Environmental biology of fishes, tập 95, số 1, 2012

Environ Biol Fish (2012) 95:1–2 DOI 10.1007/s10641-012-9983-7 Preface: feeding ecology of elasmobranchs David A Ebert & W David McElroy & William T White Received: 24 October 2011 / Accepted: 16 January 2012 / Published online: February 2012 # Springer Science+Business Media B.V 2012 Abstract Elasmobranchs are apical predators in most marine communities where they occur, often playing a substantial role in the food web dynamics of those communities However, despite their high trophic status they are often poorly studied compared to most commercially important teleosts Furthermore, despite efforts towards ecosystem-based management, elasmobranchs are still often lumped into generic categories referred to as “shark” or “skate” unclassified, with limited effort to identify individual species The role of elasmobranchs in ecosystems has never been more important to our understanding of marine ecology due to high levels of exploitation of many species Similar to other high trophic level predators, many elasmobranchs have life-history characteristics that make them D A Ebert (*) Pacific Shark Research Center, Moss Landing Marine Laboratories, 8272 Moss Landing Road, Moss Landing, CA 95039, USA e-mail: debert@mlml.calstate.edu W D McElroy Apex Predators Program, Northeast Fisheries Science Center, National Marine Fisheries Service, 166 Water Street, Woods Hole, MA 02543, USA W T White CSIRO Marine & Atmospheric Research, Wealth from Oceans Flagship, GPO Box 1538, Hobart, Tasmania 7000, Australia vulnerable to over-exploitation Elasmobranch populations are now heavily targeted in many fisheries throughout the world Increasing exploitation of this group is especially alarming because their feeding ecology is poorly studied and by extension their influence in shaping ecosystems Given recent increased attention on elasmobranchs in the scientific literature, management and conservation circles, and the general news media, researchers over the past decade have begun to more closely examine the ecological role of this important taxon of fishes Due to this increasing awareness, and the development of new and innovative methods and analytical techniques, it prompted us to organize an international symposium on the “Feeding Ecology of Elasmobranchs” The symposium was held on 10 July 2010, in conjunction with the 27th annual meeting of the American Elasmobranch Society meetings in Providence, Rhode Island Keywords Sharks Batoids Symposium Feeding ecology Trophic level Preface Elasmobranchs (sharks and batoids) are apical predators in most marine communities in which they occur, and by virtue of their abundance and widespread occurrence can play an influential role in the food-web dynamics of marine communities However, despite their high trophic role they are little studied relative to most commercially important teleosts Furthermore, despite a move towards ecosystem-based management of marine resources high trophic-level predators, such as elasmobranchs, are still often lumped into catchall categories generically referred to as “shark” or “skate” unidentified, often with minimal effort to identify the individual species in question The role of elasmobranchs in ecosystems however has never been more important due to possible declines in some populations from intensive exploitation throughout portions of the range and habitat degradation, among other causes Similar to other apical predators, many elasmobranchs have life-history characteristics that make them vulnerable to over-exploitation Regional evidence of this pattern has been documented on both coasts of the United States where elasmobranch landings have been rapidly increasing among groundfish fisheries The increasing exploitation of this group is especially alarming because little is known about their feeding ecology, as with many aspects of their life-history, and by extension their influence in shaping ecosystems Of nearly 1200 elasmobranch species currently recognized relatively few have had any quantitative dietary information reported, with most species either having only anecdotal or no information available Ebert and Bizzarro (2007) noted that of 245 known skate species at the time, less than 24% had any quantitative dietary information available, and that only 37 species had studies that examined more than 100 stomachs Furthermore, Cortés (1999) found similar results for 149 species of sharks with 45% of the species having less than 100 stomachs examined The reported trophic levels (TL) reported in these studies for skates (mean TL 3.8) and sharks (mean TL 4.0) strongly support the characterization of these taxa among the highest trophic marine-predators (Cortés 1999; Ebert and Bizzarro 2007) Given the increased awareness of elasmobranchs globally, their importance as apical predators in marine ecosystems, and fisheries management and conservation concerns, researchers over the past decade have begun to more closely examine the role of these charismatic fishes Due to increasing attention on shark and batoid feeding ecology both in the scientific and conservation communities, as well as in the public media, and the development of new and innovative study methods Environ Biol Fish (2012) 95:1–2 and analytical techniques; we were prompted to organize an international symposium on the “Feeding Ecology of Elasmobranchs” The aims and goals of this symposium were to bring together an international group of researchers to share their knowledge and experience, and stimulate discussion on the latest advancements and developments in the feeding ecology of elasmobranchs The symposium was broadly organized around several areas of research, including ecomorphology and sensory biology, stable isotope analysis, new and innovative analytical methods and models, and traditional food habit studies and the role of sharks and batoids in marine communities The symposium was held on 10 July 2010, in conjunction with the 27th annual meeting of the American Elasmobranch Society meeting in Providence, Rhode Island, with a total of 21 speakers representing six countries, presenting papers Presentations included a good mix of established professionals and graduate students presenting data and results from Masters theses or doctoral dissertations As a result of this symposium we invited all of the speakers plus a couple of additional authors, who unfortunately were unable to attend the meetings at the last moment, to submit manuscripts for consideration into this special issue After a careful peer review process, a total of 12 papers by authors from five countries were accepted for publication and are included here in this special issue We as guest editors would like to thank all of the participants who presented papers, those who refereed manuscripts for this special issue, and to Environmental Biology of Fishes (EBF), especially editor in chief, David Noakes, and managing editor Lynn Bouvier, and the staff at EBF for making this a smooth process The American Elasmobranch Society, Pacific Shark Research Center, and the David and Lucile Packard Foundation provided support for the Symposium References Cortés E (1999) Standard diet compositions and trophic levels of sharks ICES J Mar Sci 56:707–717 Ebert DA, Bizzarro JJ (2007) Standardized diet composition and trophic levels in skates Environ Biol Fish 80:221–237 Environ Biol Fish (2012) 95:3–20 DOI 10.1007/s10641-011-9959-z Breaking with tradition: redefining measures for diet description with a case study of the Aleutian skate Bathyraja aleutica (Gilbert 1896) Simon C Brown & Joseph J Bizzarro & Gregor M Cailliet & David A Ebert Received: September 2010 / Accepted: 27 October 2011 / Published online: 16 November 2011 # Springer Science+Business Media B.V 2011 Abstract Characterization of fish diets from stomach content analysis commonly involves the calculation of multiple relative measures of prey quantity (%N,%W,% FO), and their combination in the standardized Index of Relative Importance (%IRI) Examining the underlying structure of dietary data matrices reveals interdependencies among diet measures, and obviates the advantageous use of underused prey-specific measures to diet characterization With these interdependencies clearly realized as formal mathematical expressions, we proceed to isolate algebraically, the inherent bias in %IRI, and provide a correction for it by substituting traditional measures with prey-specific measures The resultant new index, the Prey-Specific Index of Relative Importance (% PSIRI), is introduced and recommended to replace %IRI for its demonstrated more balanced treatment of the relative measures of prey quantity, and less erroneous behavior across taxonomic levels of identified prey As a case study, %PSIRI was used to examine the diet of the Aleutian skate Bathyraja aleutica from specimens collected from three ecoregions of the northern Gulf of Alaska (GOA) continental shelf during JuneSeptember 2005–2007 Aleutian skate were found to primarily consume the commonly abundant benthic crustaceans, northern pink shrimp Pandalus eous and Tanner crab Chionoecetes bairdi, and secondarily consume various teleost fishes Multivariate variance partitioning by Redundancy Analysis revealed spatially driven differences in the diet to be as influential as skate size, sex, and depth of capture Euphausiids and other mid-water prey in the diet were strongly associated with the Shelikof Strait region during 2007 that may be explained by atypical marine climate conditions during that year S C Brown (*) : J J Bizzarro : G M Cailliet : D A Ebert Pacific Shark Research Center, Moss Landing Marine Laboratories, 8272 Moss Landing Road, Moss Landing, CA 95039, USA e-mail: simoncbrown@gmail.com Introduction Present Address: J J Bizzarro School of Aquatic and Fishery Sciences, University of Washington, PO Box 355020, Seattle, WA 98195-5020, USA Keywords Skate Bathyraja aleutica Diet Gulf of Alaska Food habits Index of relative importance Diet information is crucial to our understanding of species ecology, trophic interrelationships, food webs, and ultimately, the flow of energy through ecosystems Stomach content analysis remains a universal technique for sampling the diets of fishes and these studies contribute large amounts of species-specific diet data for potential use in trophic ecosystem modeling that provide ecosystem-based fishery management advice (Ainsworth et al 2010) Although there has been an arguably successful call (294 citations in Web of Science) for consistency in reporting diet compositions of elasmobranchs with percent number (%N), percent weight (%W), frequency of occurrence (%FO), and the standardized Index of Relative Importance or %IRI (Cortès 1997), there remain critical unresolved weaknesses in this widely used and accepted methodology (Cortès 1998; Hansson 1998) There are, in fact, not only weaknesses, but also serious unrecognized mathematical flaws in the presentation of diet data both graphically and by indices that have gone largely unnoticed in the published literature These methodological problems are a direct consequence of diet researchers’ incorrect mathematical understanding of diet measures A deeper understanding of the structure of dietary data and the resulting mathematical relationships between diet measures not only resolves current methodological weaknesses, but also leads to increased extraction of information about food habits in graphical displays and diet composition tables Proper metrics are necessary to characterize diets of abundant, data-poor species, to avoid faulty or incomplete conclusions about their trophic roles Skates as abundant mesopredators (Ebert and Bizzarro 2007) likely play important trophic roles in demersal fish communities and may be able to overtake the resource niche left open by depleted teleost stocks (Stevens et al 2000; Link and Sosebee 2008) For large marine ecosystems with large industrial fisheries, like the northern Gulf of Alaska (GOA) shelf, attaining trophic information on skates is beneficial for consideration in ecosystem based modeling and management In the GOA, skates contribute substantially to the bycatch (retained and discarded) of directed fisheries such as those for Pacific Halibut Hippoglossus stenolepis and other commercially valuable groundfish species (Ormseth and Matta 2009), but their ecology is poorly understood Skates of the genus Bathyraja are widely distributed in the North Pacific In the GOA the Aleutian skate Bathyraja aleutica (Gilbert 1896) is one of the most abundant skate species (Ormseth and Matta 2009) The Aleutian skate ranges from the northern Sea of Japan into the Sea of Okhotsk