FISH and FISHERIES, 2003, 4, 289^347

Taste preferences in ®sh

Alexander O Kasumyan1 & Kjell B DÖving2

1Department of Ichthyology,Faculty of Biology,Moscow State University,119992 Moscow,Russia; 2Department of Biology, University of Oslo, N-0136 Oslo, Norway

Abstract Correspondence: The ¢sh gustatory system provides the ¢nal sensory evaluation in the feeding process. Alexander O Unlike other vertebrates, the gustatory system in ¢sh may be divided into two distinct Kasumyan, Department of subsystems, oral and extraoral, both of them mediating behavioural responses to food Ichthyology,Faculty items brought incontact withthe ¢sh.The abundance of taste buds is anotherpeculiarity of Biology,Moscow of the ¢sh gustatory system. For many years, morphological and electrophysiological State University, techniques dominated the studies of the ¢sh gustatory system, and systematic investiga- 119992 Moscow, tions of ¢sh taste preferences have only been performed during the last 10 years. In the Russia E-mail: present review,basic principles in the taste preferences of ¢sh are formulated. Categories alex_kasumyan@ or types of taste substances are de¢ned in accordance with their e¡ects on ¢sh feeding mail.ru behaviour and further mediation by the oral or extraoral taste systems (incitants, sup- pressants, stimulants, deterrents, enhancers and indi¡erent substances). Information Received17July 2002 on taste preferences to di¡erent types of substances including classical taste substances, Accepted3April 2003 free amino acids, betaine, nucleotides, nucleosides, amines, sugars and other hydrocar- bons, organic acids, alcohols and aldehydes, and their mixtures, is summarised. The threshold concentrations for taste substances are discussed, and the relationship between ¢sh taste preferences with ¢sh systematic positionand ¢sh ecology is evaluated. Fish taste preferences are highly species-speci¢c, and the di¡erences among ¢sh species are apparent when comparing the width and composition of spectra for both the stimu- lants and the deterrents.What is evident is that there is a strong similarity inthe tastepre- ferences between geographically isolated ¢sh populations of the same species, and that taste preferences are similar in males and females, although at the individual level, it may vary dramatically among conspeci¢cs.What is noteworthy is that taste responses are more stable and invariable for highly palatable substances than for substances with a low level of palatability.Taste preferences as a function of pH is analysed.There is a good correspondence between development of the gustatory system in ¢sh ontogeny and its ability to discriminate taste properties of food items. There is also a correspondence between oral and extraoral taste preferences for a given species;however,there is no cor- relationbetween smelland taste preferences.Taste preferences in ¢sh show low plasticity (in relation to the diet), appear to be determined genetically and seem to be patroclinous. Fish feeding motivation and various environmental factors like water temperature and pollutants such as heavy metals and low pH water may shift ¢sh taste preferences. Com- parisons between bioassay and electrophysiological data show that palatability is not synonymous with excitability inthegustatorysystem.The chemical nature of stimulants and deterrents in various hydrobionts is outlined. The signi¢cance of basic knowledge in ¢sh taste preferences for aquaculture and ¢sheries is emphasised.

Key words feeding behaviour, ¢sh, food intake, gustatory system, taste preferences, taste stimulants

# 2003 Blackwell Publishing Ltd 289 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Introduction 291

The gustatory system in ®sh 292

Feeding behaviours mediated by the gustatory system 297

Classi®cation of taste substances 298

Methods in studies of ®sh taste preferences 299

Taste preferences to different types of substances 300 Classical taste substances 300 Free amino acids 301 Betaine 306 Nucleotides and nucleosides 306 Amines 307 Sugars and other hydrocarbons 307 Organic acids 308 Alcohols and aldehydes 308 Other types of substances 309 Mixtures 309 Thresholds 310

Speci®city of taste preferences 313 Species speci¢city 313 Population speci¢city 315 Speci¢cityamong conspeci¢cs 315 Speci¢city between sexes 317

Genetics of taste preferences 318

Ontogeny of taste preferences 319

Ecology of taste preferences 320

Internalfactors effecting taste preferences 322 Feeding experience 322 Feeding motivation 323 Environmentalfactors effecting taste preferences 323 Temperature 323 Water pollutants 324 Low pH water 325 Taste preferences as a function of pH 326 Chemicalnature of stimulantsin ®sh food organisms 327

Relation between oral and extraoral taste preferences 329

Correlation between bioassay and electrophysiological data 330

Correlation between smell and taste preferences 331

Naturaldeterrents 332

Application in aquaculture and ®sheries 334

Epilogue 335

Acknowledgements 336

References 336

290 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

also participate in the consummatory phase in ¢sh. Introduction The sense of touch or mechanoreception is function- The aim of the present review is to provide a basis for ally and structurally connected to the gustatory sys- understanding ¢sh feeding behaviour through an tem (Kanwal and Caprio 1988;Hayama and Caprio examination of the taste preference ¢sh have for dif- 1989, 1990). It has been noted, for example, that an ferent foods and the single chemical components of increased hardness or presence of acute outgrowths the food. and spines on the external integument of the food Feeding in ¢sh, as in other groups of animals, is an organisms results in a decrease in consumption of important function of life, and is a result of processes such organisms by ¢sh (Hunter 1980;Lemm 1983; that are associated with searching, aiming, accept- Stradmeyer et al.1988;Stradmeyer 1989). The number ing, seizing, oral processing and evaluation of the of such examples is limited, and thus may indicate quality of food objects. Swallowing, digestion, an auxiliary role of mechanoreception in ¢sh feeding absorption and assimilation follow these processes. behaviour. Additionally, althougth there are numer- The success of these tasks, directly related to the ous solitary chemosensory cells covering the oral satisfaction of energy requirements, is responsible cavity of ¢sh (Whitear 1992;Whitear and Moate for the rate of growth, maturation and fecundity of 1994), there are few data on the functional properties ¢sh, their social status, migratory activity, life strat- of this common chemical sense (Silver 1987;Silver egy and their resistance to the impact of unfavour- and Finger1991;Kotrschal1996), and we do not know able factors. what role this chemosensory system plays in the oral Many sensory systems contribute to ¢sh feeding evaluation of prey quality.Furthermore, there are no behaviour, but their role and signi¢cance may pro- data to indicate an obvious role of the common che- foundly di¡er at di¡erent phases of the feeding beha- mical sense in the ¢nal phase of ¢sh feeding beha- viour. The consummatory phase of the feeding viour (Kotrschal1991,1992,1995;Whitear1992). behaviour starts with the awareness of a food object Studies of the gustatory system of ¢sh began in the and terminates with either swallowing or rejection. 19th century, and for many years, morphological The various sensorycues involved inthese tasks vary, techniques dominated this area of research. In the but what determines the palatability (taste) of a food last decades, electro-physiological methods have object for the ¢sh involves predominantly gustatory come into use to evaluate the properties of the gusta- and some aesthetic qualities. Various mechanosen- tory system. Presently, there are several reviews on soryorgans, and what is termed as a common chemi- the ¢sh gustatory system, focusing on ¢sh taste bud cal sense, can mediate these latter qualities. In many morphology and distribution (Kapoor et al. 1975; instances, a ¢sh rejects a food item after it has been Tucker1983;Jakubowski andWhitear1990;Gomahr taken into the mouth cavity (Bardach et al.1959;Tes- et al. 1992;Reutter 1992;Reutter and Witt 1993), the ter1963;Atema1980;Gerhart et al.1991;Schulte and physiologyof the peripheral taste neurones (Bardach Bakus 1992). Such behaviour demonstrates that an and Atema 1971;Caprio 1982;Tucker 1983;Caprio object that is selected as a food item, based on the 1984;Caprio 1988;Hara 1994a,b), the molecular information given by any of the sensory systems mechanisms of taste receptor transduction (Brand available, is not su¤cient to guarantee its suitability and Bruch 1992) and the organisation of gustatory as food for the ¢sh. The sensory systems in question neural pathways within the central nervous system are olfaction, vision, acoustic, lateral line organ, of ¢sh (Kanwal and Finger1992). extraoral gustatory subsystem or electroreception. Methods for estimation of ¢sh taste preferences This consideration indicates the important control have been nonexistent or at best, primitive, and have functions of receptors located within the ¢sh'smouth not been elaborated until recent times.When study- cavity in evaluating the speci¢c food items. ing taste preferences in ¢sh, it is di¤cult to avoid the The gustatory sensory system provides the ¢nal in£uence of other chemosensory systems and, while evaluation in the feeding process. It is well known it is easy to deprive a ¢sh of the olfactory system, it from both scienti¢c research and the anecdotes of is unrealistic to remove the taste organs or impede ¢shermen that the taste properties of feeds or baits taste perception. Thus, many of the investigations have a dramatic in£uence on food consumption of developed with the purpose of studying taste prefer- ¢sh, growth rate and ¢shing success (Jones 1980; ences yielded results which re£ected not only the Mackie and Mitchell 1985;Takeda and Takii 1992; gustatory system, but also the olfactory system, Kasumyan 1997). However, other sensory systems which is another principal chemosensory system

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 291 Taste preferences Alexander O Kasumyan & Kjell B DÖving

that mediates important cues and induces ¢sh feed- skin of the head. The oral taste buds have an endo- ing behaviour (Atema1980). The results of these stu- derm origin (Barlow and Northcutt 1995), while the dies as a basis for evidence of ¢sh taste preferences external taste buds are claimed to be of ectodermal should be used with signi¢cant caveats, apprehen- origin (Kapoor et al. 1975). The abundance of taste sions and restrictions. A decade ago, there were only buds is another peculiarity of the ¢sh gustatory sys- limited data focusing on the taste preferences of ¢sh; tem: taste buds are more numerous in ¢sh than in however, systematic investigations of ¢sh taste pre- any other animal. For instance, channel cat¢sh, 35^ ferences have been performed during the last 39.5 cm in body length, have 680 000 Æ 36 000 taste 10 years (Kasumyan1997). buds on the entire body and ¢ns surface (Finger et al. In the present review,we summarise the appropri- 1991), which is nearly 100 times more than is found ate information on taste preferences in ¢sh, include in the oral cavity of an adult man. The taste bud den- unpublished data and use the information to formu- sity varies according to the location and ¢sh species. late basic principles in the taste preferences of ¢sh. In the gular region of some benthivorous cyprinids, We shall also evaluate the relationship between ¢sh the density of the external taste buds reaches up to taste preferences with ¢sh systematic position and 300 mm À2, but is much lower in other areas of the ¢sh ecology. Interesting aspects of the taste prefer- body surface and ¢ns. The density of external taste ences in ¢sh are variations associated with age and buds is lower in planktivorous and surface-feeding population and feeding experience in early life. The cyprinids than in the bottom feeders (Gomahr et al. individual variability within a species is another 1992). In cyprinids, the density of oral taste buds is aspect that will be considered, as will the e¡ects of between 300 and 400 mmÀ2 in the palatal organ various biotic and abiotic factors ontaste preferences. (Osse et al.1997)and30^35mmÀ2 in the area sur- There is a large number of species mentioned in this rounding the teeth of salmonids (Marui et al.1983a; review;therefore, we have listed the species with Hara et al.1993). Latin and family names in Table 1. The prey of some A typical taste bud has an ovoid or pear shape;its ¢sh is listed inTable 2. long axis is orientated vertically to the apical surface of the epithelium, which is normally not keratinised. As a rule, the base of a taste bud is situated on top of The gustatory system in ®sh a small ascending papilla of the corium (dermis). The peripheral organs belonging to the gustatory Consequently,a taste bud is often situated on a dome system are the taste buds. Taste buds constitute the on the epidermis (type I);in other cases, taste buds structural basis of the gustatory system in all are slightly elevated (type II;Fig. 1)or are not elevated gnathostomes ^ ¢sh, amphibians, reptiles, birds and (sunken;type III;Reutter et al. 1974). Marginal cells mammals. Jawless ¢sh (hag¢sh (Myxinidae) and lam- form the border or margin between a taste bud and prey (Petromyzontidae) have end buds ^ organs that the neighbouring strati¢ed squamous epithelium. resemble the teleost taste buds (Reutter1986). Phylo- Oral and extraoral taste buds consist of gustatory genetically, it seems likely that the taste buds were receptor, supporting and basal cells. Both gustatory not consecutively and progressively developed from receptor cells and supporting cells have an elongated one vertebrate class to the next (Reutter and Witt shape, run more or less parallel and follow the long 1993). The ¢rst descriptions of ¢sh taste buds were axis of the taste bud. They reach the surface of the published early in the 19th century (Weber 1827; epithelium by a pore (Fig. 1), which has a diameter of Leydig 1851). Since these times, numerous articles 2^5 mm to more than 20 mm (Jakubowski and concerned with taste bud morphology in various Whitear 1990;Reutter 1992). The number of gusta- ¢sh species have appeared (for reviews, see Reutter torycells in a taste bud varies from ¢ve, as in the goby (1986) and Jakubowski andWhitear1990). (Pomatoschistus sp.), to 67, as in the armoured cat¢sh In ¢sh, the taste buds are situated not only within (Jakubowski and Whitear 1990). Gustatory receptor the oral cavity, pharynx, oesophagus and gills, but cells terminate with large receptor microvilli, which may also occur on the lips, barbels and ¢ns, and over are conical in shape and measure up to 2.5 mmin the entire body surface in many species. These pecu- length and about 0.3 mminwidth.Thesecellsform liarities distinguish ¢sh from other groups of the receptor area or receptor ¢eld in the pore. An gnathostomes, except the amphibians, as in some apexof supportingcells bears up to10small cylindri- developmental stages, the taste buds are found in cally shaped microvilli, which are about 0.5 mm amphibians within the oral cavity and also in the long and about 0.1 mm wide. The supporting cells

292 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 1 The index of common and scienti¢c names for ¢sh species

Common name Latin name Family, order

Alluaudi cichlid Haplochromis (Astatoreochromis) alluaudi Cichlidae, Perciformes Angel®sh Pomacanthus spp. Pomacanthidae, Perciformes Angel®sh Holacanthus spp. Pomacentridae, Perciformes Angler®sh Lophius piscatorius Lophiidae, Lophiiformes Arctic ¯ounder Liopsetta glacialis Pleuronectidae, Pleuronectiformes Armoured cat®sh Corydoras sp. Callichthydae, Siluriformes Atlantic cod Gadus morhua Gadidae, Gadiformes Atlantic herring Clupea harengus harengus Clupeidae, Clupeiformes Atlantic salmon Salmo salar Salmonidae, Salmoniformes Bitterling Rhodeus sericeus amarus Cyprinidae, Cypriniformes Black molly Poecilia sphenops Poeciliidae, Cyprinidontiformes Blue-green chromis Chromis viridis Pomacentridae, Perciformes Bluehead wrasse Thalassoma bifasciatum Labridae, Perciformes Bream Abramis brama Gadidae, Gadiformes Brill Scophthalmus rhombus Scophthalmidae, Pleuronectiformes Brook char Salvelinus fontinalis Salmonidae, Salmoniformes Brown trout Salmo trutta Salmonidae, Salmoniformes Bullhead Ictalurus natalis Ictaluridae, Siluriformes Capelin Mallotus villosus Osmeridae, Salmoniformes Channel cat®sh Ictalurus punctatus Ictaluridae, Siluriformes Chinook salmon Oncorhynchus tshawytscha Salmonidae, Salmoniformes Chub Leuciscus cephalus Cyprinidae, Cypriniformes Chum salmon Oncorhynchus keta Salmonidae, Salmoniformes Common carp Cyprinus carpio Cyprinidae, Cypriniformes Crucian carp Carassius carassius Cyprinidae, Cypriniformes Dab Limanda limanda Pleuronectidae, Pleuronectiformes Dace Leuciscus leuciscus Cyprinidae, Cypriniformes Dover sole Solea solea Soleidae, Pleuronectiformes Dusky grouper Epinephelus guaza Serranidae, Perciformes European eel Anguilla anguilla Anguillidae, Anguilliformes European grayling Thymallus thymallus Salmonidae, Salmoniformes European minnow Phoxinus phoxinus Cyprinidae, Cypriniformes European sea bass Dicentrarchus labrax Moronidae, Perciformes Five-bearded rockling Ciliata mustela Phycidae, Gadiformes Flathead grey mullet Mugil cephalus Mugilidae, Mugiliformes Frolich char Salvelinus alpinus erythrinus Salmonidae, Salmoniformes Gilthead sea bream Sparus aurata Sparidae, Perciformes Goby Pomatoschistus sp. Gobiidae, Perciformes Goby Gobiosoma bosci Gobiidae, Perciformes Gold®sh Carassius auratus Cyprinidae, Cypriniformes Grass carp Ctenopharyngodon idella Cyprinidae, Cypriniformes Green chromis Chromis caerulea Pomacentridae, Perciformes Grunt Haemulon spp. Haemulidae, Perciformes Guppy Poecilia reticulata Poeciliidae, Cyprinidontiformes Humbug dascyllus Dascyllus aruanus Pomacentridae, Perciformes Jack mackerel Trachurus japonicus Carangidae, Perciformes Japanese eel Anquilla japonica Anguillidae, Anguilliformes Japanese ¯ounder Paralichthys olivaceus Paralichthydae, Pleuronectiformes Japanese pilchard Sardinops sagax melanosticta Clupeidae, Clupeiformes Lake char Salvelinus namaycush Salmonidae, Salmoniformes Largemouth bass Micropterus salmoides Centrarchidae, Perciformes Marbled rock®sh Sebasticus marmoratus Scorpaenidae, Scorpaeniformes Navaga Eleginus navaga Gadidae, Gadiformes Nile tilapia Oreochromis niloticus Cichlidae, Perciformes Oriental weather®sh Misgurnus anguillicaudatus Cobitidae, Cypriniformes Parrot®sh Scarus spp. Scaridae, Perciformes

