/1 NaCl (Wigglesworth, 1938) the osmotic pressure difference between the haemolymph and the external medium is for practical purposes the same in all these solutions.
Measurement of swelling and shrinkage (Aëdes detritus)
These were observed visually in sea water or dilute sea water after the mouth, or mouth and anus, had been sealed. They were recorded on a comparative scale in which ‘+ + +’ represented severe, and ‘+’ slight shrinkage (or swelling).
Visual observations on larvae (Aëdes aegypti)
These were made using a binocular microscope and the arrangement shown in Text-fig. 3. The cellophane served to restrict the movements of the larva and the medium was renewed from time to time to eliminate the effects of evaporation. Movements of the gut wall, and of fluid in the gut forwards and backwards from the pyloric chamber were noted; in the case of backwards movement the frequency of con-veyance of droplets to the rectum was recorded for suitable time intervals and a rough estimate was made of the size of the droplets. The droplets are moved quite rapidly back to the rectum and as the intestine has an S-bend in it and is often partly obscured by the Malpighian tubules more accurate estimations of droplet size were impracticable. Usually the droplet size stayed fairly constant but quite often varied considerably. The rate of discharge of rectal fluid was measured similarly.
Experimental temperatures
The onset of swelling (in 2 mm/1 NaCl) after application, of ligatures or pinching nervous tissue was followed in fed larvae of A. aegypti at 28°C. Visual observations using the arrangement of Text-fig. 3 were made at 26–29°C and some observations on the long-term effects of ligatures between abdominal segments IV and V were made with the larvae kept at ∼ 18°C, as were the observations on A. detritus.
RESULTS
(a) Aëdes aegypti
Apparently any operation which interrupts the nerve cord at the neck, in the thorax, or in abdominal segment I, causes the larvae to swell. The degree of swelling is roughly the same for all operations but it occurs more rapidly in larvae with large anal papillae. When swelling is complete the gain in weight is about 20% (Text-figs. 4, 5). Two fully swollen larvae (with normal papillae) are shown in Plate 1. The body becomes quite turgid so that the larvae cannot swim with their characteristic flexing movements, the thin cuticle of the neck and the intersegmental regions becomes distended, and the papillae splay apart due to distension of the thin cuticle round the anus. The larvae survive for at least 24 h as judged by movements of the heart and mouthparts and the occasional performance of an incipient flexing movement, but in extreme cases the midgut stretches and the rectum, intestine and pyloric chamber plus the proximal ends of the Malpighian tubules prolapse through the anus (Text-fig. 6).
The time course of swelling after ligaturing or interrupting (pinching) the nerve cord is shown in Text-figs. 4 and 5. These larvae were fed and sodium-deficient (with large papillae) and were taking sodium up from 2 mm/1 NaCl (Stobbart, 1970); the time courses of the sodium uptakes are shown in figs. 13 and 15 of that paper, from which we see, by comparing unoperated larvae and controls with those ligatured at the neck, that control of sodium uptake is independent of the swelling. Swelling seems to start immediately after the operation and in the case of the experimental larvae is complete in about 6 h, by which time the body weight has increased by some 20%. The initial rate of swelling appears to be about 8% of the body weight/hor 190%/d. Larvae sham-ligatured round the respiratory siphon show a much less, extensive degree of swelling—the initial rate is only about 2% of the body weight/h and after 6 h the gain in weight is only some 7%. We may conclude therefore that the initial rate of swelling due to interruption of the nerve cord is about 6% of the body weight/ h. Unoperated larvae show an even slower gain in weight (c. 2% in 12 h) which is presumably due to an increase in haemolymph volume following the uptake of sodium chloride.