and Bering Sea to the eastern GOA, but has been reported as far south as Cape Mendocino, California, U.S.A (Hoff 2002), typically occupying shelf and slope depths of 100– 800 m (Mecklenburg et al 2002) It is a relatively Environ Biol Fish (2012) 95:3–20 large skate, reaching a maximum size of 150 cm total length (TL) Diet data reported from the western North Pacific and Aleutian Islands indicate that Aleutian skates consume primarily decapod crustaceans, with fishes and cephalopods also represented (Orlov 1998; Yang 2007); however, scant dietary information is available for the Aleutian skates in the northern GOA The findings of this study represent the first detailed trophic information on the Aleutian skate population from the northern GOA shelf ecosystem To improve current data reporting methods for stomach contents analysis of elasmobranchs, our objectives for this study are to: 1) define the structure of diet data; 2) elucidate mathematical relationships between diet measures; 3) demonstrate how appropriate operationalization of these relationships can correct current flaws in compound diet indices and graphical displays and; 4) illustrate the application of this approach in reporting the diet composition on dietary data from the Aleutian skate Additionally, we provide an in-depth statistical analysis of ontogenetic, regional, and interannual variation in the diet of the Aleutian skate from the northern GOA ecosystem during 2005–2007 The structure of diet data Stomach content analysis typically incorporates measurements of numerical abundance (i.e aggregate counts of individual prey items in each designated prey category), gravimetric (or volumetric) abundance (i.e aggregate weights or volumes of prey items in each designated prey category), and the frequency of occurrence of prey categories among all stomach samples (Hyslop 1980) The resulting diet data matrix, by numerical abundance or biomass, is composed of columns of prey categories (i) by rows of individual stomach samples (j) standardized to proportion by total individual stomach content (i.e by row): prey category ðiÞ 0 0:25 0:5 0:25 7 stomach sample ðjÞ 0 0:5 0:5 0:75 0:25 This stomach sample by prey item diet matrix is well-suited for multivariate statistical analysis Environ Biol Fish (2012) 95:3–20 (de Crespin de Billy et al 2000), but also conveniently serves in calculation of diet measures The average percent number (%N) and average percent weight (%W) for each prey item are column averages of this matrix and are additive, meaning that they sum to 100% for all different prey items of a prey category For example, in the dietary data matrix provided above, the average percent value for each prey category (i.e each column) is 25%, and the aggregate values for all prey categories sum to 100% The Percent Frequency of Occurrence (%FO) can be calculated as the column averages from the same diet matrix, but translated to discrete conditions of presence/absence (i.e binary data) From the example diet matrix above, the %FO for each prey category (from left to right) is: 25%, 50%, 50%, and 75% and sums to an indeterminate value which, is 200% in this example Because different prey categories occur together within a single stomach sample, %FO is specific to each category, meaning that its value can range from >0% to 100%, independent of the %FO values of all other prey items Relationship to relative measures of prey quantity There is a specific mathematical relationship between %N and %W, with %FO embedded in the structure of diet matrices from which these measures are calculated Either a prey is absent in a stomach sample, thereby receiving a value of zero for all three diet measures, or it is present, ranging from >0 to by proportion of numerical or gravimetric abundance The meaning of the zero value is the same regardless of which diet measures it is incorporated into: a discrete condition of absence This obfuscation of discrete absences with numerical and gravimetric abundance has profound consequences for the graphical display of diet compositions and compound diet indices that incorporate two of more individual measures (e.g.,%FO,%N,%W) Relationship to graphical displays If %N or %W is plotted against %FO, as suggested for some graphical displays of diet data (Costello 1990; Cortès 1997) exactly one-half of the plot space is non-existent; it is mathematically impossible for any prey to occur at coordinates above the 1:1 line of %N, or %W, and %FO This is because a value of %N or %W cannot exceed its corresponding %FO value for any prey category The determined value of % FO represents an upper limit to %N and %W values because discrete absences are averaged into all measures, this creates a mathematical dependence between diet measures whose strength increases with the increasing frequency of zero values in a diet data matrix Amundsen et al (1996), fully realizing this graphical limitation of diet measures in constructing feeding strategy diagrams (Costello 1990), proposed a new measure termed prey-specific abundance Preyspecific abundance is defined as the percent numerical abundance of a prey item averaged over the stomach samples in which it occurs (i.e excluding zero values) Like %FO, the value of the prey-specific abundance for a prey item may take any value >0% to 100% independent of the values for all other prey items For example, from the diet data matrix above from left to right the prey-specific abundance is 100%, 50%, 50%, and 33.3% We adopt the terminology of Amundsen et al (1996) and extend it to weight (prey-specific weight) for the remainder of this study Although prey-specific abundance was originally proposed as a modification to the Costello (1990) feeding strategy diagram, it should actually be applied to correct any graphical method attempting to plot %N or %W against %FO For example, IRI diagrams, in which %N and %W are plotted against %FO to produce rectangular surfaces representing each prey category’s relative importance, where the unit area is equal to the unit IRI, represent a false geometry The measures representing each rectangle’s sides are not actually at right angles unless the diet data matrix contains no zero values, thus a rectangle is an incorrect representation, falsely implying that %FO is orthogonal to %N and %W Relationship to compound diet indices The IRI and %IRI, like the aforementioned graphical methods, are also flawed by the inherent redundancy in combining mathematically dependent measures A major weakness in the %IRI resulting from this combination is that it is non-additive across taxonomic levels (or other nested prey categories), an attribute of %IRI that has been incorrectly attributed to the inclusion of %FO in its calculation This means, despite being standardized to 100% for any given Environ Biol Fish (2012) 95:3–20 taxonomic level, there is disagreement in %IRI values from one taxonomic level to the next The %IRI therefore has an intrinsically arbitrary property since its values are dependent upon the taxonomic level or the designated prey categories chosen by a researcher, largely defeating its purpose as a standardized measure of prey importance to facilitate comparisons Because it is not likely that all prey will be consistently identified to the same taxonomic level within a study, let alone between studies, this is a serious weakness in %IRI that results from its flawed calculation The flaws in the most commonly used contemporary dietary indices can be described through simple algebra Each diet measure is defined: Prey-specific abundance (%PNi,%PWi): n P %Aij j¼1 %PAi ¼ ð1Þ ni Average percent abundance (%Ni,%Wi): n P %Ai ¼ %Aij j¼1 ð2Þ n Frequency of Occurrence (FO): FOi ¼ ni n ð3Þ where %Aij is the abundance (by counts or weights) of prey category i in stomach sample j, ni is the number of stomachs containing prey i, and n is the total number of stomachs It is evident that Eq is the multiplicative product of Eq and Eq 3: n P %Aij j¼1 ni n P ni  ¼ n %Aij j¼1 n which is proof of the theorem that the average percent abundance (%N and %W) is already a compound index containing %FO Because of their direct mathematical dependence, %N and %W should not be redundantly combined with %FO in a compound diet index such as the Index of Relative Importance (IRI) whose intended purpose is to provide a more balanced estimate of dietary importance The resulting bias in IRI from combining these measures can be isolated by mathematical expression Simple algebraic re- arrangement of the IRI using the equivalencies above results in the following: IRI ¼ %FO  ð%N þ %W Þ ¼ %FO  ½ð%PN  %FOÞ þ ð%PW  %FOފ ¼ %FO2  ð%PN þ %PW Þ It is now revealed %FO is unknowingly squared in the calculation of IRI which, among other undesired behaviors, accounts for the empirically observed bias in %IRI to over-emphasize frequently occurring prey and under-emphasize rare prey with respect to the %FO (Ortaz et al 2006) Having isolated this bias in IRI, we can now modify it by substituting %N and %W with their corresponding prey-specific abundances, %PN and %PW: %PSIRIi ¼ %FOi  ð%PNi þ %PWi Þ This Prey-Specific IRI (%PSIRI) sums to 200% and therefore dividing by results in a version of the standardized %IRI, with an important distinction: the %PSIRI is additive with respect to taxonomic levels, such that the sum of %PSIRI for species will be equal to the %PSIRI of the family containing those species, and so forth This characteristic enhances %PSIRI for comparisons between predators and studies because its values are not dependent upon taxonomic level or prey categories designated by a researcher Additionally, because of the mathematical relationships between diet measures the %PSIRI is also a generalized form of the standardized Geometric Importance Index (Assis 1996) where only %N and %W are considered: %PSIRIi ¼ %FOi  ð%PNi þ %PWi Þ %N þ %W ¼ ¼ %GII 2 It is preferable, however, to express %PSIRI in terms of prey-specific measures and %FO such that there is less obfuscation of presence/absence and abundance, or biomass If only numerical abundance, or biomass, is recorded then %N and %W may act themselves as compound indices: %N ¼ %FO  %PN and %W ¼ %FO  %PW This last expression demonstrates another useful point; if a researcher does not wish to combine numerical abundance with biomass in an index like %PSIRI, they may still present prey-specific measures Environ Biol Fish (2012) 95:3–20 with the %FO and use %N and %W as separate compound indices to summarize relative importance Material and methods Study area The study area was located on the northern GOA continental shelf, extending from the Kenai Peninsula westward to the Shumagin Islands along Alaska Peninsula The northern GOA shelf extends far offshore, encompassing Kodiak Island and ranging to 200 m depth, and is crosscut with deeper canyons and gullies (Weingartner 2005) This region contains the highest biomass of skates in the GOA and is the regional center of abundance for the Aleutian skate (Stevenson et al 2007; Ormseth and Matta 2009) Specimen collection Skate stomachs were obtained by participation aboard fishery-independent bottom trawl surveys conducted by the National Marine Fisheries Service’s Alaska Fisheries Science Center (AFSC) in June-July of 2005, and the Alaska Department of Fish and Game (ADFG) in June-September of 2006 and 2007 In 2005, haul stations were spaced across the entire northern GOA shelf In 2006 and 2007, however, haul stations were stratified into distinct regions contained within the larger study area, including: Alaska Peninsula (AKP), Shelikof Strait (SHS), and the shelf adjacent to the southeastern side of Kodiak Island (KOD) We therefore examined spatial variation in 2005 across the shelf for evidence of regional correspondence in the diet, and then made regional comparisons in 2006 and 2007 Skates were sexed and total length (TL) was measured (+/− mm) The foregut (esophagous and stomach to the pyloric sphincter) was excised, fixed in 10% neutral buffered formalin, and then preserved in 70% ethanol for subsequent analysis The stomach was not taken from any skate that showed signs of stomach eversion or regurgitation of prey in the buccal cavity Stomach contents analysis Stomach contents were sorted, identified, enumerated, and weighed An index of vacuity was used to express the number of empty stomachs encountered as a percentage of the total stomachs examined Any highly digested material that could not be placed in a taxonomic category, stomach parasites, and inorganic material were noted but excluded from further analysis Identification of prey was performed to the lowest possible taxonomic level primarily using taxonomic keys, field