ABSTRACT Pacific Sand dabs exibit postural change when introduced to flow, The rheotaxis and the underlying hydrodynamics are not well studied for flatfish. No known values for lift or drag exist. In this study force transducers were used to measure drag and lift on the sand dab, and corresponding drag and lift coefficients were calculated. The lift is remarkably variant with altered posture anddrag is virtually unchanged. It is this four fold lift reduction for head to tail flow that enables the sand dabs to stay in place in currents up to 402 to 50% faster, The sand dab is equally efficient with flows in the opposite direction; this is essential to accommodate the bidirectional surge. A mechanism for lift reduction, or more importantly for increased slipping velocity, is proposed. INTRODUCTION The Pacific Sand dab, Citharichthys sordidus is a bottom dweller in ocean depths from 30 feet to 1600 feet. The flatfish is well designed for sea floor inhabitation and would prefer to stay camouflaged on the bottom. At depths of 30 feet with wave height of 10 feet and a period of 10 seconds, wave surge can reach surprising velocities of V = 150 cm s-l. Surge of this magnitude means sand dabs are exposed to significant flows with dislodging potential. Sand dabs exibit a rheotaxis or reaction to water flows in an attempt to withstand larger flow speeds. Arnold, C 1969 observed similar behavior in plaice and Arnold and Weihs, 1978 studied the hydrodynamic factors underlying the rheotaxis in plaice. Empirically measured values of lift and drag force, and in particular the lift and drag coefficients, remain unknown for flatfish. The combination of streamline and aerfoil shapes make sand dabs a unique hydrodynamic study, and proves to help explain the function of the rheotaxis. MATERIALS AND METHODS Flow Tank Experimentation was done in a flow tank similar to that described by Vogel and LaBarbera, 1978 (figure 1). A constant flux of sea water kept the temperature in a range of 13 to 15 C. Tlow speeds reached 106 cms. These mainstream speeds were measured by timing fluorescein dye over a one meter distance. Live Fish Sand dabs were caught with hook and line and kept in a sandy bottom holding tank, receiving a constant supply of sea water. The temperature was 10 to 12°C. The sand dabs were fed freshly killed fish three times weekly. Fish length ranged from 16.5 to 21.0 cm. The fish densities were calculated from the equation (where mis the mass in air, mthe submerged mass and o is the density of sea water. The average fish density was 9-1.055 x 10 kg m-3. Model Fish Models were made of silicone rubber imbedded with wire for rigidity, A sand dab was killed in a water bath containing a high concentration of MS 222, resulting in minimal physical alterations of the fish. The fins were carefully spread and pinned down, and the fish was then frozen. Parafin wax was poured over the prepared fish forming the mould. A layer of silicone rubber was applied, followed by five wires and the final layer of rubber. Aluminum foil fins backed by fine wire were used to replicate dorsal and anal fins. The model was 17.1 cm. in length and slightly more dense than the live fish (P = 1.062 10° kg m2). The blotted mass was m = 48.7g and the submerged mass m,- 1.8 g. Maximum width measured from anal to dorsal fin was w - 8.3 cm. Transducers Twotransducers were used to measure forces on the model sand dab, Both the drag and lift transducers were made of acrylic plastic and designed to rest in the top portion of the flow tank (figures 2 and 3). The model was attached to a double-cantilever beam. Any force applied to the model resulted in a proportional displacement of the beam. These displacements were sensed by a linerly variable differential transformer and recorded on a voltmeter. After weight calibration the drag trans¬ ducer had a linear regression line with the equation F, = 1.180 V - .002, r 7 .999 where Fy is the drag force and V is the voltage reading. The lift transducer yielded the equation Fj = 2,25 V + .002, also with r».999; Fj is the lift force. Since measured lift forces approached zero the adjusted equation Fj = 2.32 V was used for voltage readings less than or equal to.008 V. The model was mounted to the moveable beam with a screw to allow easy removal and posture adjustments. When attached to the transducers and lowered into the tank, the model