Observations on swollen and swelling larvae show that the swelling is due to changes in the midgut. In unoperated larvae the tubular fluid accumulates in the pyloric chamber which is closed posteriorly by the collapsed wall of the intestine and is to some extent closed anteriorly by the contents of the midgut. After a while the pyloric chamber becomes noticeably distended in spite of the fact that some of the fluid entering it is moved forwards into the midgut by means of occasional low-amplitude retroperistaltic movements of the wall of the pyloric chamber and the rear end of the midgut (Wigglesworth, 1933 a). When the hydrostatic pressure (see below) in the pyloric chamber has reached an appropriate level the wall of the intestine suddenly dilates and moves forward to engulf the fluid in the chamber and convey it by peristalsis to the rectum. When the nerve cord is interrupted the retroperistaltic movements become more frequent, sometimes more or less continuous, so that the midgut becomes appreciably dilated. This dilation makes the midgut wall move away from the food column and this presumably makes it easier for fluid to leak from the pyloric chamber to the midgut. The result is that the hydrostatic pressure in the pyloric chamber cannot build up to a level adequate to start the peristalsis of the intestine, or can do so only slowly, consequently the output of urine is inhibited or stopped and the larvae swell. It is clear from Text-fig. 1 that a ligature between abdominal segments IV and V will restrict the tubular fluid to a small region of the midgut so that production of a given amount of tubular fluid will cause a greater increase in hydrostatic pressure in the rear of the midgut and in the pyloric chamber. If such a ligature is applied to a swollen larva the pyloric chamber dilates as in unoperated ones, pyloric fluid is conveyed to the rectum and the output of urine starts up again causing the swelling of the hind end of the larva to diminish. These results are partly illustrated in Text-fig. 7A–C and are representative of the results obtained from 20 larvae.
Although application of a ligature between the abdominal segments IV and V of a swollen larva enables urine output to start again, the effect is often not permanent but may decline after 2–3 h. This is illustrated by an experiment on ten starved larvae reared by method (ii). They were ligatured (between IV and V) at zero time and placed in tap water at room temperature. At suitable times they were examined for swelling anterior and posterior to the ligature as shown by distension of the neck (this becomes apparent sooner than distension of the intersegmental regions) and splaying of the anal papillae. The results are shown in Table 1. As is to be expected from Wigglesworth’s observations (1933 a) the swelling of the anterior part of the body is very slight. After 22·5 h, however, the posterior part was markedly swollen in half of the larvae. Examination of these swollen ones suggested that the cause of swelling was not curtailment of urine production but loss of the intestine’s ability to perform peristalsis. In one of these larvae it was found that the pyloric chamber had become grossly distended with clear fluid, and in another, that with the exception of the anus which was closed, the whole gut posterior to the ligature was relaxed and filled with fluid. In this larva the anus was seen to open briefly and emit a small quantity of fluid. A similar, though less extreme, relaxation of the gut was seen in an unswollen larva though here discharge of fluid had presumably kept pace with the osmotic inflow. Wigglesworth (1933 a) used larvae ligatured in the same way but makes no mention of any swelling of the body posterior to the ligature, perhaps because of the shorter time intervals used, though it is possible that his stock of larvae differed slightly from mine in this respect. Larvae with large anal papillae appear to be more liable to swelling in the posterior region when ligatured in this way (Text-fig. 6). Perhaps the greater osmotic inflow and consequently greater rate of tubular fluid production in such larvae causes the intestine to tire sooner.
In view of the results obtained with A. detritus (see below) some confirmation of Wigglesworth’s (1933b) observation that urine production in A. aegypti is curtailed or stopped in iso-osmotic solutions seemed desirable. The osmotic inflow may be reduced by sealing the mouths of larvae which have had their anal papillae removed (Wigglesworth, 1933 a) thus limiting water entry to that which can occur through the relatively impermeable body surface; and, as sucrose is known to penetrate the gut walland the anal papillae slowly (Ramsay, 1950; Stobbart, unpublished), it may be stopped or reversed by placing larvae in iso-or hyper-osmotic sucrose solutions made up in tap water. The addition of osmotically insignificant amounts of phenol red to the sucrose and other solutions enables one to check the efficacy of the seals or to follow the ingestion of the external medium.
Results from starved larvae reared in tap water are presented in Tables 2 and 3. Weight changes due to any salt movements would have been quite negligible and if the guts of the larvae had been quite empty the observed weight changes could have been attributed solely to the differences between osmotic inflow and urine production superimposed on a small weight loss due to metabolism. In fact the guts of starved larvae always contain appreciable amounts of detritus and the larvae probably lost some weight due to defaecation. Nevertheless Table 2 shows clearly that the weight changes resulting from curtailment of osmotic inflow are small and so it seems reasonable to conclude that the rate of urine production in the experimental larvae was considerably reduced. The results also suggest, in agreement with those of Wiggles-worth (1933 b), that the gut is involved in taking up a small amount of fluid from the iso-osmotic and hyperosmotic solutions (the former of which seems to have been slightly hyperosmotic for these larvae).
