Batoid fishes (rays, skates, sawfishes, and guitarfishes) are macrosmatic which means they rely heavily on their sense of smell for survival and reproduction. Olfactory cues provide important information for navigation and tracking, recognition of prey/predators/conspecifics, and reproductive signaling. For batoid fishes to receive an olfactory signal for sensory processing, an odorant molecule must traverse the external fluid environment, funnel into the nose, and bind with an odorant receptor. Therefore, sensory processing times will depend on odor capture and nasal irrigation efficiency. However, batoid fishes are dorsoventrally compressed with their nostrils on the ventral surface of their body. Their nostrils, called the nares, are disconnected from the pharynx and mouth, and physically separate olfaction from respiration. Therefore, there is not a direct pump-like mechanism to irrigate their nares. The ventral position and pump-less design of the batoid nare presents several challenges for odor capture and irrigation, which may have led to the expansive nasal diversity we see in this group. This dissertation explores the comparative functional morphology and fluid dynamics of the olfactory apparatus and its functions in these fishes. The first chapter of this dissertation categorizes the diversity of nare morphology across Batoidea into discrete morphotypes within a comparative, phylogenetic, and functional framework. The second chapter examines how the internal anatomy of different morphotypes influences nasal irrigation efficiency. The final chapter examines how the external anatomy of different morphotypes influences odor capture potential. The batoid nares consist of two paired, blind chambers on the medioventral surface of the fish, near the mouth and gills. There are two major anatomical components: an incurrent nostril where water is thought to enter and an excurrent nostril where water is thought to leave. There is considerable morphological diversity in the shape, size, and placement of the incurrent and excurrent nostrils. Batoid fishes possess one or more nasal flaps situated around their incurrent nostrils. Most batoid fishes also have a nasal curtain that loosely covers the excurrent channel and forms the excurrent nostril. The nasal curtain is also variable in its morphology. Housed inside the incurrent nostril is the olfactory rosette, which is composed of a longitudinal array of numerous parallel plates of tissue called lamellae. These lamellae are coated in sensory and non-sensory epithelium. To elicit an olfactory response, an odorant molecule must pass into the nose and make contact with the sensory epithelium.
Therefore, batoid fishes rely on water flow to direct odorants into their olfactory chamber for sensory processing. Water needs to be actively drawn into the olfactory chamber due to slow diffusion times and the impeding boundary layer surrounding a swimming fish. There are several hypothesized nasal irrigation mechanisms for batoid fishes. First, the beating of cilia of non-sensory cells (kinociliated cells) likely assists in internal circulation of water through the chamber by generating velocity gradients. However, these cells are unlikely to draw water into the incurrent nostril. Without an internal pump, batoid fishes must rely on harnessing external flows. These flows include the relative forward motion of a swimming fish (i.e., the motion “pump”), indirect respiratory flow (i.e., buccopharyngeal pump), pitot- and venturi-like mechanisms, and/or viscous entrainment. Actively swimming fishes may rely on the motion pump. However, the ventral placement of the batoid nares is not aligned with the freestream flow direction, which limits the motion pump. Stationary or slow swimming batoids may draw from the indirect respiratory current generated by the buccopharyngeal pump. The close proximity of the excurrent nostril with the mouth suggests that mouth suction may help to facilitate flow through the olfactory chamber. The morphology and relative position of the incurrent and excurrent nostrils may also assist in nasal irrigation by generating a secondary flow through pressures differences (pitot/venturi) or a shearing force (viscous entrainment). Additional morphological features, like the nasal flaps or nasal curtain may also have some sort of hydrodynamic function.
To understand how this unique nasal morphology influences olfaction, I first explored the functional morphology of the nares across Batoidea. In the first chapter, I develop a morphometric model to quantify the diversity in incurrent nostril shapes, sizes, and positions on the head in an ecological, phylogenetic, and functional framework. Specifically, swimming mode, lifestyle (benthic vs. pelagic), habitat, and diet were examined for correlations with nasal morphotype. Morphometric measurements were taken on all 4 orders present in Batoidea to broadly encompass nasal diversity (Rhinopristiformes 4/5 families; Rajiformes 2/4 families; Torpediniformes 4/4 families; Myliobatiformes 8/11 families). External nasal diversity was categorized into 3 major morphological groups with several subtypes, termed: flush [subtypes: circle, comma, intermediate], open, and protruding nasal morphotypes.