was inverted with its ocular side down. Use of the transducers required velocity adjustments due to the constriction of flow by the transducer. The speed adjustment for the drag transducer was U = 1.27 U, and U = 1.25 U, for the lift transducer. Uo is the original mainstream velocity and U is the adjusted velocity. The estimated error for voltage readings is se= .002 for both trans¬ ducers. Manometer Manometry was used to estimate the lift force on the model. An inverted, sloped manometer (Vogel, 1981) was filled with corn oil with a measured density of p =.912 10° kg m23. One pressure probe was stationary beneath the fish on the lateral line. The second probe measured the constant mainstream water pressure. Values were recorded for flat and fins up postures. For more information on manometry, see Vogel, Life in Moving Fluids. Inclined Plane Static friction coefficients, M., were measured with an inclined plane apparatus (figure 4). Values were calculated from an average angle of slippage, 6, using the equation - tan 0. (This is g. - mg sin e and Fy = mg cos 0. Sub¬ derived from Fg - Fy. E stitution results in P= tan 0.) The plane had an acrylic plastic surface to mimic the flow tank floor and all measurements were made while the apparatus was submerged in sea water. Experimental Procedure The rheotactic behavior of live sand dabs was observed in the flow tank, with careful note of current velocities. Fish were centered in the tank and oriented in the desired direction. Speeds were increased in 5 cm s4 increments with a minimum of 5 minutes between changes. Similar observations were recorded for the model sand dab. Transducer measurements varied with model shape and distance from transducer. Precautions were taken to consistently mimic the rheotactic response with the model. It was found necessary to remove all air bubbles trapped between the model and the transducer apparatus. The drag transducer was tested with a 1/2 inch in diameter cylinder. The drag coefficient was C = 0.58. Recorded values for similar shaped cylinders are C = 0.48 to Cq = 0.75 (Hoerner, 1965). BEHAVIOR Sand dab Rheotaxis Sand dabs undergo a remarkable series of behaviors when introduced to flow (figure 5). When no flow exists sand dabs typically settle on the tank bottom, head raised up to 2 cm off the substrate. As head to tail flow begins the head is lowered. With increasing flow most fish commence burying movements. This behavior consists of rapid body un¬ dulations coordinated with anal and dorsal fin beating. In addition to being used as a settling device while on sand, it appears to help extrude water from beneath the fish. This results in increased surface area in contact with the substrate. The burying movements seem to ac¬ complish results similar to the "posterior fin-beating response" described by Arnold, 1969. As flow increases, dorsal and anal fins are raised into a fins up posture (figure 6). This entails a characteristic raising of the poster¬ ior portion of the anal and dorsal fins. This is coupled with a notice¬ able body arch directly above the posterior region and a slight upward cambering along the length of the body. Also the fins up posture creates a water channel along the length of the dorsal side. The flow originates from a small adjustment of an anterior portion of the dorsal fin. A natural channel exists on the ventral side. This channel results from the inability of the anal fins to contact the substrate. This posture is the most intruiging behavior as these changes cause significant flow redirection enabling the fish to stay stationary longer. Flows moving from tail to head elicit a different series of reactions from the sand dabs. No burying movements are observed and the fins up posture is never noted. Some fish also slip slightly sooner when flow moves from tail to head. Neither arching nor cambering is observed as the flat posture appears most beneficial for remaining in place. THEORY Hydrodynamics Rheotaxis suggests that as flows increase fish alter their posture to postpone slippage. Sand dabs orient randomly when placed in the flow tank and with the initiation of flow, do not appear to prefer one flow direction to the other. Arnold and Weihs, 1978 theorize that the head to tail flow is hydrodynamically advantageous based on the decreased drag force on the streamlined body. But lift force is probably greatest in this