As the larvae are continuous filter-feeders they continually take in a certain amount of the external medium. When hyperosmotic sucrose solution is ingested the gut becomes greatly distended (Ramsay, 1950; Table 3) presumably due to osmotic uptake of water from the haemolymph. The distribution of phenol red between gut, haemolymph and Malpighian tubules in normal larvae in tap water, and in those with the gut distended due to ingestion of sucrose, is shown in Table 3. It appears from this Table that phenol red does not normally penetrate the wall of the midgut. However, as we find larvae in sucrose solutions with severely distended colourless guts and no colour in the haemolymph but with occasional colour in the Malpighian tubules, it would appear that the distended gut becomes preferentially permeable to the phenol red molecule which is rapidly removed from the haemolymph by the Malpighian tubules. In their ability to excrete phenol red rapidly the Malpighian tubules appear to resemble those of other insects and the kidney tubules of fish (Lison, 1941; Forster & Taggart, 1950).
(b) Aëdes detritus
The observations on this species are presented in Table 4. Larvae with the mouth or mouth and anus sealed were placed in slightly concentrated, or highly diluted sea water roughly equivalent to fresh water (Na concentrations 496 and 2·2 mm/1 respectively). As would be expected from the earlier work (Beadle, 1939; Ramsay, 1950), larvae with the mouth sealed shrank in sea water, but surprisingly such larvae also shrank in dilute sea water at a somewhat slower rate, but one which was still much higher than the rate shown by larvae in sea water with the mouth and anus both sealed. This can only mean that the rate of urine production exceeded the rate of osmotic entry or withdrawal of water, and we must conclude that urine production continues even when the normal supply of water (i.e. from drinking) is stopped. Preliminary observations suggest that the sodium concentration of the haemolymph drops more rapidly when the larvae are placed in dilute sea water if the mouth is sealed, for after 25 h in dilute sea water the concentration was 63 mm/1 with the mouth sealed, but with the mouth free it was 114 mM,/1 after 45 h—the concentration for unoperated larvae in sea water being about 155 mm/1.
DISCUSSION
(a) Aëdes aegypti
The results presented in the preceding section confirm the more extensive observations of Wigglesworth (1933a, b, 1932) and Ramsay (1950, 1953) on the movements of water (which are summarized in Text-fig. 8) within the body of the larva.
So far as regulation of the body volume is concerned the crucial point is the division of the flow of tubular fluid into streams of appropriate size proceeding to the midgut and the rectum. The present results show clearly that this division is achieved by nervous control of the activity of the rear end of the midgut wall. The following points must be borne in mind when suggesting a scheme for the control of body volume: (i) Ligatures at the neck or between thorax and abdomen start the swelling, as do pinchings or freezings of the thoracic ganglia or nerve cord in the first abdominal segment, (ii) The gut and its innervation are left intact in the pinched or frozen larvae, (iii) Intestinal peristalsis is increased by distension of the body wall (Ramsay, 1953). (iv) The effect of ligaturing swollen larvae between abdominal segments IV and V suggests strongly that an adequate hydrostatic pressure in the pyloric chamber is an important factor in triggering the peristaltic activity of the intestine; this is supported by the observation (Ramsay, 1950) that incipient retroperistalses in the intestine can occur if the anus is blocked and the rectum distended with fluid. However, other factors are probably also involved for under normal circumstances the size of the droplets may alter for no apparent reason, and a ligature between IV and V often fails to maintain long-term activity. It seems possible that the stomatogastric system (which of course is interrupted by ligatures) may be of importance here.
The scheme illustrated in Text-fig. 9 will clearly explain all the experimental observations and we may conclude that something similar to it operates in the larvae. A few comments on this scheme are, however, necessary. (1) As the head is essential for inhibition of midgut retroperistalsis, it is very likely that the effect of ligaturing at the neck is due to interruption of the stomatogastric nervous system. This system is described by Christophers (1960, p. 342) and is very similar to that of Culex (Cazal, 1948). However, in Corethra larvae, which maintain water balance in a similar way, ligaturing at the neck does not cause swelling (Schaller, 1949). (2) The effect of distension of the body wall is presumably mediated by some sort of stretch receptor (cf. Maddrell, 1964). If so it is surprising to find that larvae with the nerve cord pinched in the first abdominal segment swell as rapidly as those ligatured at the neck or between thorax and abdomen, or those with the thoracic ganglia pinched (Text-figs. 4, 5). In the former case any stretch receptors in the thoracic wall would presumably still be able to influence the stomatogastric system and we might therefore expect some inhibition of midgut retroperistalsis and a consequently slower rate of swelling.