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 293 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 1 continued

Common name Latin name Family, order

Parrot®sh Sparisoma spp. Scaridae, Perciformes Pig®sh Orthopristis chrysopterus Haemulidae, Perciformes Pike Esox lucius Esocidae, Esociformes Pin®sh Lagodon rhomboides Sparidae, Perciformes Plaice Pleuronectes platessa Pleuronectidae, Pleuronectiformes Platy Xiphophorus maculatus Poeciliidae, Cyprinidontiformes Porgy Calamus spp. Sparidae, Perciformes Puffer Fugu pardalis Tetraodontidae, Tetraodontiformes Rainbow trout Oncorhynchus mykiss Salmonidae, Salmoniformes Reticulated dascyllus Dascyllus reticulates Pomacentridae, Perciformes Roach Rutilus rutilus Cyprinidae, Cypriniformes Russian sturgeon Acipenser gueldenstaedtii Acipenseridae, Acipenseriformes Sea bream Chrysophrys (Pagrus) major Sparidae, Perciformes Sea cat®sh Arius felis Ariidae, Siluriformes Sergent-major Abudefduf saxatilis Pomacentridae, Perciformes Sharpnose puffer Canthigaster valentini Tetraodontidae, Tetraodontiformes Shore rocklings Gaidropsarus mediterraneus Phycidae, Gadiformes Siberian sturgeon Acipenser baerii Acipenseridae, Acipenseriformes Small puffer Sphoeroides sprengleri Tetraodontidae, Tetraodontiformes Smooth dog®sh Mustelus canis Triakidae, Carchariniformes Snapper Ocyurus chrysurus Lutjanidae, Perciformes Stellate sturgeon Acipenser stellatus Acipenseridae, Acipenseriformes Striped bass Morone saxatilis Moronidae, Perciformes Sun®sh Lepomis macrochirus Centrarchidae, Perciformes Tench Tinca tinca Cyprinidae, Cypriniformes Tilapia Tilapia zillii Cichlidae, Perciformes Tile®sh Malacanthus plumieri Malacanthidae, Perciformes Turbot Scophthalmus maximus Scophthalmidae, Pleuronectiformes Vendace Coregonus albula Salmonidae, Salmoniformes Winter ¯ounder Pseudopleuronectes platessa americanus Pleuronectidae, Pleuronectiformes Wrasse Thalassoma lunare Labridae, Perciformes Wrasse Thalassoma spp. Labridae, Perciformes Wrasse Halichoeres spp. Labridae, Perciformes Yellowtail Seriola quinqueradiata Carangidae, Perciformes Zander Stizostedion lucioperca Percidae, Perciformes surround and separate o¡ the gustatory receptor relate to the age of the individual cells (Jakubowski cells. Up to ¢ve basal cells are situated at the basis of and Whitear 1990). It has been suggested that both the taste bud and have desmosomal attachments to the light and the darkcells have synaptic associations both the gustatory and the supporting cells. External with the basal cells as well as with the nerves, and taste buds in some of the Gadidae species have no that theyare receptor cells (Reutter1986). basal cells (Jakubowski andWhitear1990). The solitary chemosensory cells have a close cyto- The nomenclature and interpretation of morphol- logical resemblance to the gustatory oral and extra- ogy and function of some cell types are not always oral receptor cells (Whitear 1971;Whitear and consistent. Welsch and Storch (1969) distinguished Kotrschal 1988;Whitear 1992), and they are also the elongated cells of the taste buds as `light' and innervated by the facial nerve, as are many taste buds `dark'cells. The cells with a single apical microvillus (Whitear and Kotrschal 1988;Kotrschal et al.1993; usually appeared less electron-dense than the sup- Kotrschal et al. 1998). Both the taste buds and the porting cells surrounding them;however, in shore solitary chemosensory cells have similar primary rocklings and ¢ve-bearded rockling, the supporting representation in the facial lobe and also have indis- cells in the external taste buds are consistently paler tinguishable ascending projections (Kotrschal and than the gustatory receptor cells. The variation of Finger 1996). Based on these ¢ndings, it was sug- overall electron density of gustatory cells appear to gested that the solitary chemosensory cells should

294 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 2 A list of the prey species mentioned in the review

Common name Latin name Family, order, and class

Annelide Perinereis brevicirrus Nereidae, Phylodocida, Polychaeta Blue crab Callinectes sapidus Portunidae, Decapoda, Crustacea Blue star®sh Coscinasterias tenuispina Asteriidae, Forcipulata, Asteroidea Brine shrimp Artemia Artemiidae, Anostraca, Crustacea Copepods Eucalanus crassus Eucalanidae, Copepoda, Crustacea Dino¯agellate Protogonyaulax tamarensis Dinophyta, Dino¯agellata, Flagellata Duckweed Lemna minor Lemnaceae, Angiospermae, Monocothyledones Earthworm Eisenia foetida Lumbricidae, Opisthopora, Oligochaeta Krill Euphausia paci®ca Euphausiidae, Euphausiacea, Crustsacea Mosquito Chironomus sp. Chironomidae, Diptera, Insecta Nemertide Lineus ruber Lineidae, Geteronemertini, Nemertini Opisthobranch Aplysia brasiliana Aplysiacea, Tectibranchia, Gastropoda Polychaete Phyllodoce maculata Phyllodocidae, Phyllodocida, Polychaeta Romaine lettuce Lactuca sativa Asteraceae, Angiospermae, Dicothyledones Shrimp Pandalus borealis Pandalidae, Decapoda, Crustacea Sponge Carteriospongia sp. Dysidae, Cornacuspongida, Demospongia Sponge Dysidea sp. Dysidae, Cornacuspongida, Demospongia Sponge Amphimedon compressa Niphatidae, Haplosclerida, Demospongia Sponge Axinella corrugata Axinellidae, Cornacuspongida, Demospongia Sponge Ircinia strobilina Thorectidae, Dictyoceratida, Demospongia Sponge Hyritios erecta Thorectidae, Dictyoceratida, Demospongia Squid Loligo forbesi Loliginidae, Decapoda, Cephalopoda

be considered as the gustatory subsystem (Kotrschal by the marginal cells (Reutter and Witt 1993). Glyco- 1996, 2000). Regretably, there are not su¤cient data conjugates are glycoproteins, which are protein about the function of the solitarychemosensory cells chains to which one or more oligosaccharide units to support this hypothesis. Moreover, the density of are attached.The receptor cells protect the taste buds extraoral taste buds varies with the region at the receptor ¢eld mechanically, and probably, also ful¢l body surface. In contrast, the solitary chemosensory functions involving taste perception processes.There cells cover the body surface uniformly, and there is is little information about the possible function of no relationship between ¢sh feeding ecology and the mucous substances surrounding the taste buds the mean density of the solitary chemosensory (Bannister 1974). It has been assumed that the cells. These observations support the view that the mucous presents a ¢lter system selecting particular solitarychemosensorycells have functions and biolo- substances prior to the chemoreception processes, gical roles distinct from the function of ¢sh taste which take place on the cell membrane of the micro- buds. villi on the sensorycells. Adult taste buds are not static structures.The cells Every taste bud has a network of nerves located in a taste bud undergo renewal from the epithelial between the basal cells and the basal part of the cells that surround the base of the taste bud. These receptor cells, which is comprised of unmyelinated cells undergo division with some of their daughter and myelinated nerve ¢bres (Finger et al.1991).The cells migrating into the taste bud. The average life- network of nerves is the location of synaptic contacts span of a taste cell is dependent on the ambient tem- within the taste bud. Synapses occur at the subnuc- perature: in channel cat¢sh, the taste cell turns over lear part and at the basal processes of the gustatory every 40 days at 14 8C and every 12 days at 30 8C receptor cells, and also at basal cells. Nerve ¢bres (Raderman-Little1979). innervating taste buds are derived from the facial The receptor cells and the mucous substances cov- (VII), glossopharyngeal (IX) and vagal (X) cranial ering the receptor ¢eld of ataste bud contain complex nerves. The facial nerve supplies external taste buds polymeric glycoconjugates secreted by the support- of the barbels and lips (N. facialis maxillaris), ¢ns and ing cells (Jakubowski and Whitear 1990), and possi- body surface (N. facialis recurrens) and oral taste buds bly by the receptor cells (Witt and Reutter 1990) or of rostral palate (N. facialis palatinus).The vagal nerve

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 295 Taste preferences Alexander O Kasumyan & Kjell B DÖving

The primary gustatory nuclei lie in the medulla oblongata, and ¢sh with numerous external taste buds, such as channel cat¢sh, have well-developed facial lobes in the rostral part of the medulla. Cypri- nid ¢sh with numerous taste buds in the mouth cavity also possess enlarged primary gustatory nuclei inthe vagal lobe regions of the caudal medulla. The glossopharyngeal nerve terminates in a medul- lar region between the facial and the vagal lobes. Interestingly, in the majority of the ¢sh species, neither the facial nor the vagal lobe is well developed, and the two may even be continuous and grossly indistinguishable from each other. Based on the observed anatomical peculiarities, the gustatory sys- tem may be divided into two distinct, though inter- related, subsystems: oral gustatory subsystem and extraoral gustatory subsystem (Finger and Morita 1985). The gustatory tracts from the facialand vagal lobes ascend and descend within the brain. The major sec- ondary ascending projections from the facial and the vagal lobes in the teleosts studied terminate in a large isthmic nucleus known as the superior secondary gustatory nucleus. Within the medulla, local gus- tato-motor networks and descending pathways form the neural substrate for re£exive movements such as food pick up or biting and swallowing or rejection (Kanwal and Finger1992). Presumably,the vagal lobe in cyprinids is involved in stimulus localisation and contains a representation of both gustatory and tac- tile inputs from the palatal organ in the oral cavity (Morita and Finger 1985). The palatal organ covered by the oral taste buds accomplishes selective inges- tion of food by the contraction of speci¢c sets of mus- cles, which can selectively hold on to `tasty' food Figure 1 Taste buds on the tongue. (A) Tongue tip of particles (Sibbing et al. 1986;Osse et al.1997).Ahigh European eel larvae, Anguilla anguilla (Anguillidae), percentage of bimodal taste/touch neurones are showing several papillae. (B) Part of the tongue with one found within the primary gustatory nucleus (Marui papillae. Several taste pores are seen emerging between et al. 1988;Hayama and Caprio 1989). These ¢ndings epithelial cells with microridges. (C) A taste pore between indicate that oral evaluation is based on both chemi- three epithelial cells. (Scanning electron microscopic cal and mechanical sensory input. pictures by K.B. DÖving). The functional properties of the ¢sh gustatory sys- tem have mainly been studied by electrophysiologi- innervates most of the orobranchial taste buds, cal means. Hoagland probably obtained the ¢rst whereas the glossopharyngeal nerve plays a minor electrophysiological recording of taste activity in a role inthe gustatorysupply to the oral cavity (Herrick teleost (Hoagland1933). Since his pioneering work, a 1901;Atema 1971;Finger 1976). A single nerve ¢bre number of reports concerning the physiology of the may innervate several taste buds, and each receptor gustatory system of teleost have appeared (for review, cell forms more than one synapse. This suggests a see Marui and Caprio (1992)).The ¢rst studies found high degree of convergence of input onto the sensory that the ¢sh gustatory system is highly responsive to nerve ¢bres (Kinnamon 1987), which also seems to many organic and inorganic chemicals as well as increase with ¢sh growth (Finger et al.1991). to natural food extracts. The responses from the

296 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

gustatory nerves are fast adaptingor phasic;the peak Feeding behaviours mediated by the response is reached withinthe ¢rst few seconds of sti- gustatory system mulation and returns to the baseline activity even with continued stimulation. The taste speci¢cities Various sensory organs can mediate the di¡erent across the ¢sh species studied were verydi¡erent: dif- feeding behaviours found in ¢sh, and a particular ferent species of ¢sh may have very di¡erent rank behaviour pattern can be evoked via several sensory orders of e¡ective taste stimuli.The estimated thresh- systems. For example, the snapping movement with olds for the most potent substances were less than the jaws can be released by stimulation of the lateral 10 À9 m. The dose^response relationships are var- line organ or electroreceptors, by visual, auditory or iously expressed and are di¡erent for di¡erent taste olfactory stimuli or by taste or common chemical substances;responses increase sharply with a loga- sensorystimuli. Other patterns might be more speci- rithmic increase in concentration and eventually ¢c, and in some species, a given feeding pattern can reach saturation (Marui et al.1983a).The speci¢cities be mediated via only one sensory pathway. and sensitivities of the taste buds innervated by the Both an extraoral and an oral gustatory subsystem di¡erent branches of the facial nerve are similar in mediate behavioural responses to food items brought the same ¢sh species (Caprio 1975;Davenport and in contact with the ¢sh. The behavioural responses Caprio 1982). The oropharyngeal taste buds, which are complex and comprise several elements with are innervated by both the glossopharyngeal and characteristic features. These response patterns are the vagal nerves, are generally less sensitive to speci- described and de¢ned below. ¢c amino acids by 1.5^2 orders of magnitude than Usually, the gustatory systems participate in the the taste buds innervated by the facial nerve (Kanwal ¢nal phases of the series of feeding behaviours con- and Caprio1983), although they have been shown to nected with food search, which normally ends with have similar amino-acid speci¢city. consumption.The bullhead, for example, may detect, Using the free aminoacids and theirderivatives as search and successfully ¢nd a food source based on taste substances, it has been found that there are information from the external taste receptors (Bar- rigid requirements of taste receptors for the molecu- dach et al.1967). It seems established that stimulation lar structure of the stimulus molecules. For example, of the external taste system may mediate grasping, gustatory responsestoaminoacidsarehighly stereo- biting, snapping or scraping behaviours, especially speci¢c: the l-isomers of amino acids have, for most inbottom feeders, nocturnalorcave¢sh, or¢shliving species studied, been shown to be more stimulatory in deep or turbid water. The presence of numerous than its d-enantiomers. Unsubstituted, l-a-amino external taste buds is common in these ¢sh species acids are usually the most e¤cient compounds for (Saxena 1959;Branson 1966;Kapoor et al. 1975; the gustatorysystemof ¢sh. Neutralaminoacids con- Gomahr et al.1992). Unfortunately,experimental evi- taining two or fewer carbon atoms, and having dence supporting the role of extraoral taste reception unbranched and uncharged side chains, are highly in these types of feeding behaviours is scarce. Most stimulatory. Acidic amino acids are poor gustatory studies have been concerned with the evaluation of stimuli, and basic amino acids are quite variable, the gustatory system for processing of food during depending upon the ¢sh species. The peptides tested the oral phase. so far are less e¡ective than their amino acid residue. Extraoral taste response: when a ¢sh with a well- Furthermore, the order of the amino acid residues in developed extraoral taste system touches a food a simple peptide does not signi¢cantly a¡ect its abil- object, the ¢sh makes an abrupt e¡ort to reach the ity to stimulate the gustatory system (Caprio 1978; object or to avoid it. The ¢sh can stop, turn around, Hidaka and Ishida 1985;Marui and Kiyohara 1987). turn to the side, swim backwards, initiate searching It should be noted that there are many exceptions to trajectories or make circles and S-shaped loops to these ¢ndings (Marui and Caprio1992). A synergistic reach the object. Finally, snapping with the jaws at e¡ect was observed for mixtures of some amino acids high frequency might follow a touch of a food item. inthe gustatorysystemof several ¢sh species (Hidaka In sturgeon, the series consists of snapping with the et al. 1976;Yos h i i et al.1979) and, in addition, electro- jaws two or more times. This species has no object physiological responses to some amino acids are vision, but searches the bottom surface thoroughly enhanced by the presence of another amino acid in and collects all appropriate food items using the cross-adaptation experiments (Marui and Kiyohara snapping behaviour. There is a dramatic increase in 1987). the bottom-surface snapping in various ¢sh species

