Ionic Regulation and Water Balance in the Aquatic Larva of Sialis Lutaria

The work described in this paper is a continuation of the investigations on the osmotic and ionic regulation of Sialis larvae reported by Beadle & Shaw (1950). In this work it was found that the larvae possessed no mechanism for the absorption of ions from dilute solutions, as have mosquito larvae, for example (Koch, 1938), and that chloride was retained in the larvae by the very low permeability of the cuticle. Further, variations in the chloride content of the blood were not paralleled by variations in the total osmotic pressure, and measurements of the non-protein nitrogen content of the blood suggested that changes in the salt content in the blood were compensated by changes in the concentration of the blood amino-acids. At that time very little was known of the regulatory powers of aquatic insects with respect to the blood inorganic ions and, in fact, the composition of the blood had not been determined in detail. Wigglesworth (1938) had found that in the mosquito larva, kept in saline solutions of less than the blood concentration, maintained the blood chloride at its normal level so that some regulatory mechanism appeared to be at work.

The experiments in this paper were carried out in order to throw further light on the normal blood composition of Sialis larvae and to discover how both the normal concentration of inorganic ions in the blood and the normal water content of the larvae are maintained.

Recently the knowledge of ionic regulation in the mosquito larva has been greatly advanced by the work of Ramsay (1951, 1953 a), who has described the role of the Malpighian tubules and the rectum in regulating the potassium and sodium content of the larvae, using an elegant micro-flame photometric technique.

As far as ionic regulation and water relations are concerned, Sialis larvae differ in a number of important ways from mosquito larvae. Thus Sialis larvae possess no ion-absorbing mechanism, and as a consequence there is no region of the cuticle that is especially permeable to water, so water penetration is slow (Shaw, 1955 b). In the excretory system there is no circulation of fluid into the mid-gut, as far as is known, and the excretory fluid is itself not much less concentrated than the blood.

It is, therefore, of considerable interest to compare the regulatory mechanisms found in Sialis with similar mechanisms in the mosquito larvae, where this is possible.

Samples of blood for the quantitative analysis of its inorganic constituents were collected in the manner described in a previous paper (Shaw, 1955 b). On the extracted blood and on the excretory fluid estimations were made of the concentrations of sodium, potassium, calcium, magnesium, chloride and bicarbonate.

The four cations and chloride were estimated by means of the ultra-micro methods described in another paper (Shaw, 1955 a).

Bicarbonate was determined by liberating carbon dioxide from the sample by the addition of acid and absorbing the CO₂ in standard alkali. This reaction was carried out in a micro-diffusion cell of the type described by Shaw & Beadle (1949) : the gas was liberated by the addition of 50 μ. of N sulphuric acid and absorbed in a 5 pl. hanging drop of N/20-NaOH containing phenolphthalein as indicator. The excess alkali was titrated with 0·05 N-HCI using the titration apparatus described by Shaw (1955 a). The time allowed for the diffusion was 2 hr. and the sample used for each determination contained between 5 and 10 μg. of bicarbonate.

Measurements of conductivity were made with the aid of a Mullard conductivity bridge and a conductivity cell which contained 2 ml. of fluid. The sample (about 1 μl.) was diluted to this volume with glass-distilled water and the bridge and cell was calibrated empirically with standard sodium chloride solutions prepared in the same way. At this dilution the accuracy was not greater than + 10 mM./l. NaCl.

Membrane potentials were recorded by means of a valve millivoltmeter consisting of a pair of RCA 954 valves connected as electrometers in a balanced bridge circuit. The input impedance was 100 MΩ and the measurements were made to the nearest 0·5 mV.

The total osmotic pressure of the blood of Sialis larvae has been measured by Beadle & Shaw (1950), who found it equivalent to the osmotic pressure exerted by a i % sodium chloride solution approximately. This value for the osmotic pressure of the blood is one which is found in many fresh-water and terrestrial animals (see Krogh, 1939), but generally the solute particles which give rise to this osmotic pressure consist largely of sodium and chloride ions. This is not usually the case in insects ; thus Beadle & Shaw found an average chloride ion concentration of 34 mM./l. in Sialis larvae, and this would account for only 10% of the total osmotic pressure. Similarly, Wigglesworth (1938) had found that in the mosquito larva in tap water the chloride ion contribution was about 18%. It was suggested, as a result of the large amount of non-protein nitrogen that was found in the blood of Sialis that the presence of free amino-acids might account for much of the osmotic pressure deficit. In order to have a clearer understanding of the relative contribution of inorganic ions to the total osmotic pressure of the blood a more detailed analysis has been made, and the results are shown in Table 1.

The sodium concentration (109 mM./l.) is much higher than that of the chloride and accounts for some 32% of the osmotic pressure. This is similar to recent measurements made on the mosquito larva by Ramsay (1951, 1953 a), who found an average concentration of 108 mM./l. for sodium in the blood of larvae kept in distilled water and, as in this case the total osmotic pressure was lower than that of Sialis blood, the sodium ions accounted for 42 % of the total.

The total concentration of anions and cations which have been measured is 186·5 mM’/l- and this would give rise to an osmotic pressure of only 55% of that found leaving a deficit of 152·5 mM./l. The sum of the cations—and these are probably the only ones present in any quantity—is 167 m.equiv./l. and this is very similar to the value of 152 m.equiv./l. determined from the conductivity measurements. Therefore the great majority of the cations must be present in an ionized form. This is, however, a big discrepancy in the concentration of anions as measured and as predicted from the conductivity measurements. The sum of the anions (46 m.equiv./l.) is less than one-third of the value of 167 m.equiv./l. required by the conductivity. The blood must contain, therefore, a large unknown anion fraction to which other inorganic anions such as phosphate and sulphate are unlikely to contribute very much.

It is not possible to calculate what fraction of the total osmotic pressure deficit of 152·5 mM./l. is made up of these unknown anions since their valency is not known. If the valency of these anions was one then they would contribute some 70 % to this deficit and the total electrolyte content of the blood would account for about 90 % of the total osmotic pressure. If the average valency was two, then the contribution to the deficit would only be 35 % and so on.

Proteins in solution often behave as polyvalent anions and their effect can be considered since the average concentration in the blood is known (Beadle & Shaw, 1950) to be 5·2%. This amount would have a negligible osmotic effect but could exert an effect equivalent to 5−10 m.equiv./l. on the conductivity, but this is small compared with the 121 m.equiv./l. that still has to be accounted for.

