Electrical Potential Differences and Ionic Transport in the Larva of the Mosquito Aedes Aegypti (L.)

A considerable amount is now known about the movements of ions which are entailed in the processes of ionic and osmotic regulation in the freshwater larvae of Aëdes aegypti (Wigglesworth, 1933a–c, 1938; Ramsay, 1950, 1951, 1953; Treherne, 1954; Stobbart, 1959, 1960, 1965, 1967, 1971 a–c). It has been shown that the larvae maintain ionic balance in very dilute solutions by means of sodium and chloride ‘pumps’ which are situated in the anal papillae, and which are apparently based on the operation of carrier molecules. These pumps, which seem to be under hormonal control, actively secrete Na⁺ and Cl⁻ into the haemolymph (in some circumstances in part exchange for H⁺ and HCO₃⁻ (or OH^-) respectively), and they have a high specificity and affinity for the ions which they handle. Large exchange components in the Na⁺ and Cl⁻ fluxes which they generate characterize the pumps, which seem to exhibit typical enzyme-substrate kinetics as described by the Michaelis-Menten equation. In contrast, information about the electrical events which accompany these ionic movements is very scanty, although it does demonstrate that in the larvae of both A. aegypti (Stobbart, 1965) and A. campestris (Phillips & Meredith, 1969) Na⁺ and Cl⁻ are actively transported against adverse electrochemical gradients.

This dearth of electrical data prompted me to undertake the work which I wish to report here. My major aim in this work was to obtain measurements of the potential difference between haemolymph and medium under conditions which were similar to those under which I made measurements of ionic movements in earlier investigations (Stobbart, 1959, 1960, 1965, 1967, 1971b), so that the electrical and chemical data could be combined to shed further light on the processes of ionic regulation. The effects of some metabolic inhibitors and sodium-free solutions on sodium fluxes are also reported here. Following earlier practice I use the terms ‘influx’ and ‘outflux’ to describe ionic fluxes found with radioactive tracer, and the terms ‘net loss’, ‘loss’, ‘net uptake’ and ‘uptake’ to describe net movements of ions.

The 4th-instar larvae used in this work were of stock L (Stobbart, 1967) and were reared by two methods. In the first they were reared to the 4th instar in a standard rearing medium (standard RM) containing 2 mM/1 Na and 3·9 mM/1 Cl (Table 1 and Stobbart, 1959). Such larvae were then: (a) starved for 60–72 h and then treated with de-ionized or distilled water (DW) at a density of 1 larva/2 ml for 24 h (Stobbart, 1965, 1967) and were then used for measurements of potential difference (PD) in various concentrations of NaCl, Na₂SO₄, and KC1; (b) were used in the fed or starved state (without treatment with DW) to investigate the effects of some metabolic inhibitors and sodium-free solutions on sodium fluxes. In the second method larvae were reared to the 4th instar under salt-deficient conditions (Stobbart, 1960) and were used in the fed state for measurements of PD during the uptake of ions from solutions of 2 mM/NaCl, 2 mM/1 KC1, and 1 mM/1 Na₂SO₄. Such larvae are henceforth called salt-deficient. All the experiments were performed at 28 °C, though the measurements of PD were sometimes made at slightly different temperatures.

The measurements of PD between the haemolymph and medium were made on larvae set in beds of beeswax/resin cement using the method described earlier (Stobbart, 1965), except that: (a) a Vibron electrometer (model 33B–2) was substituted for the modified pH meter, and was used with electrodes of considerably smaller tip diameter; (b) various external media were placed around the anal papillae depending on circumstances; (c) measurements of PD were taken for a period of 1 min; (d) permanent records of the variation of the PD during this time were obtained with a pen recorder which was connected to the voltmeter. The mean values of the PD during the 1 min period were obtained from these records by graphical integration. The measurements were made at temperatures between 27·5 and 31 °C.

The larvae were set in the cement and allowed to settle down for 30 min with the anal papillae immersed in the DW in which the larvae had been kept for the previous 24 h. Suitable concentrations of the salt solutions (0·00155–5 mg monovalent ion/1) were now placed around the anal papillae, and after about 10 min (20 min in the case of the two most dilute solutions) the PD was measured. These times correspond roughly to those which were used in the measurement of ionic fluxes (Stobbart, 1965, 1967). Ten larvae were used for each concentration for which the mean PD and the standard error (s.E.) were calculated.

The larvae were set in the cement and allowed to settle down for 30 min with the anal papillae immersed in the very dilute RM. The RM was then replaced by one of the salt solutions, and the PD was measured after a suitable time. As it was not possible to make the measurements at accurately predetermined times, the data were not grouped with respect to time, and are shown in the figures as measurements from individual larvae.

