1. A study has been made of water balance in two closely related species of aquatic bug, Notonecta glauca L. and N. marmorea Fabr., and the mechanism appears to be the same in both.

  2. Ammonia is a major nitrogenous excretory compound and the bulk of the ionic material in the rectal fluid is in the form of ammonium bicarbonate.

  3. Water is eliminated through the excretory system and the output of water appears to be related to the output of ammonia. The relationship is such that the concentration of ammonia in the rectal fluid tends to a limiting value of about 120 mm./l.

  4. Water is gained by uptake through the cuticle and gut. The volume of water ingested is closely related to the volume of water eliminated through the excretory system and the uptake of water through the cuticle appears to be constant at about 7 % of the body weight per day.

  5. At the time of moulting the increase in weight is in part due to retention of water in the tissues and in part due to the retention of water in the rectum.

  6. It is pointed out that the mechanism by which water balance is maintained in Notonecta may not be of wide application to other aquatic insects.

Several workers have investigated mechanisms of water balance in aquatic insect larvae. An account of the mechanism of water balance in the mosquito larva Aedes aegypti L. will be found in papers by Wigglesworth (1933) and Ramsay (1950, 1951, 1953). Beadle (1939) and Ramsay (1950) give an account of the situation in the larva of Aedes detritus Edw. The alderfly larva Sialis lutaria L. has been studied by Shaw (1955a, b); Limnephilus affinis Curt, and other trichopterous larvae by Sutcliffe (1961a, b, 1962). Aquatic imagines have attracted little attention and the work to be described in this paper goes some way towards remedying this situation.

The investigation was carried out on two closely related species of aquatic bug, Notonecta glauca L. and N. marmorea Fabr. The bugs were collected from ponds and ditches in Monmouthshire and Glamorgan. Both species occurred together in ponds lying very close to the sea; elsewhere N. marmorea appeared to be absent. N. marmorea is reported to occur mainly in brackish waters in coastal areas (Southwood & Leston, 1959), unlike N. glauca which appears to be confined to fresh waters, and it was hoped that a comparison of the two species would throw some light on these differences in distribution.

Adult specimens varied a great deal in weight, but no attempt was made to obtain uniformity in this respect. Adults of N. glauca ranged in weight from about no to 150 mg. and adults of the somewhat smaller N. marmorea ranged from about 100 to 120 mg. To facilitate comparison uptake and output measurements have been calculated in terms of a standard body weight of 100 mg.

In the laboratory the bugs were kept without food in separate containers until required, and the experiments were carried out at a temperature of about 18° C.

Collection and analysis of rectal fluid

Measurements were made on the clear fluid voided from the anus of the intact bug. In order to obtain this fluid a bug was carefully dried on filter-paper and placed in a wax-lined dish. The specimen was gently secured in order to induce the ejection of fluid which was collected in a capillary pipette and analysed without delay. For the purpose of routine ammonia measurements bugs were narcotized with CO2 in order to induce the ejection of fluid from the rectum and this had the advantage that handling was reduced to a minimum.

Ammonium, bicarbonate and conductivity measurements were made in duplicate or triplicate and 0·2μl. of fluid was used for each determination. Ammonium was measured by the diffusion method described by Shaw & Beadle (1949) and 0·01 N-HC1 was used for the titration. In order to facilitate the mixing of the sample and alkali in the diffusion chamber a clear lid was employed and the absorbing acid was placed on the floor of the chamber. Bicarbonate was measured by direct titration of the sample with 0·01 N-HCI which contained 5 ml. of B.D.H. ‘4·5’ indicator per 100 ml. A measure of the total ionic concentration was obtained by comparing the conductivity of samples diluted in about 44 μl. of de-ionized water in a Perspex conductivityy cell (Shaw & Staddon, 1958) with the conductivity of standard solutions of ammonium bicarbonate prepared in the same way. Ammonium was measured with an average error of ± 2 mm./l., bicarbonate with an error of ± 4 mm./l. and conductivity measurements with an average error of ± 2 mm./l.