This apparent inconsistency is probably due to the fact that the thorax, whose fused segments lack thin intersegmental cuticle, is less distensible than the abdomen (Plate i). (3) The swelling induced by the sham operations (Text-figs. 4, 5) is rather surprising. We might suppose that it results from the destruction of a certain number of the stretch receptors thus allowing some swelling to go unnoticed, but it occurs equally in larvae sham-pinched (along one side of the thorax) or sham-ligatured (around the inextensible respiratory siphon). Possibly therefore the swelling results from the swamping of information from stretch receptors with that coming from the wounded regions of the body wall. (4) No suggestion is made in the scheme as to how rectal discharge may be triggered off. In view of the behaviour of the intestine, however, it seems likely that hydrostatic pressure in the rectum will be an important factor, but it is unlikely to be the only one as the amount of fluid in a discharge can vary widely. Under normal circumstances most of the salts are resorbed from the rectum before discharge of the fluid—this resorption, however, is curtailed when the larvae are kept in concentrated solutions (Ramsay, 1950, 1953) and it seems likely that under these conditions the resorption of water is increased (Wigglesworth, 1933 a, b) though the rectal fluid is always iso-osmotic to the haemolymph (Ramsay, 1950, 1953). Details of the control of salt and water resorption are not known but it seems likely that salt resorption may be controlled by the same hormone as controls salt uptake by the papillae (Stobbart, 1970) as the two tissues have a common embryological origin, transport salts in morphologically the same direction, and for maximum salt conservation or uptake require to be activated at the same time. It seems reasonable to suppose therefore that under normal circumstances salt and water resorption are set near maximum and minimum respectively, and that a discharge of the rectum is started either in response to a low electrolyte concentration (or osmotic pressure) in the rectal fluid, or to an appropriate hydrostatic pressure, or to a combination of these factors.
The apparently continuous secretion of tubular fluid is not surprising in an animal subjected to a continuous osmotic gain of water, but it contrasts markedly with the hormonally induced secretion in the blood-sucking bug Rhodnius (Maddrell, 1963). As salts must be resorbed from the rectal fluid, and as this apparently cannot be achieved without some resorption of water (Wigglesworth, 1933 a, b, 1932; Ramsay, 1950, 1953) production of tubular fluid at a rate exceeding the osmotic inflow is a necessity. In the normal larva in water balance we have in fact the relationships
Rate of tubular fluid > Rate of urine production = Rate of osmotic inflow production
and
Rate of tubular fluid production. = Rate of urine production + Rate of resorption from rectum + Rate of passage of fluid tubular fluid to midgut
The final term of this latter equation presumably represents a safety factor against increased osmotic inflow, and it need not necessarily represent a recycling of metabolic wastes since these could be retained in the lumen of the midgut and be eventually passed back to the rectum with the food residues. In fact this is what would appear to happen to the molecule of phenol red, which, like metabolic wastes, is excreted by the Malpighian tubules. While it would clearly be preferable for most of the salts in the rectal fluid to be resorbed as indeed is the case (Ramsay, 1950, 1953), the effective ionic pumps in the papillae (Stobbart, 1960,1965, 1967, 1970) could in emergency presumably permit, in the interests, of extruding excess water, toleration of a temporary curtailment of salt resorption due to increased flow of fluid through the rectum.
Preliminary observations suggest that this type of control of urinary output persists in the adult. Even before a blood meal is complete the adult female starts to excrete droplets of a clear urine which contains appreciable amounts of sodium and potassium (Chambers & Stobbart, unpublished). This diuresis is stopped by a ligature at the neck or between the thorax and abdomen, and the effect of these ligatures appears not to be due to interruption of a hormone supply since extracts of heads and thoraxes prepared from animals in diuresis fail to restart diuresis in opened ligatured preparations and in isolated preparations of rectum intestine and Malpighian tubules (Schofield & Stobbart, unpublished).
