Effect of Water Turbidity on the Visual Acuity of Harbor Seals (Phoca Vitulina)

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Vision Research 46 (2006) 1777–1783 www.elsevier.com/locate/visres

Eﬀect of water turbidity on the visual acuity of harbor seals (Phoca vitulina)

Michael Weiﬀen a, Bettina Mo¨ller b, Bjo¨rn Mauck a, Guido Dehnhardt a,*

a Department of General Zoology and Neurobiology, University of Bochum, ND 6/33, D-44780 Bochum, Germany b Department of Animal Physiology, University of Cologne, Weyertal 119, D-50931 Ko¨ln, Germany

Received 17 May 2005; received in revised form 15 August 2005

Abstract

The underwater visual acuity (the angle subtended by the minimal resolvable line width of high contrast square wave gratings at a viewing distance of 2 m) of two male harbor seals was determined at diﬀerent levels of water turbidity. Starting with visual acuity angles of 5.50 and 12.70 in clear water we found visual acuity to decrease rapidly with increasing turbidity at rates of 7.40 and 6.00 per formazin nephelometric unit (FNU). Besides the individual diﬀerences in visual performance of the harbor seals tested, our results reveal a dramatic loss of visual acuity even at moderate levels of turbidity. At sites in the German Wadden Sea, where harbor seals are known to roam and forage, we measured turbidity levels exceeding 40 FNU. These data suggest that turbidity has to be considered as an important factor in the sensory ecology of pinnipeds. 2005 Elsevier Ltd. All rights reserved.

Keywords: Turbidity; Visual acuity; Harbor seals; Phoca vitulina

1. Introduction picture of the object and degrades visual resolution (Duntley, 1963). Again, the magnitude of the contrast Underwater vision in marine organisms strongly de- reduction caused by scattering depends on the degree of pends on the optical properties of their respective aquatic turbidity of the respective water. For a horizontal path of habitat. According to empirically tested theory about the sight and light of a certain wavelength the apparent con- propagation of visual information underwater (see Dunt- trast of an object decreases with distance according to a*r ley, 1963; Jerlov, 1976; Lythgoe, 1979, 1988), the visibility Cr = C0*e , where C0 is the contrast between the object of an object underwater depends on its inherent contrast, and the background space light when viewed at close dis- i.e., the difference in brightness between the object and its tance, Cr is the apparent contrast at distance r, and a is background, and the attenuation that the reflected light the narrow beam attenuation coefficient that describes the undergoes on its way from the object to the observer. attenuation of light on its way through the water due to The attenuation is strongly dependent on the presence of scatter and absorption (Duntley, 1963; Lythgoe, 1988; Par- dissolved and particulate matter in the water, like phyto- tridge & Cummings, 1999). From this equation it derives plankton, ‘‘yellow substance’’, and suspended solids that the distance range within which the object can be seen (Jerlov, 1976; Mobley, 1994). In addition, light from the depends on the spectral contrast sensitivity of the observ- object that is forward scattered at low angles, and the scat- erÕs eyes and the components of the water, which determine tering of ambient light towards the observer over the entire the narrow beam attenuation coefficient. An aquatic organ- path of sight (veiling brightness) is superimposed on the ism relying primarily on vision is strongly dependent on the transparency of the water and turbidity is therefore recog- * Corresponding author. Tel.: +49 221 9772900; fax: +49 221 97750604. nized as a relevant ecological factor in aquatic systems, E-mail address: [email protected] (G. Dehnhardt). e.g., in models of aquatic visual feeding (Aksnes & Giske,