In order to economize on time and effort these measurements were made on living larvae, used either singly or in groups of up to four. The larvae, which were at any rate initially in a steady state with respect to ions, had been previously labelled with ²²Na by rearing them (and subsequently starving them if necessary) in standard RM labelled with ²²Na at a specific activity of about 0·075 Ci/M Na. Initially the larvae were rinsed with DW and dried quickly on filter paper. They were then placed, ventral surface down, in definite positions within a marked area of a planchette (the same planchette was used throughout) and were immobilized by moving their rear ends forwards with a paint brush so that the anal papillae touched and adhered to the head capsule. The planchette was now placed beneath an end-window Geiger counter connected to suitable scaling equipment, and the radioactivity was measured over a period of about 2 min. The number of pulses registered generally fell between 2000 and 8000. The larvae were now placed in ‘infinite’ amounts of appropriate inactive experimental media, either Na-free or containing 2 mM/1 Na, from which they were removed when necessary for further measurements of radioactivity as described above. Previous experience with this sort of technique (Stobbart, 1959) has shown that the constancy of the geometry of the sample is adequate. In cases where groups of larvae were used the planchette was placed sufficiently far from the window of the Geiger counter to eliminate dead-time errors.

[Na_in] for the larvae used in these measurements was always approximately 100 mM/1 (haemolymph concentration) or 80 mM/1 in the case of papilla-less larvae (Stobbart, 1959), for even if they were placed in DW, the stay in it was short enough for sodium losses to be negligible. If the larvae are placed in an infinite amount of a sodium-free solution, equation (2) will still hold (until the larvae start to deteriorate), but now represents the declining sodium concentration in the larvae (i.e. [Na_in]), the rate of decline being proportional to ⁠. On the other hand, if larvae are placed in a solution which contains sodium and a substance which inhibits sodium influx, equation (2) will hold as a first approximation. In this situation represents the declining sodium concentration in the larvae but the rate of decline is proportional to ⁠; as [Na_out] (i.e. 2 mM/1) is low relative to 80–100 mM/1 haemolymph concentration initially ⁠. Equation (2) shows that when log₁₀ is plotted againt t, a straight line of slope K_out/2·303 will result, and this is the way in which the data have been presented in the figures.

Estimates of the surface area of the anal papillae were made by measuring the lengths and breadths of papillae: (a) in photographs of living animals (taken at a known magnification); (b) in fixed and stained preparations using a calibrated filar micrometer-eyepiece. Each papillae was then considered to be a cylinder with one end covered, and its surface area calculated accordingly, assuming breadth/2 = radius. It is difficult to estimate the accuracy of the method, but the results should certainly be of the right order of magnitude.

Anal papillae were removed from the larvae, when required, by treatment for 5 min with 5% NaCl (Wigglesworth, 1933b;,Stobbart, 1959); unless a statement is made to the contrary it is to be understood in the following sections that the larvae had intact papillae.

Measurements of the total sodium, potassium, and chloride concentrations in whole larvae were made with the techniques described earlier (Stobbart, 1967, 1971b). The lines through the points in the figures were drawn in by eye unless a statement is made to the contrary, and the usual convention of significance was observed in the statistical tests.

The PD, specimen traces of which are illustrated in Fig. 1, was found to vary considerably with time and also between one animal and another. The variation with time was most marked in solutions of NaCl and KC1, and generally took the form of drifts (which were often reversed) of the PD in one sense or the other. Occasionally the PD changed sign (serials 1, m and/), Fig. 1). Of the larvae placed in NaCl and KC1 the PD showed the greatest amount of drifting in those which were deficient in salts (serials y, z and z′, Fig. 1). In Na₂SO₄ the PD was in general much steadier, on occasion being more or less constant during the period of measurement.

In all the solutions used fairly rapid irregular oscillations of PD sometimes occurred (serials a, t and z′, Fig. 1). Occasionally these could be correlated with muscular contractions of the larvae, but often they occurred when the larvae were apparently quite still.

The larvae used in these measurements were either fed or starved, but had not had their ionic pumps stimulated by treatment with DW or by being reared in salt-deficient conditions. They were reared in RM, and 30 min after being set in the beeswax/resin mixture various salt solutions were placed around the anal papillae. The PD was then measured 10–17 min later; 10 larvae were used for each measurement except in one case. The mean values, together with their standard errors (s.E.), are displayed in Table 1.

The results are rather variable, but an analysis of variance performed on the mean values shows that there is a significant difference between the fed and starved larvae, the latter showing PD values which are significantly more positive. The analysis also shows that there are significant differences between the larvae in the different solutions. The significances of the differences between all possible pairs of mean values are displayed in Table 2. In general the PD is more positive in Na₂SO₄ and more negative in KCl, particularly in the case of starved larvae. However, the PD may also be negative in the standard RM.

As shown earlier (Stobbart, 1965, 1971b) the pretreatment with DW causes the larvae to lose small amounts of their ions, and this loss activates the sodium pump fully and the chloride pump to a very considerable extent so that the larvae can reclaim some of the sodium they lost, and attain sodium balance after some 20 h of the treatment. Starved larvae stimulated or balanced in this way were used earlier (Stobbart, 1965, 1967) for measurements of Na⁺ and C⁻ fluxes as a function of the external concentration of these ions, and in the present investigation measurements of PD have been made over a similar range of concentrations (0·00155, 0·156, 0·313, 0·625, 1·25, 2·5 and 5·0 mg monovalent ion/1).

The results (which will be considered in relation to the ionic fluxes later) are shown in Fig. 2 and Tables 9 and 11. The PD is never very large. In the case of larvae placed in solutions of NaCl or Na₂SO₄ there is a change from about − 33 mV at the lowest, to about + 30 mV at the highest concentration, while in larvae placed in KC1 the change is from −8·4 to about −26 mV (haemolymph side relative to exterior in all cases).