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 297 Taste preferences Alexander O Kasumyan & Kjell B DÖving

upon stimulation by food odour (Atema et al.1980; or to spit it out. For many ¢sh, even if an object is Kasumyan and Ponomarev 1986;Kasumyan 1999a). newly spat out, it can be recaught. This sequence This pattern of feeding behaviour is also evident in can be repeated several times. The retention time ¢sh with awell-developed visual system;for example, may be determined for each recatch, and sum of rainbow trout and brown trout (DÖving and Kasum- the retention times may be calculated. The dura- yan, unpublished). tion of a taste response is usually relatively short and Once the food item is in a proper position in rela- ends by swallowing a food object or by ignoring it tion to the mouth of the ¢sh, the ¢sh grasps or snaps when ¢sh lose interest in a food object and do not the object with the jaws. Food objects containing recapture it. aversive substances cause ¢sh to neglect the object During the retention time, a putative food can be by moving forward, as do sturgeons, or change their sorted into waste material and acceptable items. The searchingdirection, as do channel cat¢sh.While inci- food is masticated and thereafter transported via the tants increase the snap frequency, suppressants orobuccal cavity.This period can be long and is not decrease the frequency of snapping movements. necessarily followed by acceptance and swallowing. Oral taste response: food objects can be taken into As studied in detail in the common carp and other the mouth by scraping, tearing, sucking, snapping cyprinids, the food selection procedure comprises and grasping (Nikolski 1963;Wootton 1998). The repetitive positioningof the food and transport move- behaviour pattern most intensively studied in con- ments. Food is transported through the posterior side junction with feeding is the sucking and snapping of the palatalorganand the postlingualorganof¢sh, movement in ¢sh (for review,see Osse et al.1997).This where the density of oral taste buds reaches a maxi- behaviour consists of a series of movements that mum value of 820 mmÀ2 in bream (Osse et al.1997). involves a number of anatomical structures and a During the process of taste selection, food items, most sequence of ¢nely adjusted muscle activity (Fryer probably, are ¢xed between the palatal organ as roof a nd I le s 1972;L ie m 1973). and the postlingual organ as bottom, and rinsed by The snapping movements performed with the jaws the alternating £ow of water, while waste materials are extremely fast in angler¢sh, which attract preys are expelled through the branchial slits (Sibbing et al. with movable illicium (Grobecker and Pietsch 1979). 1986). In the complex mastication process, the food Furthemore, in species of other families, such as is crushed and ground to small and acceptable pieces gadids and salmonids, these movements are so fast (Sibbing 1988;Sibbing 1991). A mollusc feeder, that they are di¤cult to observe in detail without the Alluaudi cichlid, uses powerful teeth to destroy the use of suitable recording equipment. shell of the mollusc and then spits out up to 90% Motta (1982, 1988) and Coughlin and Strickler of the fragments of the shell without loss of £esh (1990) have studied particulate feeding behaviour on (Hoogerhoud1989). plankton. Recordings of the prey capture mechan- Benthivorous and herbivorous ¢sh keep the food in isms used by the blue-green chromis, while feeding the mouth for much longer time than piscivorous on calanoid copepods and brine shrimp reveals that ¢sh. Even though predators have a short retention it is capable of a ram-jaw,low-suction feeding, as well time, the gustatory system still controls the feeding as a typical suction feeding behaviour. By adjusting responses. The taste judgements made during the the degree of jaw protrusion and the amount of suc- oral examination is an obligatory phase of feeding tion used during a feeding strike, they can modulate behaviour in every ¢sh species, irrespective of the feeding strikes according to the prey type being feeding strategies used or the extent to which the encountered. The ram-jaw feeding mode enables gustatory system is developed. blue-green chromis to capture highly evasive cala- noid copepods within 6^10 ms (Table 2). Classi®cation of taste substances When a food object is in the mouth, it is subject to ¢nal sensory judgements. As a result of these oral Chemical substances can be divided into several tests, the food item canthen be rejected or swallowed. categories or types, according to their e¡ects on the The time for which the object is kept in the mouth feeding behaviour of ¢sh. The nomenclatures used before swallowing or rejection is called the retention for classi¢cations of chemical stimuli which evoke time. During the retention time, the ¢sh detect and feeding behaviour are varied (Lindstedt1971;Mackie recognise taste substances, assess the palatability of 1982;Mearns et al. 1987;Sakata 1989).We will empha- the food object and perform the decision to swallow sise the distinction between the substances that have

298 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

e¡ects on the oral and extraoral taste systems. These extraoral or the oral taste systems and are classed as systems play di¡erent roles in ¢sh feeding behaviour `indi¡erent'substances. (Atema 1980;Pavlov and Kasumyan 1990) and di¡er in their functional characteristics (Marui and Caprio Methods in studies of ®sh taste 1992;Kasumyan 1999a). The presence or absence of preferences these compounds in the diet determines whether a food item is grasped or ignored, is eaten or rejected Several methods for presenting and quantifying ¢sh and, to some extent, how much of the food is con- taste preferences have been developed as bioassays. sumed. Space does not permit us to discuss these methods at `Incitants' are substances that induce capture of length. Many of these methods, though interesting the food item via the extraoral taste system. There and valuable, have limitations because they do not are a number of di¡erent behaviours that can pre- make it possible to distinguish between the taste and sumably be evoked by incitants like: suction, grasp- the olfactory systems. Examples are the methods ing, snapping, biting, tearing or pinching. developed by Carr and coworkers (Carr 1976;Carr `Suppressants' are substances that decrease the and Chaney1976;Carr et al.1977). Likewise, Sutterlin rate of grasping food items. Like incitants, suppres- and Sutterlin (1970) used polyurethane discs impreg- sants are mediated by the extraoral gustatory system nated with various substances for studying taste pre- and control the rate of grasping food items. ferences in Atlantic salmon. Jones (1989,1990) used `Stimulants'are substances characterised by high cotton pellets soaked in solutions of test substances ingestion rate. This behaviour is evoked by the oral in attempts to study taste preferences in rainbow taste system. Stimulants promote ¢sh feeding. trout. Usually, ¢sh swallow the food items containing sti- Several authors have used a solid matrix carrying mulants at ¢rst capture. This behaviour is mediated the stimulants. A few authors have used puri¢ed by the oral gustatory system. starch gel as a carrier for test substances (Hidaka `Deterrents' are substances that make the ¢sh et al. 1978;Oh sug i et al.1978;Murofushi and Ina1981; abandon food intake and evoke food rejection. Deter- Hidaka 1982;Kasumyan 1992;Kasumyan and Kazh- rents are characterised by high rejection and low laev 1993 a). Mearns et al. (1987) developed another ingestion of food items that are caught. Often, ¢sh approach using agar gel as carrier for taste sub- feeding motivation is decreased for a short period stances. This method was elaborated by Kasumyan after a ¢sh has tested a food item that contains a and collaborators (Kasumyan and Sidorov 1993a,b, deterrent. Usually, the retention time of food items 1995 c, d;Ka su mya n a nd Mor s y 1996, 1997). The i r that contain a deterrent substance is short and the results showed that this method distinguished ¢sh does not attempt to recatch the food item. This between smell and taste (Table 3), and they proposed behaviour is mediated by the oral gustatory system. the index of palatability as a quantitative estimation `Enhancers' or potentiators are substances that, of taste preference (in percent): although not feeding stimulants per se, accentuate I ˆ 100 Á R À C†Á R ‡ C†À1 the £avour of the food and cause ¢sh to increase their pal consumption of £avoured food. In some cases, how- where R is the consumption of £avoured pellets and C ever, stimulants may be ine¡ective as feeding enhan- is the consumption of un£avoured pellets. The Ipal is cers. This behaviour is evoked by the oral taste an analogy with the index of electivity used by Ivlev system. (1961). Presumably,there should also be another category Large disk-shaped agar gel (1.5 cm in diameterand of taste substances,`detractors',which essentially are about 2.0 cm in length) were used to present test neutral substances that diminish the positive e¡ects compounds for studying taste preferences in herbi- of the stimulants. Unfortunately, appropriate exam- vorous tilapia, but experiments on anosmic ¢sh were ples of substances that belong to this categoryof taste not made using this test procedure (Johnsen and stimuli do not yet exist. Adams 1986;Adams et al.1988). Other bioassay proce- As will be shown, a single substance might have dures include incorporating various compounds into di¡erent e¡ects on the taste behaviour of the same freeze-dried pellets (Mackie and Adron 1978;Mackie ¢sh and can be both incitant and deterrent, or inci- 1982). Radiographic (Talbot and Higgins1983;Talbot tant and stimulant. A number of substances do not 1985;To f te n et al. 1995) or radioisotopic methods induce any e¡ects on ¢sh behaviour either via the (Storebakken et al.1981;Jobling et al. 1995) have also

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 299 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 3 Comparing taste responses in intact and anosmic ¢sh

Retention time (s)

Concentration Acceptance Palatability Number Number (mM) ratio (%) index (%) of snaps First capture All captures of trials

Intact fish 100 100ÃÃÃ 59 1.0ÃÃÃ 15.3ÃÃÃ 15.3ÃÃÃ 48 10 49.9Ã 31.8 1.4 9.3ÃÃ 12.4 48 1 28.3 4.6 1.5 7.6 9.8 48 0 25.8 ± 1.8 4.4 7.8 48

Anosmic ®sh 100 100ÃÃÃ 48.1 1.0 14.4ÃÃÃ 14.4ÃÃÃ 20 10 60Ã 26.3 1.8 7.9 11.7 20 1 35 0 1.4 8.2 9.0 20 035±1.94.3920

Properties of taste responses to different concentrations of L-cysteine in intact and anosmic carp Cyprinus carpio (9±12 cm body length). The experiments with anosmic ®sh were performed 10 days after operation. See text for details. Data modi®ed from Kasumyan and Morsy (1996). In this and the following tables, the signi®cance levels are indicated as follows: ÃÃÃP < 0.001; ÃÃP < 0.01; ÃP < 0.05. The control pellets were made of 2% agar and contained the red colour Ponceau 4R at a concentration of 50 mM. The Chi-square test was used for estimation of acceptance ratio and Student's t-test for all other characteristics of taste responses. been used for comparison of the palatability of food stances on smoothdog¢sh.Therehave beenalimited materials for ¢sh. A special approach was developed number of studies performed on taste preferences in for the assessment of taste preferences and a distinc- ¢sh to classical taste substances, and they are di¤- tion between oral and extraoral taste systems cult to compare, as they were obtained using various (Kasumyan 1992;Kasumyan et al. 1992;Kasumyan behavioural assays and di¡erent carriers for the che- and Kazhlaev1993a). micals. Sucrose: it has been found that the European min- now can be trained to distinguish among sugar solu- Taste preferences for different types of tions and solutions of sour, bitter and salty substances substances (Strieck 1924). Trudel (1929) was able to It has been shown that chemical substances evoke train European minnows to distinguish between sev- di¡erent types of behavioural taste responses, thus eral natural sugars, and Krinner (1934) was able to theyareperceivedas di¡erent by¢sh.The assessment train European minnows to distinguish between of palatability has been investigated for free amino sodium chloride and saccharose. Conditioned Euro- acids and their derivatives, classical taste substances, pean minnows are able to discriminate between nucleotides and nucleosides, quaternary amines, saccharose and saccharin at low concentrations organic acids and other types of substances. In the (Glaser 1966). The chemosensory system mediated following sections, we summarise data available con- ¢sh responses to a solutionof the substance, as condi- cerning ¢sh taste preferences to these substances tioned stimuli were not de¢ned in these studies. and their mixtures. Studies, where appropriate methods were used, show that sucrose is usually an indi¡erent taste sub- stance for ¢sh or evokes a positive oral taste response. Classical taste substances Sucrose is palatable for many herbivorous ¢sh like Man almost exclusively con¢ned earlier investiga- grass carp and is palatable for omnivorous ¢sh in tions on gustation in ¢sh to classical taste sub- which algae is the main component of the diet. stances, representing what has been considered the Examples are dace, roach, guppy, black molly and basic taste sensations recognised. They are consid- platy (Andriashev1944;Appelbaum1980;Kasumyan ered to be sweet, sour, bitter and salty, represented, 1997;Ka su mya n a nd Mors y 1997;Ka su mya n a nd for example, by sucrose, acetic acid, quinine and Nikolaeva1997,2002;Tables 4 and 5). Of 36 ¢sh spe- sodium chloride. von Uexku« ll (1895) made the ¢rst cies tested by di¡erent authors, sucrose was found to study of ¢sh taste preference to classical taste sub- be a stimulant for 15 species and was an indi¡erent

300 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

taste substance for 18 mostly carnivorous species. deterrents for the Dover sole (Mackie and Mitchell Sucrose evokes a deterrent taste response in one ¢sh 1982b). In gold¢sh, gelatine pellets £avoured with species only, namely the pu¡er (Hidaka 1982). quinine or ca¡eine were strongly rejected (Lamb According to Appelbaum (1980), common carp also and Finger 1995). The shore rockling rejected meat released a negative taste response to food containing soaked in warmot extract (Artemisia sp. (Asteraceae); sucrose;however, as was shown using another beha- Andriashev 1944) and wild gold¢sh rejected pellets vioural approach, sucrose is an indi¡erent taste sub- containing calcium chloride (Kasumyan and Niko- stance for common carp, at least at a concentration laeva 2002). of 0.29 m (10%;Kasumyanand Morsy1996). For humans, calcium chloride has a bitter and Acids: it was shown that food with 0.1 m HCl is astringent taste. In contrast to quinine, calcium rejected by the pu¡er (Hidaka et al. 1978) a nd t he chloride has been found to be an indi¡erent taste sub- shore rockling (Andriashev 1944). The pike, the zan- stance for17¢sh species, a deterrent for three species der and the vendace rejected food containing citric and a stimulant for seven species (Tables 4 and 5). acid (Appelbaum 1980). Atlantic salmon consumed Sodium chloride: sodium chloride is either an little food that has a sour taste (Stradmeyer1989). In indi¡erent taste substance or a stimulant for ¢sh contrast, acid food was willinglyconsumed by tilapia (Tables 4 and 5). The addition of sodium chloride to (Adams et al. 1988), Nile tilapia, European eel and starch gel or to starch gel containing synthetic clam rainbow trout (Appelbaum1980). In a number of ¢sh extract did not enhance or reduce the pellet con- species, citric acid is a highly e¤cient stimulant, in sumption by fugu (Hidaka 1982). Shore rockling others it is a strong deterrent (Kasumyan 1997; rejected meat pieces soaked in 10% sodium chloride Tables 4 and 5). In total, 37 ¢sh species were investi- (Andriashev 1944). A high level (60%) of sodium gated concerning their taste preferences to diet con- chloride made arti¢cial food unpalatable for Eur- taining citric acid or other acids. For16 ¢sh species, opean eel, but the taste response of rainbow trout acids strongly depressed food acceptance, and for was positive for sodium chloride at a very high con- the other 15 ¢sh species, the palatability of food was centration (80%;Appelbaum1980). highly increased after the addition of acid to the diet. Citric acid was an indi¡erent taste substance for eight Free amino acids ¢sh species only. It is noteworthy that acidi¢ed diets are palatable mostly for salmonids and poecilids, but The food of ¢shconsists of avarietyof preyorganisms evoke a negative taste response in acipenserids and and aquatic plants. All these feeds contain consider- in manycyprinids. able amounts of free amino acids (Dabrowski and Quinine, calcium chloride and related compounds: Rusiecki 1983;De la Noue and Choubert 1985;Holm quinine is a highly deterrent substance for all ¢sh and Walter1988;for review, see Carr et al. 1996). Fish species tested. Smooth dog¢sh rejected mackerel taste receptors are highly sensitive to free amino impregnated with quinine immediatelyafter the food acids and induce strong electrophysiological had been snapped up (von Uexku« ll 1895). Sheldon responses in recordings of the nervous activity from (1909) found that a solution of quinine hydrochloride gustatory ¢bres (for review, see Marui and Caprio produced jerk responses and other movements in 1992). Thus, free amino acids have been found to be the smooth dog¢sh, when applied to the inside of the highly e¤cient incitants or stimulants for various mouth and spiracle, but not when applied to other freshwater and marine ¢sh (Hidaka 1982;Mackie parts of the body. Evidently, this bitter-tasting sub- 1982;Mackie and Mitchell 1983;Mearns et al. 1987; stance elicited responses only from the regions that Adams et al.1988;Jones1989;Lamb and Finger1995). bear the taste buds. Quinine solution applied extrao- In some species, the mixtures of free amino acids rally had released an avoidance response in some tel- alone had the same level of palatability as the whole eost ¢sh (Scharrer et al. 1947), and Hidaka et al. extract of preferable food organisms (Fuke et al.1981; (1978) reported that quinine inhibited food consump- Mackie 1982;Takeda et al. 1984;Johnsen and Adams tion in the pu¡er. Quinine and several other bitter- 1986;Mearns et al.1987). tasting compounds like papaverine, theophylline For European eel, a mixture of neutral and acidic and ca¡eine inhibited the feeding behaviour of the amino acids as well as neutral amino acids only were turbot (Mackie 1982). Quinine, papaverine, sodium the most palatable, while aromatic and basic amino ricinoleate and sodium taurocholate, all of which acids were inactive (Mackie and Mitchell 1983). The taste bitter to humans, were found to be feeding same conclusions were made after testing various

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 301 at preferences Taste 302 Table 4 The index of palatability to classical taste substances for 27 di¡erent ¢sh species

Substances Concentration, Brown trout, Brook char, Lake char, Frolich char, Chum salmon, Common carp, Crucian carp, Goldfish, Roach, mM (%) Salmo trutta Salvelinus Salvelinus Salvelinus Oncorhynchus Cyprinus carpio Carassius Carassius Rutilus rutilus caspius fontinalis namaycush alpinus keta carassius auratus

erythrinus DÖving B Kjell & Kasumyan O Alexander

Citric acid 0.26 (5) 78.7ÃÃÃ 95.2ÃÃÃ 60.0ÃÃÃ 53.4ÃÃÃ À100ÃÃÃ 53.9ÃÃÃ À61.2ÃÃÃ À31.0ÃÃÃ À100ÃÃ Sodium chloride 1.73 (10) 43.5ÃÃÃ 60.7 48.4ÃÃÃ 12.6 12.3 22.4 0.5 À15.9Ã 32.7Ã Calcium chloride 0.9 (10) 37.7ÃÃ 87.7ÃÃÃ 46.7ÃÃÃ 17.7 À7.8 30.4ÃÃ 6.1 À14.4Ã 21.8 Sucrose 0.29 (10) À10.2 90.6ÃÃÃ 11.2 9.7 À14.7 À13.8 0 1.9 32.7Ã

Dace, Chub, European Tench, Bitterling, Grass carp, Tiger barbus, Zebra®sh, Nine-spined Leuciscus Leuciscus minnow, Tinca tinca Rhodeus Ctenopharyn- Puntius Brachydanio stickleback, leuciscus cephalus Phoxinus sericeus godon idella tetrazona rerio Pungitius phoxinus amarus pungitius Citric acid 0.26 (5) 17.1 À9.8 10.3 56.4ÃÃÃ À89.1ÃÃÃ 46.0ÃÃÃ À18.5 100 66.7ÃÃÃ Sodium chloride 1.73 (10) 24.4Ã 17.9 7.9 36.8ÃÃÃ À6.3 À5.1 11.0 100 À10.4 ÃÃÃ ÃÃÃ