It was also found by Beadle & Shaw that the blood concentration of non-protein nitrogen-containing molecules (av. 258 mg. N%) was such that if each molecule contained 1 atom of nitrogen the concentration of these molecules would be 184 mM./l. This is greater than the osmotic deficit, and therefore some of these molecules must contain more than one molecule of nitrogen and might be peptides. Owing to the large amount of these nitrogen-containing compounds which are present, it is probable that the unknown anions form part of this fraction and therefore might consist of dicarboxylic amino-acids or peptides containing these.

The potassium concentration is quite low and characteristic of the carnivorous insects (Boné, 1944) and similar to the concentration found in other aquatic insects. Thus Ramsay (1953 b) gives values of blood potassium in Aedes, Dytiscus and an aquatic Tabanid as 3, 5 and 5 mM./l. respectively. Both the divalent cations, calcium and magnesium, are present in high concentrations and the magnesium is especially high, the concentration approaching that found in the marine crabs. However, the high concentrations of these ions and the low concentration of chloride ions seem to be characteristic features of insect blood composition.

Measurements of the total amounts of some of the inorganic constituents of Sialis larvae indicated that the distribution in the tissues must be very different from that in the blood. Beadle & Shaw (1950) measured the total chloride of the larvae and showed that if the chloride was confined entirely to the blood then the blood volume was approximately 56 % of the wet weight of the larvae. This estimate was confirmed by blood-dilution experiments. From a knowledge of the blood volume and of the blood concentration of the inorganic ions it is possible to calculate their concentration in the tissues, and experiments were carried out to determine the tissue concentrations of the four cations.

A larva was first weighed (W₁), then as much blood as possible removed and the larva weighed again (W₂). The dry weight was determined by heating to constant weight at 100 °C. (W₃). The dried larva was incinerated in a small crucible at 450°C. for several hours until the organic material had been removed, the resultant ash dissolved in 50 μl. of N/5-HCI and the concentration of the cations in this solution measured. The concentration of cations in the blood sample which had been removed was also determined.

so that C₃ can be calculated.

Table 2 shows the results of these calculations for the four cations. Potassium replaces sodium as the most important ion; the calcium concentration is not very different from that of the blood but the magnesium is much higher. Nothing is known of the anion composition except that chloride is practically if not completely absent. Thus sodium and chloride ions are confined almost entirely to the blood, whereas more than 90% of the potassium of the larva is located in the cells.

The excretory fluid of Sialis larvae, which is presumably produced by the Mal-pighian tubules, collects in the rectum which gradually distends. When it is fully distended the fluid is discharged through the anus. In order to collect this fluid larvae are narcotized with CO₂, carefully dried with filter-paper and placed on a waxed slide under a binocular dissecting microscope. The end of a capillary pipette, holding about 5 μl., is brought in contact with the opening of the rectum. This opening is normally kept closed by a sphincter muscle, but if the rectum is full of fluid, the touch of the pipette tip on the anus will cause this muscle to relax and the contents of the rectum are discharged. As the fluid is discharged it is sucked up into the pipette and in this way 2 −3 μl. of fluid can usually be collected. Since the water intake of a larva weighing 100 mg. is known to be approximately 4 mg. per day at 20 °C., it is probably that the excretory fluid is discharged about once a day.

If the rectum is not full then no discharge takes place but fluid can still be collected. If a slight pressure is exerted on the abdomen by the finger, the micro-pipette can be inserted through the rectal sphincter and the contents of the rectum taken up into the capillary.

It is important to be certain that the fluid collected in this way has been derived only from the rectum and not from the rest of the gut. The mid-gut opens directly into the rectum, the opening being normally closed by a sphincter just anterior to the point where the Malpighian tubules enter and if this sphincter opened mixing of the contents might occur. However, the fluid normally present in the fore- and mid-gut is a very dark brown in colour, whereas the rectal fluid is quite colourless. If any mixing occurs the latter fluid becomes coloured so any samples of the fluid which were coloured were discarded.

The results of the analysis of the fluid derived from normal larvae are shown in Table 3. Although no measurements were made of total osmotic pressure, the conductivity determinations show that the fluid is not a dilute one, as in so many fresh-water animals, but its conductivity is as much as 65 % of that of the blood. On the anion side this is made up almost entirely of bicarbonate, and a large number of measurements of chloride failed to reveal the presence of this ion at all. Of the cations, sodium and potassium only account for a small percentage of this fraction. Dr B. W. Staddon, working in this Department, measured the ammonium ion concentration, and the author is indebted to him for the figure for the concentration of this ion (100 mM./l.) which is given in the table. This concentration is sufficient to account for the rest of the cation fraction as required by the conductivity measurements.

The excretory fluid, therefore, consists largely of ammonium bicarbonate in sufficient concentration to make the total osmotic pressure of the fluid at least two-thirds that of the blood. This is markedly different from measurements on the mosquito larva, Aedes aegypti, made by Ramsay (1950), who showed that in this species the rectal fluid of larvae kept in distilled water was very dilute (average osmotic pressure equivalent to 12 mM./l. NaCl).

The concentration of sodium in the fluid (12 mM./l.) may be too high. It was found (Shaw, 1955 b) that sodium was only lost slowly from the larvae when they were washed in running tap water for 3 weeks, and it seemed likely that this sodium was lost through the cuticle, together with chloride. If this is true then sodium could not be lost in the excretory fluid as well, and it is possible that the sodium found in the rectal fluid was in the process of being reabsorbed and that this was not complete when the fluid was removed.

In view of the fact that the excretory fluid is completely different in composition from the blood, it is of great interest to know how this fluid is formed and the part the rectal epithelium plays in its formation. In order to study the equilibrium conditions which exist across the rectal wall, measurements have been made of the potential difference across this membrane. For this purpose a larva was narcotized and the abdomen dissected open and pinned out in a shallow depression in a wax block in such a way as to expose the rectum. The electrodes consisted of glass capillaries which tapered to a fine point and were filled with 1 % NaCl solution. Fine silver wires coated with silver chloride were inserted through the wider ends of the capillaries. The two electrodes were held in micro-manipulators and one of them was moved into position so that its tip was beneath the surface of the blood ; the other was first held in a similar position so that the out-of-balance potential between the two electrodes was measured. The second electrode was then inserted through the rectal epithelium to make contact with the rectal fluid and a measurement of the potential difference between the two electrodes made. The results of these measurements are shown in Table 4. The mean value was 24 mV. but two of the readings (40 mV.) are much higher than the others. These two measurements were made on the rectum when it was contracted, whereas the others were taken on the fully distended rectum. The rectal wall is considerably stretched in the distended rectum, and it is possible that leakage around the electrode in these cases might reduce the measured potentials and that the higher values found in the contracted rectum may be nearer the true ones.

z =valency of ion, F=the Faraday, T=abs. temp., E_o—E_i =the potential difference.