Total body K concentration cannot be used to estimate haemolymph concentration (Stobbart, 1965, 1967), but the K concentration of the haemolymph in larvae treated with DW has been found to be about 3 mM/1 (Stobbart, 1967) and is likewise effectively constant over the periods of measurement. This value has been used (as above) to calculate the potassium potentials, which differ appreciably from the PD observed in KCl only in the five most dilute solutions (Fig. 2).

As K⁺ is lacking (except as traces of impurity) from NaCl, K⁺ and Cl⁻ from Na₂SO₄, and Na⁺ from KCl, the potentials of these ions (positive for cations, negative for anions) in these solutions must have been very high. No attempt has been made to estimate these potentials accurately, but if we assume a level of impurity of 0·1 % on a gram-ion basis, then the potentials would be roughly +166 mV (K⁺), +256 mV (Na⁺) and −238 mV (Cl⁻) in the 5 mg ion/1 solutions, and higher in the more dilute ones. It must be noted that the ionic potentials illustrated in Fig. 2 can only be related to PD observed in solutions containing the ion under consideration.

Fed larvae reared under salt-deficient conditions were used for these measurements. As described earlier (Stobbart, 1960, 1971b) such larvae are smaller than normal ones, have greatly hypertrophied anal papillae, and have the sodium pump fully activated; they contain about 0·3 of the normal concentration of sodium in the haemolymph, and about 0 41 of the normal total body sodium concentration, but in spite of the salt deficiency they are as active as normal larvae and seem quite healthy (Stobbart, 1960, 1971b). The measurements were made over a 5 h period for larvae in NaCl (at the end of which time Na uptake is largely complete (Stobbart, 1960, 1971b)) and over an 8 h period for larvae in Na₂SO₄ and KCl. In addition to the measurements of PD, the total body concentrations of Na, K and Cl were measured before and after the period in which ions were taken up.

The results for total body concentrations are given in Table 3 which shows (in agreement with earlier work (Stobbart, 1965, 1967, 1971a)): (i) that Na is taken up from NaCl, and more slowly from Na₂SO₄, in which there is a small loss of K (in part exchange for Na); (ii) that Cl is taken up much more slowly from KCl.

The results for the measurements of PD are illustrated in Figs. 3–5. The PD measurements made at zero time were actually made with the dilute rearing medium (Na concentration 0·001–0·002 mM/1) surrounding the anal papillae. However, as the figures show that there was no immediate change in PD when the experimental solutions were put round the papillae, these initial measurements may be regarded as the PD observed at zero time in the experimental solutions. In the case of larvae in 2 mM/1 NaCl the PD, which is initially ∼ − 43 mV, rises to about +37 mV after 1 h. After this time there is no significant correlation between PD and time, and the mean value is +37·23 mV ± 2·55 (S.E.), N (degrees of freedom) = 50. There is considerable scatter in the data, but it must be remembered that the readings are from individual larvae. For larvae in 2 mM/1 NaCl the contemporaneous changes in the concentrations of haemolymph and total body sodium are known in detail from earlier work (see Fig. 13) and in outline from the present work (Table 3) and these data have been used to calculate the sodium potentials for suitable times. These potentials (based on the earlier analyses) are also shown in Fig. 3. As it is not known how f changes with concentration inside the larvae it has been taken as 1, but little change is caused in the potentials if it is assumed to be 0·8. Strictly speaking the haemolymph concentrations should be used for calculation of the sodium potentials, but as ∼ 90 % of sodium is in the haemolymph which comprises at least 62·5 % of the body weight (Stobbart, 1965), use of the total body sodium concentrations leads to only a slight underestimation of the potential. As before we see that the observed PD is very different from the calculated sodium potential (Fig. 3, Table 4). The PD data will be considered in relation to sodium fluxes at a later stage. Earlier measurements of PD (Stobbart, 1965) in which larvae were pretreated with DW and then returned to RM, gave negative values of PD during the subsequent uptake. This apparent discrepancy has not been investigated, but the difference could be due to a preponderance of Cl uptake in RM which contains 3-9 mM/1 Cl, compared to 2 mM/1 Na; however, the earlier data were based on only a small number of larvae.

The contemporaneous changes in haemolymph chloride concentration are not known for these larvae, but Table 3 shows that there was a very considerable increase in total body chloride concentration over the 5 h period. As ∼ 75 % of the total body chloride is in the haemolymph which comprises at least 62·5% of the body weight (Stobbart, 1965, 1967) and as the fraction of chloride in the tissues seems from the present, and from earlier work, to be lower in salt-deficient larvae (Stobbart, 1965, 1967; Wigglesworth, 1938), we may reasonably use total body concentrations for obtaining minimal estimates of initial and final chloride potentials. These chloride potentials (estimated using the data in Table 3) are, as Fig. 3 and Table 4 show, also very different from the observed values.

Potassium concentration has not been measured in these larvae, but it presumably remains fairly constant at about 3 mM/1 (Stobbart, 1967; Table 3). As only very small traces of K⁺ can have been present in the external NaCl solution it is clear that the potassium potential must have been positive and very high − about +190 mV if we assume the same level of impurity as before.