guides, consultation with experts, and museum collections The number of each prey item in a stomach sample was estimated using the most conservative count when the detached components of prey were present Prey items were then weighed (0.01 g) after excess moisture was blotted away Sample size sufficiency To assess whether the number of sampled skate stomachs was sufficient to describe the diet, cumulative prey curves (Ferry and Cailliet 1996) were computed with EstimateS (Version 8.2, R K Colwell, http://purl.oclc.org/estimates) The estimated number of unique prey categories and associated 95% CIs were plotted against the cumulative number of stomach sample examined Because visual examination of prey curves for an asymptote is unreliable, the slope of the linear regression (b) through the ultimate five sub-samples was used as an objective criteria where b≤0.05 signified acceptable leveling off of the prey curve for diet characterization (Bizzarro et al 2009) Statistical analysis of diet variation For the purpose of statistical analysis, prey categories were designated for different prey types at higher taxonomic levels (e.g polychaetes) and for common prey taxa (e.g Tanner crab Chionoecetes bairdi) Highly digested fishes were often difficult to identify, as skates appeared to digest hard parts, like otoliths and bones that aid in identification Therefore, a majority of the prey category “Fish” contains highly digested fishes, but also several infrequently encountered species (e.g sockeye salmon Oncorhynchus nerka, three-spined stickleback Gasterosteus aculeatus) Because of regular occurrence of some fish species in skate stomachs, Pacific walleye pollock Theragra chalcogramma, capelin Mallotus villosus, flatfishes (Order Pleuronectiformes), Pacific sandlance Environ Biol Fish (2012) 95:3–20 Ammodytes hexapterus, dwarf wrymouth Cryptacanthodes aleutensis, eelpouts (Family Zoarcidae), and pricklebacks (Family Stichaeidae) were assigned to separate prey categories The numerical diet composition was chosen for statistical analysis because this measure best reflects feeding behavior (Hyslop 1980; Amundsen et al 1996) and there was high measurement error in the weights of prey found in various stages of digestion in skate stomachs Furthermore, distance matrices of the diet composition by number and by weight were strongly (rank) correlated indicating similar results from statistical analysis would be obtained by either measure Redundancy Analysis (RDA) was chosen as an appropriate multivariate statistical technique because it allows for components of dietary variation chosen a-priori to be examined and tested for statistical significance and was performed in the software CANOCO v4.5 (ter Braak and Smilauer 2002) Euclidean distance, upon which RDA is calculated, is not an appropriate resemblance measure for prey abundances because it is possible that the joint absence of a prey in two samples will produce less distance between the samples than two samples that share the same prey species in differing abundances (Orloci 1978) For this reason, prior to analysis, the diet data was transformed using the Hellinger transformation, in this case the square root of the proportional abundance of a prey category in a stomach sample, or: y;ij ¼ rffiffiffiffiffiffiffi yij yiþ where yij is the abundance of prey items of category j in sample i, and yi+ is the total abundance of all prey items in sample i (Legendre and Gallagher 2001) A matrix of Euclidean distances calculated on y’ij translates directly to a matrix of Hellinger distances: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u p rffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffi 2 uX y1j y2j À DHellinger ðx1 ; x2 Þ ¼ t y y 1þ 2þ j¼1 that range from a minimum distance of zero when two samples are identical in species abundances, pffiffiffi to a maximum distance of when two samples share no species in common (Legendre and Legendre 1998) Intraspecific and spatial diet variation Variance partitioning procedures were performed using RDA to differentiate significant spatial and intraspecific components of dietary variation in 2005 (Borcard et al 1992) Continuous variables included haul coordinates, skate size (TL), and depth (m) Sex (male, female) was the only categorical variable To account for potential non-linear relationships of prey with space, such as prey patches, haul coordinates (x, y) were entered as variables into a spatial matrix using the nine terms of the cubic surface trend regression (Borcard et al 1992) of the form: S ¼ x þ y þ xy þ x2 þ y2 þ x2 y þ xy2 þ x3 þ y3 Likewise, intraspecific variables were entered into a matrix along with 1st degree interaction terms: I ¼ TL þ Depth þ Sex þ TL  Depth þ TL  Sex þ Depth  Sex Forward selection of each set of explanatory variables was then performed by Monte Carlo permutations (10 000) to select only the variables explaining statistically significant components of variation in the diet Multiple partial RDAs were then used to partition the variance unique to, and shared between, each set of variables as described in Borcard et al (1992) Regional diet variation Homogeneity of multivariate dispersion in regional diet composition was statistically tested by permutation ANOVA (Anderson 2006; Oksanen et al 2011) The response variable was the dietary dispersion, calculated as the Hellinger distance between each sample and the spatial median of all samples in principal coordinates analysis (PCoA) space If statistically significant heterogeneity of variance was detected in the diet among regions, these results were interpreted qualitatively, using %PSIRI diagrams, since a permutation test on regional diet difference would fail to produce the correct null hypothesis If heterogeneity of variance in the diet among regions was not strong, then redundancy analysis (RDA) was computed with region as a factor and Monte Carlo permutations (10 000) were performed to indicate significance of the resulting model Environ Biol Fish (2012) 95:3–20 Results Sampling results In total, 270 Aleutian skate were sampled from 2005 to 2007 The index of vacuity was 6.6% indicating non-selective opportunistic foraging on various available prey resources Only the stomachs containing prey were analyzed further Females were caught more frequently than males (1.73:1.00) The size distribution was left-skewed for females (−0.657) and males (−1.028), with greater frequencies of largersized individuals (Fig 1) Male and female Aleutian skate, however, exhibited significantly different size distributions (K–S Z=2.075, 2-tailed, p=0.0003), with females being larger Smaller size classes of Aleutian skate (< ~80 cm TL) were largely absent from the sampled GOA shelf area, only beginning to appear at depths >~200 m and were predominantly male (Fig 1) Diet description Crustaceans were the dominant prey taxa of Aleutian Skate, followed by fishes (Table 1) Among the crustaceans, Pandalid shrimps and Tanner crab were Fig Size-depth distribution of female (circles) and male (crosses) skates Rug charts display the frequency distributions for total length (cm) and depth (m) of each sex (female=outside, male=inside) most frequently consumed (67.7% FO and 55.5% FO respectively) and comprised 33.7% PSIRI and 20.0% PSIRI of the total 74.6% PSIRI of all crustacean prey Euphausiids, when consumed (24.0% FO), contributed modestly (53.4% PN and 14.3% PW) to the diet composition Hippolytid and crangon shrimp were regularly consumed prey items (27% FO and 39% FO) but never occurred in great relative abundance within stomach contents Among identified fish prey, Pacific walleye pollock and capelin were the only two species consumed with any regularity (17.7% FO and 16.5% FO) The remaining identified fish prey was composed of various small forage species, i.e., Pacific sandlance, dwarf wrymouth, pricklebacks, eelpouts, Pacific sandfish Trichodon trichodon, and flathead sole Hippoglossoides elassodon A single large sockeye was recovered from the distended stomach of an individual skate Cephalopods were tertiary prey (16.9% FO) and did not contribute substantially to overall diet by weight or number (3.1% W and 2.3% N respectively) but did contribute modestly to the diet of those skates that consumed them (18.1% PW and 13.4% PN) Comparison to %IRI For comparison, %IRI was calculated at the taxonomic level of Class (%IRICLASS) and Family (%IRIFAMILY) (Table 1) These taxonomic levels represent conceivably realistic prey categories that two independent researchers might choose in order to assess relative importance in the diet of Aleutian skates The discrepancy between %IRICLASS and the sum of the corresponding %IRI FAMILY categories contained therein (Σ%IRIFAMILY) reflects the erroneous behavior of %IRI Notably, Class Crustacea displayed an increase from 85.5%IRICLASS to 94.3 Σ%IRIFAMILY, and Class Teleostei correspondingly decreased from 13.9%IRICLASS to 5.1 Σ%IRIFAMILY %PSIRI is additive with respect to taxonomic level (i.e % PSIRICLASS=Σ%PSIRIFAMILY within rounding errors) and more balanced with respect to %FO For example, the relative contributions of Class Crustacea to diet composition by %PSIRICLASS and Σ%PSIRIFAMILY (summed from Table 1.) were 74.6% and 74.5%, respectively Corresponding values of Class Teleostei were 22.1% when calculated using either taxonomic level of teleosts (Teleostei) Environ Biol Fish (2012) 95:169–183 DOI 10.1007/s10641-011-9827-x The neural control of feeding in elasmobranchs: A review and working model Leo S Demski Received: September 2010 / Accepted: April 2011 / Published online: May 2011 # Springer Science+Business Media B.V 2011 Abstract A working model of the neural control of feeding in elasmobranchs is presented and summarized in graphic form The model is based on a review of studies in sharks and batoids augmented by suggestions and comparisons from research in mammals and teleosts The focal point of the model is a proposed Hypothalamic Feeding Area (HFA) that encompasses the medial periventricular zone in the inferior lobe and a small area immediately dorsal to it Electrical stimulation in the HFA has evoked feeding in nurse sharks and neuropeptides and neurotransmitters known to influence feeding in mammals and teleosts have been localized immunocytochemically in the region in several elasmobranchs The HFA of elasmobranchs appears to be analogous to and possibly homologous with ‘hypothalamic feeding centers” in bony fishes and tetrapods Such “centers” are thought to integrate external and internal stimuli and control feeding in relation to available energy stores The HFA’s strong olfactory connections in elasmobranchs are consistent with smellinduced feeding activities In elasmobranchs, the HFA has reciprocal connections with the central pallium of the telencephalon, a region that processes visual, acoustic, mechanoreceptive and electroreceptive lateral L S Demski (*) Pritzker Marine Biology Research Center, Division of Natural Sciences, New College of Florida, 5800 Bayshore, Sarasota, FL 34243, USA e-mail: demski@ncf.edu line and possibly somatosensory information These pathways may provide multisensory control in feeding HFA connections with the cerebellum, brainstem and spinal cord most likely mediate hypothalamic coordination of the sensorimotor components of elasmobranch feeding The review and model help to identify areas for suggested research Keywords Shark Batoid Feeding Hyothalamus Brain Neuropeptides Introduction This paper reviews current understanding of the neural control of feeding in elasmobranchs It is based on a synthesis of the results of research on sharks and batoids and predictions derived from a comparative analysis of feeding control systems in mammals and teleost fishes The comparative aspects involve the identification of brain pathways in elasmobranchs that are anatomically/biochemically similar, if not homologous, to those controlling feeding in both mammals and teleosts as well as the “educated” assumption that such pathways have retained similar adaptive functions common to the primitive jawed ancestor of the three groups of vertebrates The outcomes of this review/analysis are summarized in a working model that is intended to serve as a focal point for initiating future research A synopsis of the paper’s four parts is given below as a 170 guide to its organization Parts 1–3 provide the supporting background for the model in Part An “up front” summary of the paper’s findings may be achieved by initially reading Part Part 1) reviews “classical” studies that identified the hypothalamus as a “center” controlling appetite and energy balance in mammals A “hypothalamic feeding area” (HFA) was later localized in both