There were several phylogenetically independent and statistically different morphological traits within these nasal morphotypes. Specifically, the position and angle of the nostril on the head, the width of the incurrent nostril, and the spacing of the incurrent nostrils largely distinguished these groups. These significant measurements correlated best with the swimming mode of the animal. Oscillatory swimmers (e.g., eagle rays, bat rays, manta rays) and body-caudal-fin (BCF) swimmers (e.g., guitarfishes, sawfishes) have narrower heads with larger nostrils that are positioned closer to the edge of their disc. However, while BCF and oscillatory swimmers had statistically similar traits, they occupied distinct regions of the morphospace. This is likely because BCF swimmers could be discriminated from oscillatory swimmers due to differences in the nasal curtain (or anterior nasal flap). Generally, oscillatory and BCF swimmers are some of the largest, fastest swimming batoids that operate at higher Reynolds numbers compared to their congeners.
However, correlations with morphology and ecology were not straightforward. There was no significant difference in the nasal morphology of fast-swimming, pelagic rays vs. slow-swimming, benthic rays, suggesting that morphology is not driven strictly by Reynolds number. Swimming mode may have been most influential for predicting nasal morphology because the motion in which the animal swims will directly affect how water and odorants are directed into its nose. For example, BCF swimmers move their heads in wide lateral sweeping motions (i.e., yawing) which may help entrain and flush their horizontally expanded, open nostrils. Oscillatory swimmers that swim with a vertical up and down motion (i.e., pitching) have vertically oriented, flush nostrils. Undulatory swimmers have more variable swimming movements that include pitching, yawing, and quick turns, and they have a diversity of nostril types. Overall, the open nare morphotype is seen in exclusively BCF swimmers, the protruding nare morphotype is seen exclusively in “true” punters (see Ch.1 for definition), and the flush morphotype was seen across a diversity of swimming modes.
In the second chapter, we aimed to understand how these discrete nasal morphotypes may influence nasal irrigation by visualizing flow into and through the differing olfactory chambers. We CT-scanned representative species of each major morphotype and one subtype (open, flush, protruding, and comma subtype) and 3D printed clear models of the head of each morphotype. Clear models were mounted in a water tunnel and neutrally buoyant dye was injected one cm from the leading edge of all models and filmed with a high-speed camera. To understand how relevant ecological and behavioral parameters may influence nasal irrigation, we tested models at varying Reynolds number (Re = 100, 500), angles of attack (head pitch 0◦, 8◦), and with and without mouth induced respiratory suction.
To compare morphotypes across parameters, we recorded the time it took for dye to reach important components of the olfactory system. We hypothesized that increasing Reynolds number, head pitch, and mouth suction would result in faster nasal irrigation times. However, we found that these parameters influenced nasal irrigation differently across morphotypes. Some morphotypes displayed quicker nasal irrigation at higher Reynolds numbers (open, flush), while other morphotypes (protruding, comma) were not significantly influenced by Reynolds number. These morphotypes (protruding, comma) displayed internal recirculating flow near their olfactory lamellae. This recirculating fluid could help increase the chances an odorant comes in contact with the sensory epithelium by increasing the time near the lamellae. Respiratory suction and head pitch were crucial for nasal irrigation in the morphotypes without this internal recirculation (open, flush). Because the open and flush morphotypes required certain behavioral modifications to irrigate their nares, these morphotypes were classified as “specific smellers.” Morphotypes that could efficiently irrigate their nares independent of respiration, head pitch, and Reynolds number were classified as “dynamic smellers.” The open morphotype, a specific smeller, is seen in BCF swimmers that swim at relatively fast speeds. Therefore, this morphotype may necessitate a higher Reynolds number flow and the yawing motion of its swimming mode to efficiently irrigate its nares. The flush morphotype, also a specific smeller, is seen across Batoidea, from oscillatory eagle rays to undulatory stingrays. This morphotype required a slight head pitch and benefitted from indirect respiratory flow. Fast swimming, oscillatory swimmers likely irrigate their flush nostrils with the assistance of the head pitching motion seen during this swimming mode. Slower swimming rays with undulatory or intermediate swimming modes may rely more on the respiratory current to irrigate their flush nostrils. The dynamic smellers are seen in a variety of swimming modes, further highlighting their versatile nasal morphology. Finally, the ability to change flow patterns through the nasal chamber with changes in behavior (head orientation, respiratory mode, swimming speed) suggests batoids could play an active role in their own chemoreception. This challenges the longstanding theory that fishes passively sense their chemical environment and have no active control of their chemoreception.
In the third chapter, we aimed to understand how the external anatomy of the nares influences odor capture by visualizing flow around the entrance of the incurrent nostrils. We tested three nasal morphotypes seen in batoid fishes with distinct flap-like protrusions situated around the nares (open, protruding, comma nasal morphotypes). Models of the nasal morphotypes were 3D printed and mounted in a water tunnel and the incoming flow was visualized using particle image velocimetry methods (PIV). Models were tested at varying Reynolds numbers (Re = 500, 1000, 2000, 3000, 6000), angles of attack (0◦, 8◦), and with and without respiratory induced mouth suction. We hypothesized that morphotypes with a longer sagittal nostril protrusion will have a greater nostril reach and odor capture potential. Nostril reach was measured as the expanse (width) of the time averaged streamlines that enter the incurrent nostril. Odor capture potential was measured by calculating a 2D closed loop flux integral around the incurrent nostril and comparing it to the available upstream flux.
