Loss of Water by Excretion and Defaecation in the Tsetse Fly

Substantial amounts of water are excreted by tsetse flies immediately after the ingestion of a blood meal (Lester & Lloyd, 1928), and during subsequent stages of digestion faecal matter is voided in the form of a semi-liquid paste, a process which must also entail some loss of water. The water balance of tsetse flies cannot be satisfactorily elucidated unless such losses are taken into account, and the present investigation was undertaken in an attempt to provide quantitative estimates of the amounts involved.

It will be convenient to consider the preliminary excretion of water separately from the subsequent voiding of faecal matter, since the techniques appropriate to the study of these two processes are different. The present account will accordingly be divided into two sections.

The flies used for these experiments were Glossina morsitans Westwood, emerged from puparia which had been collected at Singida in the Central Province of Tanganyika. Within 2 hr. of emergence the flies were isolated in single 3 in. x 1 in. gauze-covered tubes and exposed to different relative humidities (80, 60, 40 and 20% R.H.) at 27° C. for 42 hr. They were weighed singly in glass-stoppered tubes on a Mettler balance reading to o-oi mg., and then fed on the forearm of the experimenter in such a way that any excretion was retained in the glass tube and evaporation minimized. The tube was then re-weighed and the amount of blood ingested found by subtraction. Preliminary experiments had shown that the rate of excretion of water falls to a basal level in less than 3 hr. (see Fig. 1), and the flies were accordingly left for this period of time at 100% R.H. (to avoid loss by transpiration) and then re-weighed. The loss in weight was taken as the amount of water excreted. The size of the fly was then measured and the wet weight, dry weight and residual dry weight (dry weight minus fat) were determined separately for thorax and abdomen, all according to methods described elsewhere (Bursell, 1959, 1960). Knowing (a) the composition of blood (21.1% solids, see Hammarsten, 1914), (b) the amount of blood ingested, (c) the composition of the fly at the end of the experiment, and (d) the amount of water excreted during the experiment, the water content of the unfed fly could be calculated;^* and in this way a relation could be obtained between the quantity of water excreted and the state of water reserves at the time of feeding.

The estimate arrived at will be subject to a number of systematic errors which must be briefly discussed :

• (1) It is assumed that only water is lost by excretion during the first 3 hr. after a blood meal; in fact evaporation of the excreta to dryness leaves a deposit of crystals, but the amount was too small to be determined with the balance available, and constitutes on an average less than 0.1% of the recorded weight losses.

• (2) It is assumed that no loss of water occurs by transpiration, but in view of the difficulty of getting a truly saturated atmosphere this assumption also is questionable. However, even at humidities as low as 80% R.H. the loss by transpiration would be less than 0-2 mg. for a 3 hr. period (Bursell, 1959) and negligible in comparison with the excretory losses which range from 10 to 20 mg.

• (3) It is assumed that no loss in weight or change in composition is caused by respiratory metabolism. G. morsitans at rest at 27° C. consumes about 0-06 mg. of fat in 3 hr. with the production of an equivalent amount of water (Bursell, 1959).

Again the quantities are negligible compared with the recorded losses, and in general it may be concluded that even in combination these systematic errors are too slight to affect the conclusions drawn.

Fig. 2 shows the amount of water excreted expressed as a percentage of the total blood meal, and plotted as a function of the water content of the feeding fly (amount of water as a percentage of the non-fatty wet weight). There is clearly a very strong correlation, the excretion being more copious the higher the water content of the fly. This is in accord with the early experiments of Jack (1939) who showed that what he termed the ‘primary excretion ‘of flies exposed to low relative humidities was less than that of flies maintained at high relative humidities. It will be seen that in general about 40-50% of the water in the blood meal is excreted, a value which agrees with the figures quoted by Lester & Lloyd (1928) and by Jack (1939).

