A Simple Procedure for the Study of Ionic Regulation in Small Animals

In the study of ionic regulation in small animals, such as insects or small crustaceans, the quantities of blood or excretory fluid available for analysis are often not greater than about 1 ml. This volume might contain anything between 0·1 and 50 µg. of any one of the inorganic ions. The larger amounts can be estimated by the conventional micro-analytical methods, but measurements around the 1 µg. level require special ultra-micro-techniques.

The methods which have already been described for the determination of inorganic ions in quantities of 1 µg. or less are of two kinds : first, there are the ultra-microchemical methods which are generally adaptations to capillary technique of conventional chemical methods of analysis; and secondly, there are methods which rely on the exploitation of some particular physical property of these substances. Among these latter methods may be mentioned the technique devised by Engstrom (1946) for the quantitative estimation of a number of elements by X-ray absorption spectrography, making use of the absorption discontinuities which appear at a characteristic wavelength for every element. Most of the biologically occurring elements can be estimated in this way, and the quantities which can be measured are extremely small (10⁻³–10⁻⁶µg.). More recently, the technique of flame photometry has been developed on a micro-scale (Ramsay, Brown & Falloon, 1953) for the determination of quantities of sodium and potassium in the order of 10⁻⁵µg.

If very small quantities are to be measured then undoubtedly these physical methods are of the greatest value, but the more conventional chemical methods have the advantage that they provide a simple means of estimating all the inorganic ions and only make use of apparatus which can be easily and cheaply constructed.

A number of ultra-micro-chemical methods (operating at the 1 µg. level) have already been described for some of the important biological ions, and these methods have been reviewed recently by Glick (1949) and by Kirk (1950). In general, the ion is either titrated directly by some specific chemical reaction or precipitated from solution by some specific reagent, and then the precipitate separated from the solution, redissolved and titrated. Table 1 lists a number of these methods and shows the chemical reactions which are used for the estimation.

A wide range of chemical methods have been used which involve many different types of titrations, of which chloride, acidimetric and oxidimetric are examples. Since often the concentrations of a number of ions have to be measured at the same time the work involved in the analysis can be greatly reduced if a standard procedure is adopted for all determinations, where possible the same type of titration being used. This avoids duplication of apparatus or continual cleaning and refilling the micro-burette with different solutions and also means that it is only necessary to become accustomed to the colour change which occurs at one type of end-point. Of the available titration methods, the ultra-micro-methods for chloride are both reliable and easy to perform and this paper describes methods for the estimation of sodium, potassium, calcium and magnesium using this titration in each case. This has the additional advantage that the chloride ion is one of the ions which will probably be measured in any case.

The methods by which the other ions can be converted to chlorides are described below, but first an account is given of the procedure and the apparatus which is common to the methods for all the cations.

The sample (about 1 µl.) is introduced by means of a capillary micro-pipette into the bottom of a small glass tube made from ‘Hysil’ tubing of 2 mm. internal diameter. The tube is 20 mm. long, is sealed at one end, and at a point 5 mm. from this end the tube tapers down to give a conical base. The contents of the tube are dried at 100° C. and transferred, in a small metal rack, to an incinerating oven maintained at 450°C., and left there until the organic matter has been decomposed. This generally requires from $12$ to 1 hr., but the time varies with the amount of organic material present.

When the tubes have cooled, the precipitating agent is added to them so that the ash is covered and each tube is about half full. The mixture is well stirred with a very fine glass rod, the top of the tube is covered with a rubber cap and the tubes allowed to stand until the precipitation is complete. The supernatant is separated by centrifugation in a ‘B.T.L.’ semi-micro angle centrifuge, the normal tubes of which have been replaced by lengths of wooden dowel rod into which holes have been drilled axially to take the small tubes. A dozen tubes can easily be accommodated at once in this way. Centrifugation for 2 min. is generally sufficient, and during this time the precipitate is deposited at the bottom of the conical part of the tube and slightly displaced to one side. This displacement, which is caused by the angle of centrifugation, greatly facilitates the removal of the supernatant. This operation is carried out by sucking off the fluid through a micro-pipette with a slightly curved tip, while the tube is inclined at an angle of about 30° to the horizontal with the precipitate on the lower side of the tube. The tube is supported in this position by a small ‘Perspex’ holder, and the operation is controlled by observation under a binocular-dissecting microscope. The relative positions of the tube, precipitate and pipette are shown in Fig. 1. Suction through the pipette is either from the mouth or by connexion to a filter pump, and in this case the suction is controlled by means of a side tube of rubber which can be closed by the fingers.