teleosts and sharks in a ventral expansion of the brain designated as the inferior lobe Based on functional/anatomical similarities to the teleosts, a similar, if not homologous, HFA was proposed for sharks The HFA of elasmobranchs is the foundation upon which the model is constructed Part 2) considers the more recent findings that a group of hypothalamic neuropeptides and the neurotransmitter serotonin control feeding in both mammals and teleosts For substances with common brain localization and function in mammals and teleosts, the anatomical/functional systems they represent are tentatively considered as conserved Some of the neurochemicals have distributions in elasmobranchs similar to those of the other two groups In such cases their suggested functions in mammals and teleosts have been extrapolated to elasmobranchs Part 3) details the anatomical connections of the elasmobranch HFA as a first approximation of a “wiring diagram” of a complex feeding control network The functions of HFA-associated areas are reviewed in the context of probable feeding-related activities Part 4) merges the analyses in the preceding components of the paper into a graphic model Part 1—Establishing an elasmobranch hypothalamic feeding area (HFA) Research in the 1950’s and 60’s (see reviews by Stevenson 1969; Peter 1979; Bernardis and Bellinger 1996) determined that the hypothalamus in the basal part of the mammalian diencephalon regulated feeding through individual controls for food intake and satiety More specifically, bilateral lesions in the ventromedial nucleus resulted in obesity and those in the lateral hypothalamic area (LHA) in weight loss Electrical stimulation of the LHA region evoked feeding in free-moving animals Electrophysiological and anatomical studies connected the two areas and suggested that hypothalamic feeding control was influenced by local glucose-sensitive neurons as well Environ Biol Fish (2012) 95:169–183 as visceral sensory and olfactory inputs Other studies indicated that feeding in normal situations of choice, reward and memory utilized extensive interconnections of the hypothalamus to the telencephalic “limbic system” and the central part of the midbrain Early studies on brain control of feeding in teleosts were attempts to discover mechanisms similar to those in mammals Feeding and related activities such as snapping-up the gravel substrate were evoked from the same area of the hypothalamus in two percomorph species, the bluegill sunfish (Lepomis macrochirus) and the black-chinned mouth breeder (Oreochromis macrocephala), using techniques of chronic brain stimulation adapted from mammal studies (Demski and Knigge 1971; Demski 1973; Demski 1983) The active region was centered in a ventral expansion of the hypothalamus known as the inferior lobe (IL) The most sensitive sites were adjacent and slightly dorsal to the more medial portion of the periventricular nucleus of lateral recess of the third ventricle (see anatomical details in Demski et al 1975; Evan et al 1976a) Similar stimulation results were obtained by Savage and Roberts (1975) in a cyprinid, the goldfish (Carassius auratus) As predicted, bilateral lesions in this part of the IL in goldfish resulted in a loss of feeding (Roberts and Savage 1978) and a lowering of growth rate (Peter 1979) The stimulation and lesion experiments in teleosts confirmed that a hypothalamic area, perhaps functionally analogous and possibly homologous to the LHA of mammals, was involved with feeding Chemosensory inputs to the area in the goldfish were also similar to those in mammalian hypothalamus, i.e., 1) feeding and related responses in free-swimming fish were elicited by olfactory tract and vagal gustatory lobe stimulation (Grimm 1959) and 2) neurons in the hypothalamic area were activated by electrical stimulation of both the olfactory tracts and the vagal gustatory lobes as well as by oral injections of food extracts (Demski 1981) Based on the above, the periventricular region of the IL was designated as the teleost “hypothalamic feeding area” or HFA (Demski 1981) The establishment of an HFA in teleosts with functions similar to those of the mammalian hypothalamus is especially important for extrapolation of the findings to elasmobranchs since the inferior lobes in both groups of fishes are similar (Smeets 1998) The general anatomy of the IL is known for a few sharks and batoids (Fig 1) In the piked dogfish Environ Biol Fish (2012) 95:169–183 (Squalus acanthias) the lateral recess is welldeveloped and it has a discrete periventricular nuclear layer comparable to that of teleosts (Northcutt 1978; Smeets et al 1983) A few scattered cell groups are present in both the lateral and dorsal regions of the IL The IL in the lesser spotted catshark (Scyliorhinus canicula) is similar but the lateral and dorsal IL areas have more cells (Smeets et al 1983) The condition in the nurse shark (Ginglymostoma cirratum) follows this pattern with the exception that the lateral recess and its surrounding nucleus are relatively shorter, i.e not extending as far laterally in the IL (see C in Fig 2; unpubl obs.) In batoids, which have inferior lobes of the same general size as the sharks, the lateral portion of the ventricle is even more restricted (Northcutt 1978; Ritchie et al 1983; Smeets et al 1983; Smeets and Boord 1985; Fiebig and Bleckmann 1989; Hofmann and Northcutt 2008) Thus, the presumed primary feeding area in the medial IL seems to be generally conservative across the elasmobranchs studied The functional significance of the apparent greater differences in the more peripheral areas of IL in these sharks and batoids is not known As a beginning exploration of feeding control in sharks, electrical stimulation studies similar to those in teleosts were carried out in nurse sharks (Fig 2; Demski 1977) Consistent feeding on cut pieces of fish was only evoked by stimulation in the medial periventricular nuclear region of the IL Other components of normal feeding (i.e snapping-up the gravel substrate and moving the barbels over the substrate) were elicited Fig Lateral (upper) and ventral (lower) views of the brain of the piked dogfish (Squalus acanthias) Scale bar equals 0.5cm; Abbreviations: CB—cerebellum; IL—inferior lobe of the hypothalamus; MD—medulla; OB—olfactory bulb; OC—optic chiasma; OL—optic lobe (optic tectum); SV—saccus vasculosus; TH—telencephalic hemisphere (Smeets 1998) 171 from sites dorsal to this area while only coughing responses were elicited from the more lateral area of the IL, the nucleus lobi lateralis of Smeets (1998) The combined stimulation results indicated that the areas for the strongest evoked feeding activities in both teleosts and elasmobranchs were quite similar anatomically; i.e medial cellular zones adjacent to or slightly dorsal to the lateral recess While developmental and additional comparative studies would be needed to establish their homology, the available results were clearly consistent with this possibility (Demski 1981) This concept of a similar if not homologous HFA in teleosts and elasmobranchs remains central to the model presented in this paper Limited information is available concerning the effects of brain lesions on feeding in elasmobranchs Studies in the late 1800’s and early 1900’s report that several sharks including what appear to be members of several common genera (Mustleus, Squalus and Scyliorhinus) are able to feed following the loss of the telencephalic hemispheres (Healey 1957; Aronson 1963) Part 2—The neurochemical substrates of HFA control The discovery of a battery of neuropeptides that facilitate (orexigenic) or inhibit (anorexigenic) appetite and control energy utilization has profoundly increased the understanding of vertebrate feeding mechanisms (see reviews by Kalra et al 1999; Arora and Anubhuti 2006; Magni et al 2009; Blouet and Schwartz 2010) In situ brain localization of the peptides has been accomplished using immunocytochemistry (ICC) and in some cases the receptors and gene sequences for some of the substances have been characterized using biochemical procedures ICC localization of the monoamine neurotransmitter serotonin has also provided significant information on appetite control and energy balance The studies on “feeding neurochemicals” have firmly established the hypothalamus of both mammals and teleosts as the major integrative “center” regulating feeding and energy balance It is beyond the purpose of this paper to review the wealth of information on neurochemicals that affect feeding Instead, substances (six neuropeptides and the neurotransmitter serotonin) involved in feeding in both mammals and teleosts, which have also been localized in elasmobranch brains, are considered from 172 Environ Biol Fish (2012) 95:169–183 Fig Transverse sections of the brain indicating sites and stimulation thresholds (frequency in Hz; current in μA) from which feeding and feeding-related behaviors have been evoked by electrical stimulation in free-swimming nurse sharks (Ginglymostoma cirratum) The insert drawing represents the lateral side of the whole brain and indicates the levels of sections A and B in the telencephalon and section C in the middle of the hypothalamic inferior lobe ▲—consistent mouthing or eating food; ●—inconsistent mouthing or eating food; /—snapping-up the gravel substrate; ─—side-to-side head movements with the barbels dragging; |—coughing; \—barbell movement; ■—negative for evoked behavior; CB—cerebellum; CN-central nucleus of the pallium; IL—inferior lobe of the hypothalamus; LP—lateral pallium; OL—optic lobe (optic tectum); OTP—area of the olfactory tract projection to the lateral pallium (see text and Ebbesson and Heimer 1970); SB— superficial basal area of the telencephalon; SP—subpallium; TH—telencephalic hemisphere; TM—tegmentum of the midbrain; VT—ventricle of the telencephalic hemisphere; V3— third ventricle; modified from Demski (1977) a comparative/functional point of view Several less well-studied feeding-related neuropeptides present in elasmobranchs are discussed briefly the lateral hypothalamus; 2) the brainstem nucleus of the solitary tract which receives visceral afferent information concerning gut stretch and taste; and 3) the telencephalic limbic system which is involved in many aspects of motivation Similarly in teleosts, NPY is present in the lateral tuberal nucleus (LTN) in the basal hypothalamus and other brain areas implicated in feeding (Volkoff et al 2005; Volkoff 2006) The arcuate nucleus and the LTN are probably homologous (Yada et al 2002; Volkoff et al 2005) In elasmobranch brains, ICC localization of NPY is only available for S canicula (Fig [NYP2]; Vallarino et al 1988a) and the cloudy catshark (Scyliorhinus torazame) (Fig [NYP3]; Chiba and Honma 1992; Chiba et al 2002) In both species the greatest density of ir-cells is in the ventromedial part of the IL The position is roughly equivalent to that of the arcuate Neuropeptide Y (NPY) NPY is a member of a widespread neuroendocrine peptide/receptor system involved in several aspects of feeding in mammals (Arora and Anubhuti 2006; Magni et al 2009; Blouet and Schwartz 2010) and teleosts (Narnaware et al 2000; Silverstein and Plisetskaya 2000; Kojima et al 2009) In mammals, NPY immunoreactive (ir) neurons in the arcuate nucleus in the ventromedial hypothalamus have connections throughout the hypothalamus, including several areas involved in feeding NPY receptors are also present in several ‘feeding-related” areas, i.e., 1) Environ Biol Fish (2012) 95:169–183 173 nucleus in mammals and tuberal area of teleosts (see above) Ir-cells and fibers (axons) are scattered throughout the IL and the basal areas and subpallium (SP) of the telencephalic hemisphere that connect to the IL (see details below) These IL and telencephalic regions also express NPY receptors in the dusky smoothhound (Mustelus canis) (McVey et al 1996) Some ir-cells in the IL contact the ventricular surface in both Scyliorhinus species Chiba and Honma (1992) illustrate an NPY “liquor-contacting neuron” with a bulbous process and a cilium, both of which extend into the ventricle Electron microscopic observations on the ventral floor of the IL in blacktip reef sharks (Carcharhinus melanopterus) reveal cilia-bearing periventricular cells with dense-cored granule-filled bulbous processes entering into the ventricle (Evan et al 1976b) The combined results support the possibility that NPY is released into the ventricle Chiba and Honma (1992) demonstrate NPY-ir cells and fibers in the dorsal/central pallium of the telencephalic hemispheres (hereafter referred to as the central pallium, CP) The CP expresses NPY receptors in M canis (McVey et al 1996) and telencephalic