The most interesting result of this calculation is the close approximation of the regression coefficient by₁.₂ to unity (t for the difference from unity is 0.78, P= 0.4-0.5) which suggests that the excretion of water is accurately related to the state of tissue reserves—of the water which is removed from the blood meal an amount corresponding to the tissue deficit is retained in the body, and only the balance is excreted.

The value of the first regression coefficient (bγ_1.2 = 0.686) is such that the dehydration of the blood meal appears to be dependent on the size of the meal. The water content of the blood meal at constant tissue deficit can be calculated from the data and it would be 56% for a full meal of 25 mg., but 65% for one of 15 mg. Since such an effect might be of significance in relation to water balance, in so far as it suggests that a greater proportion of water is retained in partial meals, it seemed worth while to investigate the phenomenon in greater detail and over a wider range of meal size.

The procedure employed was the same as before except that feeding of the flies was interrupted so that partial meals were taken. The dry weight of the tissues of the abdomen was estimated by subtracting the dry weight of the blood meal from the total as determined at the end of the experiment, and, assuming that the water content of the abdominal tissues is the same as that of the thoracic tissues, the amount of water held in abdominal tissues could be calculated. This, subtracted from the total amount of water contained in the abdomen would give the amount of water in the gut content, and, knowing the amount of solids in the gut, the water content of the blood meal could be calculated. Fig. 3 shows the water content 3 hr. after feeding plotted as a function of the size of the meal.^† It is clear that the smaller the meal the smaller is the proportion of water which is removed from it. This effect seems to be independent of the tissue deficit of the feeding fly—a multiple regression was calculated with size of meal and tissue deficit as the independent variable, and the coefficient of regression with tissue deficit was found not to differ significantly from zero (bγ_2.1 = 0.502 ± 0.451; t; 1.1; P = 0.3).

The situation with regard to the excretion of water has been summarized in Fig. 4, of which curve (a) represents, for different sizes of meals, the amount of water in the blood ingested, and curve (b) the amount retained in the blood meal 3 hr. after feeding. The difference between these two curves represents the water available for replenishment of tissue reserves, the balance being excreted. 61 % is the critical level to which water content may be reduced before death supervenes (Bursell, 1959) and for flies of the size of the present series this would correspond to a tissue deficit of about 10 mg. Fig. 4 shows that this amount of water would be available from a meal amounting to no more than 70 % of a full one.

The difference between curve (b) and the abscissa represents the amount of water retained in the blood meal, the dotted line showing what would have been retained had the percentage concentration been the same for all sizes of meal. When less than half of a full meal is taken the difference is as much as 1 mg. This retention of disproportionately large amounts of water after partial feeds may be of some importance to the fly, for i mg. represents about 10% of the total water reserves of the tissues.

Following concentration of the blood meal, nitrogenous waste materials begin to accumulate as a result of the deamination of products of digestion and general catabolic activity, and these, together with indigestible portions of the blood meal, mainly haematin, begin to be removed from the hind gut. The faeces are voided periodically as a semi-liquid paste, and it was of interest to determine how much water is lost with the faecal pellets, and whether the amount depends on the state of water reserves of the fly or on the humidity of the environment. The following procedure was adopted to investigate these questions.

Flies were fed and left for 3 hr. to complete the excretion of water, after which they were lightly anaesthetized with chloroform and the dorsal surface of the thorax was fastened to a copper wire holder by means of molten paraffin wax. A side piece of the holder passed to the ventral surface of the fly and to this were fastened the meso- and meta-thoracic legs to ensure that excreta were avoided freely and not picked up by the legs during the performance of cleaning movements. The fly was then mounted with plasticine in a glass container whose rim could be immersed below the surface of liquid paraffin contained in a petri dish. Humidity could be controlled in the closed space with wet filter-paper or silica gel, and loss of water from the faeces was prevented by collection under liquid paraffin.