# Calcium chloride 0.9 (10) 9.2 44.0 14.8 33.4 À7.4 À20.9 15.7 0 À11.0 ÃÃ ÃÃÃ Ã 03BakelPbihn t,FISHadFISHERIES, E I R E H S I F and H S I F Ltd, Publishing Blackwell 2003 Sucrose 0.29 (10) 28.6 À1.3 7.5 10.6 5.6 63.4 10.0 100 À0.7

Atlantic Banded cichlid, Black molly, Guppy, Platy, Siberian Stellate Arctic Navaga, wolf®sh, Heros severum Poecilia Poecilia Xiphophorus sturgeon, sturgeon, ¯ounder, Eleginus Anarhichas sphenops reticulata maculatus Acipenser Acipenser Liopsetta navaga lupus baerii stellatus glacialis Citric acid 0.26 (5) 42.3Ã À22.0ÃÃ 1.3 95.2ÃÃÃ 77.2ÃÃÃ À89.5ÃÃ À89.5ÃÃ À11.3Ã 28.9 Sodium chloride 1.73 (10) 35.3 À73.5ÃÃÃ 33.4ÃÃÃ À100 À87.1ÃÃÃ À50.0ÃÃ À47.5ÃÃ 9.2Ã À100 Calcium chloride 0.9 (10) 31.1 21.5ÃÃÃ À19.3 72.4 23.6 À100ÃÃÃ À83.7ÃÃ 6.9 9.5 Sucrose 0.29 (10) 3.3 À13.5 53.5ÃÃÃ 83.7ÃÃÃ 74.5ÃÃÃ À12.5 À7.1 2.0 20.3

À1 The index of palatability (in percent) was calculated by the formula Ipal ˆ 100 Á (R À C) Á (R ‡ C) , where R is the number consumed of pellets containing a particular substance and C is the number of blank pellets consumed. See text for details. 4 ,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 5 Taste reactions to di¡erent substances Table 6 Stimulants and deterrents

Concentration Species Stimulants Deterrents Substances mM (%) Stimulants Deterrents

Platy (Xiphophorus maculatus)132 Alanine 0.1 12 2 Tench (Tinca tinca)120 Cysteine 0.1 9 5 Chum salmon (Oncorhynchus keta)12 2 Serine 0.1 8 2 Black molly (Poecilia shenops)811 Glutamine 0.1 7 1 Gold®sh (Carassius auratus)83 Glycine 0.1 7 1 Roach (Rutilus rutilus)80 Threonine 0.1 6 3 Dace (Leuciscus leuciscus)80 Histidine 0.1 6 4 Siberian sturgeon (Acipenser baerii)7 1 Proline 0.1 6 4 Russian sturgeon 60 Arginine 0.1 6 6 (Acipenser gueldenstaedtii) Valine 0.1 6 6 Common carp (Cyprinus carpio)6 7 Phenylalanine 0.1 5 5 Navaga (Eleginus navaga)67 Asparagine 0.1 4 2 Brown trout (Salmo trutta caspius)6 5 Methionine 0.1 4 5 Guppy (Poecilia reticulata)50 Lysine 0.1 3 4 Frolich char (Salvelinus 41 Norvaline 0.1 3 2 alpinus erythrinus) Glutamic acid 0.01 7 3 European minnow (Phoxinus phoxinus)4 3 Triptophan 0.01 6 2 Bitterling (Rhodeus sericeus amarus)4 9 Aspartic acid 0.01 6 3 Stellate sturgeon (Acipenser stellatus)3 0 Isoleucine 0.01 3 3 Grass carp (Ctenopharyngodon idella)3 17 Leucine 0.01 3 3 Chub (Leuciscus cephalus)10 Tyrosine 0.001 7 2 Crucian carp (Carassius carassius)0 0 Citric acid 0.26 (5) 11 9 Arctic flounder (Liopsetta glacialis)0 0 Sodium chloride 1.73 (10) 7 5 Calcium chloride 0.9 (10) 7 3 The number of amino acids that acts as stimulants or deterrents Sucrose 0.29 (10) 8 0 in 21 ®sh species. The number of amino acids (l-isomers) tested was 21 for all species except the Siberian sturgeon that was The assessment of the palatability of amino acids (l-isomers) and tested for 19 amino acids. classical taste substances based upon trials on different ®sh spe- cies. The number of ®sh species examined was 20 for cysteine was found for three species of the genus Acipenser and norvaline and 21 for the other amino acids, and 27 for clas- (Kasumyan and Sidorov1994;Kasumyan1999a). sical taste substances. The assessment of palatability of taste The spectra of stimulatory free amino acids are substances was made in comparison to blank pellets. highly species-speci¢c, and the list of palatable mixtures of di¡erent amino acids for European sea amino acids is di¡erent for di¡erent ¢sh species (for bass, where it was also shown that the nonamino details see section `Species speci¢city';Table 7). The acid components of synthetic squid extract wereinef- amino acids l-alanine, l-cysteine and l-serine act fective as taste stimulants (Mackie and Mitchell most frequently as a stimulant for 21 ¢sh species 1982 a). (Table 5). Inaddition, the l-alanine and l-serinewere Not all amino acids, however, are palatable for ¢sh. palatable for sea bream (Goh andTamura1980a), tila- In a series of experiments performed by Kasumyan pia (Johnsen and Adams 1986) and pu¡er (Hidaka and collaborators, the oral taste responses evoked by 1982).The amino acids l-glutamine, glycine, l-gluta- common free amino acids were investigated for 21 mic acid, l-tyrosine, l-arginine, l-histidine, l-threo- ¢sh species (reviewed in Kasumyan 1997;Table 6). nine, l-valine, l-tryptophan and l-aspartic acid The data reveal that the number of amino acids that followed as stimulants in that order after l-alanine, acted as stimulants ranged from 0 to 13. There are l-cysteine and l-serine; l-lysine, l-norvaline, l-leu- onlya few ¢sh species for which the number of stimu- cine and l-isoleucine were rarely observed as stimu- latory amino acids reaches more than 10. For 13 ¢sh lants (Table 5);however, some of these amino acids, species, the number of stimulatory amino acids was l-leucine and l-isoleucine, were stimulants for rain- six or less. On an average, for all ¢sh species tested, bow trout (Jones1989). 28% of the free amino acids were stimulatory. The The essential amino acids, which are important as range of free amino acids acting as incitants for ¢sh nutrients for most animals, are also a requirement is wide and includes 14, 15 and 19 amino acids, as for many ¢sh species (Millkin 1982). However, the

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 303 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 7 The indexof palatability to free aminoacids for 21di¡erent ¢sh species

Amino acid Concentration Brown trout, Frolich char, Chum salmon, Common carp, Crucian Goldfish, Roach, (L-isomers) (mM) Salmo trutta Salvelinus Oncorhynchus Cyprinus carp, Carassius Rutilus caspius alpinus keta carpio Carassius auratus rutilus erythrinus carassius

Alanine 100 À6.6 5.1 38.7ÃÃÃ 39.6ÃÃ À1.2 17.1 39.4ÃÃ Arginine 100 31.6 9.8 À20.9Ã À59.8Ã 026.5ÃÃ 44.5ÃÃÃ Asparagine 100 À68.8 À5.9 À6.3 À33.0 À2.3 À28.2 15.3 Cysteine 100 87.4ÃÃÃ 63.2ÃÃÃ À39.3ÃÃÃ 74.1ÃÃÃ À11.6 À5.1 À17.6 Glutamine 100 À100Ã À5.9 12.7 39.2ÃÃ À1.1 À33.0 34.3Ã Glycine 100 À100Ã 0 16.7 À26.7 À1.1 36.1ÃÃÃ 28.7 Histidine 100 58.7ÃÃ 37.9Ã 22.0ÃÃÃ À14.0 À0.6 À30.8 7.7 Lysine 100 0.8 À79.9ÃÃ À5.3 À21.0 À4.6 À28.2 23.1 Methionine 100 À100Ã 0 30.4ÃÃÃ À84.9ÃÃ À3.4 À36.0Ã 14.5 Norvaline 100 À100Ã 5.1 15.1 À38.7 À4.6 À28.2 37.0ÃÃ Phenylalanine 100 52.9ÃÃ 14.2 37.5ÃÃÃ À87.3ÃÃÃ À0.6 À42.9Ã 0 Proline 100 À0.8 9.8 22.6ÃÃÃ 55.7ÃÃÃ 0.6 À64.1ÃÃ 5.0 Serine 100 À71.1 0 27.1ÃÃÃ À87.3ÃÃÃ À1.2 26.1ÃÃ 46.3ÃÃÃ Threonine 100 À54.0 À5.9 7.3 À61.5ÃÃ À2.3 35.2ÃÃÃ 41.7ÃÃÃ Valine 100 À100Ã À12.6 17.1Ã À100ÃÃÃ À2.8 31.2ÃÃÃ 18.7 Aspartic acid 10 67.3ÃÃÃ 65.6ÃÃÃ À2.2 40.7ÃÃ À1.1 20.4 À35.8 Glutamic acid 10 29.0 65.2ÃÃÃ 14.3Ã 42.3ÃÃÃ 0.6 8.1 7.7 Isoleucine 10 À100Ã À5.9 35.3ÃÃÃ À36.1 0.6 33.5ÃÃÃ À19.0 Leucine 10 À39.8 27.8 39.5ÃÃÃ À23.0 0.6 39.2ÃÃÃ 21.7 Tryptophan 10 67.7ÃÃÃ 18.1 14.9Ã À51.5Ã 0 À13.3 35.5Ã Tyrosine 1 76.0ÃÃÃ 0 29.3ÃÃÃ À24.6 À1.1 37.4ÃÃÃ 32.1Ã

Dace, Chub, European, Tench, Bitterling, Grass carp, Black Leuciscus Leuciscus minnow Tinca Rhodeus Ctenopharyn- molly, leuciscus cephalus Phoxinus tinca sericeus godon idella Poecilia phoxinus amarus sphenops Alanine 100 64.3Ã 52.0ÃÃ 26.3Ã 79.2ÃÃÃ 11.9ÃÃÃ À49.6Ã 31.5ÃÃÃ Arginine 100 39.4 0 À54.0ÃÃ 57.1ÃÃÃ 4.9 À66.4ÃÃ À39.6ÃÃÃ Asparagine 100 31.5 À14.1 À6.4 38.2Ã 0 À100ÃÃÃ À32.6ÃÃÃ Cysteine 100 57.2Ã 33.3 À8.8 87.1ÃÃÃ À53.1ÃÃÃ 43.4ÃÃÃ À25.9ÃÃ Glutamine 100 48.4 21.2 27.9Ã 49.6ÃÃ À18.6ÃÃÃ À7.9 28.6ÃÃÃ Glycine 100 58.6Ã 0 20.0 43.2Ã À5.8 28.4ÃÃ À14.7 Histidine 100 À33.3 À23.8 À51.9Ã 35.6 À58.8ÃÃÃ À87.4ÃÃÃ À70.0ÃÃÃ Lysine 100 À2.4 0 À37.8 56.5ÃÃÃ 6.5Ã À85.4ÃÃÃ À67.1ÃÃÃ Methionine 100 31.5 36.5 13.8 54.1ÃÃÃ 6.1Ã À84.5ÃÃ 28.2ÃÃÃ Norvaline 100 65.5ÃÃ 29.8 15.4 22.5 À2.3 À86.4ÃÃÃ 29.7ÃÃÃ Phenylalanine 100 41.0 À100 À65.5ÃÃ 2.9 À26.3ÃÃÃ À84.5ÃÃ 43.7ÃÃÃ Proline 100 57.2Ã 40.0 31.2Ã 74.6ÃÃÃ À18.6ÃÃÃ À100ÃÃÃ À29.0ÃÃ Serine 100 62.3Ã 38.0 17.5 68.5ÃÃÃ 0 À74.5ÃÃ 29.7ÃÃÃ Threonine 100 50.0 37.1 À14.2 17.3 À10.1ÃÃ À100ÃÃÃ 38.9ÃÃÃ Valine 100 73.1ÃÃÃ 30.9 25.9Ã 38.0Ã À29.4ÃÃÃ À72.0ÃÃ À61.9ÃÃÃ Aspartic acid 10 2.2 À4.0 À18.1 45.1ÃÃ À14.2ÃÃÃ 36.4ÃÃ À32.6ÃÃÃ Glutamic acid 10 59.8Ã À4.0 À16.7 26.4 1.0 À66.4ÃÃ À24.3ÃÃ Isoleucine 10 44.6 À50.1 À14.2 À1.5 À1.8 À100ÃÃÃ À29.3ÃÃÃ Leucine 10 À2.4 À46.6 À2.6 0 À7.5Ã À85.4ÃÃ À25.9ÃÃÃ Tryptophan 10 39.4 10.4 1.4 À32.7 2.7 À88.4ÃÃÃ À3.0 Tyrosine 1 À35.2 À52.9 À20.5 À3.1 7.3ÃÃ À100ÃÃÃ 21.6ÃÃ

304 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 7 continued

Guppy, Platy, Siberian Russian Stellate Arctic Navaga, Poecilia Xiphophorus sturgeon, sturgeon, sturgeon, flounder, Eleginus reticulata maculatus Acipenser Acipenser Acipenser Liopsetta navaga baerii gueldenstaedtii stellatus glacialis

Alanine 100 À100 59.7ÃÃÃ À16.3Ã À19.4 58.5ÃÃÃ 3.3 72.1ÃÃÃ Arginine 100 À100 27.5Ã 12.6Ã 28.9Ã À15.4 2.6 À100Ã Asparagine 100 À100 43.9ÃÃÃ 12.5Ã 20.5 14.2 2.6 53.7ÃÃ Cysteine 100 77.5Ã À100ÃÃÃ ± 55.3ÃÃÃ 56.1ÃÃÃ À0.9 À100Ã Glutamine 100 95.3ÃÃÃ 26.1Ã À7.9 21.7 À13.8 4.3 35.8 Glycine 100 97.9ÃÃÃ 16.9 18.9ÃÃÃ 19.4 35.5 2.6 46.8Ã Histidine 100 À100 55.7ÃÃÃ 14.0Ã 8.9 57.0ÃÃÃ 5.8 À24.7 Lysine 100 77.8Ã À30.4 À11.1 25.6 30.5 À3.6 À100Ã Methionine 100 À100 À100ÃÃÃ 6.7 4.0 À3.3 2.6 37.6 Norvaline 100 0 0.4 ± À20.9 À3.0 2.6 35.8 Phenylalanine 100 À100 36.7ÃÃ 6.0 27.6 1.0 2.6 74.8ÃÃÃ Proline 100 50.0 46.9ÃÃÃ À3.1 À9.7 À15.0 2.6 À18.3 Serine 100 À100 41.4ÃÃÃ À1.6 29.4 20.1 2.6 50.3Ã Threonine 100 À100 42.2ÃÃÃ 28.2ÃÃÃ 30.6Ã 4.0 6.6 18.3 Valine 100 0 À1.5 14.6Ã 19.1 À15.0 6.6 À100Ã Aspartic acid 10 0 18.4 5.9 49.6ÃÃÃ 17.6 6.6 À100Ã Glutamic acid 10 96.1ÃÃÃ À5.1 18.0ÃÃ 36.9Ã À37.1 0.9 À100Ã Isoleucine 10 À100 46.8ÃÃÃ 3.8 À6.2 1.6 2.6 34.9 Leucine 10 À100 40.4ÃÃÃ 11.1 À3.4 À1.9 2.6 34.9 Tryptophan 10 0 37.6ÃÃÃ 4.2 39.4ÃÃ À1.3 2.6 61.1ÃÃÃ Tyrosine 1 À100 72.5ÃÃÃ À1.0 À4.5 À4.7 0 À100Ã palatable amino acids were not always among these that the l-amino acid component of a synthetic squid essential amino acids. In common carp, none of the mixture was totally inactive as taste stimulant for essential amino acids were among the palatable sub- this ¢sh, but the nonamino acid components were stances (Kasumyan and Morsy1996). highly stimulatory (Mackie1982). Ikeda et al. (1988b) The amino acids are indi¡erent taste substances showed that for jack mackerel, only l-tryptophan, for ¢sh more often than they are stimulants. In parti- among 20 free amino acids tested, had a positive cular, l-norvaline, l-leucine, l-lisoleucine, l-aspara- e¡ect on diet consumption. The amino acids fraction gin, l-lysine, l-tryptophan, l-glutamine and glycine of a synthetic krill elicited low feeding activities in are mostly indi¡erent taste substances for ¢sh marbled rock¢sh (Takaoka et al.1990) and a mixture (Tables 5 and 7). l-Alanine, l-asparagine, l-aspartic of several free aminoacids includingalanine, glycine, acid, l-glutamine, glycine, l-histidine, l-serine, l- proline, serine, leucine, valine, histidine and trypto- threonine and l-valine (all in 0.1^1.0 m) did not have phan was ine¡ective as feed enhancers for juvenile stimulatory or deterrent e¡ect for rainbow trout largemouth bass (Kubitza et al.1997). (Jones 1989), whereas tilapia, red sea bream and puf- Aminoacidsmighthaveanaversivetastefor¢sh fer (Goh and Tamura 1980a;Hidaka 1982;Johnsen and thus function as deterrents. Some of these sub- and Adams1986) had no reaction to a range of amino stances signi¢cantly decrease the palatability of acids. The majority of amino acids did not have any diets. Of the amino acids examined by Mackie (1982) e¡ect on oral taste responses in chum salmon, roach, in rainbow trout, l-proline and a mixture of l-taur- Siberian sturgeon, Russian sturgeon, guppy, frolich ine, l-alanine and l-arginine were deterrents. char, European minnow and bitterling. The number Among 21 ¢sh species investigated by Kasumyan of aminoacids classedas indi¡erent taste substances and collaborators, deterrent amino acids were found was 16 in guppy and frolich char,18 in stellate stur- for 12 species. Four of these ¢sh species responded geon and 20 in chub. Moreover, none of the amino negatively for one or two amino acids only;however, acids were found to be stimulatory for crucian carp for some ¢sh, the number of deterrent amino acids and Arctic £ounder (Table 7). This corresponds well was high, being 7^11 for bitterling, common carp, with data obtained for turbot in which it was found navaga and black molly and 17 for grass carp

