Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

On the filtration mechanisms and oral anatomy of lunge-feeding baleen whales Pinto, Sheldon James Dominick 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2011_fall_pinto_sheldon.pdf [ 15.41MB ]
Metadata
JSON: 24-1.0072276.json
JSON-LD: 24-1.0072276-ld.json
RDF/XML (Pretty): 24-1.0072276-rdf.xml
RDF/JSON: 24-1.0072276-rdf.json
Turtle: 24-1.0072276-turtle.txt
N-Triples: 24-1.0072276-rdf-ntriples.txt
Original Record: 24-1.0072276-source.json
Full Text
24-1.0072276-fulltext.txt
Citation
24-1.0072276.ris

Full Text

ON THE FILTRATION MECHANISMS AND ORAL ANATOMY OF LUNGE-FEEDING BALEEN WHALES by SHELDON JAMES DOMINICK PINTO B.Sc., The University of Alberta, 2006  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  October 2011 © Sheldon James Dominick Pinto, 2011  Abstract Here we endeavoured to quantify the filtration mechanics of rorquals and the material properties of baleen “gums” (termed zwischensubstanz) by examining and testing the baleen of a fin whale (Balaenoptera physalus). It was hypothesized that fin whales use cross-flow filtration to filter krill from engulfed seawater such that krill and other debris do not become entangled in the baleen fringes. Cross-flow filtration was proposed as an alternate mechanism to dead-end sieving since it would create a highly concentrated suspension of krill inside the mouth (potentially at the oesophageal opening) and would also not require krill to contact the baleen, eliminating clogging and filtering efficiency losses.  We tested filtration mechanisms by placing a sixty-two centimetre section of baleen from a fin whale in a circular water tank and imitating the whale’s environment through various flow scenarios and setups. It was not conclusively determined whether cross-flow filtration is the mechanism used by fin whales, but a new mechanism was proposed called centripetal filtration in which two slugs of water spiral anteriorly on the left and right side of the whale’s oral cavity. Further examination of this proposed mechanism is required.  The material properties of the zwischensubstanz that holds baleen plates together and the development of baleen plates through this zwischensubstanz were also examined. Zwischensubstanz exhibits isotropic properties similar to soft rubber in compression with an average Young’s modulus of 2.56 ± 0.60 MPa and 44.4 ± 2.4% hysteresis when compressed at 0.5 Hz, as it appears to space the baleen plates and absorb stresses  ii  translated from the plates. Through this rubbery zwischensubstanz, the baleen plates develop from conical papillae to hard, keratinized plates. The zwischensubstanz forms a matrix around the papillae and is calcified and keratinized before exiting the zwischensubstanz as a fully developed plate.  The discoveries made here with regard to centripetal filtration and the properties of zwischensubstanz are preliminary attempts at quantifying baleen whale filtration and its associated feeding structures. Such work has been rare in the literature and there are many questions left to be answered by eager scientists with regard to the greatest biomechanical event in the world.  	
   	
    iii  Table of Contents Abstract ............................................................................................................................... ii	
   Table of Contents ............................................................................................................... iv	
   List of Tables ..................................................................................................................... vi	
   List of Figures ................................................................................................................... vii	
   Acknowledgements ............................................................................................................ ix	
   CHAPTER ONE ................................................................................................................. 1	
   Introduction ......................................................................................................................... 1	
   Background ......................................................................................................................... 3	
   Feeding structures ........................................................................................................... 4	
   Baleen ......................................................................................................................... 5	
   Zwischensubstanz ....................................................................................................... 8	
   Oral cavity and tongue ................................................................................................ 9	
   Ventral groove blubber (VGB) ................................................................................. 12	
   Prey ............................................................................................................................... 13	
   Feeding methods ........................................................................................................... 16	
   Lunge-feeding ........................................................................................................... 18	
   Filtration mechanisms and fluid dynamics ................................................................... 21	
   Hypothesis..................................................................................................................... 25	
   Materials and Methods ...................................................................................................... 27	
   Baleen model in a flow tank ......................................................................................... 27	
   Fin whale baleen in a flow tank .................................................................................... 28	
   Imitating a fin whale’s feeding environment ................................................................ 31	
   Circular tank experiment............................................................................................... 34	
   Measurement techniques ............................................................................................... 39	
   Results ............................................................................................................................... 42	
   Observations from trial experiments ............................................................................. 42	
   Flow velocity measurements......................................................................................... 45	
   Observations of circular tank experiment ..................................................................... 47	
   Baleen resistance measurements and observations ....................................................... 49	
   Discussion ......................................................................................................................... 50	
   Experimental setup analysis .......................................................................................... 50	
   Cross-flow filtration ...................................................................................................... 51	
   iv  Centripetal filtration ...................................................................................................... 53	
   Baleen resistance ........................................................................................................... 55	
   CHAPTER TWO .............................................................................................................. 56 Introduction ....................................................................................................................... 56	
   Background ....................................................................................................................... 57	
   Hypothesis..................................................................................................................... 60	
   Materials and Methods ...................................................................................................... 62	
   Compression and tensile tests ....................................................................................... 62	
   Young’s modulus and hysteresis calculations .............................................................. 65	
   Microscopy ................................................................................................................... 67	
   Results ............................................................................................................................... 69	
   Compression and tension tests ...................................................................................... 69	
   Microscopy observations .............................................................................................. 76	
   Discussion ......................................................................................................................... 90 Comparison to other keratins ........................................................................................ 93	
   Microscopic analysis ..................................................................................................... 95	
   CONCLUSIONS .............................................................................................................. 97 Filtration Mechanics ......................................................................................................... 97	
   Future directions ........................................................................................................... 97	
   Proposal for a new theory of filtration .......................................................................... 99	
   Conclusion .................................................................................................................. 100	
   Zwischensubstanz Properties .......................................................................................... 102	
   Future directions ......................................................................................................... 102	
   Conclusion .................................................................................................................. 103	
   REFERENCES ............................................................................................................... 105	
    v  List of Tables CHAPTER ONE Table 1.1: Tagged fin whale data ...................................................................................... 20 Table 1.2: Cross-flow filtration measurements ................................................................. 46 Table 1.3: Dead-end filtration measurements ................................................................... 47 Table 1.4: Baleen resistance measurements...................................................................... 49 CHAPTER TWO Table 2.1: Zwischensubstanz compression testing results ................................................ 73 Table 2.2: Zwischensubstanz tensile testing results ......................................................... 76  vi  List of Figures CHAPTER ONE Figure 1.1: Baleen plate shape changes .............................................................................. 6 Figure 1.2: Sei whale baleen laminae curvature ................................................................. 7 Figure 1.3: Mysticete mouth cross-sections...................................................................... 10 Figure 1.4: Fin whale tongue development....................................................................... 11 Figure 1.5: Dead-end versus cross-flow filtration ............................................................ 23 Figure 1.6: Baleen curvature angle measurement ............................................................. 27 Figure 1.7: Bamfield flume trial ....................................................................................... 30 Figure 1.8: Vancouver Aquarium plywood box trial ........................................................ 33 Figure 1.9: Circular tank experimental setup .................................................................... 35 Figure 1.10: Baleen metal attachment plate ...................................................................... 36 Figure 1.11: Neoprene rubber baleen seal ........................................................................ 38 Figure 1.12: Baleen measurement locations and directions.............................................. 39 Figure 1.13: Plate curvature changes along baleen rack ................................................... 43 Figure 1.14: Bent fringes in cross-flow ............................................................................ 48 Figure 1.15: Centripetal filtration diagram ....................................................................... 54 CHAPTER TWO Figure 2.1: Papillae growing into zwischensubstanz ........................................................ 60 Figure 2.2: Zwischensubstanz sampling locations............................................................ 62 Figure 2.3: Compression testing setup .............................................................................. 63 Figure 2.4: Typical keratin curve ...................................................................................... 65 Figure 2.5: Hysteresis calculation method example ......................................................... 66 Figure 2.6: Zwischensubstanz microscopy sampling locations ........................................ 67 Figure 2.7: Polarized light microscopy interpretation ...................................................... 68 Figure 2.8: Antero-posterior compression loading plots .................................................. 70 Figure 2.9: Medio-lateral compression loading plots ....................................................... 71 Figure 2.10: Dorso-ventral compression loading plots ..................................................... 72 Figure 2.11: Dorso-ventral tensile loading plots............................................................... 74 Figure 2.12: Medio-lateral tensile loading plots ............................................................... 75 Figure 2.13: Antero-posterior slice of zwischensubstanz ................................................. 78 Figure 2.14: Antero-posterior slice of zwischensubstanz with filter ................................ 78 Figure 2.15: Dorso-ventral slice of zwischensubstanz .................................................... 79 vii  Figure 2.16: Dorso-ventral slice of zwischensubstanz with filter..................................... 80 Figure 2.17: Medio-lateral slice of baleen plate in dorsal zwischensubstanz ................... 81 Figure 2.18: Medio-lateral slice of baleen plate in ventral zwischensubstanz ................. 82 Figure 2.19: Dorso-ventral slice of ventral baleen plate and zwischensubstanz ............. 84 Figure 2.20: Dorso-ventral slice of baleen plate in ventral zwischensubstanz ................ 85 Figure 2.21: Dorso-ventral slice of baleen plate near ventral surface with filter ............ 86 Figure 2.22: Dorso-ventral slice of baleen plate near ventral surface ............................. 87 Figure 2.23: Magnified dorso-ventral slice of baleen plate near ventral surface ............. 88  viii  Acknowledgements First and foremost I would like to thank my supervisor, Bob Shadwick, for accepting me as a student and guiding me into the world of biology. It was a fantastic nine-hundred and ninety nine days working in the Shadwick lab and I really appreciate the opportunities that you’ve given me. The entire Shadwick lab was also a key to my academic success and without them I might not have made it. In particular I would like to thank Micha for the many hours that we spent discussing our new lives in Zoology and the trials and tribulations of having to memorize latin names and other such biological nuances. Both John Madden and John Gosline on my supervisory committee were also amazing to work with. It was an honour to learn from you guys and discuss my ideas with your knowledgeable minds. Within the Zoology family there were many more people that helped me get here including the boys in the shop who helped design and build many experimental parts, the comparative physiology group who have been wonderful to work alongside and learn from, everybody that I have TAed with/for who have opened my eyes to the delights of teaching and the administrative staff who were able to answer all my questions along the way. My family has been incredibly supportive of my academic life and without the continuous support of my parents and sister I would never have had the opportunity to take on a Master’s degree (especially in Biology!). Thanks for giving me the tools I’ve needed to get this far. My lovely wife Alys has been, without a doubt, the biggest reason I am here. Not only has she been my sugar mama (allowing us to live in Vancouver for four years), but every success I have had since I met her has been in no small part a result of her wonderfulness. Thank you for everything. Lastly, I’d also like to thank Friday doughnuts, an absolutely crucial part of my weekly routine over the past thirty-three months.  ix  Chapter One This envie is a mightie monster greate, That swims like whale amonge the little frie, Whose gaping mouth would soone consume and eate The gogions small that in small corners lye -Thomas Churchyard from The First Parte of Churchyardes Chippes  Introduction The whale has long been a creature of mystical fascination. Communities around the globe have historically adopted whales into their cultures not only as a source of food, but also as a symbol of the impressive beauty that nature has to offer. The indigenous people of Canada gave the “great whale” a most fantastic symbolism and integrated the whale’s mysteries deep into their society through art and culture. With their enormous bodies and sleek, effortless swimming whales became entrenched in Native folklore and took on many spiritual meanings dating back several hundred years. This mysticism was, and still is, largely based on the unknown elements of the whale’s lives. Out of sight in the depths of the ocean, whales spend most of their time feeding and searching for food. This insatiable appetite has led to whales becoming some of the largest animals in history (Alexander 1998) and furthered our fascination with them. Baleen whales (mysticetes), and in particular the rorquals, make up many of the largest animals in history as a combination of their environment, prey abundance and, most of all, their highly efficient feeding methods have permitted a heavy investment of energy into growth (Goldbogen et al. 2011). And so our fascination turns from awe to one of scientific wonder.  1  Baleen whales (mysticetes) are distinguished from toothed whales (odontocetes) in that they have no teeth and instead use racks of baleen to filter prey from surrounding seawater. Filter feeding is known to be energetically efficient (Rubenstein & Koehl 1977) and is used by several fish (Sanderson et al. 2001), sharks (Motta et al. 2010) and even the tiny zooplankton on which the whales feed (Hamner 1988). The filtration process and its various techniques have been well documented in most of these other species, but extensive research on whales has been prevented not only by their size, but also their elusiveness. Research on living rorquals is now highly regulated by organisations such as the International Whaling Commission due to a long history of excessive whaling that has depleted several whale species to near extinction. Unfortunately, at the time of the heavy whaling relatively little scientific research was conducted on the captured animals, and so the scientific community is left with few options in order to understand whale behaviour and even simple aspects of whale anatomy and physiology. One method that is now used in place of capturing whales is bio-logging or tagging, which provides one of the only means of observing the whales’ behaviours. The rapid advancement of tagging technology has enabled scientists to collect extraordinary amounts of data that can then be related back to animal behaviours. Information such as swimming speed, depth and body orientation can all be calculated from digital tag data, and behaviours can also be observed directly through the use of crittercams and other recording devices.  