Faeces were cleared at 08.00 and at 16.00 hr. each day; they were transferred by means of a wide-mouthed pipette to small watch-glasses made from heated coverslips. Excess paraffin was quickly removed with a fine pipette leaving the faeces covered with a thin film. The watch-glasses were weighed and then exposed over phosphorus pentoxide at 37° C. ; they were re-weighed at intervals to constant weight, the water content of the faeces being found by difference. The paraffin was now removed with benzene (30 min.) after which the faeces were dried in vacuo at 85° C. and the watch-glasses were reweighed. The faeces were removed with concentrated sulphuric acid, and the watch-glasses were washed in water, dried and weighed, the amount of solid material being obtained by subtraction.

Preliminary tests showed that known amounts of water added to watch-glasses under liquid paraffin could be quantitatively recovered in this way. Further it was shown that any change in weight of dry faecal material during a half-hour exposure to small volumes of benzene were too small to be detected with the balance available (Mettler reading to o-oi mg.). As a further check on the method a mixture of uric acid, haematin and water was made up with a known proportion of solids, and its water content was estimated as outlined above. The values obtained were all within 3 % of the true water content, a degree of accuracy which was considered sufficient for present purposes.

Fig. 5 a shows the water content of faeces voided by G. morsitans maintained at 100% R.H. (curve (i)) and at 0% R.H. (curve (ii)). In saturated air the faeces first voided have a water content of about 60%, but values rise steadily in the course of the hunger cycle to reach levels of about 90% after 4-5 days. By contrast, the water content of faeces voided during desiccation in dry air is only about 40% during the first 16 hr., and drops a further 5% in the course of the hunger cycle (the flies maintained under these conditions die, presumably of desiccation, after about 3 days, while in saturated air flies will live for 5-6 days).

The fact that a substantial difference can be detected in the water content of faeces even at the first collection suggests that the difference might be a direct reflection of the difference in relative humidity, rather than in the state of water reserves ; over the short period the water reserves of flies exposed to 0 and 100 % R.H. would not differ very greatly (see Buxton & Lewis, 1934, p. 202; Jack, 1939, table XVIII ; Bursell, 1959, table I). But further work would be required to settle this point.

Fig. 5b shows the output of solids; the rate appears to be slightly greater in dry air, but it declines more rapidly so that at the end of the third day the total amount of solid excreted is the same under the two sets of conditions, namely 2.5 mg. Taking average values of 72 and 35 % for faecal water content in saturated and dry air the corresponding amounts of water lost during the first 3 days would be 6.4 and 1.4 mg. In other words, the ability to withdraw water from the faeces under conditions of desiccation represents a saving of 5 mg. of water, which is more than 30% of the total water reserves of the newly fed fly (10 mg. held in the tissues, see p. 693 above, and 6 mg. held in the blood meal, see Fig. 4).

It is well known that in many insects water is resorbed from the gut content in the region of the hind gut, and it seems that the rectal glands are particularly active in such resorption (Wigglesworth, 1950). Well-developed rectal glands, richly supplied with tracheae, are present in Glossina (illustrated by Hoare, 1931), and there can be no reason to doubt that these organs are largely responsible for the dehydration of faecal material. The interesting point which emerges from the present results is that although the mechanism is there, the tsetse fly does not at all times conserve water to the best of its ability; full powers of conservation are exercised only when water reserves are threatened. And this applies not only to excretory losses, but seems to be a general characteristic of the water balance of this insect. The control of transpiration by spiracular regulation in both adult and pupa finds full expression only if the humidity of the environment is low, or if the water reserves of the insect have been seriously depleted (Bursell, 1957, 1958). Since water conservation of either type would call for an expenditure of energy, it is possible that the phenomenon of regulated conservation should be seen as an instance of metabolic economy.

Part of this study was carried out in East Africa under employment by the East African High Commission, and a preliminary account has appeared in the 1958 Annual Report of the East African Trypanosomiasis Research Organisation. I am indebted to the Wellcome Trust which provided a grant for equipment used to complete the investigations in Salisbury; and my thanks are due to Prof. E. B. Edney for helpful advice.