1993; Aksnes & Utne, 1997). In laboratory experiments is a significant and hitherto underestimated factor in the Miner and Stein (1996) could show that the detection dis- sensory ecology of pinnipeds. tance between bluegills (Lepomis macrochirus) and large- mouth bass (Micropterus salmoides) decreased as a 2. Method negative power function of turbidity from about 2 m in clear water to 0.23 m at 10 nephelometric turbidity units 2.1. Subjects (NTU). This is in accordance with the findings of Vogel and Beauchamp (1999) for the lake trout (Salvelinus namay- The study was conducted in our marine mammal re- cush). Miner and Stein (1996) could also show that turbidity search lab at Zoo Cologne, Germany, where we keep eight affected the habitat choice of the prey species (bluegills) that male harbor seals (Phoca vitulina). The experiments were spent more time in deep-water habitats when turbidity was carried out in accordance with the European Communities high. Such a turbidity-related change in habitat choice was Council Directive of 24 November 1986 (86/609/EEC). also reported by Abrahams and Kattenfeld (1997) for Two experimentally naive seals (Bill, three years of age; fathead minnows (Pimephales promelas). Sam, seven years of age) served as subjects. Their daily diet In foraging pinnipeds vision has been suggested to be consisted of 2–5 kg of herring supplemented with vitamin. the predominant source of sensory information (Hobson, The entire amount of fish was fed during the experiments. 1966; Lavigne, Bernholz, & Ronald, 1977; Levenson & Experiments were carried out in a freshwater pool with a Schusterman, 1999; Walls, 1942). During the last decades capacity of 300 m3 and a maximum depth of 1.7 m. The research on vision in pinnipeds has focused on their experimental pool was connected via a gate to a larger pool amphibious visual capacity with special emphasis on the of 600 m3. The water was not filtered, but water quality was adaptations to the optical properties of the aquatic envi- maintained by a continuous flow of freshwater and regular ronment. Studies on several phocid and otariid species re- water changes. To separate the subjects from the rest of the vealed that the pinniped eye, due to its morphology, animals the gate was closed during the experiments. Only retinal architecture and spectral sensitivity, is well adapted during test sessions subjects were allowed to enter the to functioning at low light levels and the narrow range of experimental pool. While one subject was working at the spectral energy that prevail underwater (for reviews see test apparatus the other seal was stationed in a hoop-sta- Dehnhardt, 2002; Supin, Popov, & Mass, 2001). Some tion several meters away from the apparatus. researchers even deny the need for non-visual explanations of pinniped underwater orientation (Levenson & Schus- 2.2. Management and measurement of water turbidity terman, 1999). However, the existence of blind but well- nourished seals in the wild (Newby, Hart, & Arnold, To achieve a continuous increase of water turbidity in 1970) and the obvious poor image transmission at high lev- the experimental pool the water was not changed during els of turbidity (e.g., at some sites in the German Wadden each series of experiments allowing algae and particulate Sea a Secchi-disk disappears from view at depths <1 m organic matter to accumulate in the water column. A total (Aarup, 2002)) are at least two points challenging this view. of ten test series was carried out between November 2001 One parameter that is used to characterize visual perfor- and December 2002. The duration of a test series ranged mance in animals is the visual acuity of its eyes. In sea from 3 to 41 days. After each experimental session the tur- lions, fur seals, and harbor seals underwater visual acuity bidity of the water in the experimental pool was measured was found to be in the range between 2.70 and 8.30, which using a laboratory turbidity photometer (NEPHLA, is the same order of magnitude as in terrestrial carnivores Dr. Bruno Lange GmbH Berlin, Germany) calibrated in like the cat (Mass, 2004; Mass & Supin, 1992, 2003; Schus- formazin nephelometric units (FNU) conforming to DIN terman & Balliet, 1970a, 1970b). However, the data pre- EN 27027 or ISO 7027 to DIN standard formazin. It sented in these studies represent the maximum visual should be noted here, that measurements of turbidity using acuity of the respective experimental animal. Although turbidity photometers just determine the attenuating or many pinniped species may have to cope with impaired scattering effect of a water sample to monochromatic light sight due to turbidity, little is known about the effect of of a certain wavelength (860 nm our case). The effect of the low or moderate levels of turbidity on the visual perfor- turbidity of the water sample on the sensory relevant spec- mance of seals. Recently, Strod, Arad, Izhaki, and Katzir tral bandwidth is not revealed by such measurements. (2004) have shown that in cormorants (Phalacrocorax After the entire water volume in the pool was changed carbo sinensis) underwater visual acuity declines linearly the turbidity of the water was 0.2 FNU increasing at a with increasing turbidity. They could also show that even mean rate of 0.2 FNU per day. The highest level of turbid- slight turbidity of less than 1 NTU has an effect on image ity measured was 7.8 FNU 35 days after the water was formation underwater. changed. The water appeared to be clear to the human To investigate the effect of turbidity on the underwater eye up to a turbidity of about 1 FNU. The increase of tur- visual acuity in pinnipeds we determined the visual acuity bidity was accompanied by a change of the color of the of two harbor seals at different levels of turbidity using psy- water, which became greenish-yellow at turbidity above chophysical techniques. Our findings indicate that turbidity 2 FNU. Due to the fact that there were only minute water M. Weiffen et al. / Vision Research 46 (2006) 1777–1783 1779 movements in the pool between test sessions, particles of higher density than water sedimented. Therefore, the experimental animals were sent to swim around in the pool at the beginning of each session, this way causing strong turbulence resulting in a more homogenous turbidity in the experimental pool. In oceanographic studies water turbidity is not routinely measured in FNU, but e.g., by means of a Secchi-disc (Aarup, 2002). To get an idea of the turbidity in the North Sea with regard to the standard formazin, water samples were collected from different depths at several sites in the German Bight where harbor seals are abundant. The turbidity of the water samples was measured immediately after they were taken using the same photometer described above. Exact positions of the sample sites were determined using a handheld GPS receiver. All samples were taken during high tide.