For a potential difference of 24 mV. the value of the exponential expression is 2·59 and the reciprocal is 0·38 at 17°C., and for the maximum recorded potential the exponential is 4·93 and the reciprocal 0·20.

Considering first the bicarbonate ion, the mean concentration in the blood is 15 mM./l. and for a potential of 24 mV. the equilibrium concentration in the rectum is 15 × 2·59 = 39 mM./l. The maximum equilibrium concentration is 74 mM./l. This is less than the measured concentration of 91 mM./l. and thus it would appear that this ion is not in complete equilibrium with the blood bicarbonate. It is noticed, however, that the potential need only be 5 mV. greater than the maximum recorded in order that the electrochemical potential of this ion should be the same on either side of the membrane. Since the accuracy of the potential measurements cannot be regarded as very great, not much significance can be attached to the apparent discrepancy. The concentration of ammonium ions in the blood is very low, and the potential would require the rectal fluid ammonium ion concentration to be even lower. Since it is, in fact, very high it is clear that this ion is not distributed in the expected manner and the rectal epithelium must be either impermeable to this ion or transport it actively against the electrochemical gradient.

The concentration of sodium ions in the excretory fluid (12 mM./l.) is lower than would be expected, since for a potential of 24 mV. and a blood concentration of 109 mM./l., the equilibrium concentration would be 41 mM./l. and for the maximum recorded potential it would be 22 mM./l. As with the bicarbonate ion the discrepancy is not a large one, and may indicate that equilibrium had not been quite reached at the.time the measurements were made.

Estimations of the chloride concentration of the excretory fluid, as described above, showed that this ion is not present. Since the mean blood concentration is 31 mM./l. the minimum equilibrium concentration would be 81 mM./l. and therefore, as in the case of the ammonium ion, the membrane must be either actively or passively impermeable to this ion.

The ratio of the mean bicarbonate ion concentrations is 6 and that of sodium ions 9, so that, remembering the large standard deviations of the means, the expression is approximately true. However, the possibility cannot be excluded that sodium ions are actively absorbed, at least to some extent, in the rectum.

These considerations give no information on the nature of the fluid produced by the Malpighian tubules or to what extent this is modified by the action of the rectal epithelium. It has been shown by Boné & Koch (1942) that chloride is contained in the secretions of the Malpighian tubules of the aquatic larvae of Limnophilus and Chironomus, and that this is largely removed in the rectum. Similarly, the recent measurements by Ramsay (1953a) on the Malpighian tubule fluid of the mosquito larva has shown that both sodium and, particularly, potassium are present and that these ions are again largely reabsorbed in the rectum.

In the case of Sialis larvae no attempt has been made to analyse the Malpighian tubule fluid directly, but some measurements have been made on the rectal fluid in the early stages of its formation, that is, soon after the distended rectum has been discharged, on the assumption that this fluid would resemble to some extent the fluid produced by the Malpighian tubules.

Measurements of the chloride concentration made in this way again failed to reveal the presence of chloride in this fluid. It is therefore assumed that in this larva that, unless chloride is reabsorbed very rapidly by the rectum, the Malpighian tubule fluid does not contain chloride. This seems a real difference from the findings of Boné & Koch since their technique also involved a collection of fluid from the anterior of the rectum. Similar measurements of the potassium concentration of this newly formed fluid, however, showed that this always contained a higher concentration than the fully formed fluid. Thus concentrations of 19, 36 and 26 mM./l. were obtained compared with the mean value of 4 mM./l. for the discharged fluid. Thus potassium is clearly reabsorbed in the rectum and, in this respect resembles the mosquito larva. The p.d. across the rectal wall is such that there is no necessity to postulate any active process being involved in this reabsorption.

If it is assumed that the fluid produced by the Malpighian tubules is isotonic with the blood as has been shown for the mosquito larva by Ramsay (1950, 1951), the simplest hypothesis for the composition of the Malpighian tubule fluid is that it consists of ammonium ions and sodium ions in the concentrations found in the rectal fluid and the rest of the cation fraction made up of potassium ions. This would require a potassium concentration of about 58 mM./l., a value which is not inconsistent with the concentrations of potassium found in the newly formed rectal fluid. There is no reason to suppose that the cations in the Malpighian tubule fluid are balanced by any other anion than bicarbonate, and therefore the fluid would be modified in the rectum by the reabsorption of potassium bicarbonate, a process which could take place by passive diffusion.

To study the regulation of the blood chloride concentration larvae were kept singly in specimen tubes containing 10 ml. of a sodium chloride solution. The solution was changed each week and the larvae were kept without food. Under these conditions the larvae would live for a very long time (2 months or more) in saline solutions not exceeding a concentration of 1 % NaCl. Several series of experiments were made and for each series a different concentration of NaCl was used and this varied between 34 and 171 mM./l. Estimations of the blood chloride concentration were made on one of the larvae every 2 days and measurements were carried on for a period of 4−5 weeks. The blood chloride was found to rise gradually and then to level off when a certain equilibrium concentration had been reached. The time taken for this concentration to be reached was usually between 3 and 4 weeks. In Fig. 1 the equilibrium values for the blood chloride concentration are plotted against the value of the outside medium concentration. When the outside concentration is greater than 130 mM./l. then blood chloride tends to attain the same value, but in solutions of a lower concentration than this the internal chloride level gradually increases above that of the outside and eventually reaches equilibrium at a concentration which may be considerably greater than that outside. Thus with an outside concentration of 34 mM./l. NaCl the equilibrium value for the blood chloride is 80 mM./l. or 2·4 times greater than the outside concentration.

This is strikingly different from the behaviour of the aquatic larva of Aedes which was studied by Wigglesworth (1938) where the blood chloride concentration is maintained almost constant in sodium chloride solutions up to about 0·75 % NaCl when the regulation begins to break down. At this concentration the external medium is almost isotonic with the blood.

Since in Sialis the blood chloride concentration rises above that of the external medium in the lower concentrations, some mechanism other than diffusion through the cuticle must be controlling the rate of chloride intake. Beadle & Shaw (1950), however, showed that Sialis larvae possessed no structures equivalent in function to the anal gills of the mosquito larvae for the uptake of ions from dilute solutions, so that the intake of chloride must, presumably occur through the gut.

This was tested by a similar series of experiments to those above, except that half of the larvae had their necks ligatured to prevent drinking. The rate of rise of blood chloride in the normal and the ligatured larvae was compared. Fig. 2 shows the results of one such series of experiments where the larvae were kept in 171 mM./l. NaCl solutions at 23°C. At this concentration the output of excretory fluid will be very small (see below) and the results will be little influenced by this. The rate of increase of the blood chloride concentration in the normal larvae is more than twice that in the ligatured larvae and equilibrium is attained in a much shorter time.