NPY increases following weeks of fasting in winter skates (Raja ocellata) (MacDonald and Volkoff 2009) The observations are consistent with NPY-mediated CP involvement in feeding (see further below) The results of the comparative analysis suggest that analogous, if not homologous, pathways exist for NPY control of feeding in mammals, teleosts and elasmobranchs Fig Transverse sections of the diencephalon of the lesser spotted catshark (Scyliorhinus canicula) (Sc) and the cloudy catshark (Scyliorhinus torazame) (St) illustrating the distributions of “feeding”-related neuropeptides; large symbols indicate immunreactive cell bodies; small dots represent immunreactive axons and terminal fields: Galanin (Sc) modified from Vallarino et al (1991); β-Endorphin [left side of section] (Sc) replotted from Vallarino et al (1989b); NPY1—neuropeptide Y [right side of section] (Sc) modified from Vallarino et al (1988a); NPY (St) modified from Chiba and Honma (1992); MCH—melanocyteconcentrating hormone (Sc) modified from Vallarino et al (1989a); α-MSH1—alpha-melanocyte-stimulating hormone (Sc) modified from Vallarino et al (1988b); α-MSH2 (St) modified from Chiba (2001) Abbreviations: IL-inferior lobe of the hypothalamus; LR-lateral recess of the third ventricle; OL-optic lobe (optic tectum); PT-posterior tuberal nucleus Galanin Galanin is an appetite-stimulating peptide widely distributed in the hypothalamus and other brain regions in mammals (Crawley et al 1990; Grundlach 2002; Arora and Anubhuti 2006) and teleosts (Volkoff et al 2005) Galanin is associated with NPY in the mammalian arcuate nucleus (Kalra et al 1999; Magni et al 2009) and its appetite-simulating effects are synergistic with those of NPY in goldfish (Volkoff and Peter 2001) Ir-galanin has been mapped in the brain of S canicula (Fig 3; Vallarino et al 1991) Positive cells 174 are located in: 1) the IL, both dorsal and ventral to the recess of the third ventricle; 2) the median eminence of the hypothalamus; and 3) the basal telencephalon, extending rostrally from the hypothalamus through the preoptic area into the SP Ir-fibers are concentrated in the areas with the positive cells The hypothalamic distribution of ir-galanin positive cells is generally similar to that for ir-NPY (Fig 3) The overlap probably has functional significance (see above) Melanocyte concentrating hormone (MCH) MCH is an appetite-enhancing peptide present in cells in the mammalian lateral hypothalamus (Magni et al 2009; Blouet and Schwartz 2010) Injection of MCH into the cerebral ventricles causes hyperphagia in rodents (Magni et al 2009) MCH is co-localized in the same neurosecretory granules with the anorexigenic agent α-MSH (see below) in rodent hypothalamic neurons (Powell and Baker 1987) In contrast, cerebral ventricular injections of MCH in goldfish suppress food intake and food restriction deceases ir-MCH in the nucleus of the lateral recess of the hypothalamus (Matsuda et al 2007; Matsuda 2009) Thus, while MCH is located in similar areas of the hypothalamus in mammals and teleosts, it appears to have significant but opposite effects on feeding in the two groups MCH ir-neurons have been localized in two areas within the boundaries of the HFA in S canicula (Fig 3; Vallarino et al 1989a) A group of positive cells is dorsomedial to the recess of the third ventricle in the posterior tuberal nucleus (PTN), an area that is partly mislabeled in the original paper Co-localization of MCN and α-MSH occurs in some of these cells Ir-cells are also located more ventrally in the LTN; some of these cells contact the ventricles Ir-fibers are prominent in the dorsal wall of the hypothalamus, the thalamus and the central midbrain (tegmentum) The similarity in the distribution of MCH in elasmobranchs with both tetrapods and teleosts is consistent with the peptide being involved in feeding in sharks and batoids Alpha-melanocyte stimulating hormone (α-MSH) α-MSH is an appetite-suppressing peptide hormone present in the arcuate nucleus in mammals (Adage et al 2001; Arora and Anubhuti 2006; Blouet and Schwartz 2010) and the LTN in teleosts (Volkoff et al 2005; Magni et al 2009) As mentioned above, these Environ Biol Fish (2012) 95:169–183 nuclei are probably homologous Along with the other melanocortins, ACTH (adrenocorticotropin) and βMSH and the orexigenic opioid β-endorphin (see further below), α-MSH is derived from the precursor molecule pro-opiomelanocortin (POMC) by tissue specific post-translational cleavage (Adage et al 2001; Volkoff et al 2005; Arora and Anubhuti 2006; Magni et al 2009) Thus, anatomically and functionally, αMSH control of feeding appears to be similar in both mammals and teleosts The findings are consistent with a conservative history for the feeding-related pathways In S canicula, ir-α-MSH is expressed in cells of the PTN (which is partly mislabeled in the original paper) where the peptide is co-localized with MCH in some of the neurons (Vallarino et al 1988b, 1989b) The presence of ir-α-MSH in the PTN was confirmed in later studies in S canicula (Molist et al 1992) and S torazame (Chiba 2001) In both species, scattered ir-cells were also identified in the walls of the “so-called” mammillary recess, a dorsocaudal prolongation of the third ventricle Some of the cells contact the ventricular surface In addition, α-MSH in S torazame is also expressed in the LTN situated in the periventricular zone surrounding the ventromedial part of the third ventricle and in the subjacent medial hypothalamic nucleus (Chiba 2001) Some of the periventricular LTN cells project to the ventricular surface The failure to report these latter ir-groups in S canicula may be due to a true species difference or more likely differences in the affinities of the antibodies used In both catshark species, α-MSH ir-axons are prevalent in the PTN and a few scattered ir-fibers are in the thalamus, midbrain and possibly other regions Chiba (2001) reports that most of the ir-fibers in S torazame extend from the hypothalamus rostrally into the medial basal telencephalon (septal and other subpallial areas) or move caudally through the brainstem The distribution of ir-α-MSH in Scyliorhinus is similar to that of both mammals and teleosts The anatomical findings suggest that hypothalamic α-MSH may also inhibit feeding in Scyliorhinus Studies of α-MSH distribution in other elasmobranchs reveal quite different patterns of distribution The NPT was immunonegative for α-MSH in the undulate ray (Raja undulata) (Molist et al 1992) Chiba (2001) was unable to find any ir-cells or fibers in the brain of the short-tail lantern shark (Etmopterus brachyurus) Rather, he illustrated ir-α-MSH in Environ Biol Fish (2012) 95:169–183 widespread glial astrocytes and tanycytes The differences in the results are hardly attributable to methodology since the same techniques and antibodies were applied to both S torazame and E brachyurus Further studies are needed to clarify the generality of these species differences β-Endorphin β-Endorphin is an orexigenic endogenous opioid in both mammals (Arora and Anubhuti 2006) and teleosts (Guijarro et al 1999) The peptide is produced in what are likely homologous cellular areas in the hypothalamus in mammals (the arcuate nucleus; Kalra et al 1999) and teleosts (the LTN; Vallarino 1985) At least in mammals, the appetite-enhancing effects of NPY and galanin may be mediated by β-endorphin release which likely provides the rewarding aspects of feeding (Levine and Billington 1989; Kalra et al 1999) Ir-β-endorphin is present in cells of the LTN and periventricular nucleus of the IL in S canicula (Vallarino et al 1989b) Ir-fibers are prevalent in the vicinity of the labeled cells as well as in the basal telencephalon and median eminence of the hypothalamus This distribution parallels those for mammals and teleosts and as such is consistent with possible similarities in functions regarding feeding behavior 175 Lethimonier et al 2004) In both teleosts and mammals, cGnRH ir-cells are present in a large nucleus near the midline in the midbrain tegmentum While ir-cGnRH II fibers are widespread in the brain including in the hypothalamus, their heaviest distribution is to the brainstem and spinal cord (Yamamoto 2003; Schneider and Rissman 2008) In skates, stingrays and sharks, the midbrain GnRH-ir nucleus is perhaps the largest group of GnRH-containing neurons found in any vertebrate (Wright and Demski 1991, 1993; Demski et al 1997) Forlano et al (2000) demonstrated that cGnRH II was the GnRH type in the midbrain nucleus in the Atlantic stingray (Dasyatis sabina) The distribution cGnRH II ir-fibers in this species is similar to that in mammals and teleosts (as above) with major projections to the midbrain, medulla and spinal cord and scattered projections to the hypothalamus and telencephalon There is a strong representation in sensory areas (the midbrain optic lobes and the primary lateral line nuclei of the medulla) The similarity of the distribution of cGnRH II in elasmobranchs with those of mammals and teleosts suggests that the peptide may also function in feeding/ reproductive co-ordination in sharks and batoids Modulation of various sensorimotor aspects of these activities is suggested by its brainstem projections Other feeding-related peptides Chicken gonadotropin-releasing hormone II (cGnRH II) cGnRH II is an anorexigenic peptide in mammals and teleosts Studies in the musk shrew (Suncus murinus), indicate that it mediates a balance between reproductive behavior and feeding by promoting the former while inhibiting the latter (Schneider and Rissman 2008) Female musk shrews have no obvious sexual cycle except as receptivity relates to food availability Given restrictions to their diet, they refuse the males; fed normally, they solicit them The key factor controlling this relationship is cGnRH II Results in goldfish are consistent with these findings Central administration of cGnRH II decreases the production and release of hypothalamic orexins, peptides that enhance feeding and inhibit sexual behavior (Hoskins et al 2008; Matsuda et al 2008; Matsuda 2009; Volkoff et al 2009) cGnRH II is conservative in its anatomical distribution in the vertebrates (King and Millar 1992, 1995; Sherwood and Lovejoy 1993; Forlano et al 2000; Cocaine- and amphetamine-regulated transcript (CART), cholecystokinin (CCK—produced by intestinal endocrine cells) and ghrelin (produced in the stomach) are three peptides involved in the central control of feeding in mammals (Kalra et al 1999; Arora and Anubhuti 2006; Magni et al 2009) and teleosts (Volkoff et al 2005; Volkoff 2006; Matsuda 2009; Riley et al 2009; Volkoff et al 2009) CART and CCK have been isolated from the brains of winter skates, Raja ocellata, (MacDonald and Volkoff 2009) and a ghrelin-like peptide has been isolated from the stomachs of C melanopterus and the scalloped hammerhead, (Sphyrna lewini) (Kawakoshi et al 2007) Functional studies are needed to determine if the peptides are involved in elasmobranch feeding Serotonin Serotonin is a neurotransmitter associated with satiety and energy balance systems in mammals (Leibowitz 176 and Alexander 1988) The anorexigenic influences are mediated through hypothalamic systems involving MCH (Heisler et al 2003), NPY, CCK and other “feeding-related peptides” (Arora and Anubhuti 2006) ICC studies indicate that serotonin-containing axons extend throughout the hypothalamus where they overlap receptors for the monoamine (Marvin et al 2010) Serotonin also inhibits feeding and influences energy homeostasis in goldfish by controlling other neuropeptides and growth hormone release (de Pedro et al 1998, 2006; Lin et al 2000; Mennigen et al 2009) The HFA in teleosts is densely innervated by serotoninergic fibers (Ekström and van Veen 1984; Batten et al 1993) The comparative results suggest similar systems for serotoninergic regulation of feeding in both groups of vertebrates The role of serotonin in elasmobranch feeding is unknown but its axonal distribution in the hypothalamus is similar to those reported above A heavy serotonin-ir innervation of IL is reported for S canicula (Carrera et al 2008), S torazame (Yamanaka et al 1990), and D sabina (Ritchie et al 1983) The similarity in the hypothalamic serotoninergic innervation in elasmobranchs with those of mammals and teleosts suggests that the neurotransmitter is probably involved in feeding in sharks and batoids Ventricular pathways in feeding control systems As indicated in previous sections, reports that feeding in mammals and teleosts is influenced by ventricular infusion of orexigenic and anorexigenic substances and observations of liquor-contacting cells immunreactive for NPY and