orientation, and the resultant slipping speed is an intricate balance between these two forces. The current initiates a drag force on the fish in the direction of flow. This drag is the sum of pressure drag and skin friction drag. The characteristicly gradually tapering tail of the streamlined sand dab delays the separation point and results in relatively low pressure drag (Blake, 1983). Blake, 1983 suggests that for a streamlined object most of the drag is a result of skin friction drag. Total drag force can be calculated from the equation Fa - 1/2 Ca.Ap U2 where F, is the total drag froce, C the drag coefficient,? the density of sea water, Ag the plan area and U is the flow speed. When the total drag force exceeds the frictional force the fish can no longer remain in place and slips downstream. The slippiay threshold can be calculated from the equation Fa -MsW, (2) where Lis the static friction coefficient and W is the effective weight of the fish. Since the streamline shape of the sand dab is also an aerfoil, a lift force is created. This force directly opposes the submerged weight (figure 7) such that W - Wo - Fy. (3) where W, is the submerged weight and Fj is the lift force. The lift force is defined as Fi - 1/2 C(.Ap U2, (4) where Cj is the lift coefficient. Combining equations 1-4 one can solve for the slipping velocities, 2 Wo 1/2 Ug - (5) A (C1 19 1. Cg and uare all unknown for live sand dabs. It is possible to calculate C, and C from forces measured on models. Mg values are also obtainable. U, can then be calculated and compared to the observed values. This serves as a check for the C and C, calculations. Since lift and drag forces increase by the square of the flow velocity, a small flow increase results in a large combination of forces acting on the fish. If the sand dab wishes to remain stationary, any posture change that can reduce either lift or drag force is quite ad¬ vantageous. RESULTS Lift and Drag The fins up posture results in a significant reduction in lift force (figure 8). The decline in lift is roughly 4 fold at all velocities. The drag force is not significantly changed by altered posture. The lift force is on the average, two times that of the drag force in the flat posture at all flow speeds. However with the fins up, the calculated lift is one half that of the drag force. Lift and Drag Coefficients The calculated drag and lift coefficients both vary over the Reynolds number range of 10’ to 10°. No substantial change in C resulted from a change in posture (figure 9). C on the other hand, dropped drammatically during the fins up posture to nearly 25% of the flat posture value (figure 10). Flow direction had no effect of C at a speed of U - 30 cm s i (approximate slipping speed) (figure 11). C however, shows substantial change with flow direction. Manometry The pressure difference between the flat posture and the fins up posture was measured to be AP = 2.59 N m2. For a plan area A = 7.5 10 ° m2 the lift reduction from the fins up posture was thus predicted to be F = .019 N. Static Friction Coefficient Sand dabs freshly anesthitized in MS 222 had a calculated static friction coefficient of M-.30. This value is similar to that of freshly killed plaice where M, -.2 to .4 (Arnold and Weihs, 1978). A, for the model was much higher due to the nature of the silicone rubber and perhaps, the absence of mucus. Values averaged M,=.84 for the flat posture and M,- .75 for the fins up posture. DISCUSSION Lift and Drag Simple observation suggests that sand dab rheotaxis alters the interaction of drag and lift: A model in the fins up posture slips at 35 cm s+ while the same model in the flat posture slips at 30 cm s 1. An anesthetized fish slips around 20 cm s1, Hydrodynamic analysis supports this conjecture. The lift reduction produced by the fins up posture increases the effective weight, W, and thereby frictional force. Since the drag force is unchanged the fish effectively postpones slippage (figure 12). From equation (5), the model in the flat posture with head to tail flow, U. - 20.7 cm s'i (Cg and C, values at Reynolds number Re = 6.5 «10“ , the approximate slipping velocity of U = 30 cm s1). For the same flow direction in the fins up posture Ug = 26.8 cm s1. Is the model a good simulation of live sand dabs? Plotting frictional force and drag force as a function of flow speed (from equation (5) with Mg-.30 