The uptake of chloride from the gut was studied under conditions such as made it possible to obtain quantitative data on the amount taken up per day and to relate this to the concentration of solution from which it was absorbed. Larvae were first weighed and then as much blood as possible was removed from a wound in the thorax, the chloride concentration of this blood being measured. The wounds were sealed with a small blob of ‘Sira’ wax, each animal was weighed again and transferred to an isotonic solution of mannitol, containing a known concentration of sodium chloride. With the blood volume reduced in this way, the larvae would take fluid into the gut at once. They remained in these solutions for 3 days and then were weighed again, the blood chloride concentration being once more measured. From these measurements and using an estimate for the blood weight of 56% of the body weight, the total amount of chloride taken up during the 3 days could be calculated.

The results of these experiments are shown in Figure 3, the uptake being expressed as HIM. ×10⁻⁶ Cl/mg. body weight/day and related to the concentration of NaCl in the external solution. The amount of chloride taken up is roughly proportional to the concentration of chloride outside. In the 171 mM./l. NaCl solution, chloride will also be entering by diffusion through the cuticle, but a correction can be made for this from a knowledge of the permeability constant of the cuticle (Shaw, 1955 b), and if this is done the uptake is reduced to about 16 mM. × 10^-8. This makes the proportionality better still and also, incidentally, indicates that the gut wall is a good deal more permeable to chloride than the cuticle—a conclusion that was reached also for the permeability to water (Shaw, 1955 b).

Thus the measured uptake of chloride can be explained by the passive diffusion of chloride through the mid-gut wall. Whether chloride can be absorbed from solutions less concentrated than 15 mM./l. or not still requires investigation.

Although it has already been established that chloride is not present in the excretory fluid of Sialis larvae living in tap water, the fact that the blood chloride concentration reaches an equilibrium value when the larvae are living in sodium chloride solutions suggests that the chloride uptake from the gut must eventually be balanced by a corresponding output in the excretory fluid. In order to test this, measurements were made of the chloride concentration of the excretory fluid of these larvae. When the larvae were living in the higher concentrations of sodium chloride the volume of the excretory fluid was considerably reduced, and therefore to make measurements on the excretory fluid of these larvae they were replaced for 2 days in tap water. This stimulated the production of rectal fluid once more and then measurements were made of the chloride concentration in the rectal fluid and in the blood. Fig. 4 shows the results of these determinations. The chloride concentrations of the excretory fluid are related to the blood concentration at the time of the measurement. The rise in chloride concentration of the rectal fluid is slow at first—when the blood concentration is twice the normal value the level has only reached about 10 mM./l. With further increases in the blood concentration the rise in rectal fluid concentration is steeper until at a blood level of about 100 mM./l. the rise becomes very steep and the concentration of chloride in the excretory fluid approaches that of the blood when the latter is about 120 mM./l. There is no evidence that the concentration ever exceeds that of the blood.

At the equilibrium concentration of chloride in the blood, the output of chloride lost must be exactly balanced by the amount of chloride absorbed. Chloride will be lost in two ways: first, in the rectal fluid if the blood chloride level is above normal; and secondly, by diffusion through the cuticle if the concentration of chloride in the blood is greater than that of the external medium. Thus if no chloride was absorbed, then the rate of loss of the blood chloride⁠, where V_r= volume of rectal fluid per day, C_r= Cl concn, of rectal fluid (function of C_b) ; C_b= Cl concn, of blood and C_o = Cl concn, of external medium. T=time constant for diffusion of Cl through cuticle and is 20 days at 14°C. W= blood weight.

This value for the loss of chloride can be computed for the equilibrium concentrations given in Fig. 3 and C_r calculated from Fig. 6. V_r = 4% of the body weight per day at 20°C. for larvae in tap water (Shaw, 1955 b). Making this calculation for external concentrations of 34, 58 and 85 mM./l. gives a rate of chloride loss as 2·64, 2·61 and 2·55 mM ×10⁻⁶/mg. body weight/day, and at equilibrium this must be balanced by an equal chloride uptake. Now comparing these rates of chloride uptake with the values obtained under conditions where drinking had been induced experimentally (Fig. 5), they indicate, particularly in the higher concentrations, that a considerably smaller amount of chloride is absorbed per day. This suggests that the amount of drinking is controlled by the larvae so that the rate of chloride absorption is kept at a fairly constant level.

The regulation of the blood sodium concentration was studied in experiments similar to those described for chloride. Larvae were kept in sodium chloride solutions and measurements made of the blood sodium concentration until the equilibrium concentration had been reached. The results of these experiments are shown in Fig. 5. In contrast to the large increases in blood chloride which occurred in the lower concentrations of the external medium, the blood sodium concentration only increases slightly in solutions of sodium chloride up to about 130 mM./l. when the two concentrations become approximately the same. At higher concentrations of the external medium the rise in the blood sodium concentration is steeper and tends to follow that of the outside solution.

Despite the difference in regulation of the chloride ion as between Sialis larvae and mosquito larvae, the behaviour of the sodium ion appears to be similar. Ramsay (1953 a) gives the mean sodium concentration of the blood of mosquito larvae in distilled water as 87 mM./l. and this rises only to 100 mM./l. when the larvae are placed in 85 mM./l. NaCl, whereas in 171 mM./l. solutions (Ramsay, 1951) the concentration rises to 208 mM./l.

In Sialis larvae, owing to the different powers of regulation of sodium and chloride ion concentrations in the blood, the ratio of sodium to chloride ions in the blood decreases markedly when the larvae are kept in saline solutions. This increase in the chloride concentration must, therefore, occur at the expense of the rest of the blood anion fraction, which as has been seen is probably made up by organic acids. Whether these substances are removed from the blood by excretion or by metabolism or whether they are temporarily stored in the tissues is not known.

The differences in regulatory powers of the larvae with respect to sodium and chloride must be due either to different rates of uptake of these ions from the gut or a differential output in the excretory fluid, or to a combination of both these processes. Both the absorption of sodium and its loss in the rectal fluid have been studied in the same way as for chloride.