Text-figs. 4 and 5 show that larvae ligatured at the neck or between thorax and abdomen, or with the thoracic ganglia or nerve cord in the first abdominal segment pinched, swell at a rate of about 8% of the body weight/h (190%/d). This rate of swelling is in fact a minimal estimate of the rate of osmotic inflow since urine production is not always stopped by the operations (Text-fig. 7 A –C) and the larvae will have lost weight by metabolism and probably defaecation. Nevertheless this estimate is much higher than that based on earlier work which was about 33% of the body weight/d. This could be due to the higher temperature used in the present work as the Q10 for penetration of water through the cuticle of the aquatic larva of Sialis is known to be high (Shaw, 1955) but it is also likely to be due to the fact that the anal papillae of the present larvae were much hypertrophied (Text-fig. 2) with a consequently greater surface area; there is also the possibility that the hypertrophied papillae are proportionally more permeable to water than normal ones as they not infrequently show cytological abnormalities (Stobbart, 1958).
The papillae of salt-water mosquito larvae are reduced in size (e.g. A. detritus). The reason for this is not very clear for although the papillae are permeable to water and salts the osmotic gradients tolerated by the freshwater forms are generally about the same as or greater than (though in the opposite direction) those tolerated by the salt-water forms. To be of use to salt-water larvae the papillae would have to transport salts outwards. The papillae are known to be derived from rectal epithelium (Christophers, 1960) which is known to secrete salts and water from the rectal lumen to the haemolymph (see Shaw & Stobbart, 1963) and it is clear that in A. aegypti the papillae transport salts in what is morphologically the same direction (Stobbart, 1960). If rectal epithelium is not capable of reversing the direction of transport then the reason for the reduction of the papillae in salt-water larvae (and in freshwater larvae reared in abnormally concentrated media—Wigglesworth, 1938) becomes apparent. The papillae of the salt-water larvae, however, are in need of reinvestigation with modern techniques, for in Aëdes campestris (Phillips & Meredith, 1969) a fortnight’s stay in a dilute medium (5 mm/1 NaCl) elicits active uptake of Na+ and Cl –through the papillae, and there is some evidence that they may be involved in extrusion of Cl-into the normal hyperosmotic medium. In this species therefore they are clearly not inert as was suggested for A. detritus by Beadle (1939). No evidence for uptake by the papillae of A. detritus was obtained from the preliminary results reported here, which suggest in fact that the gut is involved in the conservation of salts, but the larvae had not been adapted to the dilute medium.
(b) Aëdes detritus
Details of volume regulation have not been worked out here, but this species differs markedly from A. aegypti in that the rate of urine production is not dictated by the rate of osmotic inflow of water. The reasons for this seem clear. A. aegypti living in dilute waters has always to cope with an inflow of water; tubular fluid is produced continually at a higher rate than the inflow and there is presumably always sufficient water available for removal of metabolic wastes. A. detritus, however, often Uves in hyperosmotic waters and has reduced anal papillae so that water movements through the body surface will be small. Under these conditions the larvae apparently gain the necessary water by drinking the medium and excreting a concentrated urine (Ramsay, 1950) and we should expect the rate of drinking to be governed by the rate of urine production. This seems to be the case, for larvae with the mouths sealed continue to produce urine at a faster rate than water can enter or leave through the body surface and consequently shrink; it would seem therefore that it is mainly the rate of drinking which is adjusted to keep the larvae in water balance. Clearly in A. detritus conveyance of tubular fluid to the rectum is not stopped when the body wall is collapsed, and so it follows that the scheme of Text-fig. 9 cannot be operating in these larvae. This toleration of shrinkage is probably an advantage to the larvae when they are in highly concentrated brine. For under these conditions, even if they were unable to win water from the medium, they might, by excreting as concentrated a urine as possible and ceasing to drink, maintain their osmotic pressure below that of the medium at the expense of a continually reduced haemolymph volume, and so manage to survive until the medium became diluted again.
REFERENCES
EXPLANATION OF PLATE I
(A) A larva reared by method ii, 20 h after the thoracic ganglia were frozen.
(B) As above but ligatured between thorax and abdomen. 1, distended cuticle around anus; 2, midgut; 3, Malpighian tubules; 4, peritrophic membrane. Note the splayed papillae, the distended neck and intersegmental cuticle, and the relative inextensibility of the thorax. The distension of the neck in (B) is to some extent caused by the constriction of the body by the ligature.