Fig 1.
Hypothesized filtration mechanisms in suspension-feeding fishes.
(A) Dead-end sieving with flow perpendicular to the filter. Small particles exit while large particles clog pores [6]. (B) Hydrosol filtration with flow perpendicular to the filter. Small particles that are retained on the sticky filter may clog pores [6]. (C) Crossflow filtration with flow parallel to the filter. Inertial lift and shear-induced diffusion cause particle migration away from the filter [7]. (D) Vortical cross-step filtration with flow approximately parallel (tangential) to the filter, which consists of backward-facing steps forming d-type ribs [8]. Inertial lift and shear-induced diffusion cause particle migration away from the filter. Smallest particles are concentrated and transported by vortices. Enlargement of cross-step filter (box on right), showing groove aspect ratio (wh-1, slot width divided by rib height) and zones (1, 2, 3) inside slot. Circles represent particles of different diameters, with smallest circles representing the smallest particles. Illustration by M. Carly Lin.
Fig 2.
Models of fluid interaction with oral structures in suspension-feeding fishes.
Coronal (frontal) sections illustrating flow patterns during (A) crossflow filtration in a species such as the blue tilapia (Oreochromis aureus, Cichlidae) [16], and (B) vortical cross-step filtration in the paddlefish (Polyodon spathula, Polyodontidae) [8]. In both (A) and (B), mainstream flow passes parallel or tangential to the filter formed by the branchial arches (ba) and the gill rakers (gr), while filtrate passes between the gill rakers and across the gill filaments (gf) for gas exchange before exiting via the opercular cavities beneath the operculum (op). In (B), the interaction of crossflow with the backward-facing steps formed by the branchial arches generates sustained vortices in the slots (sl) between the branchial arches. The first gill slot extends from the suspensorium in the oral roof and the dentary to the first branchial arch. Illustration by Virginia Greene/virginiagreeneillustration.com.
Fig 3.
Methods to quantify gape area and gill slot area in preserved paddlefish.
Paddlefish that had been preserved in suspension-feeding position were held as if mounted in the flow tank (A) and ImageJ 1.49 was used to quantify the area inside the perimeter of the oral gape (outlined in red). The medial edges of structures along the margins of the gill slots (schematic shown in B) were then traced by hand onto a clear vinyl sheet that had been inserted inside the oral cavity to be flush with these structures. This method reproduced the outlines of the slots as a two-dimensional image (C), and ImageJ was then used to quantify the area of the gill slots. Illustration by M. Carly Lin.
Fig 4.
Sustained vortices were generated in 3D models.
Dorsal views of the computer-aided design (CAD) images (top) and 3D-printed models (bottom) in the recirculating flow tank, with mainstream flow (MF) entering the open gape at the right of each image. Red rectangles on CAD images denote locations on 3D models where vortices were recorded for analysis. Zones (1, 2, 3) are labeled inside the slot as in Fig 1D. The sustained vortex that formed in zone 2 was visualized by introducing dye into the slot between the second and third ribs via a cannula inserted through the dorsal midline of each model. The mesh that covered the lateral margins of the slots in all models is most visible in (A). (A) α = 55°, (B) α = 90°, (C) α = 110°. Scale in mm.
Fig 5.
Vortices caused particle concentration along slot margins in 3D models.
Lateral views of 3D models in the recirculating flow tank, with mainstream flow (MF) entering the open gape at the right of each image. Particles (Artemia cysts, 210–300 μm diameter) were concentrated in zones 1 and 3 of all models, while vortices scoured the particles from the mesh in zone 2. (A) 55° model, (B) 90° model, (C) 110° model. Scale bar 0.5 cm. Photos by Pablo Yañez.
Table 1.
Vortex parameters quantified for 3D cross-step models (mean ± SD, n = 5 vortices).
Fig 6.
Sustained vortices were generated inside the oral cavity of paddlefish that had been preserved in ram suspension-feeding position.
Lateral views of three paddlefish specimens (A, B, C) in the recirculating flow tank, with mainstream flow entering the open gape at the right of each image. Vortices were visualized by introducing dye into the slot between the first and second ceratobranchials via the tip of an infusion needle (ne) that was flush with the first branchial arch (ba). Scale bars 0.5 cm.
Table 2.
Vortex parameters quantified for paddlefish preserved in suspension-feeding position (mean ± SD, n = 5 vortices).
Table 3.
Flow speeds in the oral cavities of paddlefish preserved in suspension-feeding position (mean ± SD, n = 3 recordings per paddlefish per location).
Table 4.
Paddlefish gill slot area and oral gape area (cm2).
Fig 7.
Fluid exit ratio was a significant predictor of flow speed in preserved paddlefish.
Linear regressions showed significant relationships between the fluid exit ratios for the three preserved paddlefish that were used in the flow tank experiments and the flow speeds anterior of the rostrum (R2 = 0.49, adj. R2 = 0.42, p = 0.035, n = 3 recordings per fish), at the gape (R2 = 0.92, adj. R2 = 0.91, p < 0.0001, n = 3 recordings per fish), and directly dorsal of the first ceratobranchial (R2 = 0.94, adj. R2 = 0.93, p < 0.0001, n = 3 recordings per fish), as well as the tangential speed of the forced vortex generated by the first ceratobranchial (R2 = 0.37, adj. R2 = 0.33, p = 0.02, n = 5 vortices per fish). Shaded regions represent 95% confidence intervals of each regression line. Horizontal jitter was applied for better visibility of the points, and does not represent variation in fluid exit ratio within fish.
Table 5.
Paddlefish branchial arch angles (mean ± SD, n = 3 fish).
Fig 8.
Inner and outer streamlines of the forced vortex visualized using water-tracing dye in the physical model.
Red rectangle on the CAD image of the 90° model (A) denotes location of the enlargement (B) illustrating water-tracing dye released through cannulae (ca), with mainstream flow (MF) entering the open gape at the right of each image. Vortices are viewed from the exterior of the model, through the mesh, as in Fig 4. For clarity, the outer streamline of the forced vortex is shown in magenta and an inner streamline of the vortex is shown in yellow, with (C) illustrating the radius (r) and the distance (d) traveled by the yellow streamline along the slot during one complete revolution of the vortex. Scale bars 0.5 cm. Illustration by Eric Carstens; revised by M. Carly Lin.