2.3. Stimuli and test apparatus

Visual acuity was determined as the angle subtended by the minimal resolvable line width of high contrast square wave gratings at a viewing distance of 2 m. Stimuli consist- Fig. 1. Experimental setup. The seals were trained to station 2 m in front ed of gratings of black and white lines of equal width. The of the test-board. The animalÕs task was to indicate the position of the animals task was to discriminate between horizontally and positive stimulus (horizontal grating) by pressing the response target at the vertically oriented gratings of the same size. This way the appropriate side with its muzzle. The experimenter crouched on a platform seals could not discriminate the stimuli on the basis of behind the opaque screen. The inset gives an overview of the test brightness cues. apparatus with seal Bill responding to the stimulus at the left side. For this photograph the apparatus was raised above water level. Gratings of 195 · 195 mm2 (subtending a visual angle of about 5.6·5.6 at a viewing distance of 2 m) were designed using computer graphic software (CorelDRAW) able to exchange the stimuli without being observed by the and printed at 1200 dpi with a laser printer using standard seals, the windows could be covered by an opaque PVC printer paper. Gratings were shrink-wrapped in transpar- shutter. The shutter could be raised and lowered manually ent foil of 125 lm. The gratings were fixed on PVC-boards by the experimenter using a pulley system fixed to the test- of 230 · 330 mm2. This way 25 pairs of stimuli were ob- board. tained with line width ranging from 0.5 mm to 24 mm. To keep the subjects at a fixed distance to the stimuli a Contrast of the gratings in air with respect to a human frame of aluminum tubes was installed 2 m in front of the LWLB observer was 0.85 calculated as C ¼ , where LB and vertical board (Fig. 1). A plastic ball was attached in the LWþLB LW are the luminance of the gratingÕs black and white lines. middle of the frame as a stationing target for the seals. Luminance measurements were performed using a Minolta The stationing ball was 650 mm below the water surface. LS-110 luminance meter (note that luminance is a photo- The runners on the back of the test-board were connect- metric measure that is based on the human spectral ed to aluminum tubes that were mounted at both sides of sensitivity). the stationing ball. A funnel-shaped response target was at- The test apparatus (Fig. 1) was designed for the presen- tached to the end of each tube. When the seal pressed the tation of stimuli in a two alternative forced choice proce- funnel with its muzzle the runner on the back of the test- dure. The apparatus consisted of a grey PVC test-board board was folded back this way signaling the experimenter (700 mm high, 900 mm wide) that was installed vertically the animalÕs choice. During a session the experimenter in the experimental pool so that the upper edge of the crouched on a platform behind the test-board separated board was 350 mm below the water surface. Two windows visually from the animal by an opaque screen. (200 · 200 mm2) were cut out beside each other in the test- board. The windows were separated from each other by 2.4. Procedure 200 mm and their upper edges were positioned 500 mm beneath the water surface. Stimulus plates were mounted be- The subjects were trained in a forced choice paradigm to hind the two windows in vertical runners that could be discriminate between horizontally and vertically oriented folded back from their vertical position to the rear side gratings of black and white lines. The respective horizontal of the apparatus. In order to hide the stimuli from the sub- gratings were arbitrarily defined as the positive stimulus. jects during the inter trial intervals (30 s) as well as to be At the beginning of each trial the animal swam to the frame 1780 M. Weiffen et al. / Vision Research 46 (2006) 1777–1783 in front of the test apparatus and placed its muzzle on the stationing ball, while the windows of the test-board were still covered by the shutter. Then the stimuli were mounted into the runners behind the windows always beginning with the left window (from the animalÕs position). During a session, stimuli were presented at both positions of the apparatus according to pseudorandom schedules (Gellermann, 1933). A trial started when the shutter covering the stimuli was lowered, so that the stimuli became visible to the seal. Lowering the shutter served as the signal for the seal that a trial