A similar picture is shown by the larvae in 1 mM/1 Na₂SO₄, and again there is no significant correlation between PD and time after 1 h. The initial value of PD is as before ∼ − 43 mV and the mean value after 1 h is ± 5·62 mV + 3-33, N = 80. For these larvae the contemporaneous changes in haemolymph sodium and chloride are not known, but Table 3 shows that during the 8 h period over which the PD measurements were made there was a significant increase in total body sodium concentration, but no significant change in that of chloride. Using these sodium values (see above) in lieu of haemolymph concentrations, minimal estimates of sodium potentials have been calculated, and as Fig. 4 and Table 4 show they are very different from the observed values. As Cl⁻ and K⁺ are lacking in the external solution (apart from traces present as impurities) the potentials for these ions must have been very high, about − 250 mV and +190 mV respectively, assuming the same level of impurity as before.

The situation is different in larvae placed in 2 mM/1 KCl. The initial PD is again ∼ − 43 mV, but it changes gradually to ∼ −23 mV over the 8 h period. The results lend themselves to an analysis of linear regression, and there is in fact a highly significant correlation between PD and time. The regression line, illustrated in Fig. 5, is defined by the equation

The haemolymph potassium concentration has not been measured in these larvae, but earlier work (Stobbart, 1967) suggests that it would have been between 2 and 4 mM/1. A value of 3 mM/1 yields a potassium potential of +10·5 mV, which, as Fig. 5 shows, differs markedly from the observed PD. As Na⁺ is ‘lacking’ from the external solution the sodium potential must have been very high, about +252 mV assuming the same level of impurity as before.

As far as Na⁺ and Cl⁻ are concerned the observed PD is very different from the calculated ionic potentials, and therefore cannot be accounted for solely by the passive movement of these ions across the anal papillae. This conclusion is supported by the lack of immediate change in PD when salt-deficient larvae are transferred from the dilute rearing medium to 2 mM/1 NaCl and KCl, and 1 mM/1 Na₂SO₄. Likewise the observed high internal sodium and chloride concentrations cannot be maintained by the observed PD as the following considerations show.

The situation is different with respect to K⁺. Inspection of Fig. 2 shows that although the observed PD is more positive than the equilibrium PD in the three most dilute solutions, it is adequate, or more than adequate, to maintain the internal potassium concentration in the remainder (equal to or more negative than the equilibrium PD). Likewise Fig. 5 shows that in 2 mM/1 KCl the best estimate of the observed PD during the 8 h period always exceeds the value of −10·5 mV needed to maintain the internal potassium concentration. We may therefore conclude that in the more concentrated solutions uptake of K⁺ probably occurs passively down its electrochemical gradient, although it may enter the larvae through a carrier system (Guthrie & Burzynski, 1972) and in fact K⁺ uptake has not yet been demonstrated in the absence of Cl⁻ (Stobbart, 1967).

The semilogarithmic graphs of Figs. 6–9 illustrate the decline in shown by larvae (which were initially in the steady state) when they were placed in RM and various other solutions. The lines drawn through the points are the regression lines calculated for the readings obtained in a given solution. As mentioned earlier, the slopes of the lines (i.e. the regression coefficients) represent K_out/2·303. Therefore K_out may be found without trouble and its standard error may be computed using the techniques of linear regression.

Seventeen measurements of K_out into RM were made for fed larvae both before and after entry into other media, and eight were made for starved larvae. The data are illustrated in Figs. 6, 7, 8,a and in Figs. 8 b, c, 9a respectively, and are summarized in

Table 5. K_out, in agreement with earlier work, is between 0-08 and 0-09 h^-1 and is much greater in fed than in starved larvae, but the mean K_out value for starved larvae is considerably higher than that reported earlier, a fact which is probably due to a less complete starvation of the larvae in the present investigation (cf. Stobbart, 1959).

The figures demonstrate considerable variation between individual larvae, and sometimes in fed larvae (Fig. 8 A (i) and (ii), Table 6) an increase in K_out after brief treatment with DW. In the case of one fed larvae (Fig. 6 A (ii)) in which the sodium exchange was followed more or less to completion, K_out changed significantly and abruptly from 0·0757 to 0·1338 h⁻¹ after 9·7 h of observation.

The results are displayed in Table 6 and Figs. 7 A, B, D (i), 8A (i), (ii), B (i) and C, and show that in both fed and starved larvae DW caused a drastic reduction in K_out to a value which was in general about one-sixth of the original. This reduction was, within the limits of experimental error, immediate. Re-application of RM caused an immediate return of K_out to approximately the original value.

This was investigated in one fed larva (Table 6 and Fig. 7B) which was placed in RM in which the NaCl had been replaced by choline chloride. The effect was an immediate reduction in K_out to the value found in DW. However, upon return to normal RM, K_out did not return to its normal value as it did after treatment with DW, which indicates that the anal papillae had been inactivated. Inspection of the papillae showed that the 15 h of treatment with choline chloride had in fact made their cytoplasm yellowish and more granular than normal.