α-MSH, suggest that secretion of “feeding-related” peptides into the ventricles may be pathways for feeding control Bleb-like protrusions in the wall of the lateral recess in the IL of both teleosts (Evan et al 1976a) and sharks (Evan et al 1976b) could be substrates for such mechanisms Environ Biol Fish (2012) 95:169–183 available to the “hungry” fish to use in locating the food; capture of the “prey” and determining if it is indeed to be eaten or rejected; control of various aspects of digestion and metabolism; remembering the feeding activity for future reference; and regulating feeding in the context of competing behavior e.g reproductive responses, territorial defense, etc (for examples of such behaviors, see papers in this symposium and Heithaus 2004; Hueter et al 2004; Motta 2004; Wetherbee and Cortés 2004) One approach to understanding this wider control is to start with the HFA and expand outward from it by defining its connections with other areas of the central nervous system The idea is simply to build a wiring diagram as a first approximation of a widespread feeding network Looking at the non-hypothalamic connections of these “secondary” regions may also be advantageous; e.g., if an area receives direct input from a primary sensory region, it is likely that its activities are influenced by stimuli in that modality A similar strategy was used to produce models of the comparative neural control of color change (Demski 1992) Experimental anatomical studies on the IL connections utilize “neural tracers” implanted or injected into an area and allowed to remain in place in the surviving animal The tracers can be taken-up by neuron cell bodies, their axons or the terminals of the axons Depending on the conditions, the tracers can move in the cytoplasm toward and into the cell bodies (retrograde transport), toward the axon terminations (anterograde transport), or in both directions The tracers are visualized in tissue sections using histochemical reactions (see review by ten Donkelaar and Nicholson 1998) The following subsections review the anatomical bases for the association of each region with HFA and provide speculation on the functions of the “related areas” in the controlling feeding The information from the subsections is incorporated into the “model” presented in Part of the paper Central pallium (CP) Part 3—A neural network model of feeding control in elasmobranchs As in other vertebrates, hypothalamic control of feeding in elasmobranchs is not autonomous but rather finely integrated in a complex of neural systems that mediate a variety of activities including: motivation to feed; analysis of sensory information In clearnose skates (Raja eglanteria), retrograde transport from IL tracer injections labels many neurons in the CP while anterograde transport from the same implants reveals a small projection from the IL back to the dorsal pallium (Fig 4; Smeets and Boord 1985) The former connection was confirmed by anterograde transport from tracer implants in the CP in S acanthias Environ Biol Fish (2012) 95:169–183 Fig Representative transverse sections of the brain of the clearnose skate (Raja eglanteria) illustrating the transport of the neural tracer horseradish peroxidase (HRP) to the telencephalic hemispheres (A is rostral to B) from an implant in the inferior lobe of the hypothalamus (C); black dots—indicate retrogradely-filled neurons that project to the inferior lobe; dashed lines—represent anterogradely-labeled fibers (axons) of IL cells; CP—central pallium (mostly dorsal pallial areas); LP—lateral pallium; IL—inferior lobe of the hypothalamus; SP-subpallium; modified from Smeets and Boord (1985) (Smeets and Northcutt 1987) The reciprocal connections between the CP and the IL were verified in the thornback skate (Platyrhinoidis triseriata) Implants of the tracer into the CP anterogradely filled the IL with a dense fiber plexus and retrogradely filled a few cells in a small nucleus in the most dorsal part of the IL (Fig 5; Hofmann and Northcutt 2008) The above results strongly suggest that the CP exerts a powerful influence on IL functions and that hypothalamic feedback to the CP is also a feature this control Most of the telencephalon of elasmobranchs, including the CP, was traditionally considered to be dominated by second and third order olfactory inputs (Aronson 1963) More recent studies using experimental methods indicate that only the LP and SP have strong relationships with the olfactory bulbs (Hofmann and Northcutt 2008) In contrast, the CP has probably an equal if not greater representation from other senses including: 1) vision (Cohen et al 1973; Platt et al 1974; Graeber et al 1978; Bullock and Corwin 1979; Luiten 1981; Bodznick and Northcutt 1984; Bleckmann et al 1987; 177 Fig Representative transverse sections of the brain of the thornback skate (Platyrhinoidis triseriata) illustrating the transport of the neural tracer biotinylated dextran amine (BDA) to the inferior lobe of the hypothalamus and other diencephalic areas (B is rostral to C) from an implant in the central part of the pallium in the telencephalic hemispheres (A); black dots—indicate retrogradely-filled neurons that project to the central pallium; dashed lines—represent anterogradely-labeled fibers (axons) of central pallial cells that project to other areas; CP central pallium (central nucleus of the dorsal pallium); Hyd-dorsal hypothalamic nucleus of the inferior lobe; Lmn-lateral mesencephalic nucleus; IL-inferior lobe of the hypothalamus; Pl-lateral pallium; Pltposterior thalamic nucleus; Tm-tectum mesencephali (optic lobe); modified from Hofmann and Northcutt (2008) Karamürsel and Bullock 1994); 2) hearing (Bullock and Corwin 1979); 3) electroreception (Platt et al 1974; Bullock 1979; Bullock and Corwin 1979; Bodznick and Northcutt 1984; Bleckmann et al 1987); 4) mechanosensory lateral line function (Platt et al 1974; Bleckmann et al 1987); and 5) somatosensory input from the head (see cephalic control in Demski 1977) Local responses to more than one sensory modality are typical of the CP (Platt et al 1974; Bullock and Corwin 1979; Bodznick and Northcutt 1984; Bleckmann et al 1987; Hofmann 2001) The available evidence suggests that the CP provides multisensory control of behaviors including feeding via connections with the IL Lateral pallium (LP) The LP is the main target of central input from the olfactory bulb in both sharks and batoids and as such it is likely at least partly homologous to the lateral pallium of mammals, which is the general area that gives rise to the olfactory cortex of tetrapods (Ebbesson 178 and Heimer 1970; Bruckmoser and Dieringer 1973; Smeets 1983; Dryer and Graziadei 1994; Hofmann and Northcutt 2008) The LP in P triseriata has reciprocal connections to the CP, SP and the dorsal IL (Hofmann and Northcutt 2008) These pathways likely provide the IL with a tertiary olfactory input which may be modulated by feedback via the IL projection back to the LP In addition, the LP inputs to the other areas of the hemisphere (SP, CP) with connections to the IL may provide the hypothalamus with olfactory information “seen” in a multisensory context Consistent with the above possibilities are observations that reactions to feeding-related odors increased in reef sharks as a result of starvation or temporary blindness (reviewed by Tester 1963a, b; Hueter et al 2004) Subpallium (SP) The SP receives an olfactory-related input from both the olfactory bulb (Ebbesson and Heimer 1970; Smeets 1983; Dryer and Graziadei 1994; Hofmann and Northcutt 2008) and the LP (Hofmann and Northcutt 2008) It has reciprocal connections with the CP (Hofmann and Northcutt 2008) and the IL (Smeets and Boord 1985) The connections of the SP compare well to those of the septal area of the tetrapod limbic system which is involved in olfactory influenced behaviors including feeding in mammals (Magni et al 2009) They also are consistent with the observation that electrical stimulation in the CP in nurse sharks evoked feeding-related and other head movements similar to those elicited from the IL (Fig 2; Demski 1977) Cerebellum Anterograde transport from IL tracer implants reveals an IL projection to the cerebellar cortex in clearnose skates, (Smeets and Boord 1985) Various sensorimotor functions of the cerebellum in elasmobranchs have been suggested, including functions in motor coordination and postural and equilibrium control, but the mediation of these activities in itself does not adequately account for the great differences in cerebellar development found within the elasmobranchs (Northcutt 1989; Smeets 1998; Yopak et al 2007; Lisney et al 2008) Thus, the function of the IL projection to the cerebellum remains unclear Environ Biol Fish (2012) 95:169–183 Diencephalon/Brainstem Smeets and Boord (1985) identified several brainstem regions with connections to the IL in R eglanteria In this species, the lateral tegmental nucleus of the midbrain has reciprocal ipsilateral connections with the IL The structure also has a bidirectional relationship with the multisensory optic lobe (optic tectum) in both S canicula and the thornback ray (R clavata) (Smeets 1981, 1982) and receives input from the spinal cord in S canicula (Hayle 1973) The lateral tegmental nucleus could provide the IL with visual, somatosensory and perhaps other types of information useful in feeding In P triseriata (Schweitzer and Lowe 1984), the LTN in the dorsal part of the IL receives an input from a cell group in the dorsal midbrain-diencephalic junction (Nucleus “A”) that is sensitive to both lateral line mechanosensory and electrosensory stimulation (Boord and Northcutt 1983; Bodznick and Northcutt 1984; Bleckmann et al 1987) As the connection predicts, the LTN responds to electrosensory stimulation in the little skate (Raja erinacea) (Bodznick and Northcutt 1984) and to both lateral line modalities in P triseriata (Bleckmann et al 1987) Feeding in elasmobranchs is influenced by both mechanosensory and electrosensory information (Kalmijn 2000; Hueter et al 2004; Gardiner and Atema 2007; Kajiura et al 2010) Other cell groups projecting to the IL in R eglanteria include: a nucleus “F” (of unknown function) situated slightly ventral to the cerebellum and scattered cells in the reticular formation and superior raphé nucleus (Smeets and Boord 1985) Based on extrapolation from hypothalamic studies in mammals (Bernardis and Bellinger 1996; Magni et al 2009; Blouet and Schwartz 2010), these pathways to the IL in elasmobranchs may represent visceral sensory inputs (e.g., from taste and gut-stretch receptors) and serotoninergic projections that have broad modulatory effects on many behaviors including feeding Neural tracer studies reveal well-developed HFA efferent connections to lower areas of the CNS In R eglanteria, the IL projects to the brainstem (various reticular cranial nerve nuclei) via fibers in the tractus lobobulbaris (Smeets and Boord 1985) The efferent system appears to be equivalent to the tractus lobobulbaris in teleosts (Demski unpubl.; Smeets 1998) The dorsomedial periventricular nucleus of the hypothalamus is continuous ventrally with the Environ Biol Fish (2012) 95:169–183 179 Comparative considerations and direct evidence from studies in elasmobranchs support the concept that a “hypothalamic feeding area” (HFA) in the inferior lobe (IL) anchors a widespread network of brain regions controlling feeding in sharks and batoids (Fig 6) As in mammals and teleosts, the hypothalamic localization of a suite of orexigenic (NPY; MCH; galanin; and β-endorphin) and anorexigenic neuropeptides (α-MSH; cGnRH II) and/or their receptors is considered a critical feature of HFA control of food intake and energy balance in elasmobranchs NPY and its receptors, galanin and β-endorphin are also present in other areas that have well-defined connections to the HFA (see below) Serotonin is a neurotransmitter that inhibits feeding in mammals and teleosts via hypothalamic systems In elasmobranchs, serotonin-ir fibers innervate most of the IL The observations suggest that hypothalamic serotonin is involved in elasmobranch feeding Defining the anatomical connections of the HFA provides a first “sketch” of a feeding control network The functional roles assigned to the HFA-associated regions are based on their non-hypothalamic connections and, when available, the results of electrophysiologial studies Concerning the telencephalic hemispheres, the IL has direct bidirectional connections with: 1) the central pallium (CP), the major multisensory integra- Fig Schematic summary of the working model of the neural control of feeding in elasmobranchs presented in the text plotted on an outline of a