The ability of aquatic animals to discriminate between different qualities in the surrounding medium can be studied with the fluviarium technique (Höglund, 1953, 1960; Lindahl and Marcström, 1958). Using this technique the reactions of small specimens of fish (cf. Table 1, p. 12) to environmental differences in 02, pH, C02, and SWL (sulphite waste liquor) have been studied comparatively in chemically well-defined, stable, and reproducible concentration gradients of different types and angles (Fig. 1). The fish are allowed to swim freely about in a test yard which is a confined space (23.5X33 cm) of a streaming aquarium called a fluviarium. Perpendicularly to the direction of flow (1 cm/sec) ten uniformly wide (3.3 cm) concentration steps are established in different standard conditions.Gradients of various steepnesses all rising from nil at one side to a certain top concentration along the other are most frequently used. In the steepest gradients used in a series of experiments the test fish encountered concentrations well beyond the actual tolerable limits, at least in the most contaminated parts of the yard. Intact animals as well as fish with sense organs eliminated have been tested for at least 30 minutes in each standard gradient. The visits to each concentration were recorded at even intervals (usually every 30 seconds) with the aid of a film camera. The strength of preference reaction (avoidance or attraction) was measured as the lateral displacement (rv) from the median line of the test yard of the mean position value (mpv) of the distribution of records over the ten concentration steps, numbered 1 to 10. Rv and mpv are expressed in section number as unit (3.3 cm, i.e.one tenth of the width of the experimental trough; cf. above). The quantitative relationship between preference reaction and steepness of gradients is presented graphically i.a. in the form of 23 reaction curves. These were found to be characteristic of species and agents tested. They also indicate the actual preference threshold values. The essential features of the innate behaviour shown in nature are recognized in the fluviarium. The fish material used in the experiments represents different physiological, ecological, and ethological types. The courses of the reaction curves of different species must be estimated against the specific behaviour displayed in the test yard under control conditions (Fig. 24). That is, when pure water is poured through the apparatus, some 135 cyprinids as for example the roach (Leuciscus rutilus L.) continually swim about in a rather aimless way. They steer up against the current (Fig. 20) and avoid to come into bodily contacts with the walls and the bottom of the test yard. When exposed to gradients of an environmental factor which acts as a directive stimulus, the animals hesitate or change direction mainly when meeting rising concentrations. Parr stages of the Atlantic salmon (Salmo salar L.), on the other hand, exhibit a typically stationary behaviour. When confronted with pH/Pco2 gradients a more or less erratic free-swimming (appetitive) behaviour is released. This is sooner or later followed by a consummatory action, that is, seemingly upon the lack of further stimulation the parr take new resting positions on the bottom of the pure side of the test yard. Possible physiological mechanisms underlying these reactions have been discussed (Fig. 38). The test agents were chosen as representatives of environmental factors which occur in all natural waters (02, pH, C02) and those which are newly introduced in the original habitats of fish (SWL). With respect to the division of abiotic, ecological factors into “natural” and “artificial” ones, the biological significance of response which makes it possible for free-moving aquatic animals to protect themselves against adverse influences from restricted parts of environmental gradients was finally considered. The main results and conclusions arrived at as regards particular factors can be summarized in the following way. (A) Oxygen (1) In oxygen gradients rising from less than 1 mg/1, roach, salmon parr, and crayfish prefer to stay in the higher concentrations. (2) The intensity of preference reaction is due to (or) the specific behaviour in the test yard and (b) the critical oxygen tension for particular species. (3) Oxygen is a non-directive stimulus to fish and crayfish. (4) Oxygen deficiency releases an emergency reaction which is characterized, inter alia by increased swimming activity. This may be regarded as an appetitive behaviour. (5) The positive reactions (cf. p. 49) obtained in pure oxygen gradients are due to ortho-kinesis. (B) Acidity and carbon dioxide The main conclusions arrived at in the section concerning the reactions studied in combined pH/Pco2 gradients can be summarized as stated below. (1) All species studied in the present experiments are able to avoid adverse conditions in pH and combined pH/PC02 gradients. (2) Fish are able to detect and avoid C02 separately from the accompanying pH. (3) Fish show marked avoidance to molecularly dissolved C02.136 (4) Roaches show avoidance of lower pH than about 5.6 and Atlantic salmon parr to lower pH than about 5.3. (5) At the concentrations existing in the present experiments HC03~, Na+,and CH are non-directive factors in the reactions of the roach. (6) Hydrogen ions are non-directive upon the reactions of roach within the pH range of c. 5.6—10.5, and to the salmon parr from pH 5.3 to atleast 7.4. (7) Concerning pH, a certain correlation seems to exist between the tolerance limits and the directive influence upon fish. (8) Regarding various species, a certain connection is found between the avoidance reactions to C02 and the narcotic effects of C02. No corresponding connection is found as regards pH. (9) Compared with roach, salmon parr give sharper avoidance reactionsto C02, but less pronounced avoidance to pH. (10) The removal of olfactory tissues and the sectioning of the nervesinnervating the lateral organs do not essentially change the reactions ofroach, minnow, and salmon parr in pH/PCo2 gradients. (11) Acidity and carbon dioxide are either perceived by different receptorsystems or by the same receptors at essentially different thresholds. (12) The possibility is discussed that the avoidance reactions in C02gradients may be attributed to a special C02 sense connected with chemoreceptors in the gill region. (C) Sulphite waste liquor (.SWL) (1) Avoidance reactions in SWL gradients are to a great extent due to olfactory sensations. The removal of the olfactory tissues of minnows and roaches essentially extinguishes the avoidance shown by intact specimens especially in the lower concentration ranges of the reaction curves (Table 15,p. 114; Figs. 42, 43, and 45—46). A good correlation is found between the development of the sense of smell among various species and the preference thresh-old values obtained in SWL gradients (Table 17, p. 128). (2) The amount of avoidance due to odorous substances contained in the SWL diminishes in the steepest gradients. Thus no connection is obtained between the presumed incipient detrimental or lethal limits on one hand and the ability of avoiding contacts with noxious concentrations on the other. On the contrary, the sensibility 1 of the olfactory receptors or the responsiveness on the whole to sensory stimuli evidently is depressed in the steepest gradients of SWL. (3) On account of the free hydrogen ion content in SWL increasing avoidance due to free C02 which is liberated from the bicarbonate content is obtained in the steepest gradients established in hard, well-buffered waters (like Uppsala water) and to free hydrogen ions per se in steep SWL gradients 1 Cf. the foot-note on p. 106.137 in soft, poorly buffered waters (like Hölle water). The reactions to these additional factors presumably delay the descending slopes in the upper concentration ranges in the reaction curves of the minnow and the roach (Figs.39—44). (4) The discharge of SWL into the water basins may influence unfavourably upon fish life by at least two reasons, (a) In low, not toxic concentrations SWL may repel fish, especially those with well developed olfactorysenses, (b) At high, toxic concentrations the ability of fish to avoid contacts with SWL diminishes.