This was investigated in three fed larvae (Tables 6 and 7, and Figs. 6B and 7 A). At a concentration of approximately 2 mM/1 in RM, 2,4-dinitrophenol (DNP) increased the outflux of sodium, while at a concentration of approximately 1-25 mM/1 it reduced it. At these concentrations DNP immobilized the larvae within a few minutes; after about 13 h the cytoplasm of the anal papillae was yellow, more granular than normal, and had withdrawn from the cuticle – there was no heartbeat and the larvae were rather turgid.

At a concentration of 0·0125 mM/1 DNP caused an immediate reduction in K_out in the starved larva. There were no immediate gross effects on the larva (indeed separate tests showed that larvae continued to swim actively at this concentration for at least 40 h). However, by 39 h the anal papillae had dropped off revealing black scar tissue at their bases. Inspection of the data for the DNP solution (Fig. 7A) suggests that the papillae were probably inactivated immediately after entry into the DNP solution (at 24·2 h), and that increased losses of sodium occurred after 29·8 h when the papillae were presumably starting to deteriorate.

This was investigated at a concentration of 5 mM/1 in the RM in one fed larvae, and at a concentration of 0·5 mM/1 in seven starved larvae and five fed larvae. The results are presented in Tables 6 and 7 and Figs. 6 A (i), 7C (ii), 7D (ii), 8B (i), (ii) and 8C.

KCN at a concentration of 5 mM/1 immobilized the larvae within 20 min, and once immobilized there was no recovery of movement after they had been returned to RM. There were no obvious effects upon the papillae. At a concentration of 0·5 mM/1 KCN immobilized the larvae after 2–5 h, but recovery of movement occurred 10–30 min after they had been returned to RM though they generally died 24–48 h after the return. At both KCN concentrations larvae, when immobilized for long enough, became very turgid and had to be handled very carefully in order to avoid rupturing the anal papillae. This and the less extreme turgidity observed in the higher DNP concentrations were presumably caused by the inhibition of either the peristaltic movements of the hind gut, or of the production of Malpighian tubule fluid, or of both (Stobbart, 1971c).

In all the larvae used, KCN caused an immediate and drastic reduction in K_out to a level which was about one-seventh of the original, so clearly it causes a reduction in outflux similar to that brought about by DW. It also, as might be expected, inhibited influx. This inhibition is demonstrated by the data of Fig. 9 A (v) (see also Table 6). In this experiment the three larvae (starved, and in the steady state) were initially placed in the labelled RM in which they had been reared. Under these circumstances their content of radioactivity was constant. After 2·5 h they were transferred to 0·5 mM/1 KCN in the same labelled RM. As shown above the KCN must have inhibited the outflux through the papillae, therefore if the influx had not also been inhibited we should have expected the internal radioactivity to rise. In fact it declined slowly at a rate comparable with those of the other fed larvae in KCN (see Tables 6 and 7) thus demonstrating that influx had been more or less completely abolished.

Immediate recovery of outflux, and presumably of influx as well, was observed (as the data of Fig. 8B (i) and (ii) show) when larvae were returned to normal RM after being treated for about 5 h with 0·5 mM/1 KCN. The increased decline of radioactivity, shown by the second set of data 9·5h after return to normal RM, was correlated with a marked deterioration in the larvae.

The effect of DW was studied in three fed and four starved larvae, and those of 0·5 mM/1 KCN and 0·125 mM/l DNP were studied in in three and two fed larvae respectively. The results are presented in Figs. 9 A (i) and (ii), 9B(i) – (v) and in Table 8.

None of these solutions affected the outflux significantly, and so the K_out values given in Table 8 are the overall regression coefficients for each larva or group of larvae. It is of course possible that more extended data would have revealed significant effects, but clearly any effects in papilla-less larvae must be very small relative to those observed in intact ones.

From Table 8 we find that the mean K_out value for fed papilla-less larvae placed in inactive media was 0·0102 h⁻¹ which is in very good agreement with the value I reported earlier (Stobbart, 1959). However, the corresponding value for starved papilla-less larvae was almost identical (Table 8), whereas the earlier value was only about one-sixth as large. This discrepancy I think is almost certainly due to the necessarily small numbers of larvae used in the present determinations of K_out, though incomplete starvation (see 2 a) may have played a part.

Using the technique described in 2c (ii) an attempt was made to investigate the effect of 0·5 mM/1 KCN upon influx using four starved papilla-less larvae. The results are illustrated in Fig. 9 (iv) and Table 8. Following treatment with KCN there was no gain of radioactivity. We know that outflux in these larvae is largely unaffected by the KCN, therefore if influx is also unaffected the internal level of radioactivity should stay constant, whereas if influx is inhibited it should decline. In fact a slow decline was noted, but the value for the regression coefficient (= K_out) was not significant, though it would probably have been found to be so if the observations had been more extensive. Thus inhibition of influx is suggested but not proved, and more extensive data would be needed in order to settle the matter.

The difference in area between the two types of papillae is obviously significant (P < 0·001).