midsagittal section of the shark brain; black dots represent groups of cells that project to other areas of the brain via the solid lines ending in the Y’s; arrows pointing back along the lines toward the dots denote a small projection in the opposite direction; lines with Y’s on both ends indicate approximately equal reciprocal connections between two areas; probable functions in italics are based on information in the text; locations for serotonin and neuropeptides involved in feeding are given for only the primary regions discussed in text; α-MSH—alpha-melanocyte-stimulating hormone; AN—anterior nucleus of the mesencephalic-diencephalic boundary; β— Endorphin-beta-endorphin; CB—cerebellum; cGnRH II— chicken gonadotropin-releasing hormone II; CP—central pallium (central nucleus of the dorsal pallium); HFA—hypothalamic feeding area in inferior lobe of the hypothalamus; MCH— melanocyte-concentrating hormone; NPY—neuropeptide Y; OB—olfactory bulb; OC—optic chiasma; OL—optic lobe (optic tectum); LP—lateral pallium; LTN—lateral tegmental nucleus of the midbrain; RN—reticular formation nuclei of the brainstem; SP—subpallium; TH—telencephalic hemisphere periventricular nucleus of the IL and as such is within the dorsal limits of the HFA The region projects to the spinal cord in both S canicula (Smeets and Timerick 1981; Timerick et al 1992) and R clavata (Smeets and Timerick 1981) The exact positions of the axons of the spinal pathway were not mapped in these tracer studies The descending efferent pathways of the HFA most likely mediate hypothalamic control of visceral and somatic components of feeding behavior In this respect, some of the IL axons in R eglanteria appear to contact neurons in both the facial and trigeminal motor nuclei (Smeets and Boord 1985) These cells control mouth and head movements involved in feeding Part 4—A summary of the model 180 tive center of the telencephalon; 2) the subpallium (SP), an area with strong olfactory associations as well as reciprocal connections with the CP; and 3) the lateral pallium (LP), the main target of olfactory bulb input Presumably, the hemispheric pathways give the HFA priority access to primary olfactory information as well as a multisensory ‘view’ of the animal’s world IL connections with the lateral line-dominated anterior nucleus (AN) in the diencephalon/midbrain boundary provide a pathway for electrosensory and mechanosensory inputs to the HFA Visual and somatosensory input to the HFA is likely through a reciprocal relationship with the midbrain lateral tegmental nucleus The lateral tegmental nucleus has a bidirectional relationship with the optic lobe (midbrain tectum), a complex layered structure that provides primary multisensory analysis and sensorimotor integration Thus, the HFA appears to assess information from primary sensory and multisensory integrative centers which it may in turn influence via reciprocal connections Bidirectional connections with other brainstem sensory and motor nuclei provide the IL with additional substrates for controlling both somatic and visceral components of feeding behavior Speculation on the role of an IL projection to the cerebellum in feeding must await more information on cerebellar function in elasmobranchs Collectively, the findings reviewed and analyzed in this paper support a model featuring the HFA as the focal integrative center in a widespread network of neural systems involved in food intake, metabolism and energy management in elasmobranchs As hypothesized herein: the HFA “sees” both within and outside the animal and appropriately provides for current and future energy needs through appetite and metabolic controls It may also be the initiator and coordinator of the varied behavior patterns involved in feeding Acknowledgements Support for the author’s studies on elasmobranchs has been provided by the United States Atomic Energy Commission, the National Science Foundation and the Florsheim Chair endowment to the New College Foundation References Adage T, Scheurik AJW, de Boer SF, de Vries K, Konsman JP, Kuipers F, Adan RAH, Baskin DG, Schwartz MW, van Dijk G (2001) Hypothalamic, metabolic, and behavioral responses to pharmacological inhibition of CNS melanocortin signaling in rats J Neurosci 21:3639–3645 Environ Biol Fish (2012) 95:169–183 Aronson LR (1963) The central nervous system of sharks and bony fishes with special reference to sensory and integrative mechanisms In: Gilbert PW (ed) Sharks and survival D.C Heath Co., Boston, pp 165–241 Arora S, Anubhuti (2006) Role of neuropeptides in appetite regulation and obesity—a review Neuropeptides 40:375–401 Batten TFC, Berry PA, Maqbool A, Moons L, Vandesande F (1993) Immunolocalization of catechoamine enzymes, serotonin, dopamine and l-dopa in the brain of Dicentrarchus labrax (Teleostei) Brain Res Bull 31:233–252 Bernardis LL, Bellinger LL (1996) The lateral hypothalamic area revisited: ingestive behavior Neurosci Biobehav Rev 20:189–287 Bleckmann H, Bullock TH, Jorgensen JM (1987) The lateral line mechanoreceptive mesencephalic, diencephalic, and telencephalic regions in the thornback ray, Platyrhinoidis triseriata (Elasmobranchii) J Comp Physiol A 161:67–84 Blouet C, Schwartz GJ (2010) Hypothalamic nutrient sensing in the control of energy homeostasis Behav Brain Res 209:1–12 Bodznick D, Northcutt RG (1984) An electrosensory area in the telencephalon of the little skate, Raja erinacea Brain Res 298:117–124 Boord RL, Northcutt RG (1983) Diencephalic and mesencephalic electrosensory centers of the clearnose skate Raja eglanteria Am Zool 23:927 Bruckmoser P, Dieringer N (1973) Evoked potentials in the primary and secondary olfactory projection areas of the forebrain in Elasmobranchia J Comp Physiol 87:65–74 Bullock TH (1979) Processing of ampullary input in the brain: comparison of sensitivity and evoked responses among elasmobranchs and siluriform fishes J Physiol Paris 75:397–407 Bullock TH, Corwin JT (1979) Acoustic evoked activity in the brain in sharks J Comp Physiol 129:223–234 Carrera I, Molist P, Anadón R, Rodríguez-Moldes I (2008) Development of the serotoninergic system in the central nervous system of a shark, the lesser spotted dogfish Scyliorhinus canicula J Comp Neurol 511:804–831 Chiba A (2001) Marked distributional difference of alphamelanocyte-stimulating hormone (alpha-MSH)-like immunoreactivity in the brain between two elasmobranchs (Scyliorhinus torazame and Etmopterus brachyurus): an immunohistochemical study Gen Comp Endocrinol 122:287–295 Chiba A, Honma Y (1992) Distribution of neuropeptide Y-like immunoreactivity in the brain and hypophysis of the cloudy dogfish, Scyliorhinus torazame Cell Tissue Res 268:453–461 Chiba A, Oka S, Saitoh E (2002) Ontogenetic changes in neuropeptide Y-immunoreactive cerebrospinal fluidcontacting neurons in the hypothalamus of the cloudy dogfish, Scyliorhinus torazame (Elasmobranchii) Neurosci Lett 329:301–304 Cohen DH, Duff TA, Ebbesson SOE (1973) Electrophysiological identification of a visual area in shark telencephalon Science 182:492–494 Crawley JN, Austin MC, Fiske SM, Martin B, Consolo S, Berthold M, Langel U, Fisone G, Barfai T (1990) Activity of centrally administered galanin fragments on stimulation Environ Biol Fish (2012) 95:169–183 of feeding behavior and on galanin receptor binding in the rat hypothalamus J Neurosci 10:3695–3700 de Pedro N, Pinillos ML, Valenciano AI, Alonso-Bedate M, Delgado MJ (1998) Inhibitory effect of serotonin on feeding behavior in goldfish: involvement of CRF Peptides 19:505–511 de Pedro N, Martínez-Álvarez R, Delgado MJ (2006) Acute and chronic leptin reduces food intake and body weight in goldfish (Carassius auratus) J Endocrinol 188:513–520 Demski LS (1973) Feeding and aggressive behavior evoked by hypothalamic stimulation in a cichlid fish Comp Biochem Physiol 44A:685–692 Demski LS (1977) Electrical stimulation of the shark brain Am Zool 17:487–500 Demski LS (1981) Hypothalamic mechanisms of feeding in fishes In: Laming P (ed) Brain mechanisms of behaviour in lower vertebrates Society for Experimental Biology Seminar Series, Cambridge University Press, pp 225–237 Demski LS (1983) Behavioral effects of electrical stimulation of the brain In: Davis RE, Northcutt RG (eds) Fish neurobiology and behavior vol University of Michigan Press, Ann Arbor, pp 317–359 Demski LS (1992) Chromatomotor systems in teleosts and cephalopods: a levels oriented analysis of convergent systems Brain Behav & Evol 40:141–156 Demski LS, Knigge KM (1971) The telencephalon and hypothalamus of the bluegill (Lepomis macrochirus): evoked feeding, aggressive and reproductive behavior with representative frontal sections J Comp Neurol 143:1–16 Demski LS, Evan AP, Saland LC (1975) The structure of the inferior lobe of the teleost hypothalamus J Comp Neurol 161:483–498 Demski LS, Beaver JA, Sudberry JJ, Custis JR (1997) GnRH systems in cartilaginous fishes In: Parhar IS, Sakuma Y (eds) GnRH neurons: gene to behavior Brain Suppan, Tokyo, pp 123–143 Dryer L, Graziadei PPC (1994) Projections of the olfactory bulb in an elasmobranch fish, Sphyrna tiburo: segregation of inputs in the telencephalon Anat Embryol 190:563–572 Ebbesson SOE, Heimer L (1970) Projections of the olfactory tract fibers in the nurse shark (Ginglymostoma cirratum) Brain Res 17:47–55 Ekström P, van Veen T (1984) Distribution of 5-hydroxytrptamine (serotonin) in the brain of the teleost Gasterosteus aculeatus L J Comp Neurol 266:307–320 Evan AP, Demski LS, Saland LC (1976a) The lateral recess of the third ventricle in teleosts: an EM and Golgi study Cell and Tissue Res 166:521–530 Evan AP, Saland LC, Demski LS (1976b) The inferior lobe of the shark hypothalamus: a scanning and transmission EM study J Morph 150:59–78 Fiebig E, Bleckmann H (1989) Cell groups afferent to the telencephalon in a cartilaginous fish (Platyrhinoidis triseriata) a WGA-HRP study Neurosci Lett 105:57–62 Forlano PM, Maruska KP, Sower SA, King JA, Tricas TC (2000) Differential distribution of gonadotropin-releasing hormone-immunreactive neurons in the stingray brain: functional and evolutionary considerations Gen Comp Endocrinol 118:226–248 Gardiner JM, Atema J (2007) Sharks need the lateral line to locate odor sources: rheotaxis and eddy chemotaxis J Exp Biol 210:1925–1934 181 Graeber RC, Schroeder DM, Jane JA, Ebbesson SOE (1978) Visual discrimination following partial telencephalic ablations in nurse sharks (Ginglymostoma cirratum) J Comp Neurol 180:325–344 Grimm RJ (1959) Feeding behavior and electrical stimulation of the brain of Carassius auratus Science 131:162 Grundlach AL (2002) Galanin/GALP and galanin receptors: role in central control of feeding, body weight/obesity and reproduction? Eur J Pharmacol 440:255–268 Guijarro AI, Delgado MJ, Pinillos ML, López-Patiño MA, Alonso-Bedate, de Pedro N (1999) Galanin and βendorphin as feeding regulators in cyprinds: effect of temperature Aquaculture Res 30:483–489 Hayle TH (1973) A comparative study of the spinal projections to the brain (except cerebellum) in three classes of poikiothermic vertebrates J Comp Neurol 149:463–476 Healey EG (1957) The nervous system In: Brown ME (ed) The physiology of fishes vol Academic, New York, pp 1– 119 Heisler LK, Crowley MA, Kishi T, Tecott LH, Fan W, Low MJ, Smart JL, Rubinstein M, Tatro JB, Zigman JM, Cone RD, Elmquist JK (2003) Central serotonin and melanocortin pathways regulating energy homeostasis Ann NY Acad Sci 994:169–174 Heithaus MR (2004) Predator-prey relationships In: Carrier JC, Musick JA, Heithaus MR (eds) Biology of sharks and their relatives CRC, Boca Raton, pp 487–521 Hofmann MH (2001) The role of the fish telencephalon in sensory information processing In: Kapoor BG, Hara TJ (eds) Sensory biology of jawed fishes; new insights Science, Enfield, pp 255–274 Hofmann MH, Northcutt RG (2008) Organization of major telencephalic pathways in an elasmobranch, the thornback ray Platyrhinoidis triseriata Brain, Behav Evol 72:307– 325 Hoskins LJ, Xu M, Volkoff H (2008) Interactions between gonadotropin-releasing hormone (GnRH) and orexin in the regulation of feeding and reproduction in goldfish (Carassius auratus) Horm Behav 54:379–385 Hueter RE, Mann DA, Maruska KP, Sisneros JA, Demski LS (2004) Sensory biology of elasmobranchs In: Carrier JC, Musick JA, Heithaus MR (eds) Biology of sharks and their relatives CRC, Boca Raton, pp 350–358 Kajiura SM, Cornett AD, Yopak KE (2010) Sensory adaptations to the environment: electroreceptors as a case study In: Carrier JC, Musick JA, Heithaus MR (eds) Sharks and their relatives II: biodiversity, adaptive physiology, and conservation CRC, Boca Raton, pp 393–433 Kalmijn AD (2000) Detection and processing of electromagnetic and near-field