was started. For a correct response the seal had to press the funnel on the side where the positive stimulus appeared. A correct response was rewarded by a piece of cut herring. An incorrect response was not rewarded. A session consisted of 30 trials. Two sessions (at 9 a.m. and at 5 p.m.) were conducted with each animal per day. During experimental sessions the site in the pool were the apparatus was installed received only indirect natural light. Underwa- ter flicker (caustics) could not be observed. Occasional measurements of the intensity of the downwelling light using a light meter with the sensor placed in a waterproof housing at the depth were the stimuli were presented, re- Fig. 2. Minimal resolvable line width and corresponding visual acuity vealed illumination levels between 575 and 2150 lux (again, angles of seal Bill (three years of age) and seal Sam (seven years of age) at a distance of 2 m as a function of turbidity. Each data point represents the note that lux is a photometric unit that is based on the hu- mean of at least 5 (max. 25) transition points collected in up to three man luminosity function). sessions. The average rate of loss of visual acuity (slope of the regression As the turbidity in the experimental pool increased rap- line, see text) is 7.40 per FNU for Bill and 6.00 per FNU for Sam. idly we decided for a method of data recording that allowed fast estimates of visual acuity. Therefore, we used visual acuity to be almost unaffected by turbidity in both the psychophysical staircase method by decreasing the line animals. Beyond 1 FNU, visual acuity decreased rapidly width of the stimuli after each correct choice and increasing with increasing turbidity. To determine the average loss the line width after each false response. In the initial trial of of visual acuity per FNU we calculated linear regressions each test session we used stimuli that we expected to be eas- for both data sets. The average rate of loss of visual acuity ily distinguishable for the animals. Thus, a session typically over the entire range of turbidity tested is 7.40 per FNU in started with an initial run of correct choices that brought Bill and 6.00 per FNU in Sam. the line width to a range where the stimuli became indistin- To determine the maximum visual acuity of our experi- guishable for the animals and incorrect responses happened mental animals at water turbidity levels 61 FNU we calcu- to occur. The upper border of this range of line width was lated the percentage of correct choices for each line width confined by the transitions between one or more consecu- in all test sessions that were conducted in this range of turbid- tive incorrect responses, resulting in an increase of the line ity. The resulting psychometric functions are shown in Fig. 3. width, and the next correct choice after which the line Visual angles at threshold (75% correct choices) calculated width was decreased. We took the average line width of by linear interpolation are 5.50 for Bill and 12.70 for Sam. these transition points as a behavioral indicator for the The results of our turbidity measurements in the North visual acuity of the subject at the respective level of turbid- Sea are shown in Table 1. Turbidity at the different sample ity. In addition, to obtain a representative 75% threshold sites ranged between 7 FNU and 40.7 FNU at a depth of estimate for visual acuity under clear water conditions we 2 m. At depths from 6 to 7 m turbidity ranged between calculated the percentage of correct choices for each line 22 FNU and 39.8 FNU. width from all test sessions that were conducted in the range of turbidity 61FNU. 4. Discussion

3. Results Our results indicate a dramatic loss of visual acuity in harbor seals even at the moderate levels of turbidity used Fig. 2 shows the minimal resolvable line width and the in this study. The estimates for the visual acuity of our seals corresponding visual acuity angles of both experimental under clear water conditions compare well with data animals plotted as a function of turbidity. Each data point reported by Schusterman and Balliet (1970b), who deter- represents the mean of at least 5 (max. 25) upper transition mined an underwater visual acuity angle of a single harbor points determined in up to three sessions. Under apparent- seal of 8.30. In their study the animal was required to dis- ly clear water conditions (turbidity 61 FNU) we found criminate between gratings of black and white lines of M. Weiﬀen et al. / Vision Research 46 (2006) 1777–1783 1781