The fact that the equilibrium concentrations of blood sodium, shown in Fig. 5, are higher than the sodium concentrations of the external media, suggests that sodium is being absorbed from the gut in the same way as chloride. Experiments were carried out to demonstrate this and to compare the rates of uptake of sodium and chloride. Blood samples were removed from weighed larvae and the sodium and chloride concentrations measured. The larvae were placed in 85 mM./l. NaCl or in tap water as a control after being weighed again. The larvae remained in these media from 2 to 4 days, then they were weighed a third time and the sodium and chloride concentrations measured once more. The control experiments served as a further check on the estimate of the average blood volume and to demonstrate, formally, that sodium as well as chloride was not absorbed from dilute solutions. The results of these experiments are shown in Table 6. Minus signs before the figures for uptake in tap water indicate that there has been an apparent loss. This measurement serves as a control of the accuracy since there should be neither uptake or loss in these cases and thus a difference of 1−2 mμM./mg./day cannot be regarded as significant. The mean values for the uptake of chloride (10·1) and sodium (9·6) from the 85 mM./l. solution indicate that these two ions are taken up together, in approximately equal amounts, through the gut. The individual measurements, also, although different in some cases, are not sufficiently different to be significant in view of the variable results in tap water. Now the potential difference measurements shown in Table 5 indicated a mean value of 18 mV. across the mid-gut wall and the equilibrium concentration of sodium ions in the gut would be 218 mM./l. Since uptake occurs from a 85 mM./l. solution, this transport of sodium ions must involve an active process, presumably occurring in the mid-gut wall. The chloride ions which are absorbed at the same time and in the same amounts may be penetrating by passive diffusion through the mid-gut wall.

Since sodium and chloride ions appear to be absorbed from the gut in equal amounts, the differences which were found in the equilibrium concentrations of these ions in the blood could be due to different rates of excretion of these ions. To test this measurements were made of the sodium concentration of the excretory fluid in the same way as for the chloride. Fig. 6 shows the values obtained for the sodium concentration of this fluid related to the concentration of sodium in the blood measured at the same time. At the normal blood sodium level the concentration of sodium in the rectal fluid is low, but as the blood concentration increases the amount of sodium excreted rises rapidly so that when the blood contains about 170 mM./l. of sodium the rectal fluid concentration is about the same. The sodium concentration of the rectal fluid is approximately proportional to the amount by which the blood sodium concentration exceeds its normal value, and in this respect contrasts with the chloride concentration of the rectal fluid which follows a curve of progressively increasing steepness with increasing blood concentration.

This differential effect in the excretion of these two ions can account for the decrease in the Na/Cl ratio which occurs when larvae are kept in sodium chloride solutions. By comparing Figs. 4 and 6 it will be seen that at the lower blood concentrations of sodium and chloride, the concentrations of these ions in the rectal fluid will be quite different but as the blood concentrations are increased so the rectal fluid concentrations become closer together. This is illustrated by the only three cases in which both sodium and chloride were determined on the same sample of rectal fluid. The concentrations of Na/Cl were 103/48, 90/51 and 140/113 mM./l. respectively.

The normal concentration of potassium in the blood has been shown to be quite low (mean = 5 mM./l.) and the mechanisms by which this low level is maintained have been studied. First, the effect on the blood potassium concentration of keeping the larvae in solutions of potassium chloride was measured. The technique was the same as for the measurements on sodium and chloride and the results for the exposure of the larvae to solutions of 34 and 171 mM./l. solutions of KC1 are shown in Fig. 7, where the blood potassium concentration is related to the number of days in the KC1 solution.

In outside solutions of concentrations up to 34 mM./l. KCI larvae would Uve for a very long time and the measurements of the blood concentration showed that the potassium level was not increased. If the concentration of the outside medium was increased then the blood potassium concentration rose and the larvae did not live for very long. For this reason it was not possible to determine equilibrium concentrations in the same way as for sodium and chloride. In 85 mM./l. KCI solutions the larvae would live for about 2 weeks, and if the concentration was increased to 171 mM./l. KCI they would survive for little more than a week. This is in marked contrast to the behaviour of the larvae in NaCl solutions where they would live for an almost indefinite period in isotonic solutions.

Before the larvae kept in isotonic solutions of potassium chloride finally die they pass through a number of fairly well-defined states in which they show changes in their normal behaviour and these symptoms can be correlated with the concentration of potassium in the blood. Normal larvae, kept in the laboratory, are quite active—they will immediately right themselves by strenuous muscular activity if they are turned over, and if they are touched on the head they give a characteristic response whereby the whole body is jerked quickly backwards.

Major differences in behaviour show themselves when the blood potassium concentration reaches about 10 mM./l. The larvae are much more sluggish than normal and will not give the characteristic backing-away response so readily. Between blood concentrations of 10 − 15 mM./l. the larvae have difficulty in righting themselves if they are turned over, although they can maintain their normal balance if left alone. At concentrations between 15 and 20 mM./l. the larvae are usually lying on their backs, but they are just able to right themselves if they are stimulated. Above 20 mM./l. they are always on their backs and quite unable to maintain their correct position. They are quite still, although stimulation will still initiate certain reflexes, such as the bending of the legs. Any further increase in the potassium content results in the death of the larva. These physiological effects are to some extent reversible; if the blood concentration has not risen above about 15 mM./l. a return of the larva to tap water and subsequent reduction of the blood potassium level appears to restore it to normal, but if the rise has been greater than this the larvae may recover some of their former activity, although the restoration is not complete and some permanent damage has clearly been done. In as much as the physiological effects of increased blood potassium first appear as modifications of normal behaviour patterns and muscular action remains intact, it would seem that the initial effects are on the central nervous system rather than the muscular system. If this is so, then this insect, unlike the locust (Hoyle, 1953) cannot have the whole of its nervous system surrounded by a sheath acting as an effective barrier to the diffusion of potassium ions.

The uptake of potassium and chloride from the gut was studied by means of experiments similar to those described for sodium and chloride. A known amount of blood was removed from a weighed larva and the potassium and chloride concentrations measured. The larva was replaced in a solution of 34 mM./l. KC1 weighed again after 2 days and the blood concentrations measured once more. Calculations of the uptake of chloride under these conditions, in four experiments, gave the following values: 3 · 7, 3 · 8, 3 · 2 and 5 · 4 m μ M./mg. body weight/day. These values are very similar to those obtained for the uptake of chloride from 34 mM./l. NaCl solutions. Thus the chloride uptake seems to be the same from both the chloride solutions and again there is no evidence of any active process being involved. Similar calculations, however, for the simultaneous uptake of potassium by the gut give a considerably lower value—0·8, mean of three readings compared with the mean of 4 mpM./mg. body weight/day for the chloride. The production of excretory fluid, however, continues during this absorption and this fluid, as will be shown, may contain large amounts of potassium, so the value deduced for the uptake actually represents the difference between the uptake and the loss in the excretory fluid. This loss, however, is not likely to be greater than 1·6 m μ M./mg body weight/day so that it is possible that the true uptake of potassium is still considerably less than that of the chloride. Further experiments would be required in order to establish this firmly. The concentration of the outside medium in these experiments is much greater than that of the blood (34 mM./l. against 5 mM./l.) and therefore, despite the direction of the potential difference across the gut wall, the uptake could be due to passive diffusion. Again, more experiments are required before the presence of an active process can be altogether excluded, but it seems clear that if a transporting mechanism is present for potassium, it does not absorb this ion as fast as does the corresponding mechanism for sodium ions.