for the live sand dab) U - 16.6 cm si (figure 13). This value is one-half the value observed. The low prediction is explained in three ways. (1) g was calculated for anesthitized fish and found to be Us -.30. It is likely that conscious fish can physically increase through fin use and general body conformation. As already mentioned, burying movements increase surface area in contact with the substrate, and perhaps increases the value of A.. This directly increases frictional force and delays movement. (2) Although the model is a good replica of sand dab shape it does lack mucus. Daniel, 1981 reported that skin friction can be reduced up to 502 in turbulent or pulsed laminar flows. Since skin friction is an important component of total drag in streamline objects, mucus is likely to decrease total drag. A decreased drag means an increased U.. Together, these two changes could shift the slipping speed into the observed range. (3) The observed slipping velocity may also be at fault. The flow tank creates a substantial boundary layer. The measured mainstream flow was the value observed at the slipping speed. The actual flow speed acting on the fish is less than mainstream flow due to this velocity gradient. Hence, the observed U. is greater than the calculated value from equation (5). C and Since C, and C, have never been quantitatively determined for sand dabs, only rough comparisons can be made to shapes of known coefficients (figure 14). Drag of an asymmetric semi-ellipsoid body with its blunt end facing the flow depends on its fineness ratio or ratio of length to thickness. For minimum drag the optimum value is between 10 to 15 (Blake, 1983). For sand dabs the ratio is between 12 and 13. The corresponding minimum drag coefficient is C, -.03 for the semi-ellipsoid body. Drag coefficients for the model sand dab range from C = ,O1 to C -.045. Arnold and Weihs, 1978 estimated Cj and C, for plaice (Pleuronectes platessa L. ) and concluded that C was 10 to 20 times Cq. But in doing so it is necessary to assume that is constant for Re » 10', This assumption is invalid for sand dabs as both Cq and C vary over a small Reynolds number range of Re = 104 to Re = 10°. The sum is also not constant. The actual ratio of lift to drag in sand dabs is 1/10 th that theorized for plaice. A partial explanation for the difference stems from basic geometry. The plaice has a larger aspect ratio, ratio of width to length, than does the sand dab. However the former results in increased lift (Blake, 1983). This accounts for a fraction of the ten fold difference. The remaining difference is most probably due to the assumption of a constant - value. Empirical evidence (figures 9 and 10) shows this assump¬ tion to be invalid in sand dabs and perhaps flat fish in general. Lift Reduction Mechanism It is well substantiated that rheotaxis reduces lift force, and therefore increases the slipping velocity. The mechanism behind lift reduction is hypothesized to be increased flow speed beneath the sand dab. The flow originates primarily from water channels along the dorsal and anal fins. The raised fins serve as the outlet channel where water exits from beneath the body. Fluorescein dyes used to study currents support this theory. A manometer confirms the increased flow during the fins up posture The measured drop in force of F =.019 N during the fins up posture is evidence of a decrease in pressure beneath the fish. From Bernoulli's principle it follows that flow velocity must increase. So the rheotaxis is merely a hehavioral reaction to reduce lift by increasing flow beneath the body. The manometer result of F =.019 N is less than the transducer measurement of F - .025 N for lift reduction from the fins up posture. The difference is explained by rough manometer procedure. Only one point beneath the fish was measured, that being under the arched section on the lateral line. This point is posterior to the center of lift and should theoretically yield a smaller value. The sand dab has developed a clever hehavior to increase slip speed for head to tail flows. But wave surge is bidirectional. To remain in place on the bottom with flow from tail to head, the sand dab doesn't raise his fins. C =.014 for the flat position. This value is not significantly different from the Cq -.016 calculated for the fins up posture in head to tail flow.. Sand dabs theoretically could