(1970b) provides the only reliable data about the visual acuity of harbor seals so far. Together with these data our results reveal that individual visual acuities of harbor seals may differ by a factor 2. The visual acuity of our seal Sam (12.70) is the poorest ever determined in a pinniped species underwater at sufficient light conditions. Assuming that the visual acuity values of our seal Bill (5.50) and the animal tested by Schusterman and Balliet represent emme- tropic underwater vision it seems likely that SamÕs poor performance is caused by ametropia. However, the number of animals tested so far is too small to make a final judg- ment about the variability of visual acuity in harbor seals. In humans visual acuity among a young ocularly normal population ranges from 0.30 to 10 (Elliott, Yang, & Whi- taker, 1995) and thus also shows considerable variation (Coppens & Berg, 2004). While marked differences in visual acuity were found in our seals at clear water conditions (61 FNU) the mean rate of loss of visual acuity over the entire range of turbidity that we tested is similar in both animals (7.40 per FNU in Bill and 6.00 per FNU in Sam). The magnitude of this loss of visual acuity reveals that the sealÕs good visual performance in clear water may be degraded even by a slight increase in turbidity. This is in accordance with the results of Strod et al. (2004) who found the visual acuity of cormorants to be affected by low levels of turbidity. The range of turbidity that we observed in our experimental pool must be considered to be moderate in absolute terms. Fig. 3. Psychometric functions of visual acuity of both experimental The turbidity range covered by standard formazin includes 6 animals under clear water conditions ( 1 FNU). Dashed lines indicate the clear tap water of 0.02–0.5 turbidity units to raw sewage of threshold level of 75% correct choices. Line width at threshold (viewing distance: 2 m) is 7.4 mm (12.70 of visual angle) for Sam and 3.2 mm (5.50 of more than thousand turbidity units. However, reports visual angle) for Bill. about turbidity in the ocean that make use of photometers calibrated in standard formazin are rare. Our measurements in the North Sea reveal a turbidity of 7 to more than Table 1 Water turbidity at sites in the German Wadden Sea, where harbor seals are 40 turbidity units. As the sample sites were located in one abundant of the most turbid areas of the Southern North Sea (Aarup, Sample site 1 2 3 4 2002) they might represent an extreme ecological situation specific for the harbor seals of the German Wadden Sea. Latitude (North) 5404.650 5404.260 5402.710 5402.440 Longitude (East) 836.410 836.080 835.940 841.470 However, Abookire, Piatt, and Speckman (2002) measured Depth [m] 2 2/7 2/7 2/6 turbidity in a fjord in Southeast Alaska where they found Turbidity [FNU] 40.7 23.7/30.2 20.6/39.8 7/22 turbidity values ranging between 7.8 and >15 FTU at Samples were taken between 12:00 and 3:00 p.m. on the first of October depths from 3 m to 16 m. Turbidity was >10 FTU to 2002 during high tide. Turbidity was measured immediately after a sample depths of 41 m. (The units FNU, NTU and FTU (formazin was taken. turbidity unit) have the same value for a certain turbidim- eter and water sample. However, it is important to note equal width versus a gray field. In another study Jamieson that different turbidimeters employing different measuring and Fisher (1970) measured visual acuity in a male and a procedures may produce different output in the same unit female harbor seal both in air and in water. Here, the ani- even for the same sample. In addition, measurements of mals were trained to distinguish between one solid black turbidity may vary with differences in the optical properties line and two black lines separated by a gap. The minimum of the particles, which make up turbidity in the respective size of the gap that could be detected served as an indicator water sample. Therefore, the data cited here are not direct- for visual acuity. Visual acuity angles around 2.00 were ly comparable to our measurements but may represent the found in both animals in air and in water. However, as magnitude of the range of turbidity seals might find in the pointed out by the authors the animals may have detected ocean, even in the same area.) Thus, the turbidity that differences in the luminance of the stimuli rather than mak- pinnipeds meet in the ocean might be considerable higher ing decisions by using the resolving power of their visual than in our study. Many pinniped species are known to for- system. Therefore, the work of Schusterman and Balliet age in turbid inshore waters and even in estuaries, because 1782 M. Weiffen et al. / Vision Research 46 (2006) 1777–1783 of the abundance of plankton and fish in such waters. One Abrahams, M., & Kattenfeld, M. (1997). The role of turbidity as a should keep in mind that the term turbidity refers to a vari- constraint on predator–prey interactions in aquatic environments. ety of components and properties of natural water affecting Behavioral Ecology and Sociobiology, 40, 169–174. Aksnes, D. L., & Giske, J. (1993). A theoretical model of aquatic visual underwater vision. These not only include the scattering of feeding. Ecological Modelling, 67, 233–250. light that degrades image contrast and limits the visual Aksnes, D. L., & Utne, C. W. (1997). A revised model of visual range in range but also the spectral filtering, which determines the fish. Sarsia, 82, 137–147. spectral bandwidth and intensity of light that is available Coppens, J. E., & van den Berg, T. J. T. P. (2004). A new source of for vision at a certain depth. Underwater vision is therefore variance in visual acuity. Vision Research, 44, 