In solutions of KC1, potassium ions will be entering the larvae through the cuticle and through the gut. The facts that the blood potassium concentration does not rise in an outside concentration of 34 mM./l. KC1 and that in an isotonic solution of KC1 it increases only slowly compared with the increase in chloride concentration suggest that large amounts of potassium can be eliminated in the excretory fluid. This was demonstrated by measurements of the potassium concentration of the excretory fluid made on larvae living in solutions of KC1. Some of these larvae were normal, others had their mouths waxed over in order to prevent any contamination of the excretory fluid with potassium chloride solution taken in through the mouth. The results of these experiments are shown in Fig. 8 where the potassium concentration of the excretory fluid is related to the concentration of the blood measured at the same time. The rise in the concentration of potassium is very rapid—small increases in the blood concentration very quickly raise the concentration of the rectal fluid above that of the blood and rise continues, with increasing blood concentration, until a value of about 100 mM./l. is reached. This is very different from the results obtained for measurements of sodium and chloride where the concentrations of these ions in the rectal fluid never increased beyond that in the blood. When the concentration of potassium in the rectal fluid reaches about 100 mM./l. it seems to level off, but more determinations are required on larvae with high blood potassium concentrations in order to establish this.

Similar high concentrations of potassium have been found by Ramsay (1953 a) in the rectal fluid of mosquito larvae kept in 85 mM./l. KC1 for 3 weeks—the mean concentration was 90 mM./l.

Conductivity measurements were also made on the excretory fluid produced by Sialis larvae living in isotonic KC1 solutions. The results are shown in Fig. 9 where the conductivity of the rectal fluid is related to the number of days that the larvae has spent in the isotonic KCI solution. The conductivity, like the potassium concentration also rises steeply and then levels off after about 4 days at a value corresponding to a 200 mM./l. NaCl solution. Now this value is probably not much greater than the conductivity of the blood at the same time, since the normal blood conductivity will have been increased by KCI which has diffused into the blood during the 4 days. Thus the concentration of the rectal fluid increases until it is approximately isotonic with the blood and then no further increase occurs. By reference to Fig. 7 it will be seen that in 4 days the blood potassium concentration will have risen to 15 mM./l. at which level the maximum concentration of potassium in the rectal fluid is reached (Fig. 8). Thus the maximum concentration of the excretory fluid is attained when the potassium concentration also reaches its maximum.

The value for the conductivity, however, is much greater than that due to the potassium alone and the differences could be accounted for approximately if ammonium bicarbonate is present also in the rectal fluid in its normal concentration. If this is so then this very efficient potassium regulating mechanism may be limited by the fact that the rectal fluid cannot become hypertonic to the blood and still excretes the normal concentration of ammonium bicarbonate, even under conditions of extensive potassium excretion.

Just as an equilibrium concentration of any inorganic ion in the larva is established when the amount of that ion absorbed is balanced by that which is excreted, so the water content of the larva will be in equilibrium when the rate of water intake is equal to the rate of elimination of the excretory fluid.

In larvae kept without food daily weight records kept for a period of 3 weeks showed little change. Measurements of water permeability (Shaw, 1955b) indicated that drinking of water did not occur in these circumstances so that the volume of excretory fluid produced must be equal to the volume of water which enters the larva through the cuticle.

Drinking, however, does occur if the blood volume is reduced or if salt is present in the outside medium. A combination of these two conditions has been used as a method for studying the uptake of water from the gut and the relation of this to the simultaneous absorption of salts.

It has already been shown (Shaw, 1955b) that if water is taken into the gut it is absorbed osmotically and that the gut wall is much more permeable to water than the cuticle. No water is taken up from solutions of mannitol isotonic with the blood even if these are swallowed. In the following experiments the effect of drinking isotonic solutions of salts was investigated. The technique was similar to that used for the study of the uptake of salts from the gut : blood was removed from weighed larvae, they were reweighed, placed in one of the isotonic salt solutions and finally a daily weight record kept for a period of 5 − 6 days. The results of the experiments are shown in Fig. 10. Six larvae were used for each experiment but the results for only one or two are given in order not to confuse the figure. There was, however, little variation and the selected examples are typical of the behaviour of the whole group. Where necessary controls were also carried out by repeating the same experiment on a group of larvae with their necks ligatured to prevent drinking. Isotonic solutions of NaCl, KC1, CaCl₂ and NaHCO₃ were used for these experiments. In the case of the NaCl solutions, the weight of the larvae gradually increased, whereas that of the ligatured larvae did not. The increase was continued until the original blood volume had been restored and in a few cases the original weight was exceeded. The larvae were perfectly normal in behaviour during this process of blood regeneration and would continue to live for long periods afterwards. The average rate of water absorption was about 5 % of the body weight per day, and therefore was actually greater than the normal rate of water uptake through the cuticle (4% body weight/day at 20 ° C.; Shaw, 1955b).

In isotonic CaCl₂, solutions no uptake of water occurs and in this respect they resemble the isotonic mannitol solution ; in the isotonic KCI solution the weights of the larvae actually decrease. In the NaHCO₃ solution, water uptake does take place at first, but the larvae do not live long in this solution. The rate of uptake is approximately the same as in the NaCl solution.

Now it is clear that this absorption of water cannot be due to osmotic forces since the solutions are isotonic with the blood and this is confirmed by the fact that water is not absorbed from the mannitol solution. Since the uptake occurs from some salt solutions, it suggests that this water transport is associated with the simultaneous absorption of salts which is occurring. It cannot be associated with the uptake of chloride since this is certainly occurring in the KCI solutions. Active transport of sodium ions has been shown to be taking place in NaCl solutions, and since the water uptake occurs also in this solution and in the NaHCO₃ solution, the simplest explanation of these facts is that this water transport is in some way linked with the active uptake of sodium by the gut wall.

Since chloride is absorbed at the same rate as sodium from NaCl solutions, and water is absorbed at the same time, then the total absorption can be regarded as that of a solution of NaCl in water. The concentration of chloride in this solution was measured by repeating the above experiments on the regeneration of blood from an isotonic NaCl solution, but at the same time measuring the blood chloride concentration before and after the experiment and hence calculating the chloride uptake. The experiments were repeated using solutions containing different concentrations of NaCl but maintained isotonic with the blood by the addition of mannitol. By measuring the water uptake and the simultaneous absorption of chloride the concentration of chloride in the absorbed water was calculated and the results are shown in Table 7. The concentration of chloride in the water in all the experiments is roughly constant and bears no relation to the concentration of NaCl in the meditan from which it was absorbed. The mean value for the seventeen experiments is 155 mM./l. (s.D. + 24 mM./l.), and thus is considerably greater than the concentration of either chloride or sodium in the blood. It is, however, nearly isotonic with the blood, and since the experiments covered a fairly wide range of concentrations of the outside medium, it seems very likely that sodium and chloride are absorbed from the gut in the form of a solution of NaCl which is approximately isotonic with the blood.