The feeding methods of baleen whales are one such activity that is becoming more accessible thanks to this advancement in technology, giving researchers a peek at behaviours surrounding foraging excursions and a previously unavailable ability to  2  observe whales in deep ocean waters. It was, however, the observation of stranded baleen whales that initially warranted my own investigation into these filter feeding mechanisms. After examining the oral cavities of several stranded rorqual whales, Bob Shadwick and Jeremy Goldbogen started to ask the question “why is there never any krill or other debris stuck in the fringes of their baleen?”. Historically, marine filter feeders have always been assumed to sieve their prey from seawater (as in a dead-end filter in which fluid flows directly through a filter membrane). So, if the baleen of whales acts as a dead-end filter for the krill (Euphausiids) that they consume, then one would expect a few krill to be stuck in its hairy fringes and remain there at least a small percentage of the time. This conundrum was the starting point of this thesis, and rather quickly it became clear that investigating the mechanisms involved in the prey filtration of rorqual whales was going to require extensive research into many fields peripheral to biomechanics that have yet to be thoroughly studied.  Background In approaching filter feeding it is essential to understand all aspects of the involved systems and their interactions with each other and their environments. Understanding the filter feeding process requires knowledge of whale anatomy, prey composition and filter mechanics. Before expanding on these factors we must establish that baleen whales are known to filter feed using three distinct methods: continuous ram-feeding, suctionfeeding, and lunge-feeding (Pivorunas 1979, Werth 2001), which will be discussed later. The focus here is on the lunge-feeding method performed by rorquals. Lunge-feeding is in fact the primary method of feeding for all Balaenopterids such as fin (B. physalus),  3  blue (B. musculus), Bryde’s (B. brydei), Sei (B. borealis) and minke (B. bonaerensis and B. acutorostrata) whales as well as the Humpback whale (Megaptera novaeangliae).  Several anatomical features are involved in the lunge-feeding process such as the baleen, the gum-like zwischensubstanz and the ventral groove blubber (VGB) as well as the internal structures of the oral cavity like the prominent vomer bone of fin whales and the large, elastic tongue that sits on the floor of the mouth. These structures will influence the filtration of seawater and the retention and swallowing of the prey. Accounting for the size, weight and buoyancy of this prey only scrapes the surface of its effects on lungefeeding, as the living prey will also be dynamic both during the capture and filtration processes. From a mechanical perspective, the whale’s baleen and the prey-laden seawater being filtered are analogous to an industrial filtration system in which particles are separated from fluids except that most industrial filters are of a much smaller scale. Regardless, many of the concepts that govern filtration are applicable at all scales and should be considered in setting up this analysis.  Feeding structures An understanding of the composition and mechanical properties of baleen and the associated anatomical structures is essential in determining their reactions to the forces and pressures caused by water flow during feeding. These properties of the feeding structures can be used in modelling the filtration process and defining aspects of the filter such as resistance to flow. However, the mechanical properties do not answer all the questions needed to model filter feeding. One must also understand the function of these  4  structures, both for feeding and non-feeding activities, to fully incorporate their effects into a experiment. As previously stated, acquiring verifiable data on both composition and function of whale anatomy is not a simple task.  Baleen Baleen laminae form the basic structure of the filtering apparatus in baleen whales. Laminae consist of one main baleen plate and several minor plates (Figure 1.1) (Williamson 1973). The baleen plates are highly calcified keratinous plates (Szewciw et al. 2010) that are spaced about one centimetre apart along the lateral edges of the rounded rostrum of baleen whales effectively hanging down from the rostrum similar to the arrangement of the upper row of teeth in humans. In general, the plates are no more than four millimetres thick, but range in size across species and antero-posteriorly within species (Figure 1.1) (Pivorunas 1976). Maximum plate height (extending ventrally from its base in the rostrum) is about twenty centimetres in minke whales whereas plates can reach heights of four metres in right whales (Pivorunas 1979).  5  Figure 1.1: Baleen plate shape changes along the length of the right baleen rack from a 19m fin whale at Hvalur HF. Plates are shown at approximately every 40cm from the anterior tip of the baleen rack (1) to the posterior end near the temporomandibular joint (10).  On the other hand, medio-lateral width of the plates is relatively small in right whales compared to blue whales (Figure 1.3). Thus, when baleen is used to filter prey from water, the effective filtering area through which water can flow ranges from 0.15 m2 in minke whales to 5.7 m2 in right whales (from Fig. 8-6, Kawamura 1974). At the anterior tip of the rostrum, baleen plates are their shortest, and in some species, such as bowhead whales, there is no baleen at the anterior end of the rostrum, thus forming two separate baleen racks on the left and right hand side of the mouth.  From a filtering perspective, it appears that the baleen plates composing one lamina can be treated as one plate as the spacing between the main and minor plates is small and inconsequential to flow dynamics (Figure 1.2). 6  Figure 1.2: Baleen lamina curvature from a Sei whale (Balaenoptera borealis) adapted from Pivorunas (1976) showing baleen fringes on its medial edge, minor plates and zwischensubstanz at its dorsal base. The general shape of baleen laminae is similar in all Balaenopterids but varies anteroposteriorly.  Each plate is composed of small tubules aligned within the plate and cemented together with highly calcified keratin to form a matrix (St. Aubin et al.1984; Szewciw et al. 2010). The matrix at the medial edge of the plates is worn away, presumably by the tongue, to expose strands of the internal tubules, which then line the entire medial edge of each plate. This exposure of tubules on the plates’ medial edges occurs to such an extent that a fibrous matted layer consisting of intertwining tubules from each baleen plate is created along the entire baleen rack. These frayed tubules are less than 200µm in diameter and  7  thus extremely compliant, but with Young’s Moduli (a measure of the elasticity of a material) of 0.65 ± 0.03 GPa and 1.22 ± 0.06 GPa in minkes and humpbacks respectively (Szewciw et al. 2010) are also quite strong. Plates are somewhat triangular in shape (Figure 1.2) but contain curvature along multiple axes. In fin whales, plates are curved antero-posteriorly with posterior concavity, but they fluctuate in curvature magnitude along the length of the racks. The lateral edge of baleen plates also does not form a straight edge; instead the tips of the plates curl out laterally (Figure 1.2).  Only recently, with material testing by Szewciw et al. (2010), have the mechanical properties of baleen been examined in any detail. As such, tensile and flexural properties of fin whale baleen tubules as well as important factors such as flexural rigidity of entire baleen plates and entire baleen racks in situ have not been determined. These unknown quantities that define the filtration mechanisms would be useful in creating an accurate model, but require extensive research in themselves.  Zwischensubstanz Holding the baleen plates in place is a rubber-like gum base termed “zwischensubstanz” (Tullberg 1883), which is German for “in-between substance” for lack of a more Anglicised term (Fudge et al.2009). Tullberg (1883) introduces and begins to describe the zwischensubstanz as a soft keratin substance that attaches to the maxillary of the rostrum via connective tissue, but in the 128 year period since these rather sparse descriptions very little has been accomplished in the way of understanding the material properties and  8  function of the zwischensubstanz. The role of the zwischensubstanz appears to be to hold the baleen in place, but depending on its material properties, the dynamics of the baleen plates during filtration will be greatly affected. If the zwischensubstanz is more viscous, then perhaps the baleen plates can adapt to incoming flow or if the zwischensubstanz is more elastic it can help create some type of pulsatile motion to aid in filtration. The material properties of zwischensubstanz and its role in feeding are examined in depth in Chapter Two.  Oral cavity and tongue The structure of the oral cavity of baleen whales plays a large role in the fluid dynamics of lunge-feeding. As Werth (2004) discovered during his analysis of Balaenidae filter feeding, the shape of the oral cavity is a key to the efficiency of filter feeding in whales. Simply put in mechanical terms, a well-designed system for flow in a filtration system can minimize energy expenditure and increase efficiency. There are two major components to the oral cavity of Balaenopterids: the vomer and the tongue.  9  Figure 1.3: Comparison of cross-sections and profiles from three baleen whale species emphasizing the size of baleen plates and the size and shape of the vomer (adapted from Brodie & Vikingsson 2009)  The vomer is a long bony protrusion of the rostrum that extends antero-posteriorly along the centre of the rostrum (Figure 1.3 and Figure 1.4) with a parabolic cross-section that decreases in size anteriorly. The vomer is extremely prominent in Balaenopterids compared to most mammals and even compared to other mysticetes (Figure 1.3). Jutting out from the roof of the mouth into the oral cavity, the vomer of the fin whale will passively redirect flow solely on account of its rather intrusive presence. The arched rostrum prominent in all mysticetes makes the vomer even more influential in directing flow as the vomer’s parabolic cross-section can efficiently redirect flow. Without any means of observing the flow patterns in the oral cavity of whales, factoring the vomer into fluid dynamics models is necessary in determining the direction of water flow.  10  Figure 1.4: Diagram of tongue musculature and oral cavity development in a "finner whale" adapted from Pivorunas (1979) showing muscular atrophy of the tongue and the growth of the "keel of palate" (i.e. the vomer) in addition to baleen plate formation.  11  The tongues of adult rorquals are fatty and flaccid, and as such it has been assumed that the tongue is not actively involved in feeding (Pivorunas 1976, Werth 2007). At birth, the tongues of rorquals are highly muscularized in order to create a seal during suckling (Pivorunas 1979). However, as the whale transitions to filter feeding, the tongue atrophies until, in adulthood, there appears to be very little muscle remaining (Figure 1.4). It has been proposed that the tongue may serve in passively sealing the oral cavity during engulfment and may be involved in swallowing the prey (Werth 2001, Lambertsen 1983). The composition of the hyolingual apparatus in Balaenopterids has not been determined, and in fact, apart from Werth’s (2007) summary of cetacean hyolingual adaptations, the function of the tongue in adult baleen whales has not been extensively investigated.  Ventral groove blubber (VGB) Ventral groove blubber (VGB) refers to the accordion-like ventral surface of the anterior half of Balaenopterids. VGB is both highly muscularized and highly extensible (Goldbogen et al. 2010, Goldbogen et al. 2011, Orton and Brodie 1987). There are several muscle layers throughout the VGB with alternating orientations and fibre sizes, but all appear to be extremely extensible in the circumferential direction resulting in anisotropic behaviour of the VGB as a unit (Orton and Brodie 1987). Orton and Brodie’s (1987) analysis showed reversible extensibility in the circumferential direction of 4 times resting length and 1.5 times resting length along the long axis of the body. This strength and flexibility serve to control the amount of incoming water during lunge-feeding by actively regulating the rate of incoming seawater and also create the pressure required 12  during the filtration phase of the lunge-feeding process described later (Potvin et al. 2009). During non-feeding swimming, the VGB remains contracted to create the hydrodynamically efficient body form associated with whales. When the VGB is fully contracted, there is no evidence to suggest that the oral cavity contains large amounts of water with the VGB most likely forcing the tongue up against the roof of the mouth.  Prey The prey choices among mysticetes vary depending on feeding methods. Clearly, certain feeding behaviours such as the bottom-feeding, suction methods of gray whales will result in different prey being consumed than the surface-feeding skimming methods of right and bowhead whales (discussed later). The lunge-feeding method of Balaenopterids favours the collection of large zooplankton, especially euphausiids (krill) such as Euphausia superba (Pauly et al.1998). Based mostly on their stomach contents, Balaenopterids are known (to a far lesser extent) to also prey upon small fish and in fact use various feeding methods as the situation requires (Kawamura 1974). In light of the fact that fin and blue whales feed 80-100% on large zooplankton (Pauly et al.1998), our focus here is on the krill that makes up the majority of their diet. The size and weight of the krill are clearly influential factors in aquatic filter feeding, but in a more advanced model one may also consider the active actions and reactions of these animals and determine if they impact the filtration dynamics. Extremely mobile and evasive prey will force the whale to adapt its behaviour to the prey’s movements. The plasticity of the prey (i.e. their susceptibility to being squished and the extent of said squishing) may also have a large influence on the filtering process.  13  Euphausia superba have an average length of about 53 mm and an average weight of 0.45 g (Croxall and Pilcher 1984). At the lower end of the size spectrum the krill would be small enough to fit through the gaps in the baleen plates were the gaps not covered by the baleen fringes. In terms of weight, the most important factor with regard to the filtration mechanics of an aquatic prey item is not the actual mass of the krill, but rather its buoyancy. Buoyancy in most zooplankton is determined by lipid content and varies with age, time of day and season (Campbell and Dower 2003, Visser and Jonasdottir 1999, Yayanos et al.1978). Lipid content and swimming performance has not been examined in great detail for Euphausia in particular, however much attention has been directed at these aspects of calanoid copepods (Order Calanoida). Being similar species with respect to biochemical composition, many of these aspects can be applied to krill as well. In copepods, lipid content can range from 10% to 76% of dry weight depending on age, time and season (Campbell and Dower 2003, Visser and Jonasdottir 1999), and thus buoyancy will vary along with it. Because the thermal expansion of zooplankton is much greater than seawater, zooplankton go from being positively buoyant at the surface, to neutrally buoyant several hundred metres below the surface to negatively buoyant below that point (Campbell and Dower 2003). As a result of the high dependency on lipid content for buoyancy, krill are rarely (if ever) in a stable environment (Campbell and Dower 2003), forcing them to swim constantly to maintain level in the water column. This may aid or inhibit diel vertical migration, and influence their aggregation methods, which in turn affect the whale’s feeding methods.  14  Euphausia swimming data is scarce in the literature, but De Robertis et al. (2003) observed E. pacifica swimming at median speeds of 1.8 cm·s-1 on a regular basis and 2.23.5 cm·s-1 during vertical ascent. Moving obliquely (<60º from the horizontal plane) through the water column, krill perform a vertical migration at dusk during which time their aggregations become less dense. At dawn, the krill will return to their more dense aggregations at depth, travelling in a similar fashion during descent. This aggregation behaviour influences the feeding habits of Balaenopterids, as dense aggregations of krill are required to energetically justify feeding on such small organisms (Goldbogen et al. 2011) forcing them to feed primarily during daylight hours.  Escape responses of Euphausia are quite different from diel migrations, with speeds of up to 60 cm·s-1 observed in response to predators (Kils 1979). O’Brien (1987) categorized Euphausia escape responses into three response types: avoidance (primary), coordinated escape (secondary) and individual escape (tertiary). The initial reaction to human predators is to swim in the opposite direction (primary) at approximately 20 cm·s-1. If the predator continues to approach, the aggregation of krill will coordinate their movements to deter the predator (secondary). This coordination lasts longer in large aggregations, but is seen with Euphausia in any group size. If the predator continues to pursue the group, the individuals within the aggregation will begin a tail-spinning manoeuvre to confuse the predator (tertiary). These manoeuvres are used in response to all sized predators (O’Brien 1987), but are more effective when avoiding smaller predators. It also appears that the tendency of krill is to create more dense aggregations to avoid predators. Although this  15  may be effective against small fish, it is in fact helpful to a whale to have a denser patch to engulf.  Another aspect that is important with regard to krill as a source of prey is that krill identify their predators visually (O’Brien 1987). In light of this fact, whales have adapted their feeding behaviours to limit being seen when approaching. As you’d imagine, that could be a difficult task for a whale. Balaenopterids avoid approaching krill aggregations from directly above (Goldbogen et al.2007) thus limiting casting shadows on the krill and alarming them to its presence.  Feeding methods Baleen whales are amongst several species of aquatic animals that filter feed. Ranging from the tiny Euphausia on which mysticetes themselves feed to large Elasmobranchii, the mechanisms used by filter feeders range almost as much as their body size. However, mysticetes filter feed using only three distinct methods: continuous ram feeding, suction feeding, and lunge-feeding (Pivorunas 1979, Werth 2001). Continuous ram feeding is performed mostly by right whales (Eubalaena) and bowhead whales (Balaena). Feeding throughout the water column (Lowry 1993), the whale will open its mouth and swim through an area of high prey density with mouth continually agape. The whales use their large muscular tongue to passively direct water along the left and right baleen racks and the prey, consisting mostly of copepods and krill (Lowry and Burns 1980), are concentrated near the oesophageal orifice. This is most likely done via active displacement of the tongue while the seawater passes through the baleen or is actively  16  pushed out of the oral cavity through other openings (Lambertsen et al. 2005). Lambertsen et al. (2005) have made these predictions of filtration mechanics based on anatomical observations of bowheads caught by the Inupiat whale hunters. However, the actual filtration process has never been documented to verify these hypotheses. Thus the precise mechanisms of continuous ram feeding remain unverified.  Suction feeding is performed mostly by gray whales (Eschrichtius robustus), which feed predominantly on benthic organisms on the floor of the ocean, although, based on their stomach contents, they are known to be opportunistic feeders (Ray and Schevill 1974, Nerini 1984). By rapidly opening their mouth and posteriorly displacing their tongue within their oral cavity, gray whales create a negative pressure inside their mouth that draws in surrounding seawater and prey (Ray and Schevill 1974). The seawater is then actively pulsated out of the mouth through the baleen, while the prey is prevented from expulsion by the baleen, instead remaining inside the mouth for consumption. Ray and Schevill’s (1974) observations of a captive gray whale calf are the only extensive accounts of gray whale feeding, and they were somewhat concocted and unrealistic due to the limited knowledge of gray whale feeding at the time. Filtration in gray whales has been assumed to be a sieving method (Pivorunas 1979, Werth 2001), but the fluid mechanics as well as the post-filtration gathering and swallowing of the prey are not yet fully understood.  