951–958. Davis, R. W., Fuiman, L. A., Williams, T. M., Collier, S. O., Hagey, W. affected by turbidity in more than one way and the role P., Kanatous, S. B., Kohin, S., & Horning, M. (1999). Hunting these constraints played in the evolution of the pinniped behavior of a marine mammal beneath the Antarctic fast ice. Science, visual systems, e.g., with regard to the spectral sensitivity 283, 993–996. of their visual pigments or color vision (compare Peichl, Dehnhardt, G. (2002). Sensory systems. In A. R. Hoelzel (Ed.), Marine Behrmann, & Kro¨ger, 2001), remains to be clarified. How- mammal biology (pp. 116–141). Oxford: Blackwell Publishing. Dehnhardt, G., Mauck, B., & Bleckmann, H. (1998). Seal whiskers detect 0 ever, given a loss of visual acuity of more than 5 per FNU water movements. Nature, 394, 235–236. as indicated by our measurements it seems unlikely that vi- Dehnhardt, G., Mauck, B., Hanke, W., & Bleckmann, H. (2001). sion could fulfill the requirements of a long range sensory Hydrodynamic trail-following in harbor seals (Phoca vitulina). Science, system for object recognition even at moderate levels of 293, 102–104. turbidity. Duntley, S. Q. (1963). Light in the sea. Journal of the Optical Society of America, 53, 214–233. This raises the question by which sensory channels and Elliott, D. B., Yang, K. C. H., & Whitaker, D. (1995). Visual acuity neural mechanisms seals can compensate for the loss of visu- changes throughout adulthood in normal, healthy eyes: Seeing beyond al information under turbid conditions. Besides the percep- 6/6. Optometry and Vision Science, 72(3), 186–191. tion of acoustic stimuli Dehnhardt, Mauck, and Bleckmann Gellermann, L. W. (1933). Chance orders of alternating stimuli in visual (1998) suggested the vibrissal system sensitive to hydrody- discrimination experiments. Journal of Genetic Psychology, 42, 206–208. namic stimuli to fit this requirements. The persistence of Hobson, E. S. (1966). Visual orientation and feeding in seals and sea lions. hydrodynamic trails generated by fish is certainly long en- Nature, 210, 326–327. ough to be used by a seal for long range sensory scanning Jamieson, G. S., & Fisher, H. D. (1970). Visual discriminations in the (compare Dehnhardt, Mauck, Hanke, & Bleckmann, harbour seal Phoca vitulina, above and below water. Vision Research, 2001). According to Davis et al. (1999), seals detect their 10, 1175–1180. Jerlov, N. G. (1976). Marine optics. Amsterdam: Elsevier Scientific prey mainly visually and may use hydrodynamic informa- Publishing Company. tion during the last stage of prey pursuit. Our data reveal Lavigne, D. M., Bernholz, C. D., & Ronald, K. (1977). Functional aspects that in turbid waters the opposite mechanism seems more of pinniped vision. In R. J. Harrison (Ed.). Functional anatomy of probable. However, to our understanding most complex marine mammals (Vol. 3, pp. 135–173). London, New York: Academic behaviors rely on the integration of information from any Press. Levenson, D. H., & Schusterman, R. J. (1999). Dark adaptation and sensory channel that provides relevant input. The resulting visual sensitivity in shallow and deep-diving pinnipeds. Marine multimodal representation of the environment is obviously Mammal Science, 15(4), 1303–1313. so robust that it even compensates for the loss of one modal- Lythgoe, J. N. (1979). The ecology of vision. Oxford: Clarendon Press. ity due to disease or extreme ecological constraints. There- Lythgoe, J. N. (1988). Light and vision in the aquatic environment. In J. fore, to understand pinniped orientation we need to know Atema, R. R. Fay, A. N. Popper, & W. N. Tavolga (Eds.), Sensory biology of aquatic animals (pp. 57–82). Berlin, Heidelberg, New York: all sensory modalities that provide relevant information in Springer. a certain behavioral context and the mechanisms that oper- Mass, A. M. (2004). A high-resolution area in the retinal ganglion cell ate on the processing of this information. layer of the StellerÕs sea lion (Eumetopias jubatus): A topographic study. Doklady Biological Sciences, 396, 187–190. Mass, A. M., & Supin, A. Y. (1992). Peak density, size and regional Acknowledgment distribution of ganglion cells in the retina of the fur seal Callorhinus ursinus. Brain Behavior and Evolution, 39, 69–76. We are grateful to M. Dambach (University of Cologne) Mass, A. M., & Supin, A. Y. (2003). Retinal topography of the harp for his support throughout this study. This work has been seal Pagophilus groenlandicus. Brain Behavior and Evolution, 62, 212–222. funded by grants of the VolkswagenStiftung and Deutsche Miner, J. G., & Stein, R. A. (1996). Detection of predators and Forschungsgemeinschaft (SFB 509, B11) to G.D. habitat choice by small bluegills: Effects of turbidity and alternative prey. Transactions of the American Fisheries Society, 125, References 97–103. Mobley, C. D. (1994). Light and water. San Diego: Academic Press. Newby, T. C., Hart, F. M., & Arnold, R. A. (1970). Weight and blindness Aarup, T. (2002). Transparency of the North Sea and Baltic Sea—A of harbor seals. Journal of Mammalogy, 51(1), 152. Secchi depth data mining study. Oceanologia, 44(3), 323–337. Partridge, J. C., & Cummings, M. E. (1999). Adaptation of visual Abookire, A. A., Piatt, J. F., & Speckman, S. G. (2002). A nearsurface, pigments to the aquatic environment. In S. N. Archer, M. B. A. daytime occurrence of two mesopelagic fish species (Stenobrachius Djamgoz, E. R. Loew, J. C. Partridge, & S. Vallerga (Eds.), Adaptive leucopsarus and Leuroglossus schmidti) in a glacial fjord. Fishery mechanisms in the ecology of vision (pp. 251–283). Dordrecht, Boston, Bulletin, 100, 376–380. London: Kluwer Academic Publishers. M. Weiffen et al. / Vision Research 46 (2006) 1777–1783 1783