After the removal of the precipitating agent the precipitate is washed in the appropriate solution twice using the same technique. The separation completed, the precipitate is dried at 100°C. and converted quantitatively to a chloride by one of the methods described below. This reaction is carried out in the same tube and the chloride is washed out from this in a micro-pipette and transferred to the titration platform. Dilute nitric acid is used for washing out when the chloride is titrated by the Volhard method and 0·1N acetic acid for the electrometric method.

For the titration, the burette and associated apparatus described by Shaw & Beadle (1949) is used since this type of simple automatic micro-burette is very easy to construct and has proved very reliable in operation. The design of the micro-burette has been improved since the original description.

The micro-burette is made of a length of ‘Veridia’ thick-walled capillary precision-bore tubing of 0·2 mm. internal diameter, which is bent at right angles and a short tip pulled at the end of the vertical limb (Fig. 1). This gives a much more robust construction than in the previous design. Fluid runs out of the burette when the titration drop on the platform is raised to touch the tip and stops when the drop is lowered if the correct dimensions of the burette are chosen (see below). The titration platform consists of a i in. square of glass which has been painted white and then waxed over, or, more simply, a square of polytetrafluorethylene (‘Fluon’), the latter having the advantage that it can be cleaned in any organic solvent. The size of the titration drop is usually about 50 µl. and is continually stirred by means of a fine jet of air directed against one side of it.

The correct functioning of the burette depends on the length of the vertical limb ; it must be greater than the rise in this limb due to surface tension (h). Now h = 2T/rdg, where T is the surface tension, d is the density and r, the capillary radius. For water and for a capillary of 0·2 mm. diameter this length = 14·3 cm. On the other hand, the height (A) must not be so great as to cause a drop to be formed at the end of the tip in air. h₁ = 2T(r + r₁) dg, where r₁ is the radius of the tip for a hemispherical drop to be formed at the tip. Thus if the tip diameter was the same as the burette diameter then this height would be twice the minimum height. The length of the vertical limb, then, must lie between these two limits and should be nearer the minimum than the maximum for the greatest stability. If the titration drop is small then the minimum height of the vertical limb will be increased by an amount equal to the pressure excess inside the drop = 2T/Rdg, where R is the drop radius. For a 50 µl. drop this correction would be about 0·7 cm. and would become increasingly important as the drop size was made smaller. The critical dimensions are: (1) the radius of the horizontal limb in which the meniscus lies, and (2) the radius of the tip where the drop forms.

The rate of flow of fluid from the burette must also be controlled so that the meniscus in the horizontal limb of the burette moves at a convenient speed and there is no tendency to over-shoot the end-point. This rate of flow can be predicted by using Poiseuille’s equation for the rate of flow of liquid through a capillary. The volume flowing in unit time, Q = πr⁴P/8µL, where P is the pressure, µ is the viscosity and L is the length of the capillary. For the burette, P can be measured by the height of the vertical limb less the minimum height.

For a burette of 0·2 mm. internal diameter, of total length 40 cm. and with the vertical limb 21 cm. high, the rate of flow of water from it can be calculated to be 0·56 × 10⁻⁴ ml./sec. and the meniscus velocity as 1·9 mm./sec. This will be the initial speed when the burette is full; as the burette empties the rate will increase. This meniscus velocity is rather too fast and will be faster still if the capillary diameter of the burette is greater, but the velocity can be reduced by making a fine tip at the end of the vertical limb. The diameter of the capillary must be reduced by a factor of five or ten to bring about an appreciable reduction in velocity but the longer the tip the greater is the reduction. The most satisfactory way of getting the correct rate of flow is to draw off a long thin tip and then break pieces off until a satisfactory rate is reached. Another way of changing the rate of flow is to alter P, and this can be done by attaching some form of pressure-regulating device to the end of the horizontal arm of the burette. This adds to the complication of the apparatus but does provide a variable rate-of-flow regulator, if this is required.