17  Lunge-feeding Rorqual whales capture their prey using a method termed intermittent ram feeding, engulfment feeding or, more recently, lunge-feeding that is considered the “greatest biomechanical event in the world” because of the impressive dynamic processes that occur during each lunging event. As the name implies, rorquals engulf a mixture of prey (mostly krill) and seawater into their oral cavity before filtering out the seawater and retaining the prey for consumption. This process can be detailed through a few simple steps. It all starts with rorquals finding an area of high krill density and swimming towards it. This may be at the surface of the water or several hundred metres below the surface. How rorquals find these dense patches of prey is still a mystery, although it is probable that whales locate their prey either by sight or sound.  Upon encountering the patch of krill, the rorqual will drop its lower mandible, creating a gape angle of approximately seventy degrees. Gape angles of nearly ninety degrees are possible in rorquals, but an energetically optimum angle would be substantially lower than this since a ninety-degree angle creates a significantly higher drag force (Goldbogen et al. 2007). This large gape angle and the enormous size of a rorqual’s mandibles (which consist of one third of the animal’s body length) combined with an outward rotation of the lower mandibles allow large rorquals to engulf volumes of prey and seawater exceeding seventy cubic metres. Engulfment should not be confused with consumption as this volume of water only enters the oral cavity and is not ingested. When the mouth is opened and water begins to enter the oral cavity, the tongue is inverted and stretched to create the lining of a capacious sac that extends ventrally along the whale ending near the  18  umbilicus for a total extension of nearly two-thirds of the whale’s body length (Lambertsen 1983). The volume of water is accommodated through an expansion of the accordion-like ventral groove blubber (VGB) that forms the ventral surface of the body from the anterior tip of the mandibles to the umbilicus (Orton and Brodie 1987). The socalled cavum ventrale defined by Lambertsen (1983) as a slippery spacing in between the intermandibular lining and the ventral groove blubber enables this extension of the oral cavity by creating a lining that envelops the entire engulfed mixture.  Once the desired volume has been engulfed, the whale then raises its lower mandible to allow only the baleen to be exposed to the exterior of the whale thus effectively creating two distinct environments (the engulfed volume and the surrounding seawater) separated only by a membrane (i.e. the baleen racks). At this point the filtration process can begin. The engulfed volume of water is pressurized via contraction of the highly muscularized VGB, and in turn this newly created pressure differential across the baleen forces the water out of the mouth through the baleen. B. physalus take thirty to forty-five seconds to expel the seawater, leading to the calculation by Goldbogen et al. (2007) that water is exiting through the baleen at approximately 0.8 m·s-1 assuming equal filtration rates along the length of the baleen (Table 1.1). In the past it has been assumed that krill, and for that matter any other debris collected, would be captured by the baleen and consumed upon completion of the water expulsion.  19  Table 1.1: Summary of data relevant to the filtration process of fin whales captured using D-tags on animals in the Southern California Bight (from Goldbogen et al. 2007)  Parameter	
   Baleen	
  filter	
  area	
   Engulfment	
  volume	
   Prey	
  density	
   Filter	
  rate	
   Reynolds	
  number	
  for	
  flow	
  past	
  baleen	
  fringes	
   Reynolds	
  number	
  for	
  flow	
  past	
  baleen	
  plates	
   Filtering	
  flow	
  speed	
    Average	
  Value	
   3.0	
  m2	
   71	
  m3	
   145	
  g·m-­‐3	
   2.4	
  m3·s-­‐1	
   570	
   4500	
   0.8	
  m·s-­‐1	
    Range	
   -­‐-­‐	
   60	
  -­‐	
  82	
  m-­‐3	
   133.8	
  -­‐	
  156.8	
  g·m-­‐3	
   2.0	
  -­‐	
  2.7	
  m3·s-­‐1	
   480	
  -­‐	
  650	
   3800	
  -­‐	
  5200	
   0.7	
  -­‐	
  0.9	
  m·s-­‐1	
    The mechanisms behind this seawater expulsion, krill capture and actual consumption have not been examined in the literature since it is very difficult and expensive to perform in vivo experiments on animals of such size. Many hypotheses have been made based solely on theoretical calculations and the necropsies of various species of baleen whale. It has been assumed that the baleen rack is acting like a sieve (i.e. a dead-end filter), collecting the prey on its medial surface for future ingestion. Werth (2001) suggested three distinct methods for prey removal from baleen: mechanical scraping by the tongue, shaking of the head or lips and rapid “backwash” of current, all of which would serve to transport prey to the oesophagus at the posterior end of the tongue to be consumed. All of these methods presuppose that the prey does indeed become trapped in the baleen fringes during filtration, but this may not in fact be the case.  In light of what we now know about rorquals, Werth’s (2001) suggested methods of prey removal all seem unlikely. Mechanically scraping the baleen with the tongue is not possible for rorquals as their tongues are known to be flaccid and minimally muscularized in adulthood (Pivorunas 1979). Although the nonlingual intermandibular 20  lining may contain more substantial muscles (Lambertsen 1983), it is unlikely that it would be enough to dexterously manipulate the large tongue. Both “shaking of the head or lips” and “rapid backwash of current” are unlikely as they both involve behaviours that have not been observed. There are no reports of rorquals shaking their head or lips after a lunging event; nor have they been witnessed to open their mouths and allow a backwashing of current through the baleen to flush the krill from the baleen fringes.  Goldbogen et al. (2007) suggested that if the baleen acted as a cross-flow filter where water flows parallel to the filter membrane as opposed to directly through it (as in a deadend filter), then the prey may never come into contact with the baleen, and therefore never get entangled in the baleen during the feeding process. A cross-flow filter would allow seawater to exit the mouth while the prey is retained in a seawater suspension within the oral cavity before being ingested. This method of filtration is seen in other filter feeding animals such as suspension-feeding fishes (Sanderson et al. 2001) in which gill rakers are used to retain food particles that are smaller than the gaps in the filter itself and is potentially feasible in other filter feeding marine animals as well.  Filtration mechanisms and fluid dynamics Biomimetic endeavours look to nature to inspire designs and solve problems, but we can also look to industrial designs to understand biological processes. The baleen racks of mysticetes are in effect a large-scale filter, performing functions similar to any industrial filter. There are, of course, many different types of industrial filters that all use different mechanisms to separate a feed stream from a filtrate, but often these filters result in a  21  highly concentrated retentate forming a “cake layer” on the filter surface. Thus, in order to understand the mechanics of a biological filter, we can first study the components and mechanics of an industrial filter.  For the sake of simplicity, filtration can be separated (pun intended) into two categories: dead-end filtration and cross-flow filtration. A dead-end filter acts like a sieve where the incoming flow is perpendicular to the filter surface, thus flowing directly through the filter. Particles larger than the pore size of the filter are trapped on the surface of the filter while smaller particles will go straight through. Dead-end filters result in large amounts of retentate accumulating on the filter surface since all particles are flowing directly into the filter (Figure 1.5). This cake layer accumulates rapidly and significantly decreases the efficiency, and thus the effectiveness, of the filter as it creates a higher resistance to flow. When a dead-end filter is used in industry, such as in many oil filters and vacuum cleaners, the filter requires regular cleaning, maintenance and replacement to function at full capacity. In order to reduce the costs of maintenance and repair, many industries are moving towards cross-flow filtration.  22  Figure 1.5: Diagram illustrating the differences between (a) dead-end filtration where water flow directly into the filter and (b) cross-flow filtration where water flow is tangential to the filter.  Cross-flow filtration requires the incoming fluid to flow tangential to the filter (Figure 1.5). The tangential velocity of the fluid creates a pressure differential across the filter that forces the filtrate and some smaller particles through the filter. The denser particles are retained without ever coming into contact with the filter. Although a small cake layer is formed on the filter surface due mostly to concentration polarization (Song and Elimelech 1995), it is far less pronounced than with a dead-end filter, but still affects the filter’s efficiency (Lu et al., 1993). The thickness of the cake layer is governed by the velocity of the tangential flow and, more precisely, the shear stress that scours the caked surface. The theories of cross-flow filtration have been studied extensively for smallscale industrial filtration applications such as harvesting of bioreaction products from fermentation broth, recovering metal precipitates from waste water and removing 23  particles during the wine and beer making processes since the mid-1980s (Lu et al. 1993, Baker et al. 1985, Lu and Ju 1989, Ripperger and Altmann 2002).  One important factor in examining these industrial cross-flow applications is that they are all small scale. For the most part, cross-flow filtration is used to remove particles of only 0.01-1.00 µm (Murkes and Carlsson 1988). At these small scales, the governing equations are influenced more by adhesive properties and molecular level relationships than at larger scales where the Reynolds number and particle density of the turbulent flow feature more prominently. The industrial micro- and ultrafiltration applications are quite different from a whale filtering a five centimetre krill through three squared metres of baleen racks. On the whale’s scale, the influences of cake formation, cake resistance and concentration polarization (Lu et al. 1993, Blatt et al. 1970) may be far less influential. However, being as large-scale cross-flow filtration has no obvious industrial applications it has not been extensively examined.  From a biological perspective, forms of cross-flow filtration have been observed and analysed in other aquatic filter feeders. Filter-feeding fish, such as herring, anchovies and tilapia, all use cross-flow filtration as they ram-suspension feed on particles of only ~53000µm (Sanderson et al. 2001, Cheer 2001). These fish take advantage of high velocities and conically shaped oral cavities to filter water through gill rakers while retaining the tiny prey (Cheer 2001). Whale sharks (Rhinocodon typus) have also been hypothesized to use cross-flow filtration in a manner similar to bony fishes as they feed on prey smaller than dead-end sieving would allow (Motta et al. 2010). Indeed, because  24  of the efficiency of cross-flow filtration relative to dead-end filtration, many filter feeding animals including elasmobranchs (Paig-Tran et al. 2011) are now being discovered to use cross-flow filtration as it appears to be a prime example of convergent evolution.  Again, the common theme between industrial filters and these well-studied filter feeders is their small scales. Unfortunately, this means that there are no established governing rules of large-scale filtration. In industry, large scale filtration is handled by enormous mesh membranes acting as dead-end sieves that are frequently maintained and replaced, whereas in the animal kingdom only the baleen whale filter feeds on objects larger than one centimetre. One significant difference that one can presume for large scale cross-flow filtration is that there will be less (if any) “cake” deposited on the filter surface. As Sanderson et al. (2001) describe, the filtrate flux for large particles is greater than that defined by Brownian diffusion alone and so during filtration of large particles several other mechanisms are presumed to prevent retentate deposits. However, it remains plausible that some other mechanisms are at work in a fin whale’s mouth and that there is a deposition of krill on the baleen rack which is then removed in some other fashion (Werth 2001).  Hypothesis What I propose here is to investigate this theory of cross-flow filtration and whether it is a feasible filtration mechanism for lunge-feeding rorquals. I believe that an analysis of the fluid dynamics around the baleen racks may show that prey is filtered without ever coming into contact with the baleen. Goldbogen et al. (2006) managed to collect in vivo data from lunge-feeding fin whales (Balaenoptera physalus) from which the speed, angle 25  and general behaviours of the diving whales could be determined (Table 1.1). With this information and guided by a lunge-feeding model designed by Potvin et al. (2009), an experiment was designed to simulate the filtration process and quantify the feasibility of cross-flow filtration.  A more complete understanding of this filtration mechanism is important from a mechanical and biological perspective. Mechanically, cross-flow filtration is not often a technique used for anything larger than a microfilter, as engineers have found that deadend filtration is sufficient at these scales. However, baleen whales are performing filtration at enormous scales and clearly doing so efficiently, so perhaps this is a process that can be mimicked for industrial purposes. Biologically there is surprisingly little understood about the biomechanics of whales, and considering that whales have successfully found a niche in nature that is unoccupied by any other animal it implies that there is a great deal of important information that we can obtain by studying them.  26  Materials and Methods Baleen model in a flow tank In order to determine if cross-flow filtration is a feasible mechanism of rorqual filter feeding an experiment was designed to recreate the potential flow of seawater inside the oral cavity of a Balaenoptera physalus. The average girth of a fin whale at its maximum point (near the temporomandibular joint) is over 700 cm (Lockyer 1986) and would be difficult both to house in a flow tank and to acquire or recreate in the first place. It would also be difficult to recreate the curvature of baleen plates (Figure 1.2). While at Hvalur HF in Iceland, measurements were taken to quantify this antero-posterior plate curvature in an approximately 19 m fin whale. Photographs were taken with scale bars using straight edges to clarify the precise curvature (Figure 1.6) from which the curvature angles could then be measured using Image J software. This provided medial and lateral curvature angles of the baleen plates relative to the rostrum along the entire length of one baleen rack and provided an estimate of the changes in baleen curvature along a fin whale’s baleen rack.  Figure 1.6: Baleen angle measurement technique used at Hvalur HF to measure the anteroposterior curvature of baleen plates on both the medial (a) and lateral (b) edges of the baleen rack.  27  Back at UBC though, it was determined that a model of a section of baleen small enough to fit in a flow tank or flume could be used and the whale’s environment could be recreated using more accessible methods. However, an accurate baleen model is difficult to create since very little is known of the material properties of the feeding structures. As a starting point, the baleen plates could be formed of thin, flexible plastic and spaced using wooden blocks and hair, or some other hair-like material, could be used to form the fringes on the medial edge of the baleen plates. Creating a baleen model consisting of several individual parts could also allow several scenarios to be tested such as variations in plate curvature, fringe density and flow orientations in addition to variations in flow speeds. However, before any such model could be fully envisioned, a baleen whale pushed up onto the Vancouver docks.  Fin whale baleen in a flow tank On Sunday July 26, 2009 a deceased 17.3m female fin whale was brought into Vancouver harbour on the bow of a cruise ship. As required by the Species at Risk Act, the Canadian Department of Fisheries and Oceans (DFO) conducted a necropsy on the animal. According to the necropsy report, the 35-40 ton animal was deemed to have been dead for several days before impact from the ship and appeared emaciated despite a lack of clear signs as to the reasons for its poor health (DFO Case # 09/02922, FOS 5028). The right baleen rack from this whale was obtained from the DFO including not only the baleen plates but also the zwischensubstanz in which they were embedded. The rack had been separated into several sections in order to facilitate transportation (herein a “baleen section” refers to a portion of the baleen rack including the zwischensubstanz, separated  28  from the rest of the rack). A 62 cm long (antero-posterior length) baleen section from near the anterior tip of the baleen rack was chosen as the test specimen. The section contained 47 baleen plates ranging in height from 19 cm to 36 cm and a medio-lateral width ranging from 15 cm at the anterior end to 29 cm at the posterior end. The dorsoventral height of the zwischensubstanz was consistently between 3 and 5 cm throughout the section. The entire baleen rack was kept frozen at -20˚C for the duration of the experiments except when removed for preparation and testing.  Initial flow trials were conducted on a larger baleen section from a more posterior location in the same whale. As few others have performed flow experiments with baleen (Mayo et al. 2001, Werth 2004), there was little to learn from the literature in terms of experimental setups and baleen manipulation. Thus, the sole purpose of this first trial was to improve our understanding of the fluid dynamics around baleen and to come up with potential experimental setups and techniques. The flume at the Fluid Dynamics Laboratory in Bamfield, BC was used for these trials. The flume is 12 m long x 2 m wide x 1 m deep, thus being one of the few flumes in western Canada large enough to fit the metre-long baleen section. Flow speeds were tested up to 0.8 m·s-1. The baleen section was suspended in the centre of the flume using an overhead crane and a series of bungee cords strapped between the baleen plates in order to grasp the zwischensubstanz from which the baleen rack then hung (Figure 1.7). This setup allowed for testing of several orientations of the baleen in the flow.  29  Figure 1.7: Team Shadwick conducting trial experiments at Bamfield Marine Sciences Centre using the 12m long flume (a) and a rope and bungee cord apparatus rigging the baleen to a crane (b) to be suspended in the flume.  Following trials at Bamfield, some zwischensubstanz degradation was observed. In order to prevent further degradation, the zwischensubstanz was fixed in increasing doses of ethanol to prevent putrification and to increase its mechanical strength at the molecular level. The sixty-two-centimetre baleen section was first thawed for 24 hours. This was its only complete thawing during my possession in order to reduce structural changes to the plates that might occur during thawing and drying. After 24 hours of thawing, the baleen section was placed in a thick garbage bag such that the zwischensubstanz base sat at the 30  bottom of the bag. The bag was then filled with a 30% ethanol solution consisting of 2 L of 95% ethanol and 4.34 L of water. This volume of over six litres of diluted ethanol allowed the zwischensubstanz to be the completely covered in the solution at the bottom of the garbage bag. The bag was then sealed with duct tape and placed in a fume hood at room temperature for 48 hours. At the end of this period the bag was opened to check on its status. At this time the zwischensubstanz remained as soft as it had when initially thawed. The solution appeared murky as zwischensubstanz pieces had mixed in with it (most likely due to initial disturbances). As the zwischensubstanz was still easily deformable at this point, the alcohol content was increased from 30% to 40% by adding 1L of 95% ethanol to the mixture and thoroughly mixing it with the existing solution before the bag was resealed. After 96 hours in the 40% solution the baleen was removed from the bag and immediately re-frozen to reduce and prevent excessive warping. The zwischensubstanz had hardened significantly and appeared less likely to fall apart and more likely to resist plate separation relative to before the process was completed. The fixing of the zwischensubstanz did not appear to cause any changes to the baleen plates since they had not been submerged in the ethanol solution.  