The PD is presumably caused by slight excesses of cationic over anionic uptake (or vice versa) or to slight excesses of ionic influx over outflux of the appropriate counter-ion of the same sign. The PD has been measured (a) in starved larvae pretreated with DW for 24 h and then placed in different concentrations of NaCl, KCl, and Na₂SO₄, (b) in fed salt-deficient larvae during uptake of ions from 2 mM/1 NaCl, 2 mM/1 KCl, and 1 mM/1 Na₂SO₄. The results (Figs. 2–5) show clearly that the observed PD is quite inadequate to maintain the internal concentrations of Na+ and Cl⁻, let alone to support the net uptake of these ions which does in fact occur. As mentioned earlier the PD showed a considerable amount of drifting. This drifting was much more marked in solutions of NaCl and of KCl, from both or which both ions are taken up (but at unequal rates) than in Na₂SO₄, from which ‘only’ Na⁺ is taken up (Stobbart, 1967, 1971a), and it was also much more marked in salt-deficient larvae than in those pretreated with DW. It thus seems likely that the drifting is to a considerable extent due to differential variations in the rates of cationic and anionic uptake.

Before equations (12) and (13) can be used, we must have an estimate of p. This can be obtained by measuring the passive flux (outflux) of the ion for a given transmembrane concentration difference and PD, inserting the value in equation (12), and solving for p. The estimate of p can then be used in equations (12) and (13) to predict the values assumed by outflux and influx at other values of transmembrane concentration difference and PD.

If the volume of the system and the area of the membrane are known the dimensions of p will be distance/time, if the area and volume are not known, but are (as here) assumed to be constant, p will have the dimension of 1 /time.

Equations (12) and (13) strictly speaking apply only to the movement of ions through single membranes through which there is a constant gradient of electrical potential, and their application to ionic movements in whole mosquito larvae obviously calls for comment. However, the situation is better than might be expected because: (a) some 90% of the ionic fluxes occur through the anal papillae; (b) electron microscopical evidence indicates that the folded outer plasma membrane of the papillae is the site of an alkaline phosphatase, and therefore probably of the ionic pumps (Copeland, 1964; Sohal & Copeland, 1966; and see Stobbart, 1967). It thus seems possible that the single outer plasma membrane is the major factor in the transport of ions. In this case, and in the absence of any detailed information about this membrane, the constant-field theory will presumably give the best estimates available of the passive ionic movements.

In what follows it will be assumed for the sake of simplicity that both the observed fluxes and the calculated passive fluxes occur entirely through the anal papillae. This assumption will be considered again later.

Earlier work (Stobbart, 1967, 1971a) has shown that Na⁺ and C1⁺ are to a large extent taken up together from NaCl, where as Na⁺ is partly exchanged for H⁺ (and Cl⁻ for HCO3⁻ or OH⁻) during uptake from Na₂SO₄ and KCl respectively. The out-fluxes of Na⁺ and Cl⁻ were found to be lowest in DW, and to increase with increasing external concentration of these ions (Stobbart, 1965, 1967). This I interpreted as showing that at higher external concentrations a carrier-mediated component in the outflux becomes apparent. It is not practicable to measure PD across the anal papillae when they are placed in DW, but measurements have been made at the very dilute external concentration of 0·00155 mM/1, at which concentration the outfluxes are practically identical with those in DW (Stobbart, 1965, 1967) and so may, on this interpretation, be taken as almost entirely passive. Therefore, knowing that the haemolymph concentrations of Na⁺ and Cl⁻ in the larvae are about 93 and 48 mM/1 respectively (Stobbart, 1965, 1967), we can reasonably use the present PD measurements at 0-00155 UIM/1, together with the earlier estimates of outflux at this concentration (Stobbart, 1965, 1967) for calculating values of p for Na+ and Cl^-. This procedure is not ideal, but, in view of the near impossibility of measuring flux and PD simultaneously in large numbers of larvae, it is the only practicable one. The four estimates of p_Na and the three of p_c1 are shown in Table 9, and as would be expected the estimates for each ion are of the same order of magnitude. Using these p values, the expected passive sodium influxes at the different external sodium concentrations and PDs have been calculated for larvae in both NaCl and Na₂SO₄, and are displayed in Table 10. Comparison of Table 10 with the sodium influx data of Figs. 10A, it and in fig. 2,Stobbart (1965) and fig. 2 in Stobbart (1967), demonstrates clearly that as expected the calculated passive influxes are always a negligible fraction of the observed influxes. These, and all subsequent passive influxes, have presumably been very slightly overestimated, as I have assumed when making these calculations that the ionic activity coefficients are unity.

The expected passive chloride influxes for larvae in NaCl and KCl have also been calculated. These are shown in Table 11, and similarly comparison of this table with the chloride influx data of Figs. 10B, 12 and of fig. 2 in Stobbart (1967) demonstrates clearly that these calculated passive influxes are also always a negligible fraction of the observed influxes.

The observed sodium and chloride outfluxes of some of the sets of data are compared with the expected passive outfluxes^* in Figs. 10 A, B, 11 and 12 (larvae in NaCl, Na₂SO₄ and KCl). The situation can be shown to be practically identical in the remaining sets (i.e. fig. 2 in Stobbart (1965) and fig. 2 in Stobbart (1967)). In every case there is a very marked divergence, at the higher external concentrations, between the expected passive outfluxes and the observed outfluxes. The divergence is most marked in the case of Cl⁻, but even for Na⁺ the expected passive outflux at 5 mM/1 external concentration is only 13–25% of the observed outflux.