Output of water in the rectal fluid

This was deduced from measurements of the rate of ammonia output and of the concentration of ammonia in the rectal fluid. Given the rate of ammonia output (in μM./day) and the concentration of ammonia in the rectal fluid (in μ.M./ μl.) then division of one term by the other gives directly the output of water in the rectal fluid (in μl./day).

In order to measure the rate of ammonia output a bug was placed in 25 ml. of de-ionized water in a covered beaker. At the end of 24 hr. the ammonia content of the water was measured colorimetrically and for this purpose Nessler’s reagent was used. This method of collecting ammonia appeared to be reliable as shown by the fact that on removing a bug at the end of the 24 hr. period the concentration of ammonia in the water underwent little or no change during a further period of 24 hr.

Now the concentration of ammonia in the rectal fluid was measured by diffusion analysis and the output of ammonia was measured colorimetrically. Dissimilar methods of analysis may yield dissimilar results, and it was clearly necessary to check this possibility. To make this check several samples of rectal fluid were pooled and analysed by both methods. The results which were obtained by the two methods did not differ significantly.

Ingestion of water

The uptake of water into the gut was measured with the aid of amaranth. A specimen was placed in an 0·01 M solution of the dye for a period of 8 hr. It was then removed, washed with water and narcotized with CO2. The dye was extracted from the gut and for this purpose the method devised by Treherne (1957) was closely followed. The gut was dissected out in isotonic saline and then transferred to 5 ml. of phosphate buffer at pH 10. It was finely ground up with the aid of glass powder. The extract was centrifuged and the concentration of amaranth in the clear supernatant was measured with a Unicam absorptiometer at an absorption maximum of 525 mμA slight correction had to be made to each measurement to allow for the effect of soluble gut material on absorption. Given the quantity of dye in the gut (in mμM.) and the concentration in the external medium (in mμM./ μl.) then division of one term by the other gives the volume of water (in μl.) ingested in the 8 hr. period.

The procedure was tested by feeding individual bugs on a measured quantity of dye. To do this each bug was narcotized by means of CO2 and the proboscis was sealed into a small wax-lined tube with a warm mixture of beeswax and rosin. A quantity of maaranth, 10 μl. of an 0·01 M solution, was introduced into the tube which was then covered in order to reduce evaporation. The bug was left in air to consume the fluid overnight. The small traces of amaranth which remained on the sides of the tube were washed out and measured in order to obtain an accurate measure of the quantity actually ingested. The quantity of amaranth recovered from the gut was always over 97 % of that ingested which shows that there can be little, if any, absorption from the gut lumen.

The reason for selecting an experimental period of 8 hr. was to avoid the loss of amaranth from the gut by evacuation in the rectal fluid. In 8 hr. the dye may have reached but not travelled beyond the pyloric valve; in 12 hr. the rectal fluid is frequently coloured with the dye.

A more direct method was also used to measure the ingestion of water. This entailed sealing the proboscis into a capillary tube of about 1 mm. external diameter with the beeswax-rosin mixture. The tube was filled with water to a mark and then secured vertically so as to suspend the bug in water in its normal resting position just beneath the surface. The decreased volume of water in the tube was recorded 24 hr. later. This method had the advantage that simultaneous measurements of water output could be made.

(a) Excretion of ammonia

Ammonia is quantitatively the most important nitrogenous excretory product of adult bugs kept without food. Measurements of total-N and ammonia-N showed that approximately 75 % of the total-N in the rectal fluid of N. glauca is in this form. The rectal fluid is not a dilute one for the mean conductivity (92 mm./l.) of the measurements which are recorded in Table 1 is as much as 56 % of that of the haemolymph (165 mm./l. as equiv. NaCl) and the greater part of the ionic material is in the form of ammonium bicarbonate. A fluid of similar composition is produced by the larval Sialis lutaria L. (Shaw, 1955b).