Imitating a fin whale’s feeding environment Designing a flow tank experiment to replicate the feeding process of a fin whale required not only preparing the baleen, but also creating a suitable environment to imitate the oral cavity and surrounding seawater of a fin whale during a lunge-feeding event. Another, more accurate trial was performed using the Marine Mammal Research Unit’s (MMRU) swim mill at the Vancouver Aquarium. This swim mill, which is normally used for  31  metabolic research on fur seals and Stellar sea lions, was 293 cm long x 183 cm wide x 107 cm deep with side walls and floor made of steel and metal cages at either end to prevent live animals from entering the turbine area. Two variable frequency drives (VFD) were used to create the flow. When the volume of water was limited to 60 cm deep and the VFDs were operating at full capacity the water could reach velocities of 1.20 m·s-1 at the centre of the tank with significant amounts of turbulence clouding the water clarity.  During the filtration process of fin whales, two separate environments are created: the inside of the whale’s mouth and the whale’s surrounding environment with the baleen rack being the separating membrane. As such, a simplified form of this environmental distinction needed to be imitated. After a fin whale has closed its mouth and is prepared to commence filtration, the whale is in fact essentially stationary in the water, swimming at 0.5 m·s-1 (or less) as the drag caused by having its mouth agape stalls its forward momentum (Goldbogen et al. 2007). When the whale is presumed to start filtering it also begins fluking again and rotates its body back to upright orientation in order to prepare for the next lunge. In order to simplify this preliminary experiment the whale’s rotation and its increase in forward velocity were not taken into account in these experiments. So the two environments imitated were: stagnant water on the lateral edge of the baleen rack and (potentially) cross-flow along its medial edge.  32  Figure 1.8: a) Trial experiment design for plywood box to create a stagnant water section on the lateral side of the baleen section and b) the completed design being tested at the Vancouver Aquarium swim mill.  For this trial experiment the separation was created using a plywood box (Figure 1.8). The plywood box had only two sidewalls and a bottom and sat with one open end against the side of the swim mill and another open end at the downstream side. The baleen was suspended in the centre of this box such that when a hinged door at the front of the box was opened a tangential flow passed across the fringed medial edge of the baleen section,  33  while the lateral edge faced a stagnant pool of water, blocked by the remainder of the front panel of the plywood box (Figure 1.8).  A solitary trial with this experimental setup showed minimal, if any, cross-flow filtration occurring and failed to provide any quantitative data before the experiment was abruptly thwarted by some unforeseen circumstances.  Circular tank experiment Now running out of options for suitable facilities we came back to the Biological Sciences building at UBC. A circular tank in the courtyard of the Biological Sciences building was used to conduct the final filtration experiments. The tank was 183cm in diameter and 89 cm deep with a 7.30 cm inward taper from the top to the bottom of the tank. There was also a 10cm lip along the upper edge of the tank and a drain at the centre of the tank floor.  In order to again create two separate environments we used a vertical wall in which the baleen was inset and sealed around its edges. The vertical wall was placed in the tank and offset 30 cm from the centre of the tank in order to form a larger volume in which to create the required flow on the medial edge of the baleen section. The wall was created in two parts: the main stationary wall forming the left and right sides and a removable inset filling the gap in the centre. The wall’s frame was made of 1.5” stainless steel L beams onto which sheets of Acrylonitrile-butadiene-styrene (ABS) were riveted to form the outer stationary wall. The ABS panels were shaped to mimic the taper of the tank, but did  34  not form a perfect seal. Thus to improve the seal where the wall meets the side of the tank, additional panels of ABS were riveted to the edge of the wall such that when the wall was in place these additional sections pressed tightly against the sides of the tank and curved into the tank walls improving flow along the wall (Figure 1.9). The centre of the wall consisted of a removable inset into which the baleen could be fixed in place. The inset was the same height as the rest of the wall with a 25 cm high ABS panel above a rectangular opening for the baleen. The steel frame from the inset slid in between the steel frame from the main wall and was held in place using a simple counter-pivoting system with two C clamps.  Figure 1.9: Circular tank experimental setup including ABS wall with extension pieces conforming to tank walls, boat motor set up for crossflow filtration and baleen section in place.  Mounting the baleen into the wall inset required a manoeuvrable, yet stable, attachment that would allow the baleen to be fixed at various angles and orientations in the wall. First, a stainless steel plate was bolted and screwed to the base of the baleen section. The 35  baleen section was partially thawed to ensure a flat base on which to attach the plate. The plate, which covered the entire base of the baleen section, was then bolted into the baleen in four locations with 4” bolts and nuts being fastened to the bolts in between the plates until the nuts were flush with the top (ventral side) of the zwischensubstanz. To ensure a stable and reliable attachment, 3” screws also attached the plate to the base at 5 other locations (Figure 1.10). This provided a solid attachment for the baleen throughout the course of the experiments. Nearing the end of the experiments after six months of attachment, the plate was noticeably looser if the baleen was left to thaw during experiments for more than 3 hours or was left in water for more than an hour at a time. These were very rare occurrences and did not affect the experiment.  Figure 1.10: Baleen plate attachment using screws and bolts to attach the baleen to a steel plate which in turn was fixed to the wall inset with a manoeuvrable u-shaped rod.  36  With the baleen now securely attached to a metal plate, all that was needed was a manoeuvrable attachment to the wall inset. This was done using a ¼” stainless steel rod that was bent into a squared U-shape and attached to the L-beam of the inset frame and the steel plate of the baleen using metal braces. The rod slid into the braces and could be locked in position using bolts to hold the rod in place (Figure 1.10). This design allowed the baleen to be rotated about the dorso-medial edge of the baleen section to ensure that the medial side of the baleen was flush with the face of the wall and also to be translated along the length of the rod if minor adjustments were necessary. The wall was also designed to hold the baleen section such that the plates were horizontal in the water by simply attaching the U-shaped rod to the side of the inset’s frame in order to test the “centripetal filtration” hypothesis (discussed later). Thus the design was intended to address cross-flow filtration, dead-end filtration and centripetal filtration all through a simple manipulation of the baleen angle and orientation. However, due to other constraints of the experimental design, the centripetal flow scenario was not directly tested.  In an attempt to reduce any gaps around the baleen and force any water from leaking around the baleen an additional neoprene rubber seal was added to the design. A neoprene rubber sheet with a hole cut to the exact measurements of the baleen was attached to a frame and bolted to the wall inset frame so as to form a closer seal around the baleen section (Figure 1.11). The seal was not 100% effective as the shape of a baleen rack and the plates significantly limits the fit. The only way to form an ideal seal would be to mould an encasement to the specific section of baleen being used.  37  Figure 1.11: Wall inset with baleen in place showing the black neoprene rubber seal around the baleen section. The entire wall inset could then be placed in the wall (as in Figure 1.9) to perform tests.  Flow in the tank was created using a Minn Kota Endura 30 boat motor with a four-inch, two-bladed propeller. The motor was clamped to the edge of the tank from where the propeller could be rotated and locked to face any direction. The motor was powered by a twelve-volt DC power supply and had five forward speeds and three reverse speeds. In order to create the required flow velocities, only the highest speed was used during experimentation. This easy manipulation of the motor along with the simple adjustment of the baleen angle and orientation granted us the opportunity to test several flow directions and baleen orientations. Testing direct cross-flow and dead-end filtration as  38  well as hybrid angles of 30˚ and 60˚ in between the two could establish a more complete comprehension of this filtration process.  Measurement techniques  Figure 1.12: Anterior and medial views of the locations on the baleen section at which measurements were taken. Four locations on the medial side (A, B, C, D) and four corresponding locations on the lateral side (1,2,3,4) were used with arrows representing the direction in which flow was measured.  The quantitative essence of a filter is defined by the flow rate through the filter, the pressure difference across the filter and the volume of retentate captured by the membrane. In this scenario of presumed cross-flow filtration with no retentate on the filter’s surface, the focus is primarily on the flow rate across the baleen and the pressure difference over it. The primary method of flow rate measurement was a hand-held flow sensor from Höntzsch Instruments. The instrument consisted of a 55 cm long probe with a 23 mm diameter vane wheel that was connected to a hand-held electronic data logger. The data logger time-averaged readings over five seconds and displayed the averaged value on a digital display. Measurements were taken at four locations on the medial side of the baleen and at the corresponding four locations on the lateral side of the baleen 39  (Figure 1.12). On the medial side, cross-flow was defined as the velocity in the anteroposterior direction and the filtering rate was defined as the velocity in the medio-lateral direction.  Unfortunately, these measurements proved difficult, and additional techniques were required to quantify the flow through the baleen. Thus, we also attempted to visualise the flow using fluorescein. Acid yellow 73 fluorescein was injected into the flow stream immediately upstream from the baleen during cross-flow filtration using a wash bottle with a one-millimetre nozzle. The flow was also visualised using neutrally buoyant krill simulations such as gummi bears and small pieces of foam. The gummi bears were initially negatively buoyant and sank to the bottom of the tank. However, when they had remained in water for approximately thirty minutes, the gummi bears had absorbed enough water to become slightly more buoyant. They were not, unfortunately, buoyant enough to simulate the neutral buoyancy of krill at fin whale feeding depth and my dream of making this thesis about gummi bear filtration came to a disappointing end. Small pieces of bedding foam were much closer to neutral buoyancy being slightly positively buoyant. Sheets of foam were cut into approximately 8 cm3 cubes, too large to fit through the baleen plates, and then spray painted with a single coat of plasti-kote car paint to increase their mass and make them closer to neutrally buoyant. The air pockets inside the foam cubes were first squeezed out in a tub of water such that they were fully saturated with water and then inserted into the flow and allowed to circulate the tank for ten minutes to ensure complete saturation and consistent flow. The movement of the foam  40  was then observed both as a manner of observing the filtration mechanics and in diagnosing the experimental setup.  We also attempted to measure the pressure difference across the baleen using a differential pressure transducer designed for breathing measurements in turtles. However, due to an apparently low-pressure differential across the baleen that was too small to measure with our equipment we were unable to get successful results. We instead quantified the baleen’s resistance to flow without using the flow tank. Placing the baleen such that its medial side was facing directly upwards, a 7.5 cm diameter clear plastic tube was placed above the baleen with one end flush with the medial side of the baleen. The tube was filled with water to a height of thirty centimetres and the water then released and allowed to flow through the baleen. The time for the tube to empty through the baleen plates was measured using a stopwatch, but since the tube emptied virtually immediately, a camera was used to record the process and the time was measured from video analysis. The emptying rate of the tube was then compared to a tube emptying into ambient air.  41  Results Rather than immediately addressing the hypotheses by replicating the filtration process of a fin whale, this project proved also to be an exercise in understanding the complications of manipulating baleen and attempting to conduct an experiment on a large scale filter for which there is limited literature. As a result, I present here both observations on the experimental designs such that future experiments aimed at addressing the filtration mechanics of fin whales will not succumb to the same follies and also some quantitative results from the experiments themselves since this is in fact a Master’s thesis.  Observations from trial experiments Baleen curvature measurements acquired at Hvalur HF were expected to be used to recreate an accurate baleen model that accounted for the changing shape of baleen plates along a fin whale’s baleen rack. The results did show that there are significant changes in antero-posterior curvature along the length of a baleen rack (Figure 1.13). Maximum curvature is achieved at the middle of the rack with angles of almost ninety degrees while the anterior end has minimal curvature of only 20˚-30˚. The posterior end has essentially flat plates, even switching to anterior concavity. However, in the last 30-50 cm of the posterior end of a fin whale’s mouth the main plates of the laminae disappear somewhat leaving several minor plates forming the entire width of the laminae. As such, the “plate curvature” in this area is more subjective.  42  Figure 1.13: a) a simplified baleen plate curvature diagram defining the angles forming the antero-posterior curvature and b) the variation in degree of curvature over the length of a fin whale’s baleen rack measured at Hvalur HF.  43  Unfortunately, little was determined in terms of fluid dynamics from the first flume trial experiment at Bamfield Marine Science Centre. The large volume of water in the twelvemetre flow tank made flow visualisation difficult, and the rather haphazard setup made precisely controlling the experimental variables a trying experience. However, it can be noted that the shape of the baleen plates and their alignment in the rack alone did not induce any cross-flow filtration. This is important, as it confirms that cross-filtration in a fin whale would need to be a pressure and velocity driven process should it in fact be the mechanism whales are using.  Using actual baleen eliminated the potential issues of material property accuracy that would have arisen with a model, but it did introduce difficulties with preparation and experimental adaptability. For example, an important lesson was learned with regard to handling the baleen. Once unfrozen, the zwischensubstanz fails to retain its form and begins to flake away from in between the plates. Should the baleen remain unfrozen in flowing water for an extended period, the zwischensubstanz would eventually wear away until there was none remaining between the plates to hold them together. As my experiments were expected to require the baleen being immersed in water for lengthy periods at a time the zwischensubstanz was fixed in ethanol as previously described.  Trials with the Aquarium’s swim mill were also minimal, but they did provide a better understanding of working with baleen. The key observation was that it is incredibly difficult to form a seal around the baleen section. At the ventral tips of the baleen plates  44  there is little material with which to form a proper seal. In the whale’s mouth, the lower lips would lie tightly against the lateral edge of the plates with the lips and baleen racks moulded to create a reliable seal. In experimental practice, such a moulded seal is far more difficult to create as the millions of years necessary to evolve a perfect seal are not at our disposal. At the anterior and posterior ends of the baleen section the shape of the individual plates proved an especially difficult shape against which to form a seal. Ideally this experiment required the medial side of the baleen section to be flush with the wall in which it was placed. However, the medial side of the baleen section is not a flat surface (Figure 1.2) and so did not create that ideal scenario. With so much plate curvature, any seal formed around any edge of the baleen section needed to be flexible enough to conform to the shape of the baleen plates regardless of the direction of flow past or through the section. The neoprene rubber used in the circular tank experiments helped address this issue, but sealing would prove difficult throughout all the experiments.  Flow velocity measurements Filtration velocities were measured for both cross-flow filtration and dead-end filtration in the circular tank experiment. For cross-flow filtration, velocities were measured in three different directions. At a single motor speed, measurements were taken in the crossflow direction (antero-posterior) and the medio-lateral direction (through the baleen plates) on the medial edge of the baleen (Figure 1.12). At the corresponding locations along the baleen section on the lateral side, velocity measurements were taken in only the medio-lateral direction (Figure 1.12). Due to constraints of the measurement tools the medio-lateral flow velocities were not directly medio-lateral, but rather taken in a  45  direction perpendicular to the edge of the baleen (Figure 1.12). In other words, the flow probe was placed flush with the edge of the baleen plates when taking the measurement. During these measurements the tank was filled to a height of fifty centimetres, approximately five centimetres above the suspended baleen section, but twenty centimetres below the top of the wall.  Table 1.2: Velocity of water at the measurement locations defined in Figure 1.12 during cross-flow filtration  Location	
   A	
   B	
   C	
   D	
   1	
   2	
   3	
   4	
    Cross-­‐flow	
   speed	
  (m/s)	
   0.84	
   0.86	
   0.42	
   0.58	
   -­‐-­‐	
   -­‐-­‐	
   -­‐-­‐	
   -­‐-­‐	
    Flow	
  speed	
  thru	
   baleen	
  (m/s)	
   0.19	
   0.24	
   0.10	
   0.13	
   0.05	
   0.02	
   0.04	
   0.03	
    Cross-flow velocity varied markedly over the four locations tested even though the motor was kept at a constant speed, varying from 0.42 m·s-1 to 0.86 m·s-1 (Table 1.2). The corresponding flow velocities through the baleen were on average about 24% of the cross-flow velocities on the medial edge of the plate (Table 1.2). The medio-lateral velocities on the lateral side were about 24% of those on the medial side (Table 1.2).  Dead-end filtration velocity measurements were taken only in the medio-lateral direction on both sides of the baleen at the same four locations as the cross-flow measurements.  46  Widely ranging velocities were observed with velocity changes through the baleen ranging from an increase of 90% to a decrease of 89% at location 3 and location 4 respectively (Figure 1.12).  Table 1.3: Velocity of water at the measurement locations defined in Figure 1.12 during dead-end filtration  Flow	
  speed	
  thru	
   baleen	
  on	
  medial	
  side	
   (m·s-­‐1)	
  [locations	
  A,	
  B,	
   C,	
  D]	
   0.35	
   0.12	
   0.2	
   0.18	
    Flow	
  speed	
  thru	
   baleen	
  on	
  lateral	
  side	
   (m·s-­‐1)	
  [locations	
  1,	
  2,	
   3,	
  4]	
   0.35	
   0.05	
   0.38	
   0.02	