Values of p found in this way (see Table 9) must obviously be larger than those found earlier. Although, as we shall see later, it seems almost certain that ‘back-transport’ does not occur in Aëdes larvae, the idea is worth following through as it involves assuming the maximum value for p, whereas the previous postulate assumed the minimum value. The expected passive influxes calculable using the ‘back-transport’ p values given in Table 9 may be obtained from Tables 10 and 11 by multiplying the influx values calculated using the ‘non-back-transport’ p values by the appropriate ratio of p values. The expected passive outfluxes calculable using the ‘back-transport’ p values may be found in a similar way, and some of these fluxes are displayed in Table 12, where they are compared with the observed outfluxes.

Calculated in this way the passive sodium and chloride influxes are a larger, but still very small fraction of the observed influxes (generally < 1 %). With respect to the outfluxes, Table 12 shows that again the calculated and observed outfluxes do not agree, the former being very much higher in the more dilute solutions.

It therefore seems clear that in these starved larvae pretreated with DW the outfluxes of sodium and chloride, like the influxes, cannot be explained in terms of passive diffusive movements unless one postulates reductions in the value of p as the external concentration is reduced. Such reduction, however, are rendered improbable by the fact that 0-5 mM/1 KCN (in the presence of 2 mM/1 Na) reversibly reduces sodium outflux to the level characteristic of DW (see below).

These are the only fed sodium-deficient larvae for which (sodium) flux data are available, and for these larvae also the same conclusion must be drawn. The relevant data from this and earlier work are collected together in Fig. 13. This figure shows as a function of time: (i)(a) the Na influx into the whole body, (b) the whole body Na concentration and (c) the Na outflux from the whole body derived from (a) and (b); (ii) the Na concentration in the haemolymph; (iii) the PD; (iv) the calculated equilibrium concentration of Na in the haemolymph; (v) the expected passive Na outflux.

The value of p (assuming no ‘back-transport’) has been measured only roughly in larvae. It is certainly less, and probably a lot less, than 0·0136 h⁻¹. If we assume that it is in fact 0·0136 h⁻¹, and calculate the expected passive outflux as a function of time using our knowledge of the PD and the haemolymph Na concentration (this latter will yield a higher flux value than the total body Na concentration) we see that the expected outflux is only a small fraction of the observed outflux, especially at the start of the period of net uptake (Fig. 13). The expected passive influx calculated similarly is, of course, very small, and ranges from 0·022 mμM/mg/h at the start, to 0·012 at the end of the period of net uptake.

I have made no attempt to estimate a ‘back-transport’ p in these larvae. This is because the observed decline in outflux with increasing time (and at constant external Na concentration) shown in Fig. 13, and measured directly earlier (Stobbart, 1960) is definitely not to be predicted from Kirschner’s (1955) premises (Fig. 14B). It is in fact one of the lines of evidence arguing strongly against the operation of ‘back-transport’ in Aëdes larvae, and it will be considered again below.

These effects were studied in starved larvae which were initially in the steady state.

DW caused an immediate reduction in sodium outflux in both fed and starved larvae, K_out being reduced to about one-sixth of the original value. This type of flux reduction (i.e. dependent on reduced external concentration) has also been observed in isolated frog muscle (Keynes & Swan, 1959) in isolated frog skin (Kirschner, 1955) and in some other aquatic animals (Croghan, 1958; Shaw, 1959; Bryan, 1960; Sutcliffe, 1967; Kerstetter, Kirschner & Rafuse, 1970; Maetz, 1971, 1972). It is consistent with, but as Britton (1970) and Smith (1969b) point out, does not prove, the operation of an ‘exchange-diffusion’ mechanism (Ussing, 1947, 1948). ‘Exchange-diffusion’, which is described in Fig. 14,A, was originally proposed to explain the high rate of sodium exchange between isolated frog muscle and the bathing Ringer solution. According to the ‘exchange-diffusion’ hypothesis the outflux is low at low external [Na] because Na/Na exhanges are reduced, and because the Na can only leak out slowly through the carrier system for which it has a high affinity; it being assumed either that the probability of a sodium ion’s escaping from the carrier at low external [Na] is less than the probability of its being replaced by another sodium ion at high external [Na], or that there is always a fairly high [Na] in the boundary layer at the external surface of the transporting membrane. Alternatively, it may be supposed that the carrier is immobilized if it becomes empty of sodium at the external surface (cf. Britton, 1970). Although a reduced outflux at low external concentration has in some cases been regarded as indicative of an ‘exchange-diffusion’ component – a more general and hence preferable term would be ‘carrier-mediated exchange’ (Croghan, 1958; Shaw, 1959; Bryan, 1960; Stobbart, 1965, 1967; Sutcliffe, 1967; Kerstletter et al. 1970; Maetz, 1971, 1972) – it is also (see above and Fig. 14B) compatible with Kirschner’s suggestion of ‘back-transport’.