This was the first time PIV was used to visualize olfactory flow around the nare of a fish and there were several surprising results. First, olfactory flow around the nares of these morphotypes was much more complex and expansive than previously thought. Second, nostril protrusion length was inversely proportional to nostril reach. The comma morphotype, which had the smallest nostril protrusion, had the largest nostril reach and odor capture potential, with a nostril reach over three times the width of its incurrent nostril. Third, flow into two of the three morphotypes enters the incurrent nostril from multiple directions. This is the first recorded instance of multidirectional olfactory flow into the nostril of a fish. In all morphotypes, the flap-like structures appear to play an important role in protruding out of the boundary layer and locally disturbing the flow to create recirculation around the nares. In two of these morphotypes (open, comma) these recirculation regions funnel water into the incurrent nostril from behind and to the side of the nostril, greatly expanding the reach and odor capture potential of the nares. Fourth, the open morphotype had a large portion of water entering the “excurrent” nostril, suggesting that flow patterns are more complex than simply an incurrent and excurrent nostril. Finally, head pitch, respiration, and Reynolds number were found to impact odor capture differently across morphotypes.
The flap-like external nasal features seen in the open, protruding, and comma morphotypes appear to be an adaptation to increase nostril reach and odor capture potential. However, there appears to be a trade-off in nasal flap length and nostril reach. The long nasal protrusion seen in the protruding morphotype extends out of the boundary layer even at the slowest swimming speeds for this species. But, the nasal protrusion is sufficiently long to prevent flow from entering from behind the incurrent nostril, greatly decreasing the nostril reach. The nares of this morphotype are also situated the farthest from the leading edge of their disc. This morphotype is also exclusively seen in batoids that “punt,” a very slow form of underwater locomotion. For these reasons, the boundary layer at the nostrils is likely to be relatively thicker for this morphotype and it may sacrifice nostril reach to aid in odor capture in a challenging environment. Conversely, the open and comma morphotypes had shorter nostril protrusions that allowed for recirculating flow to enter the nares from multiple directions.
In summary, there are several ways batoids can irrigate their nostrils and these irrigation strategies likely represent many-to-one mapping. Specifically, batoid nasal irrigation and odor capture strategies likely differ across and within the same nasal morphotype and likely even vary situationally with the ecology and behavior of the animal. Batoid fishes with the flush morphotype that are oscillatory swimmers likely rely on the motion pump and their already positive body angle during swimming. Batoid fishes with the flush morphotype that are undulatory, intermediate, or BCF swimmers may need to behaviorally mediate their olfaction by changing their mode of respiration or head angle. Batoid fishes with nostril protrusions (open, protruding, comma) may rely less on behavioral modifications. However, the length of the nasal protrusion is important for nostril reach. Morphotypes with shorter nostril protrusions induce multidirectional flow into the incurrent nostril which could increase the likelihood that an odorant is captured. The open nare morphotype, a fast BCF swimmer, has the least pipe-like geometry and the only morphotype that lacks a nasal curtain. This morphotype has multiple nasal flaps that help disturb and redirect the flow into its nares, but it also relies on behavioral modifications to irrigate its nares at low Reynolds number flows. Finally, the comma morphotype performed the best out of all nasal morphotypes in terms of nasal irrigation speeds, nostril reach, and odor capture potential. This dynamic smeller uses both its anterior nasal flap and nasal curtain to channel a large expanse of water into its incurrent nostrils without behavioral modifications.
The comma nasal morphotype could be a possible candidate for bioinspired design. Specifically, this morphotype does not depend on an internal pump to bring water to their olfactory lamellae and is effective at irrigating its nares at low and high Reynolds numbers (Re =100, 500, 1000, 2000) and with changing orientation into a plume. The internal geometry of this morphotype recirculates fluid around the sensory structures and the external geometry passively captures water over 3 times the width of its own nostril. Therefore, this geometry could be a potential candidate geometry for chemical detection systems onboard underwater vehicles that are often limited by the power demand of the pump. By changing the geometry of the inlet, there is potential to passively increase the amount of water sampled.
This dissertation provides new insights into the unexpected complexity of fish olfaction and highlights the multifactorial nature of successful odor capture and nasal irrigation in these dorsoventrally flattened fishes and may be relevant odor-harnessing geometries for future underwater chemical detection systems.