    Observations of circular tank experiment Through the dust particles that were unintentionally dispersed throughout the tank it was possible to decipher some fluid dynamics simply through empirical observations. One primary concern was the variability of flow speeds. With water filled to five centimetres above the height of the baleen and the motor set at maximum speed, the water was first allowed to get up to speed for several minutes before measurements were taken. After flow appeared to have reached some consistency, measurements were taken with the flow probe. However, the velocities throughout the tank were not as consistent as they looked. Velocities frequently fluctuated by more than 0.20 m·s-1 over the five second averaging interval of the flow probe. Burst flows were witnessed sporadically, but frequently enough not to be ignored, and were strong enough to prevent holding the flow probe perfectly steady. 47  Flows were also observed to be going around the baleen and around the wall. At the downstream end of the wall where a gap was created around the wall using a two foot long 2” x 4”, water was observed to be flowing from the medial side of the wall into the lateral side at an average rate of 0.14 m·s-1, whereas water simultaneously flowed from the lateral side to the medial side at the downstream end of the baleen section at an average rate of 0.05 m·s-1.  Figure 1.14: Baleen fringes are seen bent due to the flow of water near the dorsal edge of the baleen section, even after water has stopped flowing for thirty minutes.  Baleen fringes were observed to move with the flow adjacent to them. During cross-flow filtration the fringes were thus bent posteriorly during the flow much the same way that 48  hair would move in flowing water (Figure 1.14). When the flow was removed and the water settled until it was stagnant, the fringes remained bent for nearly thirty minutes before returning to their original position. This was especially noticeable at the dorsal end of the baleen plates where the fringes were most dense.  Baleen resistance measurements and observations Height of the water in the plastic tube and the distance the water level dropped were recorded using a metre stick, a Timex wristwatch and a Pentax Optio W90 digital camera. A segment over which the time and height were easily read was chosen as the measuring segment, covering as large a time segment as possible. During the control experiment where the tube was emptied directly into ambient air, the water travelled 30.5 cm in 0.20 s. In both tests through the baleen, the water travelled 28.0 cm in 0.19 s. Thus the respective average velocities of the control and baleen tests were 152.5 cm·s-1 and 147.4 cm·s-1 (Table 1.4). These tests were conducted without fine-scale accuracy, and it would be unwise to draw conclusions from such a small difference in velocities.  Table 1.4: Velocities of gravity-fed water through baleen used to help quantify baleen resistance.  Distance	
   Time	
   (cm)	
   (seconds)	
   	
  Control	
   30.5	
   0.20	
   Test	
  1	
   28.0	
   0.19	
   Test	
  2	
   28.0	
   0.19	
    Velocity	
   (cm·s-­‐1)	
   152.5	
   147.4	
   147.4	
    49  Discussion As with the presentation of results, the interpretations of these experiments can be broken down into discussing the experimental setup and addressing the initial hypothesis.  Experimental setup analysis Variable flow speeds in the circular tank setup were the result of a small motor with a large volume of water. Due to the limited availability of appropriate flow tanks, we were left with this circular tank even though it was not ideal. With the boat motor’s 10 cm propeller blades attempting to move a column of water of 50 cm, a flow gradient was unavoidable. With Reynolds numbers of 570 past the baleen fringes and 4500 past the baleen plates (Goldbogen et al. 2007), fin whales are not expected to have laminar flow inside their mouths during filtration, but for experimental purposes a more consistent flow would have improved measurement reliability.  The accuracy of the flow probe in this setup can also be questioned. When measuring the flow rate through the baleen during cross-flow filtration experiments the probe itself may obstruct the readings. The probe is intended to measure uniaxial flow directly through the vane wheel and not be placed in flow parallel to the vane wheel. Thus, measuring mediolateral velocities during cross-flow most likely resulted in skewed measurements. Ideally velocity measurements should be taken directly at the face of the baleen rack’s edge with a measurement tool small enough not to influence the fluid dynamics.  50  Filtration through a membrane is driven by both pressure and velocity. Theoretically, if one has a membrane with water flowing across it on one side and stagnant water of the same density on the other side, a differential pressure will be created across the membrane causing flow from the stagnant side to the flowing side according to Bernoulli’s law. In this experiment, the pressure differential was not expected to be large due to the low velocity of the flow relative to the area and depth of the baleen, but it was clear that the experiment was influenced by the pressure differential. Based on the minimal amounts of flow measured through the baleen during cross-flow experiments, we expected that the pressure gradient was the limiting factor. However, consistently measuring flows of 0.14 m·s-1 into the stagnant side at the downstream end of the wall implies that there were no such pressure limitations. Considering that flow was measured going both into and out of the stagnant water side and that the baleen resistance was measured to be negligible in separate experiments, we don’t believe that pressure was the limiting factor preventing cross-flow filtration.  Cross-flow filtration All this being said with regard to the experiment itself, it does appear that cross-flow filtration is not a likely mechanism of fin whale filtration. Although the flow velocity measurement accuracy can be questioned, the small percentage of flow measured through the baleen implies that cross-flow filtration would not be feasible to perform in the thirty to forty-five seconds over which whales filter (Goldbogen et al. 2007). With the flow velocity through the baleen being only approximately 25% of the cross-flow speed and the velocity exiting the baleen plates at the lateral edge being only 25% of that, a cross-  51  flow velocity of approximately 12.8 m·s-1 would be required to attain the 0.8 m·s-1 exit rate predicted by Goldbogen et al. (2007). Such a velocity would require enormous force generation by the VGB to create both the velocity and the increased pressure that would be required to create those higher velocities. A unidirectional flow moving anteriorly inside the oral cavity at this velocity is also unlikely, not only because of previous observations of rebounding waves (Kot 2005), but also because the scenario would call for too large a volume of water to not pass directly by the baleen, instead moving along the whale’s medial axis.  The location of the entire slug of water inside the whale’s mouth must be taken into account. Cross-flow filtration requires the filtrate to be in close proximity to the filter as it flows parallel to the filter membrane, but a fin whale’s oral cavity does not necessarily promote this flow as the width of the cavity can be several metres wide and the only way of directing water towards the baleen is with the vomer or the tongue. Thus, in order to create cross-flow filtration the VGB would be required to contract relatively quicker along the whale’s medial axis in order to force the engulfed slug of water to flow directly along the baleen. There is neither observational nor anatomic evidence that this is indeed the case, making it an argument against cross-flow filtration (at least in this capacity).  It was also noted during experiments that a cross-flow across the baleen fringes caused the fringes to be bent and remain so for about thirty minutes after the flow was eliminated. This can also be considered evidence against cross-flow filtration in that fin whales spend almost all the daylight hours feeding (Simon et al. 2010), and were they to  52  use cross-flow filtration and bend their baleen fringes on a daily basis, one would expect that the fringes would start to fall naturally in this direction. This is not seen in fin whale carcasses nor was it observed at Hvalur HF. This does not eliminate the possibility of cross-flow filtration as it has been suggested that water rebounds forwards and backwards inside the whale’s mouth (Kot 2005), which may prevent any permanent bending in one direction, but it does create some doubt of the theory.  Centripetal filtration When it started to become evident that cross-flow filtration was an unlikely mechanism of filtration, we began considering other mechanisms of filtration. The most prominent new proposal is one of centripetal filtration. If we consider the oral cavity of a fin whale after engulfment when its mouth is closed and full of water, we can simplify the fluid dynamics by considering the entire engulfed volume to be a single slug of water. Thus, as the VGB starts to contract, this slug of water has only one way to exit the oral cavity and that is through the baleen. However, the engulfed volume of water extends inside the oral cavity all the way to the umbilicus, two-thirds of the whale’s length, while the baleen extends only to the end of the jaw, less than one-third of the whale’s body length. Thus, in order for that slug of water to exit through the baleen, its general direction would need to be anterior and dorsal inside the oral cavity over the duration of the filtration process. With this general direction of flow, the water would be flowing towards the roof of the mouth, almost directly at the vomer. It is feasible then that the vomer would passively deflect water to the left and right hand sides of the oral cavity in the direction of the  53  baleen and create two separate centripetal flow patterns on each side of the mouth (Figure 1.15).  Figure 1.15: Diagram depicting the cross-section of a fin whale’s mouth. The thick blue arrows represent the flow of water inside the mouth during theoretical centripetal filtration and the thin blue arrows represent water exiting through the baleen. Water would either flow ventrally along the edge of the baleen plates or flow at some acute angle (ø) into the baleen plates.  At the interface with the baleen, the water may flow either along the medial edge of the baleen or potentially at some angle (ø) into the baleen (Figure 1.15). If the water flows directly along the edge of the baleen fringes, we are effectively looking at a cross-flow filter with flow still running parallel to the filter surface, but now orthogonal to the original theory. If the water is flowing through the baleen at some angle (ø) then the filtration type can be considered a hybrid between dead-end and cross-flow filtration. The anterior momentum of the water may add another twist to the flow, creating a corkscrew spiral towards the anterior of the oral cavity on the left and right hand sides of the mouth. During this theoretical centripetal filtration, the krill would be expected to take one of 54  two paths: stay with the flow of the water and come into contact with the baleen or, depending on the velocity and the centripetal motion, end up in a suspension at the central axis of each left and right centrifuge in the whale’s mouth and never come into contact with the baleen. This would only occur if, at some point during filtration, the pressure gradient acting on the krill within these centrifuges was large enough to overcome the centrifugal force of the vortex (Julien 1986). In turn, this would only occur if the whale applied enormous amounts of pressure to the vortex or changed its height in the water column by a substantial amount, both of which are unlikely. The more likely scenario is that the krill are spun around the centrifuge and come into contact with the baleen as the vortex flows past the fringes.  Baleen resistance The results of the baleen resistance testing should be viewed with caution. The intent of the experiment was intended to very loosely quantify the resistance of baleen. As a result, the accuracy of the measurements was on par with the intent. Regardless of the fine scale accuracy of the experiment, it did make evident that baleen offers very little resistance to water flow. A baleen rack with low resistance would clearly minimize the amount of force required from the VGB to push water through the baleen as the pressure gradient across the baleen would also be low. The exiting flow rate predicted by Goldbogen et al. (2007) also becomes more reasonable with lower baleen resistance than expected since low resistance begets higher flow rates.  55  Chapter Two Towards thee I roll, thou all-destroying but unconquering whale; to the last I grapple with thee; from hell’s heart I stab at thee; for hate’s sake I spit my last breath at thee. Sink all coffins and all hearses to one common pool! and since neither can be mine, let me then tow to pieces, while still chasing thee, though tied to thee, thou damned whale! Thus, I give up the spear! -Herman Melville from Moby Dick  Introduction The oral cavity of a mysticete is truly a place of mystery. Not only is the mouth of a fin whale an area rarely seen in a living animal, but the scientific literature provides very little analysis of this most crucial area of anatomy. It is, after all, the oral cavity and its many structures that allow rorquals to efficiently lunge-feed, thus providing a significant contribution to their enormous size (Goldbogen et al. 2011). One of these structures is termed the zwischensubstanz (German for in-between substance) (Tullberg 1883 via Fudge et al. 2009) because it lies in between the baleen plates at the (dorsal) base of the plates, fixing them to the rostrum. In a basic sense, the zwischensubstanz-to-baleen relationship is similar to that of gums-to-teeth in humans in that both the zwischensubstanz and gums attach the maxillary of the rostrum (via connective tissue) to the embedded baleen plates. As a result, the main function of the zwischensubstanz has been assumed to be to space the baleen plates and provide the foundation from which the plates can grow. 56  The material properties of zwischensubstanz have only been addressed once in any substantial way in the literature and that was by the man who came up with the term “zwischensubstanz”, namely Tycho Tullberg (1883). Thus we attempt to perform some preliminary analysis of zwischensubstanz in this feeding-centric thesis for two main reasons. First, one of the functions of zwischensubstanz is to maintain the spacing of the baleen plates, and in this way (however distant) it is a contributing factor to feeding. In fact, the zwischensubstanz analysis could provide confirmation or rejection that baleen plates are moving or vibrating during filtration based on their elasticity as proposed by Werth (2011). Second, since a few years have passed since 1883 and the literature on the subject has progressed so little (read: not at all), there is much to contribute being as we are in possession of fin whale zwischensubstanz, even if it is only a preliminary analysis of its material properties.  Background Tullberg (1883) took the first steps in understanding the properties of baleen and the zwischensubstanz from which it grows. By examining the morphology and development of baleen from embryonic blue whales at a remarkable level of detail considering the era during which he lived, Tullberg established a few basic facts about zwischensubstanz from which we can start our analysis. The most useful information that Tullberg passed on regarding zwischensubstanz was in the form of his hand-drawn schematics that summed up with fine detail of the many observations that he made. Tullberg showed that the zwischensubstanz in fact has no distinct stratum corneum, but that it is possible to  57  differentiate between an inner mucous layer and an outer more keratinized layer. He also states that the “keratinized part of the covering layer [of the baleen] gradually thins until it disappears entirely at the base of the connective tissue plate”. With these two observations, although without explicitly stating it, Tullberg has implied that the keratin layer of the baleen plates is formed in the zwischensubstanz. In fact, the schematics and description of the zwischensubstanz also led the translators (Fudge et al. 2009), to remark that zwischensubstanz appears to be a remarkable soft α-keratin that consists of many more cell layers than the typical soft α-keratin of mammalian skin, implying that zwischensubstanz could be the source of keratin that forms the baleen plate’s hard matrix.  This concept of zwischensubstanz being a soft keratin is an intriguing one. Keratin is a scleroprotein that often functions to structurally support or protect epithelial cells from mechanical and non-mechanical stresses (Coulombe and Omary 2002). The baleen plates of all mysticetes are keratinized and exhibit varying levels of tensile strength across species (Szewciw et al. 2010). Thus if baleen is formed in the zwischensubstanz, it follows that the zwischensubstanz is the source of the keratin. The mechanical properties of zwischensubstanz have never been analyzed, but judging by the analysis of Tullberg (1883) and Fudge et al. (2009) it appears that we should expect some anisotropic behavior in which stiffness is higher along one axis as the baleen plate development is uniaxial in the dorso-ventral direction within the zwischensubstanz. However, to the touch, zwischensubstanz feels like a rubbery substance, with no obvious anisotropy.  58  Ranging in dorso-ventral thickness along the baleen rack, the zwischensubstanz in a fin whale is approximately 5 cm high on average. In between each plate the antero-posterior thickness of zwischensubstanz remains approximately constant throughout the baleen rack at 0.6 cm (Goldbogen et al. 2007), as it serves to create and maintain the gap between plates. The white colour of the zwischensubstanz also remains constant throughout the baleen rack, seemingly unaffected by the variations in pigmentation of the baleen plates. A closer look at the material properties and structure of the zwischensubstanz may elucidate the formation of baleen plates and give us some insight into its involvement in the feeding process.  The focus of a material property analysis begins with two key factors: Young’s Modulus (E) and hysteresis. These two factors help to define a material’s responses to stress by quantifying their stiffness and visco-elastic energy retention respectively. The elastic modulus of zwischensubstanz quantifies how much the baleen plates can move in response to water flow or other stresses that they encounter and to what extent they retain their formation. Hysteresis defines how much energy is lost to heat when zwischensubstanz extends or contracts in response to stresses and as such is a basic method of defining a material’s visco-elasticity.  Baleen formation is a process that remains unknown due to the limited access to the developing stages of baleen that scientists have had to endure throughout history. However, based on observations from whales captured at Hvalur HF, we can at least form some hypotheses on the matter of baleen growth, if not its development.  59  Hypothesis Figure 2.1: Baleen from a fin whale has been sliced to show the papillae (red tissue strands) entering the dorsal base of the white zwischensubstanz (a). In (b) one can see the papillae entering long holes into the zwischensubstanz. The papillae appear longer and thinner than they would in vivo as they have been stretched when rotating the baleen section.  Examining a section of baleen and pulling the zwischensubstanz away from the connective tissue that attaches it to the rostrum reveals Figure 2.1. The thin strands of  60  pink tissue in this picture are termed the papillae and appear to penetrate into the zwischensubstanz through aligned holes. These holes can be traced to the ventral surface of the zwischensubstanz where the holes are instead replaced by baleen plates of the same medio-lateral width as the holes. This leads us to the obvious conclusion that the papillae are the origins of the tubules that are eventually surrounded by a keratin matrix and form the baleen plates (Szewciw et al. 2010). The microscopic analysis presented here of zwischensubstanz and the baleen plates forming within it are intended to clarify the process of baleen plate formation and test the hypothesis that tissue from the zwischensubstanz forms the keratinous matrix of baleen plates. The material property analysis simply aims to gain a quantitative understanding of the mechanical properties of zwischensubstanz. 	