Of the other substances tested both choline chloride and DNP caused an immediate reduction in outflux (apparently permanent in the case of choline chloride) but also they had clear gross effects upon the anal papillae. In view of the uncertainties raised by these results, these substances will not be considered further other than to note that the ‘permanent’ flux inhibition caused by choline chloride might be due to persistent occupation by the choline ion of Na⁺ sites on the proposed carrier molecule.

In contrast dilute KCN has no clear gross effects on the papillae, its effects are reversible, and it too causes an immediate and drastic reduction in the sodium fluxes. This reduction (which is comparable to that caused by DW) demonstrates that the fluxes are dependent on aerobic metabolism. Such a reduction in the fluxes by a metabolic inhibitor (cf. Shaw, 1959) is not in agreement with the hypothesis of ‘back-transport’, which under these circumstances predicts a reduction in influx and an increase in outflux (see Fig. 14). In addition, the great increase in both influx and outflux of sodium observed (at constant external concentration) when steady-state larvae are fed (Stobbart, 1959, 1960) is not explicable in terms of ‘back-transport’, nor is the decline in sodium outflux which occurs (again at constant external concentration) as sodium uptake proceeds (Fig. 13, and Stobbart, 1960).

However, this line of argument should not be pushed too far – the data in fact only relate the ionic fluxes to the external concentration, and the Michaelis relationship is in any case only obeyed approximately. In view of the fact that the Na⁺ and Ci⁻ carriers have graded affinities for other cations and anions, and that Na⁺ and Cl⁻ are taken up in part exchange for H⁺ and HCO₃⁻ (or OH⁻) respectively, it seems reasonable to suggest that the removal of Na⁺ and Cl⁻ from the carriers may be due to temporary configurational changes in the carriers which reduce their affinities for these ions (Stobbart, 1967, 1971a).

The increase in fluxes and net uptake observed earlier as a consequence of (i) feeding the larvae, or (ii) depleting them of salts, may be interpreted in terms of an increased synthesis of carrier molecules. In the latter instance, at any rate, the synthesis is probably under hormonal control (Stobbart, 1959 –67, 1971b).

The effect of KCN is accommodated by the model if we suppose that KCN inhibits the combination of the ion with the carrier, or if it inhibits the movement of the carrier through the osmotic barrier. Either way, in demonstrating the dependence of the fluxes upon aerobic metabolism, the effect rules out a simple ‘exchange-diffusion’ model of the type proposed by Ussing (1947, 1948) – see Fig. 14A. In addition, the effect of KCN suggests that the passive (i.e. diffusive) movements of sodium through the anal papillae are very low, and that passive movements are therefore mainly associated with the general body surface and the urine. This point is illustrated by the data of Table 13 which have been taken from Tables 6–8 and from earlier work. It lends further support to the contention that the outfluxes occurring through the anal papillae at high external concentrations are carrier-mediated, and of course suggests that the passive influxes through the papillae are very much smaller than they have been supposed to be.

No data are available concerning the effect of KCN on chloride fluxes. However, since chloride, like sodium, is actively transported, and since the chloride and sodium pumps are similar, it seems likely that the effect would resemble closely the effect on the sodium fluxes.

Knowing the approximate area of the anal papillae and (since the fluxes were expressed as m/m/mg/h) the weight of the larvae, the ionic fluxes occurring through the papillae may be calculated in terms of M × 10⁻¹²/cm²/sec. The steady-state fluxes and those occurring during net uptake are compared in Table 14 with fluxes found in other cells and tissues under approximately natural circumstances (i.e. roughly normal ionic concentrations on both sides of the transporting membranes, and any PD not shorted out). The fluxes in the papillae are much the highest, but as the outer plasma membrane of the papilla is much folded (Copeland, 1964; Sohal & Copeland, 1966) the fluxes may be some ten times smaller when expressed per unit area of membrane. Such a correction would bring the steady-state fluxes into line with those observed in other organisms, but the fluxes associated with net uptake in fed larvae still remain high.

It seems likely that, in view of the great concentration differences involved, the part played by PD in the movement of ions will prove (as in Aëdes larvae) to be unimportant in the vast majority of freshwater animals. However, the PD may be more important in animals regulating hypo-osmotically in the sea and in saline water (Smith, 1969a, b; Maetz, 1971). In the brine shrimp Artemia, for example, an asymptotic relationship between passive sodium outflux and external concentration (agreeing well with the relationship actually observed) can be predicted on the basis of the Goldman constant-field theory (Smith, 1969a, b; Thuet, Motáis & Maetz, 1968). Given a limited amount of ionic carrier in the transporting membrane, and a situation where PD is unimportant, some sort of asymptotic relationship between influx (and probably also between outflux) and external concentration is to be expected. Whether the outflux results from the interplay between diffusion and ‘back-transport’, or whether it is carrier-mediated, can only be determined by studying the effects of metabolic inhibitors upon the outflux. It is at least possible that carrier-mediated exchanges will prove to be a widespread (though not necessarily an obvious) characteristic of ionic pumps, for if a pump is to cause no carrier-mediated exchange at all, it will presumably be necessary for the carrier molecule to be completely divested, at each cycle, of the ion which it carries.

I am grateful to Professor J. Shaw for some helpful discussions. Some analyses of Newcastle tap water were kindly provided by Dr P. Greenaway.