acoustic signals in elasmobranch fishes Phil Trans Royal Soc London Ser B, Biol Sci 355:1135–1141 Kalra SP, Dube MG, Pu S, Xu B, Horvath TL, Kalra PS (1999) Hypothalamic appetite-regulating pathways in the hypothalamic regulation of body weight Endocr Rev 20:68–100 Karamürsel S, Bullock TH (1994) Dynamics of event-related potentials to trains of light and dark flashes: responses to missing and extra stimuli in elasmobranch fish Electroenceph Clin Neurophysiol 90:461–471 Kawakoshi A, Kaiya H, Riley LG, Hirano T, Grau EG, Miyazato M, Hosoda H, Kangawa K (2007) Identification 182 of a ghrelin-like peptide in two species of shark, Sphyrna lewini and Carcharhinus melanopterus Gen Comp Endocrinol 151:259–268 King JA, Millar RP (1992) Evolution of gonadotropin-releasing hormones Trends Endocrinol Metab 3:339–346 King JA, Millar RP (1995) Evolutionary aspects of gonadotropin-releasing hormone and its receptor Cell Mol Neurobiol 15:5–23 Kojima K, Kamijo M, Kageyama H, Uchiyama M, Matsuda K (2009) Neuronal relationship between orexin-A- and neuropeptide Y-induced orexigenic actions in goldfish Neuropeptides 43:63–71 Leibowitz SF, Alexander JT (1988) Hypothalamic serotonin in control of eating behavior, meal size, and body weight Biol Psychiatry 44:851–864 Lethimonier C, Madigou T, Muñoz J-A, Lareyre J-J, Kah O (2004) Evolutionary aspects of GnRHs, GnRH neuronal systems and GnRH receptors in teleost fish Gen Comp Endocrinol 135:1–16 Levine AS, Billington CJ (1989) Opiods are they regulators of feeding? Ann NY Acad Sci 575:209–219 Lin X, Volkoff H, Narnaware Y, Bernier NJ, Peyon P, Peter RE (2000) Brain regulation of feeding behavior and food intake in fish Comp Biochem Physiol A 126:415–434 Lisney TJ, Yopak KE, Montgomery JC, Collin SP (2008) Variation in brain organization and cerebellar foliation in chondrichthyans: batoids Brain Behav Evolution 72:262–282 Luiten PGM (1981) Two visual pathways to the telencephalon in the nurse shark (Ginglymostoma cirratum): II ascending thalamo-telencephalic connections J Comp Neurol 196:539–548 MacDonald E, Volkoff H (2009) Neuropeptide Y (NPY), cocaineand amphetamine-regulated transcript (CART) and cholecystokinin (CCK) in winter skate (Raja ocellata): CDNA cloning, tissue distribution and mRNA expression responses to fasting Gen Comp Endocrinol 161:252–261 Magni P, Dozio E, Massimilano R, Celolli F, Masini MA, Prato P, Broccoli M, Mambro A, Massimo M, Strollo F (2009) Feeding behavior in mammals and humans Ann NY Acad Sci 1163:221–232 Marvin E, Scrogin K, Dudás B (2010) Morphology and distribution of neurons expressing serotonin 5-HT1A receptors in rat hypothalamus and the surrounding diencephalic and telencephalic areas J Chem Neuroanat 39:235–241 Matsuda (2009) Recent advances in the regulation of feeding behavior by neuropeptides in fish Ann NY Acad Sci 1163:241–250 Matsuda K, Shimakura S-I, Miura T, Maruyama K, Uchiyama M, Kawauchi H, Shioda S, Takahashi A (2007) Feedinginduced changes of melanin-concentrating hormone (MCH)-like immunoreactivity in the goldfish brain Cell Tissue Res 328:375–382 Matsuda K, Nakamura K, Shimakura S-I, Miura T, Kageyama H, Uchiyama M, Shioda S, Ando H (2008) Inhibitory effect of gonadotropin-releasing hormone II on food intake in the goldfish, Carassius auratus Horm Behav 54:83–89 McVey DC, Rittschof D, Mannon PJ, Vigna SR (1996) Localization and characterization of neuropeptide Y/ peptide YY receptors in the brain of the smooth dogfish (Mustelis canis) Regul Pept 61:167–173 Environ Biol Fish (2012) 95:169–183 Mennigen JA, Harris EA, Chang JP, Moon TW, Trudeau VL (2009) Fluoxetine affects weight gain and expression of feeding peptides in the female goldfish brain Regul Pept 155:99–109 Molist P, Rodríguez-Moldes I, Anadón R (1992) Immunocytochemical and electron-microscopic study of the elasmobranch nucleus sacci vasculosi Cell Tissue Res 270:395– 404 Motta PJ (2004) Prey capture behavior and feeding mechanics of elasmobranchs In: Carrier JC, Musick JA, Heithaus MR (eds) Biology of sharks and their relatives CRC, Boca Raton, pp 165–202 Narnaware YK, Peyon PP, Lin X, Peter RE (2000) Regulation of food intake by neuropeptide Y in goldfish Am J Physiol, Regul Integr Comp Physiol 279:R1025–R1034 Northcutt RG (1978) Brain organization in the cartilaginous fishes In: Hodgson ES, Mathewson RF (eds) Sensory biology of sharks skates and rays Office of Naval Research, Arlington, pp 117–193 Northcutt RG (1989) Brain variation and phylogenetic trends in elasmobranch fishes J Exp Zool Suppl 2:83–100 Peter RE (1979) The brain and feeding behavior In: Hoar WS, Randell DJ, Brett JR (eds) Fish physiology vol VIII Academic, New York, pp 121–159 Platt CJ, Bullock TH, Czéh G, Kovačević N, Konjević Dj K, Gojković M (1974) Comparison of electroreceptor, mechanoreceptor, and optic evoked potentials in the brain of some rays and sharks J Comp Physiol 95:323–355 Powell KA, Baker BI (1987) Ultrastructural demonstration that melanin-concentrating hormone-like and α-melanocyte stimulating hormone-like immunoreactive molecules co-exist in the same neurosecretory granules Neurosci Lett 268–274 Riley LG, Walker AP, Dorough CP, Schwandt SE, Grau EG (2009) Glucose regulates ghrelin, neuropeptide Y, and the GH/IGF-I axis in the tilapia Comp Biochem Physiol A 154:541–546 Ritchie TC, Livingston CA, Hughes MG, McAdoo DJ, Leonard RB (1983) The distribution of serotonin in the CNS of an elasmobranch fish: immunocytochemical studies in the Atlantic stingray, Dasyatis sabina J Comp Neurol 221:429–443 Roberts MG, Savage GE (1978) Effects of hypothalamic lesions on food intake in the goldfish (Carassius auratus) Brain Behav Evol 15:150–164 Savage GE, Roberts MG (1975) Behavioural effects of electrical stimulation of the hypothalamus of the goldfish (Carassius auratus) Brain Behav Evol 12:42–56 Schneider JS, Rissman EF (2008) Gonadotropin-releasing hormone II: a multi-purpose neuropeptide Integr Comp Biol 48:588–595 Schweitzer J, Lowe D (1984) Mesencephalic and diencephalic cobalt-lysine injections in an elasmobranch: evidence for two parallel electrosensory pathways Neurosci Lett 44:317–322 Sherwood NM, Lovejoy DA (1993) Gonadotropin-releasing hormone in cartilaginous fishes: structure, location, and transport Environ Biol Fish 38:197–208 Silverstein JT, Plisetskaya EM (2000) The effects of NPY and insulin on food intake regulation in fish Am Zool 40:296–308 Smeets WJAJ (1981) Efferent tectal pathways in two chondrichthyans, the shark Scyliorhinus canicula and the ray Raja clavata J Comp Neurol 195:13–23 Environ Biol Fish (2012) 95:169–183 Smeets WJAJ (1982) The afferent connections of the tectum mesencephali in two chondrichthyans, the shark Scyliorhinus canicula and the ray Raja clavata J Comp Neurol 205:139–152 Smeets WJAJ (1983) The secondary olfactory connections in two chondrichthyans, the shark Scyliorhinus canicula and the ray Raja clavata J Comp Neurol 218:334–344 Smeets WJAJ (1998) Cartilaginous fishes In: Nieuwenhuys R, ten Donkelaar HJ, Nicholson C (eds) The central nervous system of vertebrates, vol 1, Springer-Verlag Berlin Heidelberg, New York, pp 551–654 Smeets WJAJ, Boord RL (1985) Connections of the lobus inferior hypothalami of the clearnose skate Raja eglanteria (Chondrichthyes) J Comp Neurol 234:380–392 Smeets WJAJ, Northcutt RG (1987) At least one thalamotelencephalic pathway in cartilaginous fishes projects to the medial pallium Neurosci Lett 78:277–282 Smeets WJAJ, Timerick SJB (1981) Cells of origin of pathways descending to the spinal cord in two chondrichthyans, the shark Scyliorhinus canicula and the ray Raja clavata J Comp Neurol 202:473–491 Smeets WJAJ, Nieuwenhuys R, Roberts BL (1983) The central nervous system of cartilaginous fishes: structural and functional correlations Springer-Verlag, Berlin Heidelberg, New York Stevenson JAF (1969) Neural control of food and water intake In: Haymaker W, Anderson E, Nauta WJH (eds) The hypothalamus Charles C Thomas, Springfield, pp 524–621 ten Donkelaar HJ, Nicholson C (1998) Notes on techniques In: Nieuwenhuys R, ten Donkelaar HJ, Nicholson C (eds) The central nervous system of vertebrates, vol 1, SpringerVerlag Berlin Heidelberg, New York, pp 551–654 Tester AL (1963a) Olfaction, gustation, and the common chemical sense in sharks In: Gilbert PW (ed) Sharks and survival D.C Heath Co., Boston, pp 255–282 Tester AL (1963b) The role of olfaction in shark predation Pac Sci 17:145–170 Timerick SJB, Roberts BL, Paul DH (1992) Brainstem neurons projecting to different levels of the spinal cord of the dogfish Scyliorhinus canicula Brain Behav Evol 39:93–100 Vallarino M (1985) Occurrence of ß-endorphin-like immunoreactivity in the brain of the teleost, Boops boops Gen Comp Endocrinol 60:63–69 Vallarino M, Danger JM, Fasolo A, Pelletier G, Saint-Pierre S, Vaudry H (1988a) Distribution and characterization of neuropeptide Y in the brain of an elasmobranch fish Brain Res 448:67–76 Vallarino M, Delbende C, Jegou S, Vaudry H (1988b) Alphamelanocyte-stimulating hormone (α-MSH) in the brain of the cartilaginous fish: immunocytological localization and biochemical characterization Peptides 9:899–907 Vallarino M, Andersen AC, Delbende C, Ottonello I, Eberle AN, Vaudry H (1989a) Melanin-concentrating hormone (MCH) immunoreactivity in the brain and pituitary of the 183 dogfish Scyliorhinus canicula: colocalization with alphamelanocyte-stimulating hormone (α-MSH) in hypothalamic neurons Peptides 10:375–382 Vallarino M, Tranchand Bunel D, Delbende C, Ottonello I, Vaudry H (1989b) Distribution of the pro-opiomelanocortin-derived peptides, alpha-melanocyte-stimulating hormone (α-MSH), adrenocorticotropic hormone (ACTH), and beta-endorphin in the brain of the dogfish Scyliorhinus canicula: an immunocytological study J Exp Zool Suppl 2:112–121 Vallarino M, Feuilloley M, Vandesande F, Vaudry H (1991) Immunochemical mapping of galanin-like immunoreactivity in the brain of the dogfish Scyliorhinus canicula Peptides 12:351–357 Volkoff H (2006) The role of neuropeptide Y, orexins, cocaine and amphetamine-related transcript, cholecystokinin, amylin and leptin in the regulation of feeding in fish Comp Biochem Physiol A 144:325–331 Volkoff H, Peter RE (2001) Interactions between orexin A, NPY and galanin in the control of food intake of the goldfish, Carassius auratus Regul Pept 101:59–72 Volkoff H, Canosa LF, Unniappan S, Cerdá-Reverter JM, Bernier NJ, Kelly SP, Peter RE (2005) Neuropeptides and the control of food intake in fish Gen Comp Endocrinol 142:3–19 Volkoff H, Xu M, MacDonald E, Hoskins L (2009) Aspects of the hormonal regulation of appetite in fish with emphasis on goldfish, Atlantic cod and winter flounder: notes on actions and responses to nutritional, environmental and reproductive changes Comp Biochem Physiol A 153:8– 12 Wetherbee BM, Cortés E (2004) Food Consumption and Feeding Habits In: Carrier JC, Musick JA, Heithaus MR (eds) Biology of sharks and their relatives CRC, Boca Raton, pp 225–246 Wright DE, Demski LS (1991) The midbrain GnRH-system in sharks and rays J Comp Neurol 307:49–56 Wright DE, Demski LS (1993) Gonadotropin hormonereleasing hormone (GnRH) pathways and reproductive control in elasmobranchs Environ Biol Fish 38:209–218 Yada T, Moriyama S, Suzuki Y, Azuma T, Takahashi A, Hirose S, Naito N (2002) Relationships between obesity and metabolic hormones in the “cobalt” variant of rainbow trout Gen Comp Endocrinol 28:36–43 Yamamoto N (2003) Three gonadotropin-releasing hormone neuronal groups with special reference to teleosts Anat Sci Int 78:139–155 Yamanaka S, Honma Y, Sana Y (1990) Immunohistochemical demonstration of serotonin neuron system in the central nervous system of the Japanese dogfish, Scyliorhinus torazame (Chondrichthyes) J Hirnforsch 31:385–397 Yopak KE, Lisney TJ, Collin SP, Montgomery JC (2007) Variation in brain organization and cerebellar foliation in chondrichthyans: sharks and holocephalans Brain Behav Evol 69:280–300 [...]... excursions off of the sea floor Examination of spatial variation in diet across the northern GOA shelf elucidated the spatial dependence of important prey species corresponding to geographic sub-regions of known ecological significance (Piatt and Springer 2007) Further examination of dietary variation within these ecoregions revealed Shelikof Strait to be unique from other regions because of the presence of. .. the periods of diet sampling during 2005–2007 17 Fig 8 Time averaged series of the temperature of the bottom layer (200 m) of the Alaska coastal current from the GAK1 mooring station (data from http://www.ims.uaf.edu/gak1/) Shading indicated periods of diet sampling consisted of forage fishes and small individuals of other demersal fish species Even though smaller size classes (