Peichl, L., Behrmann, G., & Kro¨ger, R. H. H. (2001). For whales and seals amphibious and turbid conditions. Current Biology, 14, the ocean is not blue: a visual pigment loss in marine mammals. R376–R377. European Journal of Neuroscience, 13, 1520–1528. Supin, A. Y., Popov, V. V., & Mass, A. M. (2001). The sensory physiology Schusterman, R. J., & Balliet, R. F. (1970a). Conditioned vocalizations as of aquatic mammals. Boston Dordrecht London: Kluwer Academic a technique for determining visual acuity thresholds in sea lions. Publishers. Science, 169, 498–501. Vogel, J. L., & Beauchamp, D. A. (1999). Eﬀects of light, prey size, and Schusterman, R. J., & Balliet, R. F. (1970b). Visual acuity of the turbidity on reaction distances of lake trout (Salvelinus namaycush)to harbour seal and the Steller sea lion under water. Nature, 226, salmonid prey. Canadian Journal of Fisheries and Aquatic Sciences, 56, 563–564. 1293–1297. Strod, T., Arad, Z., Izhaki, I., & Katzir, G. (2004). Cormorants keep Walls, G. L. (1942). The vertebrate eye and its adaptive radiation. New their power: visual resolution in a pursuit-diving bird under York: Hafner Publishing.