The output of ammonia fluctuates a little from day to day as shown by the results of measurements which are recorded in Table 2. It is not known what causes these fluctuations in ammonia output but small fluctuations in the output of rectal fluid may be partly responsible.

(b) Relationship between water output and ammonia output

In view of the fact that ammonia is the major end product of protein catabolism in Notonecta adults it is of interest to know whether the output of water in the rectal fluid is closely related to the output of ammonia. In Fig. 1 measurements of water output have been plotted against measurements of ammonia output. Although the scatter is great it would appear from inspection of Fig. 1 that the output of water is broadly related to the output of ammonia, and over a wide range of ammonia output the degree of correlation is high (r = 0·88; P < 0·01; n = 36). In Fig. 1 the straight line is the calculated regression line and this was determined utilizing the measurements obtained for both species. The regression coefficient of water output on ammonia output is 8·5 and the ordinate of the intercept is 8·0. A separate calculation for N. glauca yielded a regression coefficient of 8·4 and intercept of 8·7. Comparatively few measurements were obtained using N. marmorea but the water output in this species does not differ significantly from that in N. glauca at comparable rates of ammonia output. On the assumption that the relationship between water output and ammonia output is a linear one then the calculated value for the intercept of 8 μl./100 mg./24 hr. provides a provisional estimate of the osmotic uptake of water through the body surface.

In Fig. 1 measurements made in September have been distinguished from measurements made in May to July. The output of ammonia was significantly lower in September, as was the output of water. The reason for the difference in the rate of ammonia output is not known but it would not be surprising to find that the level of protein catabolism undergoes regular seasonal fluctuations.

(c) Relationship between the concentration of ammonia in the rectal fluid and the rate of ammonia output of ammonia output

In Fig. 2 measurements of the concentration of ammonia in the rectal fluid have been plotted against measurements of ammonia output. A relationship between the two variables can be calculated if it can be assumed that the water output is a linear function of the ammonia output. Let y be the rate of water output (in μl./100 mg./day), x the rate of ammonia output (in μM./100 mg./day) and z the concentration of ammonia in the rectal fluid (inμM/μl) Given y = mx + c = x/z then z = x/(mx + c). According to this function the concentration of ammonia in the rectal fluid tends to a limiting value which is 118 mm./l. using the constants m = 8·5 and c = 8·0, the regression coefficient and intercept respectively, previously calculated for the regression of water output on ammonia output. In Fig. 2 (in which measurements of ammonia concentration are given in mm./l.) a curve for m = 8·5 and c = 8·0 has been drawn for comparison with the experimental results.

(d) Uptake of water through the cuticle

Holdgate (1956) devised a method for measuring the uptake of water through the cuticle of an aquatic insect and used it to measure the uptake in adult specimens of Notonecta obliqua. The animal is killed by H2S and weighed before and after a period of 24 hr. in a desiccator. In order to distinguish between the increase in weight due to the adherence of a film over the surface and that due to the absorption of water through the cuticle the animal is first blotted with filter-paper, weighed and then transferred to a desiccator. Thereafter weighings are carried out at half-hourly intervals. During the early period of desiccation the animal loses weight rapidly and this is considered to be due to the removal of superficial moisture. But after 30 min. a steady rate of weight loss is attained and this js considered to be due to transpiration through the cuticle. By extrapolation of this steady rate back to the time of removal from the water a figure can be obtained for the actual amount of water taken up by the animal. In the present work a slightly modified version of this method was used to measure the uptake of water through the cuticle of N. glauca. The bugs were killed by ammonia gas and immersed in water for a period of only 8 hr. They were weighed to the nearest 0·1 mg. The mean value for the uptake of water in six specimens was 7·34 mg./100 mg./24 hr., or 9·2 mg./specimen/24 hr., and the standard deviation was ±1·5 mg./100 mg./24 hr. Holdgate obtained a value of 11·2 mg./specimen/24 hr. for the rate of water uptake in N. obliqua which compares favourably with the value for the uptake in N. glauca.