The chloride-ion concentration is estimated by either one of two methods. The first of these makes use of the Volhard titration, which was first employed as an ultra-micro method by Wigglesworth (1938), but in this case the burette and associated apparatus described in the preceding section is used. A similar procedure to that described by Wigglesworth for the addition of nitric acid, silver nitrate and indicator is used, and the end-point facilitated by using a white titration platform illuminated with a ‘daylight ‘electric lamp. Using this arrangement and estimating amounts of chloride of about 1 µg., it was quite easy to keep the standard deviation for the estimation of identical quantities down to the ± 1 % level. The method can, in fact, be used to determine much smaller quantities of chloride than this with little loss of accuracy. By reducing the tip diameter of the burette and the size of the titration drop, and by observing this drop during the titration under a binocular microscope, it is possible to reduce the scale so that quantities as small as 0·01 µg. chloride can be measured to ±2%.

The other technique that can be used with the same apparatus is an electrometric method similar to that of LinderstrØm-Lang, Palmer & Holter (1935) in which the chloride is titrated directly with silver nitrate, and the end-point is indicated by the change in potential between two electrodes. The burette is the same as used in the first method except that a piece of platinum wire is fused into the vertical limb of the burette, which together with the silver nitrate in the burette served as one of the electrodes and gave a fixed potential. The other electrode consisted of a fine silver wire which dipped into the titration drop (see Fig. 1). The potential between these two electrodes depends on the concentration of silver ions in the titration drop. This potential was measured with a direct-reading valve voltmeter, and at the start of the titration with 0·1 N silver nitrate in the burette a reading of about 210 mV. is obtained. During the course of the titration this gradually falls to about 170 mV., and then suddenly drops to 70 mV. as the endpoint is reached. The titration should be stopped when the potential has fallen to 120 mV. Towards the end of the titration fluid from the burette must be added slowly in order to give time for the new potential to be established and to avoid overshooting the end-point. This is best done by means of a pressure regulator used in the way described above since contact between the titration drop and the burette tip should not be broken during the titration.

The accuracy at the 1 µg. chloride level is ± 1 %, and therefore comparable with the Volhard method. The latter method is the simpler of the two, for the electrometric method requires a pressure regulator and a valve voltmeter, but the electrometric method has the advantage that the end-point detection is an automatic one.

To the incinerated sample in the precipitation tube is added 50 µl. of a 5% solution of ammonium oxalate adjusted to pH 5 to prevent the co-precipitation of magnesium. The mixture is stirred and allowed to stand for 1 hr. and then the precipitate separated in the manner already described. The precipitate is washed with 50 µl. of concentrated ammonium hydroxide solution which has been diluted 10 times with distilled water. The pecipitate is dried at 100°C. and then heated at 450°C. for half an hour to convert the oxalate to carbonate. The carbonate is dissolved by the addition of 50 µl. of N/5-HCl, and finally heated to dryness at 100µ C. The excess HCl evaporates off and calcium chloride is left behind. This is transferred, by washing, to the titration platform and the chloride content measured.

Magnesium is precipitated from solution by the use of 8-hydroxy quinoline. In order to avoid interference from calcium, the precipitation is carried out on the supernatant from the calcium separation described above. The ammonium oxalate supernatant, together with the washing solution, are transferred directly to a precipitation tube. To this tube also is added 1 µl. of a 5 % solution of 8-hydroxyquinoline in alcohol and the mixture stirred with a fine glass rod. The tubes are capped, placed in the 100°C. oven for a few minutes to start the precipitation, and then left for half an hour at room temperature for it to be completed. The precipitate is separated in the usual way and washed also with the ammonium hydroxide solution. The presence of the relatively high concentration of ammonium salts does not seem to interfere with precipitation by the hydroxyquinoline, but these salts can be removed, if desired, by evaporating the solution to dryness and heating to 450°C. for a few minutes to decompose them.

The washed precipitate is heated at 450°C. for $12$–1 hr. which decomposes the organic compound leaving behind magnesium oxide. This is dissolved in dilute HCl and the solution evaporated to dryness in the same way as in the calcium method. The magnesium chloride which remains is measured as before.