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 305 Taste preferences Alexander O Kasumyan & Kjell B DÖving

(Table 6). Deterrent taste properties are characteris- trimethylamine oxide (TMAO) for turbot (Mackie tic for l-valine, l-arginine, l-phenylalanine, l- and Adron1978),TMA-HCl,TMAO and NH4Cl for jack methionine, l-cysteine, l-lysine, l-histidine and l- mackerel (Ikeda et al.1988b), inosine for Atlantic cod proline (Table 5). Many of the deterrent amino acids (Johnstone and Mackie 1990), TMA-HCl, TMAO-HCl, are highly aversive taste stimuli. Pellets containing hypoxanthine, inosine, and IMP and l(‡)-lactic acid these substances were rejected by ¢sh in100% of the for European eel (Mackie and Mitchell 1983). More- trials. Pellets contained arginine, cysteine, lysine, over, it was found that omission of betaine,TMA and valine, aspartic acid, glutamic acid and tyrosine, TMAO increased food acceptance in Atlantic cod when fed to navaga, and glutamine, glycine, methio- (Franco et al.1991). The main ingredient of the com- nine, norvaline, valine and isoleucine, when fed to mercial product `Finnstim' (Finnish Sugar Co. Ltd.) is brown trout. Chubs rejected phenylalanine, while betaine, but at a concentration of 1.5%, it did not act valine was rejected by common carp;asparagine, as a stimulant for stellate sturgeon and Siberian stur- proline, threonine and isoleucine were rejected by geon (Kasumyan et al. 1995). Betaine alone at a con- grass carp and cysteine and methionine were centration of 9.7 g of betaine per1 kg of diet had little rejected by platy (Table 7). e¡ect on freeze-dried diet intake by Dover sole juve- The taste properties of amino acids are highly niles with a mean weight of 2.5 g, but did have a sti- stereospeci¢c. d-isomers of amino acids are usually mulatory e¡ect on intake for Dover sole with a mean deterrent or indi¡erent types of taste substances, weight of 50 g (Mackie et al.1980).Thisstimulatory while the corresponding l-forms might be highly activity of betaine was con¢rmed for Dover sole palatable for the same ¢sh species, as shown for rain- (Mackie and Mitchell1982b). Based on electrophysio- bow trout (Adron and Mackie 1978), plaice (Mackie logical studies, the activity in taste nerves induced 1982), European sea bass (Mackie and Mitchell by stimulation of betaine was found to vary among 1982a) and European eel (Mackie and Mitchell1983). ¢sh species (Yoshii et al. 1979;Goh and Tamura The stereospeci¢city of amino acids as gustatory sti- 1980a;Kiyohara et al.1981;Marui et al.1983a,b; muli was shown by electrophysiological methods in Ishida and Hidaka1987;Hara et al.1999). many studies (Caprio1978;Yoshii et al. 1979;K iyohara Betaine has important synergistic properties et al.1981;Marui et al. 1983a,b;Marui and Kiyohara with amino acids or other substances. In plaice, 1987);however, as was shown for the sea cat¢sh, omission of betaine completely abolished activity of there is a group of taste ¢bres in the facial taste sys- nonamino acid components of synthetic squid mix- tem that is more responsive to d-alanine than to its ture, although this substance itself was without feed- l-enantiomer (Michel and Caprio1991). ing stimulant activity (Mackie1982). The palatability of alanine orglycine for red sea bream increased after adding taste-indi¡erent betaine into the diet (Goh Betaine and Tamura 1980a). In extracts of blue crab and £at- Betaine (glycine betaine, trimethylglycine) is widely head grey mullet, a solution containing either distributed in ¢sh food organisms (Konosu et al. betaine alone or the amino acids alone was only 2^ 1966;Konosu and Hayashi 1975;for review, see Carr 9% as e¡ective as the extracts themselves for pig¢sh et al.1996).The gustatory system of many ¢sh species and pin¢sh. Mixtures containing betaine and amino is highly sensitive to this substance (Yoshii et al. acids were 9^16times as e¡ective as anyof the other 1979;Goh and Tamura 1980b;Marui et al.1983a,b; components used alone (Carr and Chaney1976;Carr Ishida and Hidaka1987;Hara et al.1999). et al. 1977;Carr 1982). Unfortunately, the chemosen- Using bioassays, betaine has been identi¢ed as a sory system that mediated ¢sh feeding activity was stimulant for pu¡er (Hidaka 1982) and Dover sole not de¢ned in this study. (Mackie et al.1980;Mackie and Mitchell1982b). How- ever, many ¢sh do not show positive taste response Nucleotides and nucleosides to betaine. Betaine was an indi¡erent taste substance for red sea bream (Goh and Tamura 1980a), plaice Inosine, inosine-50-monophosphate (IMP), 1-methy- (Mackie 1982), rainbow trout (Jones 1989) and large- linosine and inosyl- (30-50)-inosine are feeding stimu- mouth bass (Kubitza et al. 1997). Betaine had no sig- lants for the turbot (Mackie and Adron 1978). ni¢cant e¡ect in chinook salmon intake when tested Inosine and IMP are feeding stimulants for the brill as a single taste substance (Hughes1991,1993), or in (Mackie and Mitchell 1985). IMP, but not inosine, mixtures together with trimethylamine (TMA) and was a feed enhancer for juvenile largemouth bass

306 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

(Kubitza et al.1997). IMPhas also been reported to act Amines as a feeding stimulant for juvenile yellowtail, the e¡ect being potentiated by amino acids (cited by The adding of TMA to a basal ¢sh diet caused a

Hosokawa et al.;Takeda et al.1984).IMP-Na2 was decrease inthe food consumption bychinook salmon highly palatable for jack mackerel;however, other fry (2 g in body weight;Hughes1993), a result similar nucleotides tested (adenosine-50-monophosphate di- to that found for ¢rst-feeding chinook salmon 0 sodium salt, AMP-Na2;adenosine-5 -diophosphate, (Hughes1991).The same type of response was appar- 0 ADP-Na2;adenosine-5 -triphosphate, ATP-Na2;ino- ent when TMA or its oxidation product, TMAO, was sine) were indi¡erent taste substances (Ikeda et al. added to the diet for turbot (Mackie and Adron1978). 1988a). Guanosi ne -5 0-monophosphate (GMP), uri- In contrast, omission of TMAO greatly reduced the dine-50-monophosphate (UMP), uridine-50-diphos- feeding stimulant activityof the nonamino acid com- phate and uridine-50-triphosphate were also found pounds for plaice (Mackie1982). to be stimulants for jack mackerel;however, other Trimethylamine oxide, creatine and creatinine nucleotides, and some of the nucleosides and related were all ine¡ective as taste additives in food for jack compounds (inosine-50-diphosphate, inosine-50-tri- mackerel (Ikeda et al.1988a), and red seabream show- phosphate, guanosine-50-diphosphate, guanosine- ed low taste preference for guanidino compounds ^ 50-triphosphate, inosine-30-monophosphate, uri- glycocyamine and creatine. The mixture of the dine-30-monophosphate, 2-deoxy-IMP;allylthio- organic bases betaine, choline chloride, homarine- IMP;xanthosine-5 0-monophosphate, adenosine, HCl, TMAO, TMA-HCl and DMA-HCl hardly induced guanosine, uridine, hypoxanthine) were ine¡ective any feeding activity (Fuke et al.1981).Dimethylthetin, (Ikeda et al.1991).The mixture of several nucleotides, on the other hand, was found to be a taste stimulant

AMP-Na, IMP-Na, ADP-Na, ATP-Na2 and UMP-Na2, for Dover sole;however, various other quaternary was ine¡ective as a taste additive for red sea bream amines and related compounds were indi¡erent taste (Fuke et al.1981). substances. Examples are dimethylpropiothetin, The nucleotide fraction of krill synthetic extract arsenobetaine, DL-carnitine, trigonelline, homarine, was the most e¡ective in increasing the daily feeding betaine aldehyde and N,N-dimethylglycine. TMA, rate for marbled rock¢sh. Most e¡ective among cadaverine and putrescine, products of decomposi- nucleotides were ADP, AMP and IMP;less e¡ective tion of amino acids and quaternary amines, were all were inosine and UMP and the least e¡ective was inactive as were 2,6-dibromo-phenol and rutin (30- ATP (Takaoka et al. 1990). The nucleotide fraction of 40-5,7-tetrahydroxy-£avone-3b-d-rutinosine;Mackie anannelid (Table 2) was inactive or repellent for juve- and Mitchell1982b). nile Japanese eel. This fraction from the annelid has nearly the same composition as the nucleotide frac- Sugars and other hydrocarbons tion of krill: AMP,UMP-Na2,IMP-Na2, ADP-Na2 and

ATP-Na2 (Takeda et al.1984). Sugars can be stimulants, deterrents or indi¡erent A mixture of the nonamino acid components of stimuli for ¢sh. Jones (1990) found that four sugars, synthetic squid extract had high palability for the d-fructose, sucrose, d-ribose and d-glucose, were plaice. However, omission of AMP, lactic acid or palatable for rainbow trout, with some sugars being TMAO in the mixture greatly decreased the feeding active only at a concentration of 1.0 m. Two other stimulant activity (Mackie1982). sugars, d-galactose and d-xylose were indi¡erent It is knownthat the process of conversion of ATP to taste substances (Jones 1990). Maltose and glucose ADP,AMPand adenosine is the pathway for autolytic were ine¡ective as taste additives for red sea bream degradation of animal £esh, and it has been shown (Fuke et al.1981), and the fraction of Romaine lettuce that these substances evoke searching activity and that contained sugars did not have any e¡ects on the increase the food intake in crustaceans, according feeding of tilapia (Johnsen and Adams1986). to their ranked order of Adenilate Energy Charge Sucrose is a stimulant for di¡erent poecilids (for (AEC;Zimmer-Faust1987): details see section `Classical taste substances'). Two species belonging to the family Poecilidae ^ the black AEC ˆ ATP ‡ 0:5ADP† ATP ‡ ADP ‡ AMP†À1 molly and the platy ^ were used to study taste prefer- Unfortunately, nothing is known about the relation- ences of ¢sh for various sugars and their derivatives. ship betweenthe palatabilityof ¢sh diet and the value In total,15 di¡erent sugars were tested. It was found of AEC. that11and 5 of these sugars were palatable for platy

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 307 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 2 Palatability index for sugars.The palatability indexes for di¡erent sugars at 0.1 m for two Poeciliidae ¢sh species, black molly, Poecilia sphenops and platy, Xiphophorus maculatus.1, Sucrose;2, fructose;3, galactose;4, saccharin;5, glucose;6, arabinose;7, maltose;8, lactose;9, ramnose;10, sorbose;11, mannose;12, sorbitol;13, mannitol;14, xylose;15, ribose.The assessment (Chi-square test) of the palatabilityof taste substances tested was made in comparison fora blankagar (2%) pellets contained red colour dye, Ponceau 4R (0.05 mm) (Kasumyan, unpublished). In all ¢gures in the review,the signi¢cance levels are indicated as follows: ÃÃÃP < 0.001; ÃÃP < 0.01 and ÃP < 0.05. and black molly, respectively. Among these sub- stimulant and 13 other organic acids were deterrent stances, only four were the same for both the species. substances. Oxalic acid and a-keto-glutaric acid have Four sugars ^ arabinose, lactose, ribose and maltose the strongest deterrent taste for bitterling (Table 8). ^ evoked negative taste responses in black molly; The value of Spearman's rank correlation coe¤cient however, none of the sugars tested were deterrents for these19acids was negative and highly signi¢cant for platy. There was signi¢cant correlation between (rs ˆÀ0.91; P < 0.001) between these two cyprinid taste preferences for sugars between these two clo- species. sely related species (rs ˆ 0.52; P ˆ 0.043;Fig. 2). In Electrophysiological recordings in common carp general, the result con¢rms the conclusion that taste showed that the taste responses to monocarboxylic preferences in ¢sh are highly species-speci¢c (for acids reached a peak for compounds having between detail see section`Species speci¢city'). three and ¢ve carbons, whereas dicarboxylic acids gave maximal taste responses for compounds con- taining ¢ve to six carbon atoms (Marui and Caprio Organic acids 1992). These ¢ndings were not matched in beha- The succinic, malic and citric acids were ine¡ec- vioural taste responses of tench or bitterling to car- tive as taste additives for red sea bream (Fuke et al. boxylic acids, even though these three species are 1981). Rainbow trout reacted positive by e-aminoca- considered to be closely related and belong to the proic acid (1.0 m). This acid, as well as l-a-amino- same family (Cyprinidae). In bitterling, for instance, butyric acid and g-aminobutyric acid, were active at monocarboxylic acids containing between three and a 10-fold dilution (Jones 1990). Fractions of Romaine ¢ve carbon atoms, which include propionic acid, n- lettuce containing malic and citric acids, which butyric acid and n-valeric acid, were indi¡erent taste are the main constituents in this lettuce, were substances, and dicarboxylic acids containing ¢ve to ine¡ective taste stimuli for tilapia. Malic acid showed six carbons, such as a-keto-glutaric acid and adipic no activity when examined at 0.01 m, but citric acid acid, were deterrents. In tench, the palatability level was a stimulant at this concentration (Johnsen and of adipic acid was much less than in oxalic, malonic Adams 1986). and succinic acids containing two, three and four The palatability of 19 organic acids was deter- carbon atoms, respectively. mined for two cyprinid species, the tench (Kasumyan and Prokopova 2001) and the bitterling (Kasumyan Alcohols and aldehydes and Prokopova, unpublished). It was found that 17 organic acids were stimulants for tench, the most There arelittle dataonthe palatabilityof alcohols and palatable being malic acid, a-keto-glutaric acid and aldehydes for ¢sh. Some alcohols such as n-hexanol oxalic acid. Cholic acid was an indi¡erent taste sub- and n-octanol were palatable for rainbow trout at a stance for tench. In bitterling, only cholic acid was a relatively low concentration (10À2 and 10À3 m), but

308 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 8 The index of palatability to organic acids at a 1948;Cantoni and Anderson1956), which generates concentration of 0.1 m for two cyprinid ¢sh, tench (Tinca the odour of the sea in humans (Iida et al. 1985;Kasa- tinca) and bitterling (Rhodeus sericeus amarus) hara and Nishibori 1987). This substance appears to be formed in alga (Iida et al. 1985) and higher plants Substances Tench Bitterling (Larher et al. 1977;Van Diggelin et al. 1986). Remark- ably, levels of DMPT up to 440 mg kgÀ1 and DMS up À1 Maleic acid 97.2ÃÃà À85.1 to14 mg kg were found in various shell¢sh species a-Ketoglutaric acid 97.2ÃÃà 100ÃÃà (Iida andTokunaga1986).The dietary administration Oxalic acid 97.1ÃÃà 100ÃÃà of degradative compounds such as DMS, acrylic acid ÃÃà Tartaric acid 96.9 À77.3 of DMPT, oxidative compounds like dimethyl sulfox- Malic acid 96.7ÃÃà À86.7 Citric acid 96.6ÃÃà À74.7 ide, dimethyl sulphone, dimethyl sulphite of DMS, Malonic acid 96.5ÃÃà À88.0 various other dialkyl sulphides was much less e¡ec- Glycolic acid 95.4ÃÃà À38.6 tive than was DMPT when tested as feed to gold¢sh. Succinic acid 91.5ÃÃà À73.7 The stimulatory e¡ects of DMPT were much stronger ÃÃà Fumaric acid 90.0 À74.0 than those of various closely related compounds, Caproic acid 87.9ÃÃà À49.4 Adipic acid 88.5ÃÃà À58.3 which bear shorter or longer methylene and dialkyl Valeric acid 85.0ÃÃà À14.8 carbon chains attached to the sulphur atom com- Butyric acid 85.0ÃÃà 0.2 pared with that on the DMPT molecule (Nakajima Ascorbic acid 79.4ÃÃà À58.3 et al.1989b). Both DMPT-enhanced and -unenhanced à Formic acid 69.8 4.0 diets are ingested in almost the same amounts with Propionic acid 69.8à À19.2 Acetic acid 60.6 2.2 ¢sh (Nakajima et al.1989a), indicating that the e¡ect Cholic acid 0 15.5ÃÃà of DMPT may be connected with metabolic e¡ects rather thantaste attractiveness.There is abriefreport that DMT, but not DMP, increases food intake in in most othercompounds tested, the activity waslost. Dover sole (Mackie and Mitchell1983). Other alcohols tested such as n-propanol and n-buta- nol were indi¡erent taste substances as were the Mixtures aldehydes ^ propionaldehyde and butyraldehyde (Jones1990). When stimulants are given together in an experi- mental pellet ordiet, there is oftena synergistic e¡ect. This synergism has been shown for pig¢sh, for which Other types of substances a synthetic mixture of betaine plus 19 amino acids It was shown that dimethyl-b-propiotethin (3- was prepared so that the concentration of each sub- dimethyl-3-thiopropanol;DMTP) and dimethyl-thio- stance was identical to that determined in a shrimp proprionic acid (2-carboxy-ethyl dimethyl sulpho- ¢ltrate, and was as stimulatory as the ¢ltrate itself. nium bromide;DMPT) are very e¡ective in increasing Both betaine and certain amino acids contributed to the number of snaps of test ¢sh directed at suspended the activity of the synthetic mixture. A mixture con- food paste. Presence of these substances in food taining the 19 amino acids alone was only about increases growth of red sea bream, yellowtail and 28% as e¡ective as the shrimp ¢ltrate, whereas a Japanese £ounder (Nakajima et al.1990).Thesesub- solution of betaine alone was only 39% as e¡ective stances also stimulate feeding and growth of some as the ¢ltrate (Carr1976). It is important to note that freshwater ¢sh. Among various sulphur-containing it is still unclear what chemosensory system ^ gusta- organic compounds, DMTP and, to a lesser extent, tion or olfaction ^ mediates behavioural responses of dimethylthetin, dipropyl(di)sulphide, dimethylsulfox- pig¢sh to ¢ltrate or mixtures. ide and dimethylsulphone, promoted snapping beha- Takeda et al. (1984) found that uracil-50-monopho- viour in gold¢sh, crucian carp and common carp. phate (5-UMP) enhanced the feeding activity of the The e¡ect of DMPT was highest at the concentration Japanese eel to amino acids, while inosine-50-mono- of 10À3 m and decreased when the concentration phophate (5-IMP) did not. Synergistic e¡ects of 20 was di¡erent from this value (Nakajima et al. free amino acids and 6 nucleotides were found for 1989a,b;Nakajima 1991). A tertiary sulphonium jack mackerel (Ikeda et al. 1988a). A mixture of amino compound, DMPT, has been shown to be a precursor acids alone or a mixture of nucleotides alone had no of dimethyl sulphide (DMS;Challenger and Simpson e¡ect on diet consumption;however, the mixture of