(e) Ingestion of water

In Fig. 3 measurements of water uptake into the gut through the mouth have been plotted against measurements of water output through the excretory system. In accordance with expectation the uptake is closely related to the output (r = 0·914; number of observations 26). The calculated regression line has been drawn through the points. The regression coefficient of water uptake on water output is 0·81 and a coefficient of unity predicted on the assumption that the rate of water absorption through the body surface is constant falls within the 95 % confidence limits of ± 0·34.

In all pairs of variates the calculated output of water was greater than the uptake and the difference between the means, 9·9 µl./100 mg./24 hr., is highly significant. The standard error of this mean difference is 0·69 μl./100 mg./24 hr. and the 95 % confidence limits are 8·5 and 11·3 μl./100 mg./24 hr., respectively. Apart from the Absorption of ingested water through the gut the only known means of gaining water from the environment is by absorption through the body surface. In the previous section the rate of water absorption through the cuticle was estimated and the value obtained, 7·3 μl./100 m g./24 hr., is not markedly lower than the value 9·9μl/100 mg./ 24 hr., the difference between the mean output of water and mean uptake of water through the gut.

In Fig. 4 measurements of water uptake into the gut have been plotted against measurements of ammonia output in the rectal fluid and the correlation coefficient is highly significant (r = 0·591; number of measurements 28). The straight line in the figure is the calculated line of regression of water uptake on ammonia output. The regression coefficient is 7·2 and the intercept is – 0·6. A coefficient of 7·2 agrees well with that of 8·5 previously calculated for the regression of water output on ammonia output.

(f) Retention of fluid in the rectum at the time of moulting

At the time of moulting water is swallowed and retained, much of it in the rectum which becomes enormously distended. An attempt has been made to distinguish between the increased weight due to retention of water in the rectum and that due to retention of water in the tissues. A number of specimens were weighed shortly before and after the final moult. The rectum was carefully dissected out of the newly emerged adult and the moisture adhering to it was removed with the aid of filter-paper. The wall of the rectum was punctured in order to release the rectal fluid which was weighed on a torsion balance after absorption on a piece of weighed filter-paper. The results are shown in Table 3. The mean weight increase at the time of moulting was 25·5 mg. The quantity of fluid retained in the rectum was 13·9 mg. and this is 56 % of the total weight increase if it can be assumed that the rectum was empty before moulting took place. This assumption is not unreasonable for the rectum of the last instar larva contains very little fluid just prior to the moult.

It seems clear that by retaining fluid in the rectum at the time of moulting an aquatic insect can achieve an increase in bulk without diluting the tissue fluids to a proportionate extent although the quantity retained may not be sufficient to obviate the problem entirely.

From the results presented in this paper the mechanism by which water balance is maintained in two closely related species of aquatic Heteroptera, Notonecta glauca and N. marmorea, appears to be as follows. Water is eliminated through the excretory system and the output of excretory fluid is closely related to the output of ammonia. The relationship is such that the concentration of ammonia tends to a limiting value of about 120 mm./l. The output of water is balanced by uptake of water through the body surface and gut. The volume of water ingested is closely related to the volume of water eliminated through the excretory system and the uptake of water through the cuticle appears to be constant at about 7 % of the body weight per day.

This conception of water balance may not have widespread application to other aquatic insects. Shaw (1955 a), for example, working on the larva of Sialis lutaria, showed that the output of water through the excretory system normally balanced the water intake through the cuticle. The larva does not appear to drink when deprived of food. This is of interest for the Sialis larva excretes ammonia and the concentration attained in the rectal fluid is quite high (Staddon, 1955; Shaw, 1955 b).

The cuticle of Notonecta appears to be appreciably permeable to water, a conclusion also reached by Holdgate (1956). But a value of about 7 % of the body weight per day for the osmotic uptake through the cuticle appears to be a surprisingly high one for the rate of uptake is no greater than this in several insect larvae. The rate at which water is taken up by osmosis through the cuticle of the Sialis larva was shown by Shaw (1955 a) to be of the order of 4% of the body weight per day at 20° C. Sutcliffe (1961 a) measured the osmotic uptake in three species of caddis larvae and in all cases the uptake was approximately equivalent to 7 % of the body weight per day. The situation appears to be all the more remarkable in Notonecta for a film of air limits the area of cuticle in contact with the water.