    61  Materials and Methods Compression and tensile tests The baleen from an approximately 18m fin whale was collected from Hvalur HF in July 2010. A 100 cm section from the middle of the right baleen rack was frozen and sent back to UBC campus in Vancouver. At UBC the section has remained frozen at -20˚C except when partially thawed for sample collection. From this section several pieces were removed for sampling using a sharp knife to cut the zwischensubstanz from the baleen section. Strips of zwischensubstanz were cut from the dorsal base at the posterior end of the baleen section at both the medial end and the centre of the section (Figure 2.2). These zwischensubstanz pieces were then cut into approximately 1.0 cm3 pieces to perform compression tests. All pieces that were cut from the baleen section were stored in water at 4˚C prior to, during and in between tests. The zwischensubstanz pieces did not appear to degrade or deform during the entire process.  Figure 2.2: a) Base of the baleen section sampled for its zwischensubstanz showing the location of sampling for compression tests and (b) a close-up of the sampling location showing the size of the pieces removed.  62  Compression testing was performed with an MTS 858 Mini Bionix (Figure 2.3). Samples were individually placed between two flat plate grips. Both the upper and lower grip surfaces were made of fine sandpaper to prevent slipping. The upper grip was attached to a 50 N load cell and data was recorded using the MTS’ associated software Teststar IIs Station Manager Version 3.5C 1817.  Figure 2.3: Compression testing setup with the MTS 858 Mini Bionix. The white zwishchensubstanz cube is labeled for posterior (P), medial (+) and ventral (-) surfaces to facilitate testing.  Specimens were first compressed uniaxially in a sinusoidal manner arbitrarily at 0.5 Hz to a maximum displacement of 1.75 mm for 20 cycles with force, time and displacement recorded at 100 Hz. An initial offset of 3.00 N of compression defined the starting point for each test in order to create consistency between trials. This cyclic testing was performed on each specimen along all three axes (dorso-ventral, medio-lateral and antero-posterior) in the same manner. The compression did not constitute yielding of the 63  material, but showed significant bulging (increase in cross-sectional area) at compressive strains of approximately 20% in part due to the high friction nature of the sandpaper grips.  Strips of zwischensubstanz were obtained in a similar manner to be used in tensile tests. Samples were cut from the posterior end of the baleen section and stored in water at the same 4˚C. Test specimens were cut from the removed samples into approximately 35-45 mm long strips with 1.0 cm2 cross-sectional areas. Tensile tests were also performed with the MTS 858 Mini Bionix with a 50 N load cell and stainless steel grooved, clamping grips. Specimens were removed from the 4˚C water and the ends were dried using a paper towel to reduce slippage in the clamps. The specimen was clamped at a sufficient pressure to eliminate slippage without affecting the zwischensubstanz integrity and preloaded to a 1.00 N tension to eliminate slack. Specimens were then pulled in a sinusoidal fashion arbitrarily at 1.0 Hz to a maximum displacement of 5.00 mm for 20 cycles with force, time and displacement recorded at 100 Hz. Immediately following the cyclic tests, without removing the specimen from the MTS, a yield test was performed. Specimens were stretched at 2.00 mm·s-1 until they broke. On three of the specimens cyclic data was collected, but they were not pulled to their breaking point due to constant slippage in the clamps that lead to a degradation of the specimens and prevented accurate data from being collected from them thereafter.  64  Young’s modulus and hysteresis calculations There were two properties of interest calculated from the compression and tension tests: Young’s modulus (E) and hysteresis. As zwischensubstanz is a keratinized material, we were expecting a standard keratin stress-strain curve with three linear phases: hookean, yield and post-yield regions (Figure 2.4). Each of these regions has a distinct Young’s modulus that defines the material at various strains. Young’s modulus is defined as   !=  !!! !! ∆!  where F is the axial force, Lo and ∆L are the initial and change in length respectively and Ao is the initial cross-sectional area. It was calculated by plotting stress versus strain for each specimen and manually calculating the slope in each of the three regions.  Figure 2.4: Typical Stress versus Strain curve for a hydrated hard α-keratin with approximately 1000x the stiffness of zwischensubstanz depicting the three distinct phases: Hookean, Yield and Post-Yield (adapted from Fudge 2002)  65  Hysteresis was calculated by using more primitive methods than one would expect in today’s tech-friendly world. For each cyclic test, force versus displacement curves were plotted on a standard sheet of white paper. The areas between the loading and unloading curve segments (Figure 2.5 red) and below the unloading curve (Figure 2.5 blue) were cut out from the sheet with an exacto knife and the respective weights of the paper segments were measured with a Mettler PM460 DeltaRange scale. Hysteresis was then defined as the ratio of the mass of the area between the curves (red area) to the total mass under the loading curve (red plus blue areas) (Figure 2.5).  Figure 2.5: A hypothetical Force vs. Displacement diagram explaining the method used for hysteresis calculation. The black line represents the loading and unloading curve of a specimen, the red area is the energy lost to heat and the blue area is the energy retained by the specimen. Hysteresis is then defined as: Red/(Red+Blue).  66  Microscopy Samples for microscopic analysis were taken from the same posterior end of the baleen section. Samples were taken at the dorsal and ventral ends of the zwischensubstanz as well as the middle. Figure 2.6 shows the seven locations that were analysed: three samples from the centre of the zwischensubstanz (once in each direction; slices 2, 3 and 6 in the antero-posterior, dorso-ventral and medio-lateral directions respectively), one medio-lateral slice at each of the ventral and dorsal ends of the zwischensubstanz (slices 5 and 7 respectively) and one dorso-ventral slice at each of the ventral and dorsal ends of the zwischensubstanz (slices 1 and 4 respectively).  Figure 2.6: Simplified diagram of a baleen section showing a lateral and posterior view of the seven microscopy locations. The numbered red bars represent the edges of the slices sampled. Slices 1, 4, 5 and 7 were taken through the baleen plates while slices 2, 3 and 6 were taken only through the zwischensubstanz.  67  Forty micron slices were then cut at these different locations with a Sartorius Werke AG microtome with a Feather C35 carbon steel blade. The slices were placed immediately on slides and viewed with a Leitz Orthoplan polarizing microscope with crossed polarizers. The use of polarized light allowed for fibre orientations to be visualized as the light passing through birefringent materials was refracted so as to elucidate its orientation. The analyzer and polarizer were consistently oriented in the same direction throughout the experiments (Figure 2.7). A first order red filter was used to intensify birefringence and highlight areas of interest in the zwischensubstanz. With consistent orientations of the analyzer and polarizer, birefringent materials appeared blue when oriented along the blue axis and yellow when oriented along the yellow axis while non-birefringent materials appeared pink when viewed with the first order red filter (Figure 2.7). Images were captured with a colour Q Imaging Micropublisher 3.3 RTV camera and its associated Q Capture 2.90.1 software. Figure 2.7: Sample microscope image viewed with a first order red filter displaying the orientation of the Polarizer and Analyser used for all images. Fibres oriented along the blue axis appear blue and fibres along the yellow axis appear yellow. The pink areas represent areas that lack birefringence.  68  Results Compression and tension tests Compression test results were analysed by plotting both Force v. Displacement and Stress v. Strain curves for each axis of compression (Figures 2.8-2.10). From these plots, hysteresis and Young’s modulus were calculated respectively (Table 2.1). Compressive tests did not display the typical keratin curves (Hookean, yield and post-yield regions), instead displaying only one distinct slope throughout all strains tested. The loading and unloading curves were consistent from cycle to cycle with no evidence of property changes during the testing. Young’s Modulus was not significantly different between directions with Moduli (± standard deviation) of 2.11 ± 0.44 MPa, 2.64 ± 0.11 MPa and 2.93 ± 0.72 MPa in the medio-lateral, antero-posterior and dorso-ventral directions respectively (Table 2.1). Hysteresis was also similar in all directions at approximately 44% (Table 2.1).  69  40	
   35	
    Force	
  (N)	
    30	
   25	
   20	
   15	
   10	
   5	
   0	
    a  0	
    0.25	
    0.5	
    0.75	
    1	
    1.25	
    1.5	
    1.75	
    2	
    Displacement	
  (mm)	
   0.6	
    Stress	
  (MPa)	
    0.5	
   0.4	
   0.3	
   0.2	
   0.1	
   0	
    b  0	
    0.05	
    0.1	
    0.15	
    0.2	
    0.25	
    Strain	
    Figure 2.8: One cycle from a cyclic sinusoidal compression loading of zwischensubstanz in the antero-posterior direction where (a) the force vs. displacement plot was used to calculate hysteresis and (b) the stress vs. strain plot used to calculate Young’s Modulus  70  25	
    Force	
  (N)	
    20	
   15	
   10	
   5	
   0	
   0	
    a  0.25	
    0.5	
    0.75	
    1	
    1.25	
    1.5	
    1.75	
    2	
    Displacement	
  (mm)	
   0.35	
    Stress	
  (MPa)	
    0.3	
   0.25	
   0.2	
   0.15	
   0.1	
   0.05	
   0	
    b  0	
    0.05	
    0.1	
    0.15	
    0.2	
    Strain	
    Figure 2.9: (a) Force versus displacement and (b) Stress versus strain curves from one cycle of a cyclic sinusoidal compression loading of zwischensubstanz in the mediolateral direction.  71  35	
   30	
    Force	
  (N)	
    25	
   20	
   15	
   10	
   5	
   0	
    a  0	
    0.25	
    0.5	
    0.75	
    1	
    1.25	
    1.5	
    1.75	
    2	
    Displacement	
  (mm)	
   0.6	
    Stress	
  (MPa)	
    0.5	
   0.4	
   0.3	
   0.2	
   0.1	
    b  0	
   0	
    0.05	
    0.1	
    0.15	
    0.2	
    Strain	
    Figure 2.10: (a) Force versus displacement and (b) Stress versus strain curves from one cycle of a cyclic sinusoidal compression loading of zwischensubstanz in the dorsoventral direction.  72  Tensile tests did display the distinct regions (Figure 2.4) of Young’s modulus for a keratinized material (Table 2.2). In the dorso-ventral direction average Young’s modulus in the Hookean, yield and post-yield regions were 2.91 ± 0.43 MPa, 0.79 ± 0.07 MPa and 1.22 ± 0.04 MPa respectively (Figure 2.11).  Table 2.1: Young’s Modulus and hysteresis measured from compression tests showed little variation between directions (n=5)  Direction	
  of	
  compression	
   Medio-­‐lateral	
   Antero-­‐posterior	
   Dorso-­‐ventral	
    Young's	
  Modulus	
  (MPa)	
   2.11	
  ±	
  0.44	
   2.64	
  ±	
  0.11	
   2.93	
  ±	
  0.72	
    Hysteresis	
   0.45	
  ±	
  0.03	
   0.44	
  ±	
  0.03	
   0.44	
  ±	
  0.02	
    In the medio-lateral direction Young’s modulus could only be calculated for one specimen due to limited clamping ability with the MTS setup. In the Hookean, yield and post-yield regions the Young’s moduli were 4.30 MPa, 1.08 MPa and 1.38 MPa respectively in the medio-lateral direction (Figure 2.12).  73  20	
   18	
   16	
   14	
    Force	
  (N)	
    12	
   10	
   8	
   6	
   4	
   2	
   0	
   0.0	
    a  5.0	
    10.0	
    15.0	
    20.0	
    25.0	
    30.0	
    35.0	
    Displacement	
  (mm)	
   0.8	
   0.7	
    Stress	
  (MPa)	
    0.6	
   0.5	
   0.4	
   0.3	
   0.2	
   0.1	
    b  0	
   0.00	
    0.10	
    0.20	
    0.30	
    0.40	
    0.50	
    0.60	
    0.70	
    0.80	
    Strain	
    Figure 2.11: (a) Force versus displacement and (b) stress versus strain curves for the tensile loading of a zwischensubstanz specimen in the dorso-ventral direction. The curve ends where the specimen broke at a strain of approximately 0.68 (b)  74  16	
   14	
   12	
    Force	
  (N)	
    10	
   8	
   6	
   4	
   2	
   0	
   0.0	
    a  2.0	
    4.0	
    -­‐2	
    6.0	
    8.0	
    10.0	
    12.0	
    14.0	
    16.0	
    18.0	
    Displacement	
  (mm)	
    1	
   0.9	
   0.8	
    Stress	
  (MPa)	
    0.7	
   0.6	
   0.5	
   0.4	
   0.3	
   0.2	
   0.1	
    b  0	
   0.00	
   -­‐0.1	
    0.10	
    0.20	
    0.30	
    0.40	
    0.50	
    0.60	
    0.70	
    0.80	
    Strain	
    Figure 2.12: (a) Force versus displacement and (b) stress versus strain curves for the tensile loading of a zwischensubstanz specimen in the medio-lateral direction. The curve ends where the specimen broke at a strain of approximately 0.71 (b)  75  Table 2.2: Young’s Modulus in the hookean, yield and post-yield regions and hysteresis measured from the tensile tests. (* denotes that there was only a single test performed and therefore no associated statistics; all other results n=4)  Direction	
  of	
  tension	
   Dorso-­‐ventral	
   Medio-­‐lateral	
    Young's	
  Modulus	
  (MPa)	
   Hookean	
   Yield	
   Post-­‐yield	
   2.91	
  ±	
  0.43	
   0.79	
  ±	
  0.07	
   1.22	
  ±	
  0.04	
   4.30*	
   1.08*	
   1.38*	
    Hysteresis	
   0.44	
  ±	
  0.01	
   0.36	
  ±	
  0.10	
    The size of each region was fairly consistent across all tensile tests with the Hookean region from approximately zero to 3% strain, the yield region from 3-50% strain and the post-yield region at greater than 50% strain. The breaking point for most specimens was between 60-70% strain at stresses ranging widely between 0.40-0.85 MPa. Hysteresis did not vary significantly between the two tensile directions at 0.43 ± 0.01 J and 0.36 ± 0.10 J in the dorso-ventral and medio-lateral directions respectively.  Microscopy observations In the antero-posterior slice (slice 2 from Figure 2.6) through the zwischensubstanz alone there are distinct dorso-ventrally oriented tubules that appear as long illuminated formations in Figure 2.13 with the same formations appearing blue under a first order red filter in Figure 2.14. These formations ranged from about 90 to 250 µm at their widest points and were 1.0 to 2.0 mm long. In the brightly illuminated tubule areas the fibres were oriented approximately 40˚ off-axis medio-laterally with no such fibres evident in the darker sections. When the direction of the analyzer-polarizer was rotated by 45˚ the dark sections remained dark showing no evidence of birefringence and implying areas of less oriented structure. These bright fibres are slightly birefringent and also appeared to be in a similar arrangement in the medio-lateral plane suggesting that they are oriented  76  circumferentially about the same dorso-ventral axis rather than in a single plane. Thus, the medio-lateral plane (not shown here) is similar to Figure 2.14. Viewing the dorsoventral plane shows the circumferentially-oriented fibres more clearly (Figure 2.15 and 2.16; slice 3 from Figure 2.6). Dark, circular 25-40 µm holes appear in these slices as tubules penetrating the viewing plane surrounded by concentric, circular fibres creating a hatch-like pattern. The radius of the hatch-like pattern is 40-80 µm corresponding to the size of the tubule that it surrounded thus creating gaps of about 80 – 160 µm between tubules. The hatch pattern is also the result of the birefringent properties of zwischensubstanz since the slice can be rotated under the microscope without substantially changing the orientation of the dark bands that make up the hatch pattern (Figure 2.15). Under the first order red filter the zwischensubstanz showed unidirectional fibre orientation between the tubules implying a single direction of growth and anisotropic behaviour (Figure 2.16).  77  Figure 2.13: Antero-posterior slice of zwischensubstanz viewed at 3.2x magnification with tubules oriented East-West (dorso-ventrally) in the image and fibres oriented diagonally in the bright areas. (slice 2 in Figure 2.6)  Figure 2.14: The same antero-posterior zwischensubstanz slice from Figure 2.12 viewed at 3.2x magnification with a first order red filter. (slice 2 in Figure 2.6)  78  Figure 2.15: Dorso-ventral slice of zwischensubstanz showing the hatch pattern visible around dark circular tubule ends viewed at 10x magnification. (slice 3 in Figure 2.6)  79  Figure 2.16: Dorso-ventral slice of zwischensubstanz at 10x magnification with a first order red filter showing some blue birefringence of fibres between tubules oriented unidirectionally. (slice 3 in Figure 2.6)  80  Figure 2.17: Medio-lateral slice of the baleen plate (with antero-posterior axis oriented East-West) embedded in zwischensubstanz near the ventral edge of the zwischensubstanz when the plate’s hard matrix has yet to form. The white strands on the left are the papillae of the inner volume of the baleen plates, the highly oriented vertical fibres in the middle form one outer edge of plate (the other edge is not in the image) and the fibres oriented at 45˚ on the right form the zwischensubstanz. (slice 7 in Figure 2.6)  81  Figure 2.18: Medio-lateral slice of the baleen plate embedded in zwischensubstanz near the dorsal surface with a first order red filter showing the matrix formed around the oval shaped slice through the papillae (and antero-posterior axis oriented somewhat NorthSouth). The linearly oriented fibres at the top of the image are the outer edge of a baleen plate whereas the fibres on the inside of the plate are oriented around the oval slice through a papilla. The sharp edge on the right is the edge of the specimen. (slice 5 in Figure 2.6)  Within the zwischensubstanz the baleen plates begin to form right from the dorsal base of the zwischensubstanz (Figures 2.17 and 2.18; slices 7 and 5 from Figure 2.6). The baleen structure appears to have two distinct parts at this stage of development: a highly oriented fibrous layer on the outside of the papillae and an inner less structured layer (Figure 2.17). The outer edge is highly oriented in the dorso-ventral direction, exhibiting very distinct orientation whereas the inner layer, although also oriented dorso-ventrally, appears less structured and wavier. Fibres within the zwischensubstanz are oriented at 45˚ 82  from the antero-posterior axis. Most of the black gaps appearing in Figure 2.17 are a result of slicing the zwischensubstanz on the microtome and not actually part of the structure of the papillae. During slicing, the inner layer of the baleen plate was easily separated from the outer edges while the outer edges, remained intact with the adjacent zwischensubstanz.  