The Reactions of Pygosteus Pungitius L. to Toxic Solutions

The gradient tank method of experimentation was devised by Shelford & Allee (1913), who used it to observe the reactions of fish to dissolved oxygen and carbon dioxide. A number of other studies on the same lines followed ; Wells (1913,1915 a, b) also dealt with oxygen and carbon dioxide, with gradients of acidity and alkalinity, and with the chlorides, nitrates and sulphates of ammonium, calcium and magnesium. Later the work extended to pollutant substances of a toxic nature; Shelford (1917, 1918) studied the reactions of several species of fish to the various waste products associated with the manufacture of coal-gas. This study included ammonia and ammonium salts, aniline, hydrogen sulphide, sulphur dioxide, carbon bisulphide, acetone, phenol, cresols, naphthalene, carbon monoxide and tar acids; and Wells (1918) contributed a further study with carbon dioxide and carbon monoxide. A description of the apparatus used is included in Shelford’s book on Laboratory and Field Ecology (1929, p. 85).

These investigators claimed that fish recognize and avoid certain substances, such as carbon dioxide, quinoline, high concentrations of sulphur dioxide and water of high acidity, but that they fail to detect many substances of high toxicity, swim into the solution, become ‘intoxicated’ and thereafter avoid the water end of the tank. In a long list of ‘positive’ substances they include ammonia, carbon monoxide, paracresol, phenol, xylene, acetone and pyridine.

The method of study is open to some measure of criticism. The use of a gradient seems advisable when the fish is required to select the optimum concentration of a substance, but in dealing with substances of high toxicity, of which extremely low concentrations may be ultimately fatal, it is preferable to employ an arrangement in which water and solution are sharply differentiated so that the fish is given every opportunity of discriminating between water and a definite concentration of the substance tested. Furthermore, the time during which the movements of the fish were recorded was generally limited to 10 or 12 min., and often bears no relation whatever to the degree of toxicity of the solution employed; thus Wells (1918) describes an experiment to determine the reactions of the black bullhead (Ameiurus melas) to carbon monoxide; the chart records the movements of the fish for 12 min. only, the concentration employed is 0·5 c.c. CO/1., but according to his table (p. 562) an 11·7 c.c./l. solution is only fatal in 9 hr. 55 min. Nevertheless, the fact that the fish spent most of its time at the solution end of the tank is taken as evidence that fish are ‘positive’ to carbon monoxide. Although many control experiments were run, with water at both ends of the tank, these control experiments do not seem to have been conducted with the particular fish whose reactions to the solutions were recorded. The control experiments of the writer, and those of Shelford (1917, chart 1), show that the behaviour of individual fish under control conditions may vary considerably.

Lastly, the gradient tank apparatus seems to have been somewhat unsatisfactory. Wells (1915b, p. 246), in his experiments with salts, found that with the solution running at the standard rate a sharp vertical gradient is also produced; at one end of the tank the solution appears to have been six times more concentrated at the bottom than at the top. Consequently it is extremely difficult to assess the concentration differences the fish encounters in its movements to and fro.

The writer (1935,1938, 1939, 194a, 1947) has investigated the degree of toxicity of heavy metal salts to fish and the nature of their lethal action, but no attempts appear to have been made to ascertain whether fish can recognize and avoid these substances. The present study deals with the reactions of the ten-spined stickleback to mercuric chloride, zinc sulphate and copper sulphate, and for comparative purposes a number of experiments were run with chloroform, alcohol and formalin. The apparatus provides a sharp differentiation between water and a known concentration of the toxic solution, enables the reactions of the fish to be checked by rapid reversals of the direction of flow of water and solution, the periods of observation are considerably lengthened, and in every experiment water was run in at both ends of the apparatus for the first 10 or 15 min. so that control observations were made on all the fish used. Pygosteus pungitius was found very suitable material as it is plentiful locally, and its small size obviates the necessity for cumbersome apparatus and the preparation of large volumes of solutions.

The general scheme of the apparatus employed is shown in Fig. 1, which is largely self-explanatory. Aspirators A and C are filled with tap water, B and D with the solution to be tested; solution and water are fully aerated. The ‘gradient tank’ consists of a horizontal glass tube, 30 mm. internal diameter, 58 cm. long, capacity about 400 ml. The apparatus is disconnected at one end, the tube opened and the fish introduced. Then the apparatus is reassembled and by temporarily tilting the tube and opening clips 3 and 4 it is completely filled with water. When all air is expelled clip 3 is closed. The outlet tubes are also filled and a sufficient quantity of water introduced into the bubble-traps E and F.