The idea that the rate of water output is related to the ammonia output is an interesting one in that it would appear automatically to limit the concentration of ammonia attainable in the rectal fluid. It is possible, of course, that the total concentration of solutes in the rectal fluid is the important factor and that no particular significance need be attached to ammonia other than as an indication of this.

Of the two species of Notonecta studied N. marmorea is reported to occur mainly in brackish waters, unlike N. glauca which appears to be restricted to fresh waters. The mechanism of water balance appears to be the same in the two species, however, and differences in distribution can hardly be associated with differences in this respect.

Beadle
,
L. C.
(
1939
).
Regulation of the haemolymph in the saline water mosquito larva Aedes detritus Edw
.
J. Exp. Biol
.
16
,
346
62
.
Holdgate
,
M. W.
(
1956
).
Transpiration through the cuticles of some aquatic insects
.
J. Exp. Biol
.
33
,
107
18
.
Ramsay
,
J. A.
(
1950
).
Osmotic regulation in mosquito larvae
.
J. Exp. Biol
,
27
,
145
57
.
Ramsay
,
J. A.
(
1951
).
Osmotic regulation in mosquito larvae: the role of the Malphighian tubules
.
J. Exp. Biol
.
28
,
62
73
.
Ramsay
,
J. A.
(
1953
).
Exchanges of sodium and potassium in mosquito larvae
J. Exp. Biol
.
30
,
79
89
.
Shaw
,
J.
(
1955a
).
The permeability and structure of the cuticle of the aquatic larva of Sialis lutaria
.
J. Exp. Biol
.
32
,
330
52
.
Shaw
,
J.
(
1955b
).
Ionic regulation and water balance in the aquatic larva of Sialis lutaria
.
J. Exp. Biol
.
32
,
353
82
.
Shaw
,
J.
&
Beadle
,
L. C.
(
1949
).
A simplified ultra-micro Kjeldahl method for the estimation of protein and total nitrogen in fluid samples of less than 1·0 μl
.
J. Exp. Biol
.
26
,
15
23
.
Shaw
,
J.
&
Staddon
,
B. W.
(
1958
).
A conductimetric method for the estimation of small quantities of ammonia
.
J. Exp. Biol
.
35
,
85
95
.
Southwood
,
T. R. E.
&
Leston
,
D.
(
1959
).
Land and Water Bugs of the British Isles
.
London and New York
:
Warne and Co. Ltd
.
Staddon
,
B. W.
(
1955
).
The excretion and storage of ammonia by the aquatic larva of Sialis lutaria (Neuroptera)
.
J. Exp. Biol
.
32
,
84
94
.
Sutcliffe
,
D. W.
(
1961a
).
Studies on salt and water balance in caddis larvae (Trichoptera). I. Osmotic and ionic regulation of body fluids in Limnephilus affinis Curtis
.
J. Exp. Biol
.
38
,
501
19
.
Sutcliffe
,
D. W.
(
1961b
).
Studies of salt and water balance in caddis larvae (Trichoptera). II. Osmotic and ionic regulation of body fluids in Limnephilus stigma Curtis and Anabolia nervosa Leach
.
J. Exp. Biol
.
38
,
521
30
.
Sutcliffe
,
D. W.
(
1962
).
Studies on salt and water balance in caddis larvae (Trichoptera). III. Drinking and excretion
.
J. Exp. Biol
.
39
,
141
60
.
Treherne
,
J. E.
(
1957
).
Glucose absorption in the cockroach
.
J. Exp. Biol
.
34
,
478
85
.
Wigglesworth
,
V. B.
(
1933
).
The function of the anal gills of the mosquito larva
.
J. Exp. Biol
.
10
,
1
15
.