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 309 Taste preferences Alexander O Kasumyan & Kjell B DÖving

these two main components of synthetic krill extract rejected at concentrations higher than 10À2.5 m evoked a positive response. Ikeda and coworkers (Lamb and Finger 1995). Deoxycorticosterone (21- showed that among 20 free amino acids, tryptophan hydroxypregn-4-ene-3,20-dione) at oral doses of was the only substance that increased diet consump- 660 mg(2Â10 À6 m) per pellet caused 94% inhibition tion. Nevertheless, a diet containing a mixture of all in the acceptance of arti¢cial food pellets in sun¢sh. amino acids, including tryptophan, was ine¡ective. At the same molar dosage, pregn-4-en-20a-ol-3-ene Among six nucleotides, only IMP was a stimulant, inhibited food consumption by 58%, while its epimer, whereas a mixture of nucleotides was ine¡ective. pregn-4-en-20b-ol-3-ene did not inhibit feeding Adding the ine¡ective nucleotides and other com- signi¢cantly (Gerhart et al. 1991). The level of deoxy- pounds (TMAO, creatine, creatinine or ammonia) corticosterone in the treated pellets was similar to into the diet, together with IMP,signi¢cantly reduced the levels that can occur in aquatic beetles in the food consumption (Ikeda et al.1988a,b). It seems plau- family Dytiscidae (Schildknecht1971).The minimally sible that the presence of inactive taste substances active concentration at which stevensine deterred together with a stimulant in a mixture may mask feeding of the bluehead wrasse was between the positive e¡ect of a stimulant. 5.5 Â10 À3 and 6.0 Â 10 À3 m (2.0^2.25 mg mLÀ1; Wilson et al.1999). The taste response of red sea bream to alanine at Thresholds 10 À3 m concentration was doubtful, but 2 Â10 À3 m The minimal concentrations needed to give a positive l-alanine was e¡ective to induce the preference or negative response of a ¢shtowards agiventaste sti- response. The behavioural response to l-alanine mulus is an important parameter when it comes to increased with increasing stimulus concentrations £avouring food or bait. Systematic behavioural stu- in the range between10À3 and10À2 m and was satu- dies of threshold parameters have been performed rated at 10À2 m (Goh and Tamura 1980a). In tench, for a number of species. For the most common gusta- the acceptance ratio of pellets containing malic acid tory substances, the threshold concentrations were increased six times in the range between 10À2 and quite high, but for a few substances, the threshold 10 À1 m. The dose^response relationships for the was relatively low. acceptance ratio for two ¢sh species are presented in The threshold concentration for stimulants is Fig. 3. The retention time for keeping the pellets in usuallyaround10À2 to10À4 m (Table 9). For common the mouth increased with concentration, an example carp, the threshold concentrations for stimulants of which is shown in Fig. 4. were10À2 m for l-cysteine,5 Â10 À3 m for citric acid Fora single species, namely tench, the tastethresh- and 0.9 m forcalcium chloride (KasumyanandMorsy old has been found to vary with the method used to 1996). In the brown trout, the threshold concentra- assess the palatability;thus acceptance ratio, num- tions of highly palatable amino acids were in the ber of snaps at pellets, retention time of the pellet at range of 10À2 to 10À3 m (Kasumyan and Sidorov ¢rst snap and total retention time during a trial give 1995c). In the Siberian sturgeon, the threshold con- di¡erent threshold values (Table 10). centrations for oral taste were 5 Â10 À2 m for citric Based on threshold concentrations determined by acid, 9 Â10 À2 m for calcium chloride and 1.7 m for behavioural assays, it is possible to estimate the sodium chloride (Kasumyan and Kazhlaev 1993a). amount of substance in one pellet, which is su¤cient The threshold for a range of stimulants for the pu¡er to release a signi¢cant taste response.These amounts and tilapia was shown to be about 10À2 m (Hidaka were 4.27 and 3.39 mgforl-cysteine and citric acid, 1982;Adams et al. 1988). In the rainbow trout, a few respectively (common carp), 0.236 mgforl-alanine substances were palatable at a concentration of (European minnow) and 0.353 mgforl-aspartic acid 10 À3 m and only l-proline was e¡ective at the con- (frolich char). These values correspond to 10À9 to centration10À4 m (Jones 1989,1990). Quinine, papa- 10 À10 mole or 1015 À1016 molecules per pellet. How- verine, sodium ricinoleate and sodium taurocholate, ever, all substances except those that are situated in all taste bitter to the human tongue and were found the surface layer in a pellet do not stimulate the gus- to be feeding deterrents for the Dover sole at a con- tatory receptors. The substance part inside the pellet centrationof 5 Â10 À6 mol gÀ1ina moist diet (Mackie is unavailable to taste receptors.Therefore, the actual and Mitchell 1982b). In the gold¢sh, quinine-£a- quantities of the substance su¤cient for eliciting a voured pellets were rejected at concentrations higher signi¢cant taste response would be lower than calcu- than 10À5 m, and ca¡eine-£avoured pellets were lated by one to two orders of magnitude.

310 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 9 Taste thresholds

Substances Threshold concentration (mM) Species Source

Incitants Citric acid 5 Acipenser baerii Kasumyan and Kazhlaev (1993b) 5 Acipenser stellatus Kasumyan and Kazhlaev (1993b)

Suppressants Calcium chloride 90 Acipenser baerii Kasumyan and Kazhlaev (1993b) 90 Acipenser stellatus Kasumyan and Kazhlaev (1993b)

Stimulants L-Alanine 1 Phoxinus phoxinus Kasumyan, unpublished 2 Chrysophrys (Pagrus) major Goh and Tamura 1980 100 Fugu pardalis Hidaka (1982) L-Aspartic acid 1 Salvelinus alpinus erhytrinus Kasumyan and Sidorov, unpublished 10 Salmo trutta caspius Kasumyan and Sidorov (1995) Betaine 10 Fugu pardalis Hidaka (1982) L-Cysteine 5 Salvelinus alpinus erhytrinus Kasumyan and Sidorov, unpublished 10 Cyprinus carpio Kasumyan and Morsy (1996) 25 Tinca tinca Kasumyan and Prokopova (2001) L-Glutamic acid 5 Salvelinus alpinus erhytrinus Kasumyan and Sidorov, unpublished L-Glutamine 10 Phoxinus phoxinus Kasumyan, unpublished Glycine 100 Fugu pardalis Hidaka (1982) L-Leucine 10 Oncorhynchus mykiss Jones (1989) L-Phenylalanine 10 Salmo trutta caspius Kasumyan and Sidorov (1995)

10 Oncorhynchus mykiss Jones (1989) L-Proline 0.1 Oncorhynchus mykiss Jones (1989) 10 Phoxinus phoxinus Kasumyan, unpublished

10 Fugu pardalis Hidaka (1982) L-Serine 10 Fugu pardalis Hidaka (1982) L-Tyrosine 0.1 Xiphophorus maculatus Kasumyan, unpublished

1 Salmo trutta caspius Kasumyan and Sidorov (1995) L-Valine 10 Phoxinus phoxinus Kasumyan, unpublished Citric acid 5 Cyprinus carpio Kasumyan and Morsy (1996) 1.0±10 Tilapia zillii Adams et al. (1988) 10 Thymallus thymallus Kasumyan and Sidorov, unpublished 50 Acipenser baerii Kasumyan and Kazhlaev (1993b) 50 Acipenser stellatus Kasumyan and Kazhlaev (1993b)

Calcium chloride 90 Acipenser baerii Kasumyan and Kazhlaev (1993b) 90 Acipenser stellatus Kasumyan and Kazhlaev (1993b) 900 Cyprinus carpio Kasumyan and Morsy (1996)

Sodium chloride 1700 Acipenser baerii Kasumyan and Kazhlaev (1993b) Sucrose 3 Ctenopharyngodon idella Kasumyan, unpublished Xylose 10 Leuciscus leuciscus Kasumyan and Fokina, unpublished Succinic acid 10 Thymallus thymallus Kasumyan, unpublished a-Ketoglutaric acid 10 Thymallus thymallus Kasumyan, unpublished Maleic acid 25 Tinca tinca Kasumyan and Prokopova (2001) IMP 2 Â 10À6a Trachurus japonicus Ikeda et al. (1991) Inosine 6 Â 10À8a Scophthalmus maximus Mackie and Adron (1978)

Deterrents Quinine 0.1 Carassius auratus Lamb and Finger (1995) Caffeine 50 Carassius auratus Lamb and Finger (1995) Stevensine 5.5±6.0 Thalassoma bifasciatum Wilson et al. (1999)

The threshold to different types of taste substances in ®sh. Values marked by an superscript letter are in mM gÀ1 dry diet.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 311 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 3 E¡ect of concentration on acceptance ratio. Dose^response relationship for acceptance ratio of l-cysteine in Frolich char, Salvelinus alpinus erythrinus, juveniles and for acceptance ratio of l-Glutamine in nine-spined stickleback, Pungitius pungitius (Gasterosteidae). Control ^ blank pellets.Vertical columns and T-bars are the mean and the SEM (Kasumyan, Sidorov and Fokina, unpublished).

The threshold concentrations for substances limited number of taste receptors are stimulated determined using bioassays are much higher than because normally the pellet is relatively small ^ the thresholds determined by electrophysiological 2.5 mm in length and 1.5 mm in diameter. Further, recordings from the teleost ¢sh. Recordings from the in the electrophysiological studies, the solution is facial, glossopharyngeal or vagal taste nerves applied for a long period (5 s;Hara et al. 1999). In revealed threshold concentrations between 10À7 behavioural assays, the time that a ¢sh spends in and10À9 m for free amino acids and some other sub- making a decision about the taste of the food item stances (Konishi and Zotterman 1961;Kaku et al. and consequently either swallowing or rejecting 1980;Kiyohara et al. 1981;Davenport and Caprio it, is generally shorter than in the electrophysiolo- 1982;Kanwal and Caprio 1983;Marui et al.1983a,b; gical experiments (Kasumyan and Sidorov 1993a,b, Marui and Caprio1992;Hara et al.1999). There might 1995 c). be several reasons for the discrepancy between the The discrepancy between the threshold concen- threshold concentrations obtained by behavioural trations obtained in the behavioural and electrophy- and electrophysiological methods. It was supposed siological experiments could be associated with that the discrepancy might be related to the methods peculiarities of functional relationship between the used to stimulate the taste receptors (Kasumyan and taste and the mechanosensory system. In beha- Sidorov 1995c;Kasumyan and Morsy 1996). In the vioural assays, both the chemoreceptors and the electrophysiological studies, a relatively larger area mechanoreceptors are stimulated simultaneously. with taste receptors in the ¢sh mouth or barbel is irri- This aspect might be of signi¢cance and there might gated by the test solution. In the behavioural assay,a be a functional interaction between the gustatory

Figure 4 E¡ect of concentration on pellet retention time. Dose^response relationship for pellet retention time contained l-glutamine in nine-spined stickleback Pungitius pungitius. Control ^ blank pellets.Vertical columns and T-bars are the mean and the SEM (Kasumyan and Fokina, unpublished).

312 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 10 Taste responses to di¡erent concentrations

Mean retention time (s)

Concentration Acceptance Mean number Number Substance (mM) ratio (%) of snaps First capture All captures of trials

Maleic acid 100 90.7ÃÃÃ 1.1ÃÃÃ 19.8ÃÃÃ 20.2ÃÃÃ 150 50 53.3ÃÃÃ 1.3ÃÃÃ 13.3ÃÃÃ 14.9ÃÃÃ 150 10 14.7ÃÃÃ 1.3ÃÃÃ 5.3ÃÃÃ 6.1ÃÃ 204 1 3.7 1.4ÃÃ 2.2 3.3 54 0 2.2 2.6 2.3 4.1 225

L-Cysteine 100 90.1ÃÃÃ 1.2ÃÃÃ 18.6ÃÃÃ 19.7ÃÃÃ 91 50 61.5ÃÃÃ 1.5 12.8ÃÃÃ 14.8ÃÃÃ 91 25 37.8ÃÃÃ 1.5 6.3ÃÃÃ 7.3ÃÃÃ 90 1 7.8 1.6 3.1Ã 4.1Ã 115 0.1 7.9 1.6 2.3 3.2 76 0 4.0 1.5 2.1 2.9 75

Behavioural taste response of tench, Tinca tinca for agar pellets containing maleic acid or L-cysteine in different concentrations (modi®ed from Kasumyan and Prokopova 2001). and the mechanosensory systems. It has been sug- a food particle of taste buds is essential for mediation gested that ¢sh taste buds (basal cells) are mechano- of the behavioural taste response. In other words, it sensitive (Jakubowski 1983;Reutter 1986;Reutter might be that the mechanosensory system facilitates 1987;Reutter 1992). The taste and tactile inputs are the transmission of taste information to the motor topographically closely connected in gustatory cen- centres in the ¢sh brain. It is also conceivable that tres located in the medulla oblongata. In channel cat- the mechanosensory system can desensitise this ¢sh, the facial gustatory centre, a facial lobe, is pathway,if notactivate it.This hypothesis is supported organised into lobules for representation of taste and by the observation that some anosmic ¢sh loose their tactile inputs from the barbels, lips, snout and £ank. ability to react even at high concentrations of food In cyprinids such as gold¢sh, the centre for intraoral extract, in spite of an intact taste system (Kasumyan taste, the vagal lobe, is organised anatomically into and Ponomarev1986, 1989;Kasumyan and Kazh laev layers. This structure contains a representation of 1993b;Kasumyan and Devitsina1997), although the taste and tactile inputs from the palatal organ in the anosmic cat¢sh of the genus Ictalurus does not lose oral cavity (Morita and Finger 1985;Kanwal and its ability to ¢nd food (Bardach et al.1967). Caprio 1988;Marui et al. 1988;Hayama and Caprio 1990;Kanwal and Finger 1992). There is evidence Speci®city of taste preferences that chemo- and mechanosensory systems are tightly knit in the oral cavity.In bottom-feeder cypri- Species speci¢city nid ¢sh, the local contraction of the palatal organ musculature, which is entirely dependent upon the It has been suggested that taste stimulants are simi- local characteristics of the gustatory stimulation, lar within a ¢sh family (Mackie and Mitchell 1985). traps the desirable food particles between the palatal This conclusion is based on the results obtained for organ and the £oor of the oral cavity. Once the food only a few ¢sh species and can therefore be consid- particles have been trapped, the waste is £ushed out ered preliminary in nature. During the last decade, from the oral cavity and subsequently,the food parti- there has been signi¢cant progress inthe ¢eld of taste cles are swallowed (Sibbing et al. 1986;Osse et al. preferences of ¢sh. A numberof di¡erent species have 1997). In sturgeon, the food particles containing inci- been tested and more extensive methods have been tants were snapped at after preliminary touching of used;thus the basis for the classi¢cation of ¢sh spe- food items by barbels covered by external taste buds cies by their taste preferences and their relationships (Kasumyan and Kazhlaev1993a). to ¢sh taxonomy has increased. The number of spe- It might be suggested that the stimulation of cies studied is still narrow;however, a comparative mechanoreceptors associated with the ¢rst touch by analysis of these scanty data is possible because the