This was not the same situation near the ventral surface of the zwischensubstanz where the baleen plate (that the dermal papillae had now formed) remained intact during slicing. In fact, the two distinct layers that formed the papillae at the dorsal end of the zwischensubstanz were no longer distinct at the ventral end (Figure 2.18). The highly structured outer edges adjacent to the papillae remained highly birefringent on the outer edges, but now also filled in the spaces directly adjacent to the papillae holes. Instead of a wavy inner structure to the baleen plates there appeared 550 µm long oval holes where papillae had been sliced at an acute angle while now surrounded by a solid matrix. Ovals viewed with the first order red filter displayed a ring of blue hued fibres inside a ring of yellow-orange fibres (Figure 2.18) implying an inner layer to the papillae matrix oriented at less than ninety degrees to its outer matrix at the ventral end of the zwischensubstanz.  83  Figure 2.19: Dorso-ventral slice of outer edge of baleen plate (fibres oriented East-West) embedded in zwischensubstanz (hatched area) near the ventral surface of the zwischensubstanz with the antero-posterior axis oriented somewhat North-South. (slice 4 in Figure 2.6)  84  Figure: 2.20: Dorso-ventral slice of a distorted baleen plate near the dorsal base of the zwischensubstanz with highly oriented outer edges and large circular papillae distorted by the cutting procedure. In vivo, the outer edges run parallel to each other and the papillae in between are presumably circular and continuously unbroken. (slice 4 in Figure 2.6)  85  Figure 2.21: Dorso-ventral slice of papillae within fibre matrix oriented circularly around each papillae near the ventral surface of the zwischensubstanz viewed with a first order red filter and antero-posterior axis oriented North-South. (slice 1 in Figure 2.6)  86  Figure 2.22: Dorso-ventral slice of a baleen plate near the ventral surface of the zwischensubstanz showing circular papillae surrounded by keratin matrix and highly oriented outer edges of the baleen plate (antero-posterior axis is oriented Northwest to Southeast in the image). The bright areas are locations exhibiting high birefringence. (slice 1 in Figure 2.6)  87  Figures 2.23: Dorso-ventral slice of a baleen plate near the ventral surface of the zwischensubstanz at 10x magnification showing the outer layer of the baleen plate (at the bottom of the image) forming the keratin matrix concentrically around each papilla. (slice 1 in Figure 2.6)  Observing a baleen plate in the dorso-ventral plane at the dorsal end of the zwischensubstanz (slices 4 from Figure 2.6) again shows highly oriented, birefringent outer edges oriented medio-laterally (Figure 2.19). The inner portion of the baleen plate consists of large (500-550 µm) hollow circles formed by thin (15-50 µm) bands of fibres (Figure 2.20). This inner portion is easily disturbed, and Figure 2.20 shows circles that have been broken and shifted during slicing. A dorso-ventral view near the ventral surface of the zwischensubstanz shows an all-together different inner portion (Figures 2.21-2.23; slices 1 and 5 from Figure 2.6). Where the zwischensubstanz and the outer 88  edges of the baleen plate appear to remain consistent from the dorsal to ventral ends of the zwischensubstanz, the inner portion of the plate has formed a more solid structure as the baleen plate is closer to being fully formed. The papillae holes are now only 85-350 µm in diameter (as opposed to 500-550 µm at the dorsal end of the zwischensubstanz) as each one is surrounded by concentrically oriented fibres around each hole with radii equal to each respective hole’s diameter (Figure 2.21). A closer look at the attachment of the inner to outer portions of the baleen plate reveals that the highly-oriented outer edges are indeed forming the concentric matrix about the inner circles as there is no longer a distinction between the inner and outer portions of the plate (Figure 2.23).  89  Discussion When analysing the material properties of zwischensubstanz one has to consider foremost the function of zwischensubstanz. It can be assumed from previous literature and from observation that the main purpose of zwischensubstanz is to form a solid base from which baleen can grow and be held in place (Tullberg 1883). Because of this location within the oral cavity, the zwischensubstanz is also presumably not subject to any large direct forces. It is essentially protected dorsally by the maxillary bone (and connective tissue) to which it attaches and anteriorly and posteriorly by the baleen plates that develop from within the zwischensubstanz itself. Its main source of stress may actually be caused by the baleen since the baleen plates face several different forces throughout the whale’s activities of daily living. Each baleen plate is embedded in the zwischensubstanz as its only means of support, thus creating a cantilevered plate in which forces will be translated from the plates to the zwischensubstanz, creating areas of high stress at the point where the plate leaves the zwischensubstanz. The forces acting on baleen and their effects on zwischensubstanz have not been examined in the literature, but one can speculate as to potential sources of stress by analysing the behaviours of lunge-feeding whales.  Lunge-feeding constitutes a large portion of a whale’s time, and during this process the baleen racks may face several different forces. One large force would be incoming water flow when the whale’s mouth is agape. Upon opening its mouth, a burst of water will hit the anterior face of the whale’s left and right baleen racks forcing them to deflect backwards. The zwischensubstanz is fixed to the rostrum not only dorsally but also at the 90  anterior and posterior-most ends of the baleen racks. As such, when a force acts on the anterior face of a plate, the zwischensubstanz undergoes compression immediately posterior to the plate and tension immediately anterior to the plate at its ventral surface. As there is a series of baleen plates, each of these forces will interact with the adjacent plates’ forces. However, since the anterior-most plates will be subject to greater forces, the zwischensubstanz will most likely experience compression throughout its entire ventral surface. In this way, the compressive stiffness of the zwischensubstanz can aid in reducing pressure drag since, to some extent, the baleen plates can adjust (by bending) to the sudden burst of water pressure caused by mouth opening at high speeds. The Young’s Moduli exhibited by the zwischensubstanz in all three directions (Table 2.1) were similar to those of a soft rubber which typically ranges from 0.7 – 4.0 MPa thus providing the required elasticity to allow baleen plates to be pushed away from their resting angle. Elasticity is crucial in this manner as the pressure drag caused by the baleen can be minimized if the plates deflect posteriorly and form a more hydrodynamic shape with respect to the oncoming flow. However, the stiffness of the zwischensubstanz will maintain the spacing between the plates and prevent baleen rack deformation.  One would expect that this is the largest force that the surrounding seawater subjects the baleen plates to since the water pressure on the face of the plates during lunge-feeding would be far greater than the water pressure at any other time. However, if this were the main source of force to which the zwischensubstanz is subjected, one would expect anisotropic strength within the zwischensubstanz where the antero-posterior strength and stiffness would be significantly higher than the other directions. We don’t see any  91  anisotropic behaviour in this sense as there is no significant difference in the compressive Young’s Moduli in any direction and in fact the antero-posterior Modulus of 2.64 MPa is lower than in the other dorso-ventral direction (Table 2.1). It is possible that the Young’s Modulus is lower in the medio-lateral direction because it is not restrained by any other structure in this direction, whereas antero-posteriorly it is constrained by the adjacent baleen plates and dorsally it is constrained by the rostrum. The lack of anisotropic compressive elasticity does suggest that there are sources of stress for the zwischensubstanz that may be as large as the lunge-feeding pressure drag that was initially presumed to cause the largest stress.  One such source of stress on the zwischensubstanz may be from the lateral edge of the baleen plates pressing tightly against the lower mandibles when the whale’s mouth is closed as well as during filtration. Such a pressure on the cantilevered plates may result in tension in the lateral edge of the zwischensubstanz with a progressive gradient toward compression at the medial edge assuming equal pressure from the mandible along the entire lateral length of the baleen rack. The low stiffness of the zwischensubstanz would then allow baleen plates to conform to the mandible shape and reduce stress on the baleen plates.  Under these circumstances the tension and compression would be experienced in both the dorso-ventral and medio-lateral directions of the zwischensubstanz. Indeed, this may be the only way that zwischensubstanz can experience tension in any direction due to its socalled “protection” by the rostrum and baleen plates. The tensile properties of  92  zwischensubstanz in the dorso-ventral and medio-lateral directions reflect this protection as the Young’s modulus in the stiff Hookean regions is similar to a lightly cross-linked rubber. It was observed at Hvalur HF that the lateral edge of the baleen plates presses hard enough into the lower mandibles to leave faint indentation marks in the skin of the mandibles.  The hysteresis measured from the cyclic tensile and compression tests also suggests that zwischensubstanz is a visco-elastic material that dissipates energy when stressed. Hysteresis remained fairly consistent across all cyclic tests and across all twenty cycles within each test at approximately 44% energy loss (Tables 2.1 and 2.2). This viscoelasticity appears to serve as a dampener for forces translated from baleen plate movement. A perfectly elastic material would result in vibrating baleen plates and more stress being assumed by the plates as the plates would be required to bend more with incoming water flow. On the other hand, a perfectly viscous material would not be able to maintain the spacing between the plates that is required for filtration. The forty-four per cent hysteresis measured here would provide the adequate plate spacing and stress absorption required for lunge-feeding whales.  Comparison to other keratins Comparing Fudge’s (2002) stress-strain curve for a hydrated hard α-keratin (Figure 2.4) to the stress-strain curves for zwischensubstanz (Figures 2.11b and 2.12b), there are some noticeable similarities and differences. Zwischensubstanz displays the typical Hookean region from zero to three percent strain during which there is a reversible deformation of 93  the α-helices’ bonds (Fudge 2002). Clearly the yield point, marking the transition from Hookean to yield regions and signalling the start of a transition from alpha-helices to Beta-sheets, is not as sharp in zwischensubstanz. This is most likely due to the fact that our test piece of zwischensubstanz was not an ideal single keratin fibre such as a single hair or strand of wool. As such there are many factors like interactions with other molecules or fibres oriented in other directions that will reduce the sharpness of that transitional point. This may also be a result of zwischensubstanz being more similar to soft α-keratins, which do not transition as sharply as hard α-keratins. It also appears that zwischensubstanz remains in the yield region for longer and does not substantially stiffen (post-yield region) before breaking, again showing similarities to soft α-keratins. The post-yield region of a keratin stress-strain curve symbolizes a complete, irreversible transition to Beta-sheets creating a much stiffer material before those bonds are permanently broken when the material yields. Zwischensubstanz does not appear to enter a stable Beta-sheet phase, more likely forming unstable Beta-sheet bonds before breaking. This may again be the result of an imperfect piece of zwischensubstanz or evidence that zwischensubstanz is a soft α-keratin. The three regions (Hookean, yield and post-yield) are all distinct in zwischensubstanz, but could be more differentiable if a solitary fibre was used for testing as opposed to a large piece. The compression tests should also have exhibited the Hookean region and start of the yield region as strains of 20% were tested, but most likely did not due to the same irregularities (Figures 2.8-2.10). Further testing would be required to confirm the properties measured here.  94  Zwischensubstanz also has a Young’s Modulus in the range of soft α-keratins and hagfish slime threads (Fudge 2002). On the other hand, hard α-keratins have Young’s moduli on the order of 1000 times greater than zwischensubstanz. In dry hagfish slime threads the yield region of the stress-strain curve extends past the typical 30% strain of hard αkeratins up to around 70% before entering the post-yield phase while in zwischensubstanz we observed the yield region extending to intermediate strains of 50% suggesting that, although similar to soft α-keratins, zwischensubstanz may behave slightly different. In hard α-keratins the α-β transition occurs around 80%, but in zwischensubstanz specimens broke before 80% strain again suggesting different structural mechanics of zwischensubstanz relative to both hard and soft α-keratins.  Microscopic analysis A microscopic look at zwischensubstanz reveals structures similar to other keratins that hint at how the baleen plates are formed. Keratinized structures in many animals such as horse hoof, human hair and rhinoceros horn are formed by living tissue most often made up of conical connective tissue fibres that are termed papillae (Hieronymous et al. 2006, Ekfalck 1990). In whales, as one would expect, these papillae appear to be larger than other animals, with papillae diameters of nearly 1 mm at its thickest point (Figure 2.20) compared to about 200 µm in horse hoof (pers. comm. with JM Gosline). The papillae grow conically into the base of the zwischensubstanz; it would appear extending up through the entire height of the zwischensubstanz since the papillae holes still appear in the slices from the ventral end of the zwischensubstanz (Figure 2.22). As the papillae extend into the zwischensubstanz the cellular layer around these conical papillae divide 95  ventrally along its length. During differentiation, the cells are cornified and calcified leading to the material properties of baleen described by Szewciw et al. (2010), St. Aubin et al. (1984) and Pautard (1963). These cells form the matrix around the papillae that we see increasing between Figures 2.20 and 2.22. Thus, the fibres oriented dorso-ventrally in Figure 2.16 and concentrically around the papillae in Figures 2.21, 2.22 and 2.23 are formed by these cellular divisions along the papillae and are also so strongly oriented as a result of its seemingly unidirectional growth. Although the images do not show it, it is also possible that the papillae do have some internal structure that could not be captured with this process, most likely due to the microtome slicing method or because it is not at all birefringent.  96  Conclusions They say Jonah, he was swallowed by a whale, But I say there’s no truth to that tale I know Jonah, he was swallowed by a song -Paul Simon from Jonah, “One-Trick Pony”  Filtration Mechanics Future directions Regrettably, this project was not able to fully address the hypothesis and the theories of filtration mechanisms of lunge-feeding whales in general, but it is the beginning of what can be some very novel research. In order to determine the filtration mechanics of fin whales, the first area that needs to be addressed is the direction of flow inside the oral cavity. Regardless of the material properties of each component of the filtration system, a computer model can be created to simply simulate the shape and contours of the rostrum, baleen and oral cavity of a fin whale. Creating a virtual whale mouth in a computer simulation with the appropriate ventral groove pressures, baleen spacing and internal and external environments, one can determine the most likely direction of the water in the whale’s mouth. This simulation could confirm or reject the hypothesis of centripetal filtration, and once the flow direction is established one can start to incorporate material properties into the simulation or create a physical model based on the results.  Several anatomical features affecting the filtration process of fin whales have also yet to be studied. Two of these areas in particular should be studied first: the coefficient of 97  friction of both the baleen fringes and the tongue. I believe that both of these factors will greatly affect the movement of the krill in the oral cavity. If the computer simulation shows that the water is moving in a centripetal-like motion, then how is the krill interacting with the baleen and tongue? Are the krill bouncing off the baleen, sliding along it or getting stuck in it? The tongue, although fatty and loose, has a sticky surface similar to the surface of a human tongue and probably has a high coefficient of friction. If this is the case, does the krill stick to it and if so, how is the krill then transported to the oesophagus for consumption?  Baleen resistance should also be quantified more accurately. One method of testing the baleen resistance is to pull a baleen section through a tow tank with a uniaxial force gauge. This resistance can then be translated into an expected pressure difference across the baleen. One can compare pressures calculated from this method to theoretical models calculated from Potvin et al. (2009) and speculate as to possible flow patterns that will have the pressures equate and thus match the measured values from Goldbogen et al. (2007). Care must be taken in dealing with the baleen shape as a tow tank experiment would be trickier than it sounds.  More advanced experiments can account for several other factors after preliminary experiments like the ones listed above are completed. Factors such as krill escape responses, whale body rotation and forward acceleration would all require epic experimental setups. Of course, there is always the alternative that restrictions on whale-  98  human interactions will be lifted and some lucky scientist can experience being in the “belly of the whale” and answer all these questions once and for all.  