combat the bidirectionality of wave surge then, by subtle changes in posture. For head to tail flow the fins up posture is best and for the reverse flow the flat posture is most advantageous. Since no single flow direction is hydrodynamically preferred, the sand dabs can avoid constant reorientation to changing flow directions. SUMMARY (1) Sand dabs exibit a behavioral response to flow. (2) The rheotactic response is characterized by the fins up posture where posterior anal and dorsal fins are raised, the body arches in this same region, a water channel is created on the dorsal side, and there is a slight upward cambering the length of the body. (3) This posture change reduces lift by a factor of four, while drag forces are unchanged. The result is an increased slipping velocity. (4) The slipping velocity prediction for live sand dabs suggests that either a) M.is larger or highly variable for the live fish. b) the drag force and coefficients are slightly high. c) the observed slip¬ ping velocities are too high, a result of the velocity gradient of the flow tank. (5) Drag and lift coefficients are not constant over a Reynolds number range of Re = 104 to Re = 105. (6) For head to tail flow the fins up posture is hydrodynamically favor¬ able. For opposite flow the flat posture is the most effective. The sand dab feels essentially equal forces in either flow direction. (7) The proposed mechanism for lift reduction is an increase in flow beneath the body thereby decreasing the pressure beneath the fish. ACKNOWLEDGEMENT I feel fortunate to have studied under a perfect blend of Man faculty. ks to Bill Gilly, Chuck Baxter, Mark Denny, and Freya Sommer for endless kno adge and the time to share it. I am deeply indebted to Mark Denny whose ingenuity and tolerance reined me through my entire project. And lastly to the class... what a sense of humor! REFERENCES Alexander, M.R. (1983). Animal Mechanics. Blackwell Scientific Publications, Oxford. PP 183-233. Arnold, G.P. (1969). The Reactions of the Plaice (Pleuronectes platessa L.) to Water Currents. Journal of Experimental Biology. 51, 686-697. Arnold, G.P. (1978). The Hydrodynamics of Rheotaxis in the Plaice (Pleuronectes platessa L.). Journal of Experimental Biology. 75, 147-169. Blake, R.W. (1983). Fish Locomotion. Cambridge University Press, Cambridge. Daniel, T.L. (1981). Fish Mucus: In Situ Measurements of Polymer Drag Reduction. Biological Bulletin. 160, 376-382. Hoerner. (1965). Fluid- Dynamic Drag. Self-published. Vogel, S. (1981). Life In Moving Fluids. The Physical Biology of Flow. Willard Grant Press, Boston. Vogel, S. and LaBarbera, M. (1978) Simple Flow Tanks for Research and Teaching. Bioscience. 28,638-643. FIGURE LEGENDS Figure 1. Flow tank made of PVC pipe and acrylic plastic. The filters are used to facilitate laminar flow. See Vogel and LaBarbera, 1978. Figure 2. Drag transducer. The model is attached to a double¬ cantilever beam. Any force acting on the model results in displace¬ ment of the beam, and is recorded by the linearly variable differential transformer. Figure 3. Lift transducer. Figure 4. Inclined plane apparatus used to measure the frictional coefficient of live and model fish. Figure 5. Sand dab reactions to flow. Figure 6. Fins up posture. Characterized by raised fins, arched body, dorsal side water channels, and an upward cambering of the body. Figure 7. Forces acting on the sand dab experiencing flow, Figure 8. Lift reduction through postural change. flat posture ———- fins up posture The drag coefficient is not constant. Postural change Figure 9. results in no alteration of the drag coefficient flat posture ———- fins up posture Figure 10. The lift coefficient is not constant. Postural change results in a drammatic change in the lift coefficient. flat posture — fins up posture Figure 11. Drag is unchanged with flow direction and posture. Lift shows signigicant change with both flow and posture. Figure 12. Slip speeds (Ug) and lift speeds (U,) for the model sand dab. fins up posture ——- flat posture Figure 13. Predicted slip speeds (Ug) and lift speeds (Uj) for a live sand dab with effective weight of 1.8 g and u. - ,30. Figure 14. A comparison of drag coefficients at Re = 102 for plan area. C —1 20 cn FIGURE 1. 6ig 11 . — I — O 1 O O OOO 5 G D 7 2 5 LIFT FORCE ,N — O --1 )- — 8 * 6 0 — 3 0 9 G) 8 O S 8L---- -c 6 8 FORCE , N 8 FORCE „N 17 N 5 8 G) 0 *