ON THE DEPTH AT WHICH FISH CAN SEE GEORGE L. CLARKE Biological Laboratories, Harvard University, and Woods Hole Oceanographic Institution 1 ' At what maximum depth fish can see in the ocean and in lakes has remained a matter for speculation up to the present time, first because little was known regarding the strength of the illumination at various depths and second because no adequate tests of the sensitivity of the fish eye had been made. During the past 10 years, however, the penetration of daylight into natural waters has been carefully studied by a number of investigators in various parts of the world. Furthermore, the recent excellent work by Grundfest ('32) on the sensibility of the sun-fish, Lepomis, gives us a reliable yardstick for one species of fish. In view of the fact that these two sets of data are now available it occurred to me that it would be of interest to calculate the greatest depth in several different types of water at which the sun-fish could just discern an object of a certain size at a given distance. Although the validity of certain assumptions must be granted in order to make such a calculation, it seems probable that we can arrive at the right order of magnitude for this fish. Whether the photosensitivity of other fish is similar to that of Lepomis can be decided, of course, only by the performance of suitable experiments. The vision of fish with " telescopic " eyes and of other highly specialized species is undoubtedly on an entirely different scale. 2 In the experiments on Lepomis, which were carried out for quite another purpose, Grundfest made use of the rheotropic response. The fish to be tested was confined in a cylindrical glass jar around which was placed a cylindrical screen composed of equal and alternate vertical bars and spaces and illuminated to any desired intensity by a band of light, 2 mp. wide, from any part of the spectrum. When this screen was rotated in either direction, the fish responded by a sudden movement in the direction of the rotation, provided that the illumination was above a certain threshold the value of which depended upon the part of the spectrum being used. By repeated trials of 1 Contribution No. 100. 2 A correlation undoubtedly exists between the degree of enlargement or of degeneration of the eyes of deep-sea fish and the depth (and consequently the illumination) at which they live. The issue is confused, however, by important differences which apparently depend upon whether the fish is bottom-living or pelagic and whether or not it possesses light organs. This problem has not yet been satisfactorily elucidated (Murray and Hjort, '12, p. 680). 452
July, 1936 ON THE DEPTH AT WHICH FISH CAN SEE 453 this sort the spectrum was explored and the spectral sensibility curve determined. The fact that the sensibility curve for Lepomis shows a sharp peak between wave-lengths 530 and 550 m,u is of interest because it is this general part of the spectrum of daylight which penetrates most effectively into most lakes and into coastal ocean water-the part of the sea in which fish occur in the greatest abundance. The water of the Sargasso Sea, on the other hand, is more transparento blue light, while in the case of certain lakes and ponds the water contains stains which absorb the shorter wave-lengths so rapidly that the yellow or red component of daylight is found to predominate. H. R. James ('36) has calculated on the basis of laboratory experiments that of the energy of daylight between 300 and 800 mp. which would reach a depth of 100 meters in distilled water 60 per cent would fall in the blue band and only 23 per cent would be in the green. At the surface 17.7 per cent of the energy is in the blue region and 12.3 per cent in the green. These results raise the question of the possibility of a shift in the sensitivity of the eye of a deep water fish toward the blue end of the spectrum. In Grundfest's experiments the minimum brightness at which the fish would respond for the 2 mp. band in the most effective part of the spectrum corresponded to a brightness of 1 X 10-6 millilamberts for white light as judged by the human eye. Although, theoretically, allowance should be made for the relative effectiveness of other parts of the spectrum since in nature the illumination to which the fish is exposed is obviously not confined to a narrow band, this unfortunately cannot readily be clone because of the method used to measure brightness. However, it is extremely doubtful that a large correction would be necessary and since we are attempting to find merely the order of magnitude, this difference may be neglected. The assumption has tacitly been made that when the fish ceases to respond it no longer " sees " the revolving screen. The further question arises of whether the visibility of a series of bars and spaces moving across the