Proposal for a new theory of filtration All of these proposals for future experiments to solve the filtration mechanics of lungefeeding whales are based on my new hypothesis of centripetal filtration that has been developed both through these experiments and through an analysis of the literature. After a fin whale has closed its mouth it commences contracting its VGB and pressurizes the volume of water and krill inside the oral cavity. There is no evidence to suggest that the VGB is not contracted equally antero-posteriorly, but it seems likely that rorquals do have control over the contractions and can actively limit its expansion (Potvin et al. 2009). For now we can assume that the contraction is uniform. Through this contraction the water inside the oral cavity is pushed upwards and forwards towards its only exit, the baleen. Upon reaching the roof of the mouth, water is directed to the left and right hand sides of mouth along the smooth, parabolic vomer. In this way, the water anterior of the temporomandibular joint begins to take on a spiralling motion on both the right and left hand sides of the mouth with water spiralling anteriorly. In this manner most of the water ends up flowing ventrally along the edge of the baleen plates, in line with the baleen fringes in a type of centripetal cross-flow filtration. Thus the krill in this mixture also end up flowing along the baleen fringes and are possibly directed into the baleen depending on the exact direction of the flowing water.  99  Based on the texture of the baleen fringes and the tongues of fin whales that I observed at Hvalur HF in Iceland, I believe that krill will in fact slide along the fringes of the baleen or bounce off of it, again depending on the direction of flow. With a high velocity of water and considering the density of the fringes, it is unlikely that much krill will get stuck in the fringes themselves. Fringes are actually quite smooth, with low friction along their long axes. As the water exits the mouth when it flows by (or into) the baleen, the krill concentration inside the mouth becomes increasingly large. When there is very little water left inside the mouth I believe that the krill will stick to the surface of the tongue due to the sandpaper-like surface of the tongue, leaving about ten kilograms of krill spread out along the whale’s enormous tongue. At this point the whale needs to get this krill to its oesophageal opening at the back of the tongue. As the tongue pushes up against the vomer when the mouth is closed, the krill that are stuck to the tongue may be squeezed between the slippery surface of the vomer and the sticky tongue. It is then possible that this squeezing occurs in such a manner to force all the krill to the back of the tongue where they can be ingested with very little seawater consumption. At this point the whale lowers its jaws for another lunge and the process begins again with very little chance of krill remaining stuck in the baleen fringes after each lunge.  Conclusion While the filtration mechanics of lunge-feeding rorquals were not confirmed, the experiments conducted here did prove to be key stepping-stones on the path to determining the techniques used by these whales. Cross-flow filtration into a stagnant chamber does not appear to provide the flow rates through the baleen plates that would be  100  required to match the observed data (Goldbogen et al. 2007), but it cannot theoretically be rejected at this time. My experimental design should be improved before making a definitive statement. As such, it remains plausible that either dead-end filtration, crossflow filtration or centripetal filtration are being used to filter the krill through their baleen, but it is more likely that a hybrid of centripetal filtration and dead-end filtration are being used. Centripetal filtration is a new term that has yet to be fully defined in the filter feeding literature, but based on what we have determined as the most likely direction of water flow inside a lunge-feeding whale’s mouth it appears to be a legitimate technique.  The concept of centripetal filter feeding involves water spinning in an anteriorly spiralling motion on the left and right hand sides of a whale’s oral cavity. As the water flows along the baleen fringes during this motion, the water flows out through the baleen plates, while the krill slide along or bounce off of the fringes and remain inside the mouth for consumption. This hypothesis should be further examined by first determining the direction of water flow inside the oral cavity through a computer simulation. It is also crucial to examine the material properties of all involved feeding structures including the tongue, baleen fringes and vomer to determine their involvements in the feeding process. With luck, the baleen whale population will return to previous levels and scientists may have an opportunity to examine the filtration mechanics first hand and appreciate with more understanding “the greatest biomechanical event in the world”.  101  Zwischensubstanz Properties Future directions Here we have presented a preliminary analysis of the material properties of zwischensubstanz and a microscopic look at its internal structures, but it only marks the beginning of work that could be done to elucidate the structure and function of zwischensubstanz and baleen plates. The most obvious next step would be to continue taking slices of the baleen plate not just through the zwischensubstanz, but along the entire length of the baleen plate once it has left the zwischensubstanz. One would then have a complete picture of the formation of the papillae into baleen plates and the resulting fringes. A continuation of material property analysis of the zwischensubstanz can also determine the precise origins of the calcium that helps to harden the baleen plate’s keratin matrix (Szewciw et al. 2010). In addition, this analysis may inform us whether the baleen fringes are formed through a chemical process, wear from tongue abrasion or by some other mechanism.  In this thesis we have also mentioned some probable sources of stress for the baleen. It is possible that the computer simulations mentioned in Chapter One (pg 56) will unearth additional sources of stress for both the baleen and the zwischensubstanz. Once the largest sources of stress have been identified they can then be tested in a more realistic scenario. Observing baleen plates, fringes and zwischensubstanz individually can inform us of their individual properties, but how they function in situ during a whale’s actual activities would bring about more relevant data. 102  In addition, examining the development of zwischensubstanz and baleen from birth, where baleen whales show only the beginnings of baleen plate function, until adulthood, where baleen and zwischensubstanz have grown into materials with completely different properties (Pivorunas 1979), would expand on Tullberg’s (1883) work and provide extensive amounts of information on baleen whale development.  Conclusion Zwischensubstanz is an integral part of not only the baleen racks, but also of the baleen whales feeding structures as a whole and has historically not been given its fair share of scientific analysis. This experiment aimed to get the ball rolling and develop a basic understanding of the material properties of this “in-between-substance” and in this manner it proved quite successful. We determined that zwischensubstanz has material properties similar to rubber and displays largely isotropic behaviour. The purpose of this isotropic behaviour is not clear and requires a further investigation of the functions of zwischensubstanz and the forces to which it is subjected.  Thanks to its visco-elastic material properties zwischensubstanz appears not only to function as a spacer for baleen plates, but also as a stress absorber for the baleen plates. However, its primary function may in fact be in its use in developing baleen plates. While baleen plates begin to form from papillae that originate in the connective tissue attached to the zwischensubstanz our microscopic analysis shows that it is within the zwischensubstanz itself that the keratin matrix forms and cornifies before exiting the  103  zwischensubstanz as a fully formed plate. Thus, the zwischensubstanz plays a most crucial role in creating the filter with which baleen whales have been able to feed on enormous masses of prey and grow into the giants of the ocean that we know and love.  104  References Alexander, R. (1998). All-time giants: The largest animals and their problems. Palaeontology, 41, 1231–1245. Baker, R., Fane, A., & Fell, C. (1985). Factors affecting flux in crossflow filtration. Desalination, 53, 81-93. Blatt, W., Dravid, A., Michaels, A.S., Nelsen, L. (1970). Solute polarization and cake formation in membrane ultrafiltration: causes, consequences and control techniques, in: J.E. Flinn (Ed.), Membrane Science and Technology, Plenum, New York, pp. 47– 91. Brainerd, E. (2001). Caught in the crossflow. Nature, 412(6845), 387–388. doi:10.1038/35086666 Brodie, P., & Vikingsson, G. (2010). On the Feeding Mechanisms of the Sei Whale (Balaenoptera borealis). Journal of Northwest Atlantic Fishery Science, 42, 49–54. doi:10.2960/J.v42.m646 Campbell, R., & Dower, J. (2003). Role of lipids in the maintenance of neutral buoyancy by zooplankton. Marine Ecology Progress Series, 263, 93–99. Cheer, A. (2001). Computational Fluid Dynamics in the Oral Cavity of Ram SuspensionFeeding Fishes. Journal of Theoretical Biology, 210(4), 463–474. doi:10.1006/jtbi.2001.2325 Coulombe, P., & Omary, M. (2002). 'Hard' and “soft” principles defining the structure, function and regulation of keratin intermediate filaments. Current Opinion in Cell Biology, 14(1), 110–122. Croxall, J., & Pilcher, M. (1984). Characteristics of Krill Euphausia-Superba Eaten by Antarctic Fur Seals Arctocephalus-Gazella at South-Georgia. British Antarctic Survey Bulletin, 117–125. De Robertis, A. (2003). Acoustic observations of the swimming behavior of the euphausiid Euphausia pacifica Hansen. ICES Journal of Marine Science, 60(4), 885– 898. doi:10.1016/S1054-3139(03)00070-5 Ekfalck, A. (1990). Amino acids in different layers of the matrix of the normal equine hoof. Possible importance of the amino acid pattern for research on laminitis Zentralblatt für Veterinärmedizin. Reihe B. Journal of veterinary medicine. Series B, 37(1), 1–8.  105  Fudge, D.S. (2002). Biomechanics of Intermediate Filament Based Materials: Insights from Hagfish Slime Threads. PhD thesis, University of British Columbia. Fudge, D.S., Szewciw, L.J., & Schwalb, A. N. (2009). Morphology and Development of Blue Whale Baleen: An Annotated Translation of Tycho Tullberg's Classic 1883 Paper. Aquatic Mammals, 35(2), 226–252. doi:10.1578/AM.35.2.2009.226. Goldbogen J.A., Calambokidis J., Shadwick R.E., Oleson E.M., McDonald M.A., Hildebrand J.A. (2006). Kinematics of foraging dives and lunge-feeding in fin whales. Journal of Experimental Biology, 209(7), 1231–1244. doi:10.1242/jeb.02135. Goldbogen, J. A., Calambokidis, J., Oleson, E., Potvin, J., Pyenson, N. D., Schorr, G., & Shadwick, R. E. (2011). Mechanics, hydrodynamics and energetics of blue whale lunge feeding: efficiency dependence on krill density. Journal of Experimental Biology, 214(4), 698–699. doi:10.1242/jeb.054726. Goldbogen, J. A., Potvin, J., & Shadwick, R. E. (2010). Skull and buccal cavity allometry increase mass-specific engulfment capacity in fin whales. Proceedings of the Royal Society B: Biological Sciences, 277(1683), 861–868. doi:10.1098/rspb.2009.1680. Goldbogen, J., Pyenson, N., & Shadwick, R. (2007). Big gulps require high drag for fin whale lunge feeding. Marine Ecology Progress Series, 349, 289–301. doi:10.3354/meps07066. Hamner, W. (1988). Biomechanics of Filter Feeding in the Antarctic Krill EuphausiaSuperba - Review of Past Work and New Observations. Journal of Crustacean Biology, 8(2), 149–163. Hieronymus, T. L., Witmer, L. M., & Ridgely, R. C. (2006). Structure of white rhinoceros (Ceratotherium simum) horn investigated by X-ray computed tomography and histology with implications for growth and external form. Journal of Morphology, 267(10), 1172–1176. doi:10.1002/jmor.10465. Julien, P.Y. (1986), “Concentration of Very Fine Silts in a Steady Vortex,” JHR, Proc. IAHR, 24(4), 255–264. Kawamura, A. (1974). Food and feeding ecology in the southern sei whale. Scientific Reports of the Whales Research Institute, Tokyo 26:25–144. Kils, U. (1979) Swimming speed and escape capacity of Antarctic krill, Euphausia superba. Meeresforchung 27: 264–266. Kot, B. W. (2005). Rorqual whale surface-feeding strategies: biomechanical aspects of feeding anatomy and exploitation of prey aggregations along tidal fronts. M.Sc. thesis, University of California, Los Angeles. 106  Lambertsen, R. (1983). Internal mechanism of rorqual feeding. Journal of Mammalogy. Lambertsen, R., Rasmussen, K., Lancaster, W., & Hintz, R. (2005). Functional morphology of the mouth of the bowhead whale and its implications for conservation. Journal of Mammalogy, 86(2), 342–352. Lockyer, C. (1986). Weights and anatomical measurements of northeastern Atlantic fin (Balaenoptera physalus, Linnaeus) and sei (B. borealis, Lesson) whales. Marine Mammal Science. Lowry, L. F. (1993). Foods and feeding ecology. In The Bowhead Whale, pp. 201–238. Ed. by J. J. Burns, J. J. Montague and C. J. Cowles. The Society for Marine Mammalogy, Special Publication Number 2. Lowry, L.F. and Burns, J.J. (1980). Foods utilized by bowhead whales near Barter Island, Alaska, autumn 1979. Marine Fisheries Review 42(9-10):88-9. Lu, W., Hwang, K., & Ju, S. (1993). Studies on the Mechanism of Cross-Flow Filtration. Chemical Engineering Science, 48(5), 863–872. Lu W.-M. and Ju, Shang-Chung, (1989). Selective particle deposition in crossflow filtration. Sep. Sci. Technol. 24 7/8, pp. 517–540. Mayo C.A., Letcher B.H., Scott S. (2001). Zooplankton filtering efficiency of the baleen of a North Atlantic right whale, Eubalaena glacialis. J Cetacean Res Manage Spec Issue 2: 225–229. Motta, P. J., Maslanka, M., Hueter, R. E., Davis, R. L., la Parra, de, R., Mulvany, S. L., Habegger, M. L., et al. (2010). Feeding anatomy, filter-feeding rate, and diet of whale sharks Rhincodon typus during surface ram filter feeding off the Yucatan Peninsula, Mexico. Zoology, 113(4), 199–212. Elsevier GmbH. doi:10.1016/j.zool.2009.12.001. Murkes, J. and Carlsson, C-G, (1988). Crossflow Filtration, Theory and Practice, Wiley, New York, NY. Nerini, M. (1984). A review of gray whale feeding ecology. In The Gray Whale: Eschrictius robustus, pp. 423–450. Ed. By M. L. Jones, S. L. Swartz and S. Leatherwood. O’Brien, D. (1987). Description of Escape Responses of Krill (Crustacea, Euphausiacea), with Particular Reference to Swarming Behavior and the Size and Proximity of the Predator. Journal of Crustacean Biology, 7(3), 449–457.  107  Orton, L., & Brodie, P. (1987). Engulfing Mechanics of Fin Whales. Canadian Journal of Zoology-Revue Canadienne De Zoologie, 65(12), 2898–2907. Paig-Tran, E. W. M., Bizzarro, J. J., Strother, J. A., & Summers, A. P. (2011). Bottles as models: predicting the effects of varying swimming speed and morphology on size selectivity and filtering efficiency in fishes. Journal of Experimental Biology, 214(10), 1643–1654. doi:10.1242/jeb.048702. Pauly, D., Trites, A. W., Capuli, E., & Christensen, V. (1998). Diet composition and trophic levels of marine mammals, 1–15. Pautard, F. G. E. (1963). Mineralization of keratin and its comparison with the enamel matrix. Nature (London) 199, 531–535. Pivorunas, A. (1976). A mathematical consideration of the function of baleen plates and their fringes. Scientific Reports of the Whales Research Institute, 28, 37-55. Pivorunas, A. (1979). The feeding mechanisms of baleen whales. American Scientist, 67, 432-440. Potvin, J., Goldbogen, J. A., & Shadwick, R. E. (2009). Passive versus active engulfment: verdict from trajectory simulations of lunge-feeding fin whales Balaenoptera physalus. Journal of The Royal Society Interface, 6(40), 1005–1025. doi:10.1098/rsif.2008.0492. Ray G.C., Schevill W.E. (1974). Feeding of a captive gray whale. Eschrichtius robustus. Mar Fish Rev 36: 31–38. Ripperger, S., & Altmann, J. (2002). Crossflow microfiltration - state of the art. Separation and Purification Technology, 26(1), 19–31. Rubenstein, D. I., & Koehl, M. A. R. (2008). The Mechanisms of Fliter Feeding: Some Theoretical Considerations. The American Naturalist, 111, 981–994. Sanderson S.L., Cheer A.Y., Goodrich J.S., Graziano J.D., Callan W.T. (2001) Crossflow filtration in suspension-feeding fishes. Nature 412:439–441. Simon, M., Stafford, K. M., Beedholm, K., Lee, C. M., & Madsen, P. T. (2010). Singing behavior of fin whales in the Davis Strait with implications for mating, migration and foraging The Journal of the Acoustical Society of America, 128(5), 3200–3210. doi:10.1121/1.3495946. Song, L., & Elimelech, M. (1995). Theory of Concentration Polarization in Cross-Flow Filtration. Journal of the Chemical Society-Faraday Transactions, 91(19), 3389– 3398.  108  St. Aubin, D. J., Stinson, R. H., & Geraci, J. R. (1984). Aspects of the structure and composition of baleen, and some effects of exposure to petroleum hydrocarbons. Canadian Journal of Zoology, 62, 193-198. Szewciw, L J, de Kerckhove, D. G., Grime, G. W., & Fudge, D S. (2010). Calcification provides mechanical reinforcement to whale baleen -keratin. Proceedings of the Royal Society B: Biological Sciences, 277(1694), 2597–2605. doi:10.1098/rspb.2010.0399. Tullberg, T. (1883). Bau und Entwicklung der Barten bei Balaenoptera sibbaldii [Structure and Development of the Baleen of the Blue Whale, Balaenoptera musculus]. In Nova Acta Regiae Societatis Scientiarum Upsaliensis, Series III, Volume 11, part 3, pp. 1-36, with seven plates of figures. Visser, A. (1999). Lipids, buoyancy and the seasonal vertical migration of Calanus finmarchicus. Fisheries Oceanography. Werth, A. J. (2001). How do mysticetes remove prey trapped in baleen? Bulletin of the Museum of Comparative Zoology, 156, 189-203. Werth, A. J. (2004). Models of hydrodynamic flow in the bowhead whale filter feeding apparatus. Journal of Experimental Biology, 207(20), 3569–3580. doi:10.1242/jeb.01202. Werth, Alexander J. (2007). Adaptations of the cetacean hyolingual apparatus for aquatic feeding and thermoregulation. The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology, 290(6), 546–568. doi:10.1002/ar.20538. Werth, A.J. (2011), Flow-Dependent Porosity of Baleen, Society for Integrative and Comparative Biology 2011 Annual Meeting, P3.135. Williamson, G. R. (1973). Counting and measuring baleen and grooves of whales. Scientific Reports of the Whales Research Institute, 25, 279-292. Yayanos A.A., Benson A.A., Nevenzel J.C. (1978). The pressure volume-temperature (PVT) properties of a lipid mixture from a marine copepod, Calanus plumchrus: implications for buoyancy and sound scattering. Deep-Sea Res 25: 257–268.  109  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072276/manifest

Comment

Related Items