Sodium is precipitated by a saturated solution of zinc uranyl acetate as the triple salt, sodium zinc uranyl acetate. The first attempt at converting this acetate to a chloride was by simply heating it with HCl and evaporating the solution to dryness. It was hoped that the acetic acid would distil off followed by the excess HCl and leave behind the chloride. However, trials showed that the yield of chloride was low and variable, possibly due to the formation of basic chlorides. Experimenting with simpler uranyl salts it was found that a much more consistent sample of uranyl chloride could be prepared by the action of dry HCl gas on the oxide (UO₃). This oxide could not be prepared directly from uranyl acetate by heating as it decomposed at a lower temperature than was required to decompose the acetate, but it was found that the oxide could easily be prepared by heating the acetate below 400°C. with a small quantity of ammonium nitrate. This substance itself decomposes at this temperature and also oxidizes the acetate to leave behind the red oxide.

Tests on a standard solution of uranyl acetate were carried out. The oxide was formed in the usual precipitation tube and converted to chloride by passing dry HCl from an HCl generator into the tube by means of a glass capillary. Formation of the chloride was very rapid and could be seen by a change in colour from red to yellow-green. The tubes then were heated to 100°C. to drive off any HCl and the chloride content measured. The results shown in Table 2 indicate a variation of less than ±2%, but the yield of chloride is only 83% of that expected if the acetate were converted quantitatively to UO₂C1₂. The experiments were repeated with standard solutions of sodium zinc uranyl acetate and a large number of determinations (see Table 2) again demonstrated that the variation between readings was small (between + 1 and 2 %). The yield for this salt also was between 82 and 83 % of that expected, but since this was constant and since the composition of the uranium chlorides is not fully known, the fact that the amount fell short of the theoretical was not very serious. The reaction was used as the basis for the sodium method and a correction applied in the calculation to compensate for the low yield.

In the complete method 50 µl. of a saturated solution of zinc uranyl acetate is added to the incinerated sample in the precipitation tube and the mixture stirred. The tubes are allowed to stand for half an hour for the completion of the precipitation. The separation is performed in the usual way, and the precipitate washed first with 50 µl. of ethyl alcohol and then with ethyl alcohol which has been saturated with sodium zinc uranyl acetate. The separated precipitate is dried at 100°C. together with a small drop of saturated ammonium nitrate solution. This was heated to 300-400°C. until the oxide was formed and the chloride produced in the manner described above. The sodium content is calculated on the assumption that 7·4 atoms of chloride were equivalent to one of sodium. This figure might be different if the reaction is carried out under different conditions, and should be checked for each new series of sodium measurements.

Potassium is precipitated as the chloroplatinate and free chloride ions liberated from this salt by the reduction of the platinum salt with sodium formate (Cunningham, Kirk & Brooks, 1941b). A slightly different method was necessary for the preliminary incineration of the sample since contact between the ‘Hysil ‘tube and the sample at high temperatures invariably lead to serious losses of potassium. The sample can be placed on a loop in the middle of a piece of platinum wire which stands upright in the tube so that the sample does not come in contact with the glass. After the incineration is complete, sufficient of a 4% solution of chloroplatinic acid in 80% alcohol is added to the tube to cover the loop in the wire, and the solution is stirred by the wire itself so that the ash comes off the wire and the precipitate is formed in the tube. The wire is removed, and after $12$ hr. the precipitate is separated in the usual way and washed with absolute alcohol saturated with potassium chloroplatinate. 50 µl. of 0·2M sodium formate is added to the precipitate, and the mixture evaporated to dryness at 100°C. and the chloride content estimated as before.

Some results obtained by the applications of these methods to the estimation of the four cations, both from solutions of their chlorides and from blood samples, are shown in Table 3. The measurements of potassium, calcium and magnesium all show standard deviations of between ± 1 and 2 % for both the inorganic samples and for the blood samples. Estimations of mixtures of calcium and magnesium indicate that the interference between two is of a low order. With the sodium determinations the error is much greater—the standard deviations are as much as ± 4%. The larger error in these measurements arises largely during the process of separating the precipitate of sodium zinc uranyl acetate from the supernatant, and this is due to the fact that the precipitate and the supernatant are similar in chemical composition and in density. It is possible that the accuracy might be improved by the adoption of a filtering technique rather than centrifugation for the separation of the precipitate. As the method stands, at least two estimations of sodium concentration must be made on each sample so that the error can be reduced to the same level as in the other cation measurements.