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 313 Taste preferences Alexander O Kasumyan & Kjell B DÖving

taste preferences for the majority of species were stimulant for common carp, grass carp and tench, based upon similar methods, using the same sub- but a deterrent for crucian carp, roach and bitter- stances and the same carrier. ling. Sodium chloride is a stimulant for tench, roach Comparative analyses of results obtained for more and dace, but a deterrent for wild gold¢sh. Calcium than 20 ¢sh species reveal that oral taste preferences chloride is a stimulant for common carp, chub and are highly species-speci¢c. The di¡erences among tench, but a deterrent for wild gold¢sh (Table 4). ¢shspeciesareapparentatacomparisonofthewidth There are examples of similarities in taste of spectra for stimulants and deterrents. The results responses in closely related ¢sh species. This is more presented in Table 6 demonstrate that few of the 21 evident among classical taste substances than aminoacids testedare e¡ective taste stimuli for chub, among free amino acids. For instance, all four classi- stellate sturgeon, guppy and frolich char. In Arctic cal taste substances possess the same taste properties £ounder and crucian carp, all free amino acids were for Siberian sturgeon and stellate sturgeon (genus indi¡erent. In grass carp and platy, many amino Acipenser). Sucrose was a stimulant for three Poecilii- acids were either stimulants or deterrents. The ratio dae species tested. Citric acid was a stimulant for lake between the number of stimulant amino acid to the char, brook char and frolich char and it was a deter- number of deterrent amino acids is 3 : 17in grass carp rent for crucian carp and wild gold¢sh. Calcium and 13: 2 in platy (Kasumyan and Sidorov 1995d; chloride was a deterrent for brook charand lake char, Kasumyan and Nikolaeva 1997, 2002;Kasumyan and was an indi¡erent stimuli for frolich char 1999 a). (Table 4). Thespectraofstimulantanddeterrentaminoacids The palatabilityof substances is di¡erent in sympa- di¡er not only in the width but also in composition. tric ¢sh that inhabit the same water, and have the In many cases, the same amino acid evokes dramati- samebiotopesandsimilar feedingecologyandfeeding cally di¡erent oral taste response in ¢sh, including behaviours. For example, the taste preferences for closely related species from the same familyor genus. a number of amino acids did not coincide amongst l-Valine, l-serine, l-threonine, l-arginine, l-gluta- European minnow,dace and chub.The experimental mine and l-proline cause opposite taste responses specimens of these ¢sh were caught in the early in common carp versus wild gold¢sh; l-cysteine, autumn of their ¢rst year of life in the same stretch l-methionine, l-valine, l-isoleucine and l-serine of a small stream, and were used for bioassay after in chum salmon versus brown trout; l-arginine, being kept for10 months together in a general aqua- l-asparagine, l-histidine, l-methionine, l-proline, rium and being fed with the same feeds. The palat- l-leucine and l-isoleucine in black molly versus ability of numerous substances was di¡erent in platy; l-alanine in stellate sturgeon versus Siberian common carp, crucian carp, wild gold¢sh and tench, sturgeon (genus Acipenser)andl-cysteine, l-lysine ¢shthat inhabit the same waterand have similardiets and l-glutamic acid in black molly versus guppy (Kasumyan and Morsy 1996;Kasumyan and Proko- (genus Poecilia). pova 2001;Kasumyan, unpublishedobservation). The number of amino acids that evoke the same The di¡erences in taste preferences are con¢rmed taste response, is much smaller in the species men- by the correlation analysis of amino acid palatability tioned above. Only l-methionine and l-phenylala- for 21¢sh species tested (Table 11). A positive correla- nine evoke the same taste responses in both tion was found for10 (4.8%) out of 210 possible pairs, common carp and wild gold¢sh; l-histidine, l-pheny- a negative correlation was found for six pairs only lalanine and l-tyrosine in chum salmon and brown (2.9%). All these pairs had a low level of relationships. trout and l-alanine, l-cysteine, l-phenylalanine, l- Eight ¢sh species did not show any signi¢cant corre- serine, l-threonine and l-tyrosine in black molly lation with any other species, and13¢sh species had and platy.There is no single amino acid that evokes a signi¢cant relationship with one to four other spe- the same oral taste response in stellate sturgeon, cies. In most cases, the positive relationship was Siberian sturgeon and Russian sturgeon, all of which between species that are distant in systematic posi- belong to the same genus Acipenser (Table 7). tion, ecology and feeding type. These pairs were: The responses to classical taste substances di¡er common carp^guppy, guppy^grass carp, black among the ¢sh species tested (Table 6). Citric acid molly^navaga, navaga^chum salmon and chum sal- was a stimulant for the salmonids browntrout, brook mon^platy. A positive correlation was not found in char, lake char and frolich char, but a deterrent related species such as guppy, black molly and platy, for chum salmon. Among cyprinids, citric acid is a all belonging to the same Poecilidae family and the

314 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

same genus as cruciancarp and wild gold¢sh, orRus- cium chloride, sucrose and many of the free amino sian sturgeon, Siberian sturgeon and stellate stur- acids were indi¡erent taste substances (Kasumyan geon (genus Acipenser). Moreover, correlation and Sidorov 1995a,c). A positive correlation was between taste responses to free amino acids was found between ¢sh in the taste preferences for 25 negative in the pair guppy^platy. As seen from substances (21free amino acids and 4 classical taste Table 11, the number of positive correlations in tasts substances;Table 12). preferences between closely related ¢sh species is Caspian brown trout belong to the`Danube' phylo- verylimited and seems to be an exception rather than genetic group of Salmo trutta populations, and Baltic the rule. For instance, a positive correlation was Sea brown trout and White Sea brown trout belong foundinonly4casesamongthe10cyprinid¢shspe- to the `Atlantic' phylogenetic group. These groups of cies tested, and in one case between three salmonid brown trout populations were separated at least ¢sh species. One positive correlation was found bet- 500 000 years ago, according to a genetic study (Osi- ween acipenserids and between poeciliids. nov and Bernache 1996;Bernache 2001). The most Thus, taste preferences are characterised bystrong recent connection between the Caspian Sea basin species speci¢city. These data clearly demonstrate and both the Baltic Sea and White Sea basins has an important and undoubtedly leading role of gusta- occurred during the last glacial period (12 000^ tory reception in the feeding selectivity in ¢sh, and 15000 years ago). There have been many thousands in their capacity to consume appropriate food items of generations of brown trout since this time, but that are speci¢c to them. still, taste preferences remain similar. The results Considerable species di¡erences in the range of indicate that taste preferences do not have impres- amino acids that are stimulatory have been revealed sive interpopulation speci¢city in ¢sh. It has been by electrophysiological studies (for review,see Marui shown for humans that persons of di¡erent national- and Caprio1992). ities have similar taste preferences for chemical substances. Population speci¢city Speci¢city among conspeci¢cs Brown trout is a polytypic ¢sh species, which has developed di¡erent geographically isolated subspe- Given the similarity between di¡erent ¢sh popula- cies overavast area. Fish from three widely separated tion of the same species, it is noteworthy that there is catchments were used to compare taste preferences variability in taste preferences between individuals in specimens belonging to di¡erent populations. of a population. Individual variation in diet has been Brown trout juveniles (5^6 cm in body length) were found in ¢sh that inhabit the same natural water obtained from Terek River (Caspian Sea basin), Luga body, but also in ¢sh maintained in experimental River (Baltic Sea basin, the Gulf of Finland) andVoro- conditions (Bryan and Larkin 1972;Ringler 1985; bejv stream (White Seabasin, Kandalaksha Bay).Fish Bridcut and Giller 1995). It was noted that taste pre- were delivered to the Moscow State University and ferences at the individual level might vary dramati- were used for trials after being kept for 1^2 months cally among conspeci¢cs. In rainbow trout, the in a common aquarium. The trials were performed number of specimens with distinctive taste prefer- in di¡erent years with ¢sh delivered from di¡erent ences is about1per 25 ¢sh (Jones1989) or 3 per11¢sh water basins. Four classical taste substances and 21 (Mearns et al. 1987). The particular specimens used common free amino acids were used as taste stimuli by these authors either consistently rejected both (Kasumyan and Sidorov1995a). the blank and the shrimp-£avoured pellets or, con- It was found that taste preferences were similar in versely,swallowed both of them. brown trout juveniles from all three populations A study of individual variability of taste prefer- (Fig. 5). Citric acid and l-cysteine were the most pala- ences was performed on the common carp (Kasum- table substances for Caspian Sea, Baltic Sea and yan 2000). Oral taste responses were assessed for White Sea brown trout. Fish in all trials swallowed twenty-four 1-year-old juveniles. Amino acids l- agar pellets containing these substances after the cysteine (0.1 m)andl-glutamic acid (0.01 m)were ¢rst snap. The retention time correlated positively used as taste stimuli, and blank pellets were used as with the acceptance ratio of pellets. l-Isoleucine, l- the control. It was found that the average acceptance glutamine, l-valine and glycine were the strongest rat io wa s 97.3 % fo r l-cysteine and 41.8% for l-gluta- deterrent taste stimuli, while sodium chloride, cal- mic acid. The acceptance ratio for blank pellets was

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 315 at preferences Taste 316 Table 11 Correlation of taste preferences between species lxne auyn&KelBDÖving B Kjell & Kasumyan O Alexander # 03BakelPbihn t,FISHadFISHERIES, E I R E H S I F and H S I F Ltd, Publishing Blackwell 2003 4 ,289^347

The values of Spearman rank correlation coef®cient for taste preferences to 21 free amino acids (19 for Siberian sturgeon) between different ®sh species. Numbers in frames represent correlation coef®cients for species of the same family. Salmonidae 1±3, Poecilidae 4±6, Acipenseridae 7±9, Cyprinidae 10±19. Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 5 Taste preferences to free amino acids and classical taste substances in brown trout.Taste preferences in juveniles of brown trout (Salmo trutta) from di¡erent sea basins (modi¢ed from Kasumyan and Sidorov1995).1 ^ l-Cysteine,0.1 m; 2^l-tyrosine,0.001 m;3^l-tryptophan,0.01 m;4^l-aspartic acid,0.01 m;5^l-histidine,0.1 m;6^l-phenylalanine, 0.1 m;7^l-arginine,0.1 m;8^l-glutamic acid,0.01 m;9^l-lysine,0.1 m;10 ^ l-proline,0.1 m;11 ^ l-alanine,0.1 m;12 ^ l-leucine, 0.1 m;13 ^ l-threonine,0.1 m;14 ^ l-asparagine,0.1 m;15 ^ l-serine,0.1 m;16 ^ glycine,0.1 m;17 ^ l-Norvaline,0.1 m; 18 ^ l-methionine,0.1 m;19 ^ l-valine,0.1 m;20 ^ l-isoleucine,0.01 m;21 ^ l-glutamine,0.1 m;23 ^ citric acid, 0.26 m; 24 ^ calcium chloride,1.73 m;25^ sodium chloride,0.9 m;26 ^ sucrose,0.29 m;22 and 27^ respective controls (blank pellets; Kasumyan and Sidorov, unpublished).

17.0%.The value of the coe¤cient of variation for the Statistical analysis showed that three specimens acceptance ratio was low for highly palatable l- had an acceptance ratio for l-glutamic acid that was cysteine (CV ˆ 5.14%), while it was much higher for more than 1 SD from the mean value for the 24 ¢sh less palatable l-glutamic acid (56.4%), and especially tested. Nine specimens showed a similar deviation for blank pellets (89.0%). All specimens tested from the mean in their reaction to blank pellets. showed a positive taste response to l-cysteine, with These results show an interesting aspect of taste seven specimens showing positive taste responses to preferences. Taste responses are more stable and l-glutamic acid. l-glutamic acid was found to be a invariable for highly palatable substances than for deterrent for one specimen among the 24 carps substances with low level of palatability. This result tested.The ¢sh rejected £avouredpellets signi¢cantly might explain the food specialisation between con- more often than the blank pellets. Many of the ¢sh speci¢cs, which is often found among wild ¢sh popu- preferred pellets £avoured by l-cysteine and showed lations (Bryan and Larkin 1972;Ringler 1985; an indi¡erent response for pellets containing l-gluta- Bridcut and Giller1995) and the disproportional arti- mic acid. There were several specimens with a rela- ¢cial food intake in cultured stocks (Storebakken tively low taste preference for l-cysteine and a very and Austreng 1988). Individual variability in taste high taste preference for l-glutamic acid. preferences has been found in other groups of ani- mals (Stevens 1990) and also in humans (Prescott and Stevenson1995). The source of individual di¡er- Table 12 Comparing populations ences is still poorly understood, and it has been sug- gested that these individual patterns of taste Baltic Sea White Sea preferences are genetically determined (Kasumyan Origin of 2000).The individual variabilitycan be used in aqua- population FAA FAA ‡ CTS FAA FAA ‡ CTS culture to promote strains that have particular taste preferences. Caspian Sea 0.49à 0.55Ãà 0.68Ãà 0.76ÃÃà Baltic Sea ± ± 0.47à 0.60Ãà Speci¢city between sexes The values of Spearman's rank correlation coef®cient for taste A study of sex variability in taste preferences of ¢sh preferences to free amino acids (FAA) and classical taste sub- was performed on adult guppy (Nikolaeva and stances (CTS) between three populations of brown trout (Salmo trutta) juveniles from Caspian Sea, Baltic Sea and White Sea Kasumyan 2000). Citric acid, sucrose, calcium chlor- basins. See text for details. ide, glycine, l-glutamic acid and l-histidine were

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 317 Taste preferences Alexander O Kasumyan & Kjell B DÖving

used as taste substances. It was found that the males Genetics of taste preferences and females of guppy have similar taste preferences for the substances tested. Sucrose, glycine and l-glu- It seems obvious that taste preferences in ¢sh as in tamicacid were stimulants forboth maleandfemale. other vertebrates are genetically determined. How- Citric acid, sodium chloride, calcium chloride and l- ever,onlyone studyhas beenperformed withthe pur- histidine had an indi¡erent taste for ¢sh.The value of pose of investigating the genetic nature of the food the Spearman's rank correlation coe¤cient was 0.97 preferences in ¢sh (Kasumyan and Nikolaeva, (P < 0.01) for acceptance ratio between specimens of unpublished study). This study was performed by di¡erent sexes. The acceptance ratio for every sub- comparing taste preferences in wild female gold¢sh, stance was higher in females than in males (Fig. 6). male common carp and hybrids between these two The results indicate that there are no di¡erences cyprinid species. There were two main reasons to between male and female guppy in taste preferences. use these species ^ the palatability for some sub- The revealed coincidence of taste preferences agrees stances is di¡erent in these two species, and gold¢sh well with the data on the feeding of guppy in natural and common carp readily yield hybrids. All speci- water bodies. It was shown that females and males mens used for these experiments were reared on chir- use the same groups of food objects, which include onomid larvae and were maintained in water with greenalgae and diatoms, benthos, detritus andlarvae similar temperature, oxygen supply and light cycle. of insects, as food (Dussault and Kramer1981). It was found that the taste preferences for the clas- Sex dimorphism with respect to taste preferences sical taste substances were similar in carp and is not a characteristic attribute in ¢sh;however, the hybrids. However, the rank according to the prefer- guppy does show sex di¡erences in its behavioural ence of these substances was signi¢cantly di¡erent taste response pattern. The intensity of some of the between hybrids and wild gold¢sh (Fig. 7). Responses elements in the taste response like retention time of of the hybrids were particular in that none of the pellets in the mouth and number of snaps at pellets amino acids had any e¡ect on pellet consumption, was lower in males than in females. Intotal, the dura- indicating that none of the amino acids were stimu- tion of the taste response was shorter among males lants. This is a striking observation because six and than among females. Observations carried out on eight amino acids are stimulants for carp and wild guppy show that males spent their time courting gold¢sh, respectively.Therewas apositive correlation rather than feeding (Farr and Herrnkind 1974;Dus- between the palatability of all substances tested, sault and Kramer 1981), which indicates that the which include four classical taste substances and 21 male behaviour strategy is orientated towards repro- free amino acids for carp and hybrid (rs ˆ 0.53; duction, with nonreproductive behaviour being P < 0.01). However, no signi¢cant relationship exist- minimised. ed between wild gold¢sh and hybrids (rs ˆÀ0.01;

Figure 6 Taste preferences in males and females.Taste preferences in male and female of guppy (Poecilia reticulata)to classical taste substances and for some of the free amino acids (modi¢ed from Nikolaeva and Kasumyan 2000).1 ^ Sodium chloride,0.9 m;2^ calcium chloride,1.73 m;3 ^ citric acid,0.26 m;4^l-histidine,0.1 m;5 ^ sucrose,0.29 m;6^l-glutamic acid, 0.01 m;7^glycine,0.1m;8^Chironomidae larvae water extract,75 g(wet weight) LÀ1. Control ^ blank pellets.Vertical columns and T-bars are the mean and the SEM (Nikolaeva and Kasumyan 2000).

318 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 7 Genetics of taste preferences. Palatability index for classical taste substances and l-proline in common carp (Cyprinus carpio), wild gold¢sh (Carassius auratus) and hybrids between these two species.1 ^ Citric acid,0.26 m;2^sodium chloride,0.9 m;3 ^ calcium chloride,1.73 m;4 ^ sucrose,0.29 m;5^l-proline,0.1 m (Kasumyan and Nikolaeva, unpublished).The assessment (Chi-square test) of the palatability of taste substances tested was made in comparison for a blank agar (2%) pellet containing red colour dye, Ponceau 4R (0.05 mm).