There are, then, three ways in which water may enter the body of the larva. First, water is taken up osmotically through the cuticle at a rate of about 4% of the body weight per day at 20° C. Secondly, water may be absorbed osmotically through the gut wall when water or dilute solutions are swallowed, and intake by this method may be large by reason of the high permeability of the gut. Finally water may be transported across the gut wall by a process which is probably linked to the active absorption of sodium ions.

The only known route for the loss of water from the larvae is by means of the excretory fluid. Since, as has been demonstrated in the preceding section, the water intake varies according to circumstances it is interesting to find to what extent the larvae are able to regulate the volume of the excretory fluid. Of the possible factors which might be involved in controlling rectal fluid volume, two have been investigated experimentally. The first was effect of the reduction of the blood volume which had been found to stimulate water uptake and the second the direct effect of altering the water intake.

In the first series of experiments, blood was removed from weighed larvae in the usual way and they were replaced in tap water, with their necks ligatured to prevent drinking which would normally occur under these circumstances. Daily weight recordings were taken and the results are shown in Fig. 11 where the weight changes expressed as percentage of the initial body weight are related to the time after they were placed in tap water. No significant weight changes occur, and this must mean that the normal volume of excretory fluid is still being produced since the water intake through the cuticle will not be affected by the experiment. This is confirmed by the fact that when a ligature is tied round the last abdominal segment of these larvae to prevent excretion then the weight increases in an amount corresponding approximately to the expected water uptake through the cuticle. Therefore one can conclude that a reduction of the blood volume has no effect on the production of excretory fluid if the normal rate of water intake is maintained.

If the same experiment is repeated on larvae whose necks have not been ligatured then drinking occurs, and the increase in body weight is very large; the blood volume is practically restored in one day (Shaw, 1955b)-In this case the water intake is increased beyond the normal level, but this does not stimulate an equivalent output of excretory fluid. It should be noted, however, that this increased uptake of water need not necessarily lead to a dilution of the blood because of the operation of the compensatory mechanism (Beadle & Shaw, 1950), whereby the blood osmotic pressure can be maintained, probably by the addition to the blood of amino-acids.

The effect of reducing the water intake on the production of excretory fluid was studied. Osmotic uptake of water was prevented by putting larvae into solutions isotonic with the blood and weight records were taken. Some larvae were also kept in liquid paraffin in which they would live for many weeks and in this way all contact with external water was prevented. In all experiments larvae with and without neck ligatures were used, and the results obtained for the former group are shown in Fig. 12, where as before one record only is presented in each case. The record shown, however, is typical of the behaviour of each group.

The loss of weight of the larvae in isotonic KCI and mannitol occurs at a rate of about 6 % of the body weight per day, although this rate tends to fall off as the blood volume is reduced and there is no doubt that, at least at the start, excretory fluid is being produced at the normal rate despite the fact that there is no water intake. In the case of the liquid paraffin the rate of weight loss is somewhat less, but further experiments would be required to establish the significance of this difference. However, in the isotonic NaCl solution the behaviour of the larvae is quite different, for in all the animals (20) no large fall in weight occurred, and in many the initial weight was maintained. In this solution, therefore, the production of excretory fluid must have practically ceased and this conclusion receives support from the fact that whereas it is easy to collect a sample of rectal fluid from larvae kept in isotonic KCI solution it is never possible to collect very much from the specimens in the NaCl solution.

In the experiments with unligatured larvae the results were approximately the same except again in the case of NaCl : here the larvae actually increased in weight in the first few days and in this case there is little doubt that intake of water through the gut wall has been taking place by means of the transport mechanism discussed above.

The effect of the NaCl solution in reducing the flow of excretory fluid is not immediately reversed when the larvae are returned to tap water. Larvae kept in isotonic NaCl reach a constant weight after a few days ; if they are now transferred to tap water (Fig. 13) then the weight increases again during the first day or two, then remains constant or falls again. This weight increase must be due to the water taken up through cuticle which has not been excreted as rectal fluid.

As with uptake of water from the gut in isotonic NaCl, this reduction in excretory fluid volume appears to be due to some activity on the part of the sodium ions. The only obvious difference between the isotonic NaCl and KCI solutions is that in the former case sodium will be diffusing in to the larva. It is possible that it is the increase in sodium ion concentration of the blood which brings about the reduction in excretory fluid production—this is supported by the fact that if the sodium concentration of the blood is raised by keeping the larvae in NaCl solutions then the flow of rectal fluid does not start again immediately after the larvae have been replaced in tap water (Fig. 13)—the blood concentration may have to be reduced again before the fluid can be produced once more.

Water output, therefore, in the form of the excretory fluid normally balances the water intake through the cuticle. The volume of this fluid is not altered by changes in the blood volume nor is it affected by a reduction in the normal intake, nor does it keep pace with large increases in the rate of water uptake. It is, however, reduced by the influx of sodium ions into the blood and may be controlled by the concentration of sodium in the blood.

Sialis larvae live for long periods in NaCl solutions which are isotonic with, or hypotonic to, the blood. In more concentrated solutions than this, however, they will not live so long. For example, in solutions which are 50% greater than isotonic the larvae only live for a fortnight or so, and in solutions twice as concentrated as the blood, they die in a few days. The water relations of the larvae living in these solutions have been studied. Weight changes have been recorded for larvae, both with and without neck ligatures, placed in solutions containing 0, 85, 171, 256 and 342 mM./l. NaCl, and Table 8 gives the average daily weight changes for a period of 3 days.

In tap water larvae with their necks ligatured show a small loss of weight, and this rate of loss is found also in similar larvae in 85 and 171 mM./l. NaCl solutions. In contrast the normal larvae lose no weight in tap water or in 85 mM./l. NaCl, and in the isotonic solution a gain in weight is found due to the uptake of water from the gut and the reduction of the rectal fluid volume. In hypertonic solutions, the normal osmotic uptake of water is replaced by an osmotic loss. This is demonstrated in the larvae with ligatured necks in the 256 and 342 mM./l. solutions, where increased losses in weight are found. The difference between these losses and the loss of the ligatured larvae in tap water are approximately equal to the expected loss of water by osmosis through the cuticle.