Sufficient time is now allowed for the fish to become accustomed to their new surroundings. If they are gently handled and not subjected to any great temperature change they become quiet in 10−15 min. and either rest perfectly still on the bottom or swim lazily up and down. Then clips 1 and 4 are opened and screw-clip 6 is released so that water flows in at both ends and out at the outlets G. The rate of discharge at H is regulated to about 300 ml./min.

The stop-clock is started, and every half minute or minute, according to the duration of the experiment, the approximate positions of the fish in the tube are recorded. Two parallel lines on graph paper represent the length of the ‘gradient tank’, and minutes are marked off on a vertical scale. Water is run in at both ends for 10 or 15 min. The behaviour of the fish under these conditions varies somewhat; all may rest perfectly still at various points on the bottom, or they may swim up and down and eventually show a preference for one end. The reason for this preference could not be discovered, as in one experiment the fish might congregate on the left side, while in another experiment conducted immediately afterwards with identical lighting conditions, etc., they might show an equal liking for the right. A uniform white background was provided, and the only illumination was a single 75 W. light suspended $312$ ft. above the middle of the tube. It was soon discovered that the fish showed little reaction to external lights and objects but reacted sharply to bubbles or other objects within the tube, and to vibrations of the experiment bench. Hence such disturbances were carefully avoided.

After 10 or 15 min., by opening and closing the appropriate clips, the water flow is stopped on one side and solution run in instead. The traps E and F prevent the entry of bubbles and serve to render the replacement of water by solution less sudden. The positions of the fish are recorded periodically, and notes made on their behaviour. In general, the solution is run in on the side for which the fish show a preference.

With a flow of 300 ml./min. at H an extremely sharp differentiation is obtained, one-half of the tube being filled with water, the other half with solution. This can be checked with dyed solution, when the distance between the two outlets (G) can be adjusted for best results. Energetic swimming causes temporary mixing in the middle of the tube, but the sharp separation of water and solution is almost immediately restored. The use of dyes also shows that a complete reversal of water and solution can be effected in 3−4 min. with the standard rate of flow. A little difficulty was encountered with concentrated solutions (e.g. N/10 CuSO₄) whose high specific gravity tended to make them spread along the water half of the tube, but here again experiments with dyes showed that this could be overcome by running water and solution at a slightly increased rate. No dye was used, of course, in the experiments in which the movements of the fish were recorded.

The fish used were obtained from a local ditch and kept in the laboratory for 2−7 days before use; 25−28 mm. was found the most suitable size, though 30 mm. fish could turn round in the tube with no difficulty. They usually showed no reaction to the current, though when much enfeebled by the toxic effect of the substances studied they tended to float on the current towards the outlets. The experiment was usually discontinued at this stage.

The room temperature was maintained at 15° C. Solutions and tap water were fully aerated and allowed to stand 3−4 hr. before the experiment was begun in order to ensure equality of temperature. At temperatures above 15° C. the fish tend to be unduly restless.

A number of experiments were run with alcohol, chloroform and formalin, substances that may be presumed to have an unfamiliar taste and smell to fish. In most of these five fish were placed in the apparatus and the experiment conducted as just described, but a series was also run in which single fish were used and all their movements recorded by the method of Shelford and Wells. Some typical results are given in Fig. 2, which also serves to illustrate the varied behaviour displayed by different fish with water running in at both ends ; thus in the second experiment the fish remained motionless until the solution was admitted, whereas that in the fourth was exceptionally restless.

On running in 4% ethyl alcohol the fish is almost immediately intoxicated, its breathing becomes slow and irregular; later the fish falls over on its side and is incapable of movement so that a pseudo-positive reaction is obtained. On restoring the flow of water (26 min.) it rapidly recovers, at 35 min. is upright and breathing vigorously, in 38 min. is swimming almost normally. With 1 % alcohol a typical negative reaction is shown; note the somewhat hesitant return to the left side of the apparatus on replacing the solution by water.