field is the same as that of a single object of the same diameter as the width of the bars. Hecht and vvald (' 34, p. 520, footnote) found that slow movement, such as was used in all the experiments with this method, does not affect the visual acuity curve in man. Therefore, since the bars were 2 mm. wide and at a distance of 1 em. from the fish's eye (thus subtending an angle of about 11 ), we may assume that the fish could just see under an illumination of 1 X lq- 6 millilamberts an object 2 nun. in diameter at a distance of 1 em., or a 2 em. object at 10 em., or a 20 em. object at lmeter, and so on. Since the sun-fish eye is so much more sensitive to radiation between 500 and 600 mp. than to other regions of the spectrum it is most appropriate to use data for the intensity of submarine irradiation obtained by a photometer, such as one of those employed by Oster and Clarke (' 35), with a maximum sensitivity in that region. According to their results the irradiation from 490 to 620 mp. at the surface of the sea ott the coast of Massachusetts during
454 GEORGE L. CLARKE Ecology, Vol. 17, No.3 the middle of a bright day in summer reaches an intensity of about 20,000,uw/cm. 2 In the Sargasso Sea the irradiation at the surface was only slightly greater, having a magnitude of about 25,000 ftw /cm. 2 Now the minimum brightness for the vision of the sun-fish of 1 X 10-6 millilamberts is equivalent to an irradiation of 1 X 10-6 X 1.5 microwatts/cm. 2 = 1.5 X IQ- 6,uw jcm. 2 Since an intensity of 20,000 uw jcm. 2 is approximately 1Q- 10 times this, it is apparent that the fish can see under an illumination of 10-10 times the maximum value of daylight-or a reduction of ten logarithmic units. This is approximately the same as for the human eye. The instruments which have been found practicable for light penetration investigations have not been sufficiently sensitive to follow the reduction of light in the water for more than three, or at most four, logarithmic units. However, if we assume that the transparency of the water below the maximum depth reached by the photometer is the same as that in the surface strata, the data may be extrapolated. This assumption is justified, to some extent at least, by the fact that in many cases (and in all the cases considered here) there is no radical change in transparency within the depths actually measured. Accordingly, the approximate depth at which the minimum intensity for vision exists (during the hours of bright daylight at the surface) for each type of water investigated has been calculated by finding the average number of meters of water causing a reduction in intensity of one logarithmic unit and multiplying this value by ten. In so far as it is reasonable to assume that the vision of our common marine fishes is similar to that of Lepomis, which inhabits fresh-water ponds and lakes, the figures in the upper part of table I, based on the data of Oster and Clarke, give the extreme depths at which vision would be possible in different types of water. According to these approximations fish would be able to see objects on the bottom in coastal waters. In the case of localities where the " depth of minimum illumination " is much greater than the depth to the bottom, such as Georges Bank, it appears that vision would be possible for fish during most of the day because the intensity of daylight is within one logarithmic unit of its maximum for several hours before and after noon (Clarke and Oster, '34, fig. 6). Although taste, smell, and touch are undoubtedly used in certain cases for the location of food, most fishes apparently depend chiefly upon sight ( Bigelow and Welsh, '24, p. 416 and p. 451). The general conclusion that vision is possible for fish on the bottom in coastal waters is supported by the following observation made by Mr. R. A. Nesbit of the United States Bureau of Fisheries. Mr. Nesbit found that white hake which he took in a bottom trawl at a depth of 132-154 m. off Bodie Island, North Carolina, were gorged with butterfish and other small active fish which presumably the hake could not have caught unless they had been able to see at that depth. In order to compare these data on fishes with the ability of divers to see under water, I made inquiries of Mr. H. V. Greenough who had been working