The behaviour of the normal larvae in 256 and 342 mM./l. solutions is interesting. In the former solution the larvae are just able to maintain their body weight and the uptake of water from the gut must be just sufficient to balance the loss by osmosis through the cuticle and also through the gut wall itself. The eventual death of these larvae is probably due to the rise in blood osmotic pressure caused by the penetration of salt. In the 342 mM./l. solution the loss of weight is very great and the larvae die in a few days, presumably from dehydration. The weight loss is much greater than from the ligatured larvae in the same solution, and at this concentration the mechanism of water uptake from the gut must break down so that the recorded loss is due to exosmosis through the gut, in addition to that through the cuticle.

The composition of the blood as far as the cations are concerned is not very different from that found in many fresh-water and terrestrial animals, for sodium accounts for a large proportion of the cation fraction and potassium is present only in relatively small amounts. The high concentrations of calcium and, particularly, of magnesium are the most unusual feature, although this seems to be a characteristic of insect blood. The anion composition is peculiar, however, in that chloride represents only a small fraction, and this again seems to be a common feature among insects. It would be of great interest to know what substances make up the rest of this fraction and in what way their concentration in the blood is regulated. Results already discussed suggest the presence of organic acids, which may well be dicarboxylic amino-acids. These acids have already been found in a number of terrestrial insects—for example, Auclair & Dubreuil (1953) found quantities of glutamic acid varying from 10 to 200 μ g. per 100 μ. and aspartic acid from 4 to 100 μ g. per 100 μ l., in the blood of a variety of insects. Little is known of the organic composition of the blood of aquatic insects: Raper & Shaw (1948) described the most important amino-acids in the blood of the aquatic larva of Aeschna cyanea. No appreciable quantities of these dicarboxylic acids were found, but in these larvae the chloride concentration is unusually high (94 mM./l.).

The maintenance of the normal composition of the blood, at least in respect of the three ions which have been studied, is brought about by the activity of the Malpighian tubules and rectum in producing an excretory fluid of varying composition. In the formation of this fluid in unfed larvae it would appear that a system of potassium secretion by the Malpighian tubules and subsequent reabsorption in the rectum, as described by Ramsay, first for the mosquito larva (1953 a) and then for a variety of other insects (1953 b), operates, although the concentration of potassium in the tubules may be rather lower in Sialis than in the other insects. This circulation still takes place despite the fact that a concentrated excretory fluid is being produced. It is possible that some sodium is reabsorbed in the rectum, but there is no evidence that chloride is present in the fluid at all. Whether this is related to the high concentration of bicarbonate present is not known, but it would be very interesting to have comparable data on the regulation of the chloride ion in other aquatic insects.

If the regulatory powers of Sialis larvae with respect to the three ions are compared, differences are noticed which may well be correlated with the physiological requirements of the larvae for these ions. Thus for the potassium ion, the excretory system is well adapted for its speedy removal from the blood if the blood concentration rises. This can be associated with the very marked detrimental effect that a small rise in the blood concentration of potassium has on the neuro-muscular system of the larvae. The removal of the ion is brought about in two ways : first, the volume of rectal fluid does not decrease with increasing concentration of potassium in the blood, despite the fact that at the same time the blood osmotic pressure may be increasing ; and secondly, the concentration of the rectal fluid is rapidly increased (so that it becomes roughly isotonic with the blood) by the addition of large amounts of potassium which raise the concentration of this ion in the fluid to a level well above that of the blood.

The regulation of the chloride ion concentration of the blood, although effected by the same system, is of a different nature. Increases in the blood concentration of over 100% of the normal value only result in a small output in the excretory fluid and as a result, if chloride is present in the outside medium, chloride tends to be accumulated in the blood. Now large increases in the blood chloride level have no adverse effect on the larvae, and it would appear that, for the chloride ion, the regulatory system is adapted for the conservation of chloride. This idea also receives support from the fact that chloride appears to be completely absent from rectal fluid of the unfed larvae. This emphasis on conservation of chloride may well be associated with the absence from these larvae of any structure capable of absorbing chloride from dilute solutions. It is interesting to find that in the mosquito larva, where such structures do exist (Koch, 1938), that the regulation of the chloride concentration of the blood of larvae in saline solutions is much more rigid (Wigglesworth, 1938).

The regulation of the sodium ion concentration occupies, in some ways, an intermediate position between that of potassium and that of chloride. Regulation is effected by excretion of sodium in the rectal fluid and the concentration of sodium in this fluid rises more rapidly than does the concentration of chloride, for equivalent increases in the blood concentration, but does not increase nearly as quickly as the concentration of potassium. Now increases in the blood sodium concentration have no adverse physiological effect on the larvae but will lead to an increase in the total osmotic pressure of the blood. Larvae can only tolerate relatively small increases in this, and therefore they are unable to store sodium in the blood in the same way as chloride, although conservation of sodium is as much a problem as the retention of chloride. Sodium storage can be effected, however, by a different mechanism which makes use of the special effect of the sodium ion on the water balance of the larvae. If sodium is absorbed from the gut in the form of an isotonic solution and the influx of sodium into the blood somewhat reduces the rectal fluid volume, then this solution could be retained in the blood by an increase in the blood volume. Since the solution is isotonic, the osmotic pressure will not be raised, and since the sodium concentration of the solution is not much greater than that of the blood, the blood sodium concentration will also only be raised by a small amount. The behaviour of the excretory system in this respect, recalls a similar behaviour of the vertebrate kidney: if isotonic NaCl is injected in the blood of a mammal no diuresis occurs, whereas an immediate diuresis would follow an injection of water or KC1 (Smith, 1937, P. 155).

In all fresh-water animals losses of salts are inevitable ; the losses are small from Sialis larvae because of the relative impermeability of the cuticle. In the absence of an external ion-absorbing mechanism, these losses must be made good by absorption of salts from the food and, in these larvae, this can be associated with the well-developed mechanism for the uptake of NaCl found in the gut.

Finally, the fact that Sialis larvae are unable to live in saline solutions more concentrated than the blood can be explained from a knowledge of their powers of ion and water regulation. The necessary physiological mechanisms are present but are not well enough developed as, for example, they must be in Aedes detritus (Beadle, 1939; Ramsay, 1950). Thus drinking is necessary to make good water losses if the blood is hypotonic to the external medium; drinking does occur but only up to concentrations of the outside medium 50% greater than the blood. This mechanism resembles the drinking occurring in marine Teleosts (Smith, 1930). Since uptake of salts through the gut and diffusion through the cuticle occurs in saline solutions a hypertonic excretory fluid would have to be produced. Although the larvae are able to concentrate salts in the rectal fluid they are not able to produce a hypertonic fluid.