The results with chloroform resemble those obtained with alcohol. A typical negative reaction is displayed towards 1/5000 and 1/10,000 solutions, but high concentrations (1/2000) may anaesthetize or kill so rapidly that a ‘positive’ reaction appears. Formalin in concentrations of 0·1−0·4% (v/v dilution taking the concentrated solution as 40%) also gives a negative reaction with very obvious irritating effects.

A representative series of the results obtained with mercuric chloride is set out in Fig. 3. The reaction to a 0·003N solution is violent; on running in the solution the fish make a frantic exodus to the water end, and thereafter, should a fish venture towards the solution, immediately it reaches the water-solution junction it moves jerkily with gobbling respiratory movements and retreats. An almost equally marked effect is produced by a 0·001 N solution, but at 0·0003 M though the salt is highly toxic at this concentration (survival time 31 min.) the reaction is much slower, for it will be noted that 9−10 min. went by before all the fish departed, and the frantic swimming and gobbling respiration seen at the higher concentrations was no longer evident. A number of experiments were carried out at 0·0001 N with somewhat indefinite results. The fish would show a negative reaction if, under control conditions, they had shown no particular liking for either end of the apparatus, but when a preference for one end existed they would not leave the solution if run in on that side. One experiment was carried oui at 0·00004 N, that shown in Fig. 3. For the first 30 min. no reaction developed, the fish kept swimming into the water but returned, swimming into the solution with no hesitation. Later a vague tendency to prefer the water zone developed, but by this time the fish were showing obvious signs of respiratory distress and the experiment was discontinued.

At the end of each experiment the fish were removed, washed in two changes of water and placed in an aquarium. A high proportion died within 24 hr., for mercuric chloride has such a rapid astringent or corrosive action upon the gill membranes that brief exposure to high concentrations may ultimately prove fatal.

Eleven experiments were made with zinc sulphate solutions and four are recorded in Fig. 4. A prompt negative reaction is shown at 0·04 N, the fish almost immediately select the water zone and avoid the water-solution junction. Occasionally a fish enters the salt zone where it swims with great agitation until it reaches the watersolution junction ; here it hesitates for a moment and then proceeds into the water.

At 0·003N a negative reaction is evident but is much less prompt; the fish show little sign of discomfort and seem to discover the water end of the apparatus by trial and error. The result of reversing the direction of flow of water and solution are evident, and the improved response on the second reversal will be noted. Three experiments were run at o·ooo6N, and in every case a negative reaction was observed though it might take 18−20 min. to develop, and so experiments limited to 10 min. would indicate fish as positive to this concentration. At 0·0003 N (approximately 10 parts Zn/million) a negative reaction is still evident, though water-solution reversals are necessary to show it up clearly and the movements of the fish are very leisurely. A solution of this concentration has no taste, even when held in the mouth for some time, and cannot be distinguished from tap water. It is interesting to find that the fish can detect and avoid in a few minutes a solution in which they can survive for 7 hr. Bull (1937) has shown that teleosts can detect very small salinity differences ; thus Gasterosteus aculeatus will develop a conditioned response to a salinity change of only 0·23 part/1000; but the same author (1930) records that when Blennius pholis is presented with an unfamiliar substance (trinitro-butyl-toluene) it fails to recognize it; at least over fifty trials failed to produce any conditioned response.

At 0·0001 N (about 3 parts Zn/million) Pygosteus displays little reaction, but as Fig. 4 indicates, the fish do not become ‘positive’ to the solution and avoid the water, for the water-solution reversals at 46 and 65 min. produce no result. The survival time at this concentration is about 15 hr., and it is possible that the fish would shun this very low concentration if given sufficient time. At the end of each experiment the fish were transferred to aquaria and all survived.

Experiments were also carried out to find whether Pygosteus can recognize and select water after a period of immersion in zinc sulphate solution. Two are recorded in Fig. 5. In the first 0·01 N ZnSO₄ was run in at both ends of the apparatus for 18 min., in the second for 36 min. In both cases the fish congregated in the water and avoided the solution almost immediately the change was made. In the second experiment the toxic process was well advanced, for marked respiratory distress had become evident.

No experiments have been carried out as yet to test whether repeated exposures to dilute concentrations of zinc sulphate will produce an improved negative response, but the writer hopes to pursue this line of investigation shortly. It may be noted that the concentrations to which Pygosteus will react when quite unfamiliar with metallic pollution may be exceeded under natural conditions, for the writer (19406) has shown that in 1939 the Frongoch stream, a tributary of the river Ystwyth, contained 57 mg./l. of zinc in solution (about 0·0018 A). Trout occur in confluent tributaries, but whether they enter the polluted stream is not known.