July, 1936 ON THE DEPTH AT WHICH FISH CAN SEE 455 with conventional diving equipment on a wreck in 29 m. of water in Vineyard Sound about five mhes distance from the point off Gay Head where Oster and Clarke had made measurements. Mr. Greenough told me that at that depth tools could be clearly seen at arm's length during several hours in the middle of bright days in July. On one occasion sufficient light existed to permit work on the wreck to continue until 7 o'clock in the evening. The value arrived at for the Sargasso Sea of 430 meters is of interest in comparison with Beebe's observation from the Bathysphere that complete darkness for the human eye was reached at a depth of between 520 and 580 m. Dr. Beebe has informed me (private communication) that without artificial illumination he was able to see a non-luminous fish four inches long ( Cubiceps) at a depth of 210m., a shark at 240m., and a jelly-fish (Aurelia) at 350 m. although entangled luminescent organisms may have aided vision in the last case. In another part of the Sargasso Sea, Helland-Hansen ('31) found that a highly sensitive photographic plate was slightly blackened after an exposure of 80 min. at a depth of 1000 meters. If we assume that a deep sea fish can see as well in blue light as Lepomis can in green light, then the maximum depth for vision in the Sargasso Sea would be about 750 meters since an average of 75 m. of this type of water are required for a reduction of one logarithmic unit in the blue region of the spectrum (Clarke, '33). The depths for the limit of vision for the sun-fish have been similarly calculated for three of the large number of lakes in northern Wisconsin whose transparencies have been carefully studied by Birge and Juday ('31, '32) and for one lake in Minnesota investigated by Erikson ('33), although the different methods used in the measurement of light penetration make comparisons difficult. The values arrived at for these four lakes are set forth in the lower part of table I. Data on the depth at which the energy is reduced to 10 per cent of its surface value for a large number of other lakes are available from the laboratory measurements of James ('36). Crystal Lake is the clearest of all those investigated and Adelaide Lake is one of the most TABLE I. M a.rimum depth for vision of fish similar to Lepomis in various types of water Depth of Thickness of stratum minimum illumination for vision of fish Depth of causing reduction similar to Lepomis water to Types of water of 1log unit (10 X column 1) the bottom Deep Basin of Gulf of Maine 23 meters 230 meters 165 meters Georges Bank Woods Hole Harbor 18 7.5 180 75 60 20 Off Gay Head Sargasso Sea 13 43 130 430 30 4500 Crystal Lake, Wisconsin 11 110 21 Adelaide Lake, Wisconsin 2.8 28 22 Trout Lake, Wisconsin 6 60 35 Gunflint Lake, Minnesota 7 70 50
456 GEORGE L. CLARKE Ecology, Vol. 17, No. 3 turbid with the exception of those which are deeply stained. Trout Lake is the deepest of the Wisconsin lakes studied. In all of these and even in Gunflint Lake which is still deeper, the calculations indicate that, just as in coastal waters, vision would be possible for fish similar to Lepomis all the way to the bottom. SUMMARY It has been calculated from the data of Grundfest ('32) and Oster and Clarke ('35) that the minimum iliumination for vision for the sun-fish, Lepomis, is approximately lq- 10 times the maximum value of daylight. The depth at which daylight would suffer a corresponding reduction in different types of ocean water and lake water has been found by extrapolating from existing measurements. The results indicate that for fish whose visual sensitivity is similar to that of Lepomis vision would be possible at the bottom in coastal regions and in lakes and at a depth of at least 430 m. in the Sargasso Sea. LITERATURE CITED Bigelow, Henry B., and W. W. Welsh. 1924. Fishes of the Gulf of Maine. Bull. U. S. B11r. Fisheries 40, Document No. 965. Birge, E. A., and C. Juday. 1931-1932. Solar radiation and island lakes. Trans. Wis. Acad. Sci., Arts, and Letters 26: 383-425; 27: 523-562. Clarke, G. L. 1933. Observations on the penetration of daylight into mid-atlantic and coastal waters. Biol. B~tll. 65 (2): 317-337. Clarke, G. L., and R. H. Oster. 1934. The penetration of the blue and red components of daylight into Atlantic coastal waters and its relation to phytoplankton metabolism. Biol. B~tll. 67: 59-75. Erikson, H. A. 1933. Light intensity at different depths in lake water. I. Opt. Soc. Am. 23: 170-177. Grundfest, H. 1932. The sensibility of the sun-fish, Lepomis, to monochromatic radiation of low intensities. I. Gen. Physiol. 15: 307-328. --. 1932. The spectral sensibility of the sun-fish as evidence for a double visual system. I. Gen. Ph) siol. 15: 507-524. Hecht, S., and G. Wald. 1934. The visual acuity and intensity discrimination of Drosophila. I. Gen. Physiol. 17: 517-547. Helland-Hausen, B. 1931. Physical ocea,nography Sars N. Atlant. Deep-sea Exp., 1910 1: 43. and meteorology. Rept. Michael James, H. R. 1936. Laboratory study of absorption of light by lake waters. Trans. Wis. Acad. Sci.. Arts, and Letters 30: (in press). Murray, J., and J. Hjort. 1912. The depths of the ocean. Macmillan and Co. 819 pp. Oster, R. H., and G. L. Clarke. 1935. The penetration of the red, green, and violet components of daylight into Atlantic waters. I. Opt. Soc. Am. 25: 84-91.