The effect of copper sulphate solutions was studied over the concentration range 0·1−0·001 N, and three typical results are set out in Fig. 6. At 0·1 N a fairly definite negative reaction is evident, but it is possible that here the effect is at least partly due to the colour, acidity and high osmotic pressure of the solution. A 0·1 N solution cannot be brought to the pH of the tap water for copper hydroxide is precipitated on the addition of alkalis. At 0·04A the reaction is vague; it will be seen that with water running in at both ends of the apparatus the fish demonstrate a preference for the left side ; on running in the copper sulphate this preference is lost but the fish persist in swimming into the solution. At 0·01 N all sign of negative reaction has vanished; on the contrary, the fish congregate at the solution end of the apparatus, and a similar result is obtained at 0·001 N.

The explanation of this curious positive reaction was forthcoming when fourteen experiments had been run at 0·001−0·01 N and the behaviour of the fish carefully watched. The fish are totally incapable of recognizing the solution and swim across the water-copper sulphate junction in either direction without the slightest hesitation. (See the recordings for single fish in Fig. 7.) In the solution they tend to become stupefied and rest motionless on the bottom ; if they succeed in reviving a little and visit the other end of the apparatus the water seems to stimulate them, and they keep up erratic movements which eventually land them back in the solution where they spend another period of stupor. The charts in Fig. 7 show how the fish nearly always persist in swimming at the water end and remain still in the solution for periods of several minutes. In the five fish experiments at any particular time when the toxic process is well under way most of the fish are lying still in the solution while one or two are swimming in the water.

The ‘positive’ result therefore is not due to the fish deliberately selecting the solution, or becoming so intoxicated by it that they avoid the water. The whole explanation lies in the fact that they cannot perceive the toxic substance so that it acts as a trap. A parallel is to be found in the experiments of Jennings & Moore (1902) with Paramecium and Chilomonas; it was found that these organisms are apparently attracted to a drop of M/180 HC1, for they rapidly form a dense aggregation around it. Loeb (1918), reviewing their work, has shown that the explanation of this positive effect is that the drop of acid acts as a trap into which the organisms swim to be paralysed or killed, so that the density of the organisms around the drop increases until a dense ring of them is formed.

A 0·.01 N solution of copper sulphate has a very distinct and highly unpleasant metallic taste, and it is interesting to find that fish cannot perceive it when they can detect such great dilutions of zinc sulphate. Furthermore, copper sulphate seems to destroy the stickleback’s ability to distinguish other substances, for the usual negative reactions are not displayed to chloroform, mercuric chloride and zinc sulphate solutions to which copper salts are added. Thus Fig. 8 shows that a 1/10,000 chloroform solution containing copper sulphate produces no reaction and the fish become anaesthetized, and that copper chloride inhibits the reaction to mercuric chloride with the result that the fish are all dying after running the solution for 16 min. Copper chloride was used in this experiment as mercuric sulphate is an unstable salt. In the experiment with zinc sulphate and copper sulphate the reaction is ‘positive’.

A number of experiments were also carried out in which the fish were immersed in 0·01 N copper sulphate for periods of 5−15 min., water run through the apparatus for 3 min., and then chloroform or formalin solution on one side. In every case either no negative reaction was evident, or it was much delayed.

The present study does not aspire to be an exhaustive study of the reactions of fish to heavy metal salts, for it is obvious that many more remain to be tested, and there is no reason to suppose that other species of fish will react in exactly the same way, but certain conclusions can be drawn. It is obvious that with any particular substance the reaction of the fish depends on the concentration difference maintained at the water-solution junction. Secondly, it will be noted that the reaction displayed bears no relation to the general degree of toxicity of the salt, for zinc sulphate, a heavy metal salt of comparatively low toxicity, is detected at great dilution, while copper sulphate, one of the most toxic, is detected only at very high concentrations. Lastly, it appears that when we employ mixtures of substances the presence of one may impair or inhibit completely the reaction usually displayed towards the other when present alone, so that the study of the reactions of fish to effluents containing some variety of lethal substances becomes one of great complexity.