P > 0.05).This study shows that the taste preferences tem and the ability to discriminate taste properties are determined genetically and seem to be patrocli- of food items. Experiments performed on larvae and nous. juveniles of Siberian sturgeon and stellate sturgeon A recent studyon di¡erent strains of rainbow trout have shown that, up to the time when it becomes investigated the gustatory responses from the pala- necessary to estimate the gustatory qualities of food, tine nerve to10 substances, mainly amino acids. The ¢sh have a well-expressed ability to discriminate results indicated that the strains could be dividedinto taste stimuli. Sturgeon larvae of `start feed'age show two response groups: one group including the Kam- the same oral and extraoral taste behavioural loops, Manx, Mount Lassen, SunndalsÖra,Tagwerker responses to the classical taste substances repre- andVictoria strains, and the second group including sented by sucrose, sodium chloride and calcium the Nisqually, Nikko and steelhead strains. It was chloride as do the older juveniles. Citric acid was an concluded that the functionalvariabilityof the gusta- indi¡erent substance for sturgeon larvae, but evoked tory system may be of genetic origin (Hara et al.1999). conspicuous oral and extraoral taste responses in sturgeon juveniles. Citric acid was a strong incitant for sturgeon juveniles and induced a high capture Ontogeny of taste preferences rate of pellets. Nearly all pellets £avoured with citric The development of taste preferences is another acid and touched by a ¢sh barbel were captured, but aspect of ¢sh chemoreception that has been scantily they were immediately rejected (Kasumyan 1992; explored. Most studies concerning ontogeny of the Kasumyan and Kazhlaev1993a). gustatory system in ¢sh have been focused on the The conclusion that larvae do not respond to morphology of the taste buds. Several studies have amino acids, which are taste stimuli for older juve- determined the stage at which taste buds appear in niles, has been con¢rmed by the comparisonof extra- the mouth or on the body surface (Appelbaum et al. oral and oral taste responses of Russian sturgeon 1983;Pevzner 1985;Kawamura and Washiyama (Kasumyan et al. 1992). A smaller number of sub- 1989;Devitsina and Kazhlaev 1993;Hansen et al. stances acted as incitants and stimulants in larvae 2002). According to these studies, the taste buds in (21^25 mm body length) than in juveniles (60^ ¢sh appear 1^2 days or even only hours before ¢sh 70 mm body length). For larvae,11and1amino acids larvae start to catch food items, which is at the begin- were incitants and a stimulant, respectively, and for ning of exogenous feeding. juveniles, 16 and 6 amino acids were incitants and It has been found that there is good correspon- stimulants, respectively (Table 13). There was a posi- dence between the development of the gustatory sys- tive correlation between extraoral and oral ranges of

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 319 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 13 Comparisons of extraoral and oral taste preferences

Extraoral taste preferences Oral taste preferences

Amino acid (L-isomers) Concentration (M) Larvae Juveniles Larvae Juveniles

Cysteine 0.1 20ÃÃÃ 35ÃÃÃ 780ÃÃÃ Histidine 0.1 20ÃÃÃ 31ÃÃÃ 928 Glutamine 0.1 17ÃÃÃ 25ÃÃÃ 736 Asparagine 0.1 14ÃÃÃ 28ÃÃÃ 435 Lysine 0.1 13ÃÃÃ 29ÃÃÃ 239 Glycine 0.1 13ÃÃÃ 14 10 34 Methionine 0.1 11ÃÃÃ 21ÃÃ 19Ã 25 Serine 0.1 10ÃÃÃ 24ÃÃÃ 742 Threonine 0.1 9ÃÃ 30ÃÃÃ 10 43Ã Phenylalanine 0.1 6 23ÃÃÃ 441 Alanine 0.1 6 22ÃÃÃ 716 Norvaline 0.1 6 7 5 15 Arginine 0.1 6 29ÃÃÃ 642Ã Proline 0.1 6 7 5 19 Valine 0.1 6 14 2 34 Glutamic acid 0.01 14ÃÃÃ 36ÃÃÃ 650Ã Aspartic acid 0.01 11ÃÃÃ 38ÃÃÃ 769ÃÃÃ Isoleucine 0.01 6 15Ã 520 Tryptophan 0.01 6 23ÃÃÃ 353ÃÃ Leucine 0.01 4 16Ã 426 Tyrosine 0.001 4 14 6 21 Control ± 5 10 7 23

Results from experiments on larvae (21±25 mm in body length) and juveniles (60±70 mm in body length) of Russian sturgeon, Acipenser gueldenstaedtii (modi®ed from Kasumyan et al. 1992). Extraoral taste preferences are given as the mean number of snaps of agar pellets containing an amino acid. Oral taste preference is the mean percentage of pellets that were consumed in relation to number of snaps. 50 pellets were offered for 20 larvae or 10 juveniles. After 5 min remained pellets were counted and the percentage of consumed pellets per number of snaps was calculated. Pellets containing only dye (Cr2O3, 0.3%) were used as control, see text.

amino acids in sturgeon juveniles (rs ˆ 0.79; Nevertheless, taste perception in ¢sh larvae is less P < 0.001), while in sturgeon larvae, the extraoral well developed than in adult ¢sh. The limited taste and oral range of amino acids was not correlated perception may be one reason for mass mortality of

(rs ˆ 0.36; P > 0.05). The extraoral range of free ¢sh larvae and early postlarvae among capelin and amino acids was similar in larvae and juveniles Atlantic herring, which is caused by a dino£agellate

(rs ˆ 0.67; P < 0.01;Fig. 8). Morphological investiga- that is toxic to the ¢sh (Gosselin et al. 1989). The toxin tions show that the taste buds appeared at an earlier (saxitoxin) content of dino£agellate reaches more stage of sturgeon development in the barbels and than 10À4 mg cellÀ1 (Cembella et al. 1988). That lips than in the oral cavity (Devitsina and Kazhlaev amount of saxitoxin is enough to kill a ¢rst-feeding 1995). ¢sh larva. The older ¢sh are highly sensitive towards The results of behavioural assay have shown that the toxins, which enable them to avoid poisonous ¢sh larvae at the `start feed'stage have the ability to food (Yamamori et al. 1988). discriminate among taste substances and release an adequate behavioural taste response.The early devel- Ecology of taste preferences opment of taste perception has been shown in many mammals (Formaker and Hill 1990) and in humans In spite of the intensive investigations of taste percep- (Beauchamp et al. 1986, 1994). It has been suggested tion in di¡erent groups of animals, mainly mammals, that the development of gustatory sensitivity to var- there is no clear concept explaining an animal'staste ious substances may be conditioned by speci¢c fea- preference to some substances and the rejection of tures of feeding and metabolism (Mistretta 1991). others. It has been suggested (Kassil 1972) that the

320 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 8 Oral and extraoral taste preferences.Values of Spearman's rank correlation coe¤cient between oral and extraoral taste preferences for free amino acids in larvae and juveniles of Russian sturgeon, Acipensergueldenstaedtii (modi¢ed from Kasumyan1999a).

palatability of substances (for mammals) might be of Romaine lettuce that contained sugars did not related to the role these substances play in the meta- have any e¡ects on the consumption of £avoured bolic processes. Nevertheless, such relationships agardisks byherbivorous tilapia (JohnsenandAdams seem to be very indirect and unpredictable. 1986). Moreover, some of the omnivorous and carni- Another explanation for taste preferences is vorous ¢sh have positive taste response for sucrose based on the animal feeding ecology. According to and other sugars. Such a preference was shown for general rules formulated with reference to numer- shore rockling (Andriashev 1944), rainbow trout ous observations of the diets of various animals, (Jones 1990), European eel, pike, zander, vendace herbivores manifested a preference for sugars more (Appelbaum 1980) and brook char (Kasumyan and often than did nonherbivores (Harborne 1993). This Sidorov 1998). However, for many omnivorous and rule seems to be applicable to ¢sh also. It was carnivorous ¢sh, sucrose was an indi¡erent taste shown that sucrose is highly palatable for grass carp, substance. None of the ¢sh species tested had been a herbivore ¢sh, and for chub and roach (Kasumyan found to perceive sucrose as a deterrent (Table 4). and Morsy 1997;Kasumyan and Nikolaeva 2002). Based on a comparison of taste preferences in her- Macrophytes, which are the main food for grass carp bivorous and carnivorous ¢sh, it was suggested that (Fischer 1973;Fowler et al. 1978) and ¢lamentous these two groups of ¢sh have speci¢c amino acids algae are the preferable food for chub and roach that stimulate feeding (Johnsen and Adams 1986). during summer time (Boikova 1986;Specziar et al. The authors claimed that amino acids found to be sti- 1997;Hor ppi la et al. 2000). Sucrose is a stimulant mulants for carnivores failed to elicit feeding for guppy (Kasumyan and Nikolaeva 1997), platy responses inthe herbivores. However,this suggestion and black molly (an omnivorous poecilid ¢sh that does not seem to be correct as grass carp, a herbivore feeds on various aquatic invertebrates and plants; ¢sh, shows deterrent responses for substances such Dussault and Kramer 1981;Bailey and Sandford as l-glutamic acid, l-lysine, l-alanine, l-serine, 1999).When cultivating guppy, it is recommended to which were highly palatable for tilapia (Johnsen and include carbohydrate components like algae, leaves Adams 1986). L-Cysteine and glycine, the indi¡erent of lettuce or spinach and vegetative £akes into its diet taste substances for tilapia, were stimulants for grass to secure growth and reproduction (Pausan 1984; carp (Kasumyan and Morsy 1997;Kasumyan and Stuart1996). Morsy1998b). Many carnivorous ¢sh show taste pre- Not all herbivorous ¢sh have a preference for ferences for the same amino acids as do the herbivor- sucrose and other sugars. For example, the fraction ous tilapia or grass carp (Table 7).

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 321 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Many ¢sh such as chub, brown trout and lake char It was found that after 6 months, the grass carp are insectivores during late springand summertime. juveniles of the `herbivorous' and the `carnivorous' For these ¢sh, calcium chloride is a stimulant, but is groups had similar taste preferences for the four clas- a well-known bitter compound for humans (Kasum- sical taste substances and the 21amino acids tested yan and Sidorov 1993a;Kasumyan 1997;Table 4). (Fig. 9). Pellets containing sucrose, cysteine, citric The taste preference for calcium chloride may be acid, aspartic acid or glycine were highly palatable related to the important role that £ying insects have substances. Seventeen of the 21 free amino acids as a food for these species. However, there are ¢sh tested were deterrent substances for grass carp sib- such as chum salmon and frolich char that also take lings of both the groups. A highly signi¢cant positive £ying insects as food and at the same time, do not correlation was found for taste preferences between show a taste preference for calcium chloride (Kasum- the herbivorous and carnivorous groups of ¢sh to ya n a nd Sidorov 1993 b, 1995 d;Ka su mya n 1997). the taste substances tested (rs ˆ 0.74; P < 0.001). Furthermore, the common carp and tench prefer the However, the taste preferences of the ¢sh were not taste of calcium chloride, but do not feed on £ying identical, and di¡erences in taste preferences insects. Insects contain substances considered bitter between the two groups were found. Pellets contain- by humans, but are readily eaten by insectivorous ing calcium chloride or sodium chloride were con- birds (Harborne 1993). The natural stimulants and sumed more readily in ¢sh of the`herbivorous'group deterrents in insects are not known. Both for the than in ¢sh of the`carnivorous'group. The di¡erence insectivorous birds and ¢sh, the chemical nature of between the acceptance ratios of control and £a- stimulants or deterrents is unknown. voured pellets was statistically signi¢cant only for the `herbivorous'group. The di¡erences in taste pre- ferences of calcium chloride and sodium chloride Internalfactors effecting taste preferences between`carnivorous'and `herbivorous' ¢sh were not signi¢cant. In each group of grass carp, pellets con- Feeding experience taining the water extract of conditioning diet were It is well known that ¢sh release strong and obvious preferred to those containing the extract of uncondi- behavioural responses to feeding signals emanating tioning diet (both 75 g LÀ1). Pellets with water from common food objects. This has been shown for extracted from mosquito larvae were consumed food stimuli mediated by olfaction and vision 1.27  more readily than pellets £avoured with (McBride et al. 1962;Atema et al. 1980;Colgan et al. water extracted from Romaine lettuce leaves in `car- 1986;Kasumyan and Ponomarev 1986;Stradmeyer nivorous' ¢sh and 1.22 less readily in `herbivorous' and Thorpe 1987;Meyer 1988). However, when it ¢sh. The consumption of pellets £avoured with mos- comes to taste preferences, the dependence upon quito larvae or plant extracts di¡ered signi¢cantly feeding experience has been shown to be less promi- between these two groups (P < 0.001 and P < 0.01, nent (Kasumyan and Morsy 1997, 1998b). In these respectively). studies, grass carp siblings were used. Grass carp lar- These observations suggest that conditioning with vae at the `start feed'stage were obtained from a ¢sh a long-term rearing on a selected food leads to only a farm (Krasnodar district) and were raised on plank- small shift in taste preferences and only a slight tonic Cladocera and the arti¢cial feed `TetraMin' increase in palatability for the substances in the food (Tetra Co., Germany) up to 3 months of age at which on which they were reared. In conclusion, the results the ¢sh had attained a body length of 2.5 cm. Then, from these experiments provide evidence that taste the ¢sh were divided into two groups, each group preferences show low plasticity in ¢sh. In other being fed a distinct diet for 6 months. A`carnivorous' words, taste preference does not depend upon the ¢sh group of grass carp siblings was reared on live mos- diet. The ¢sh diet might be re£ected in taste prefer- quito larvae, while a `vegetarian' group was reared ences of ¢sh to food extracts, which have a more com- on duckweed and Romaine lettuce leaves. Both types plicated composition. The results show that taste of conditioning diets were adequate for grass carp preference in ¢sh is under strict genetic control. juveniles because in nature, zoophagy is combined Based on these results and the previously discussed with macrophytophagy (Fischer 1973;Fowler et al. data, it seems reasonable to conclude that taste pre- 1978). At the end of the rearing period, the 9-month- ference in ¢sh isacharacteristicfeatureof the species old grass carps in both the groups were 6^9 cm in and is independent of population, sex and previous body length. feeding experience.

322 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 9 E¡ect of feeding experience ontaste preferences.Taste preferences to free aminoacids, classical taste substances and food extracts ingrass carp (Ctenopharyngodon idella) juveniles reared ondi¡erent diets.1^ l-Alanine,0.1 m;2^l-arginine, 0.1 m;3^l-asparagine,0.1 m;4^l-aspartic acid, 0.01;5 ^ l-valine,0.1 m;6^l-histidine,0.1 m;7^glycine,0.1 m;8^l-glutamine, 0.1 m;9^l-glutamic acid,0.01 m;10 ^ l-isoleucine,0.01 m;11 ^ l-leucine,0.1 m;12 ^ l-lysine,0.01 m;13 ^ l-methionine,0.1 m; 14 ^ l-norvaline,0.1 m;15 ^ l-proline,0.1 m;16 ^ l-serine,0.1 m;17 ^ l-tyrosine,0.001 m;18 ^ l-threonine,0.1 m; 19 ^ l-tryptophan,0.01 m;20 ^ l-phenylalanine,0.1 m;21 ^ l-cysteine,0.1 m;22 ^ sucrose,0.29 m;23 ^ citric acid, 0.26 m; 24 ^ calcium chloride,1.73 m;25^ sodium chloride,0.9 m;28 ^ water extract of Romaine lettuce (Lactuca sativa)leaves, 75 g (wet weight) LÀ1;29 ^ Chironomidae larvae water extract,75 g (wet weight) LÀ1;26 and 27^ controls (blank pellets) for `free amino acid'series and for `classical taste substances and food extracts'series, respectively (modi¢ed from Kasumyan and Mors y 1997 b).

In the `warm'series, 19 of 21free amino acids tested Feeding motivation were incitants for the extraoral gustatory system of The ¢sh feeding motivation strongly in£uences taste stellate sturgeon, and caused signi¢cant increase in preferences. It was found that starvation during 17^ the number of snaps at agar pellets (Table 15). The 18 hours caused a widening of the extraoral taste three amino acids, l-alanine, l-histidine and l- spectrum in 4^6 cm long Siberiansturgeonjuveniles cysteine, were stimulants and caused a considerable in comparison with ¢sh which had food ad libitum increase inthe numberof pellets consumed inrelation 1.0^1.5 hours before behavioural trials. The number to the number of pellets snapped. A signi¢cant corre- of substances e¡ective for oral taste reception chan- lation was found between extraoral and oral taste ged relatively little (Table 14). The water temperature responses (r ˆ 0.49; P < 0.05). In the`cold'series,17of was18 8C during the experiments (Kasumyan1997). 21 free amino acids were incitants for ¢sh, and two amino acids, l-histidine and l-valine, were stimu- lants. No signi¢cant correlation was found between Environmentalfactors affecting taste extraoral and oral taste responses of ¢sh in the `cold' preferences series (r ˆ 0.20; P > 0.05). No signi¢cant relationship between oral taste responses of ¢sh in the `warm' Temperature and the `cold' series was found (r ˆ 0.20; P > 0.05), The ambient water temperature is one of the most although extraoral taste responses obtained at potent abiotic variables a¡ecting vital functions in di¡erent water temperatures were highly correlated exothermic animals like ¢sh. Temperature also has (r ˆ 0.92; P < 0.001). The mean number of pellets an e¡ect on taste preferences in ¢sh. The grass carp snapped and swallowed was signi¢cantly higher in shows di¡erent levels of relative food preference for the`warm'series.These results con¢rmthoseobtained the same objects in warm (20 8C) and cold (13 8C) by other authors concerning the e¡ect of water water (Adamek et al. 1990). A study on the e¡ect of temperature on ¢sh feeding activity (Knights 1985; water temperature on ¢sh taste responses was per- Metcalfe et al. 1986;Stradmeyer and Thorpe 1987). formed on juvenile stellate sturgeon 4^5 months in The di¡erence in water temperature between the age and 6^8 cm in body length (Kasumyan et al. `warm'and `cold' series (7.5 8C) is enough to change 1992). Two groups of ¢sh were held separately in dif- the behavioural responses to chemical signals in ferent water temperatures for 5 days before starting ectothermic animals (Yamashita 1964;Flerova and the bioassays. In the`warm'series, the water tempera- Gdovskii 1976;Jones 1980;Jones 1983;Van Damme ture was 20 8C and in the `cold'series, it was 12.5 8C. et al. 1990). The extraoral taste responses in the

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 323 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 14 The e¡ect of starving on extraoral and oral taste preferences

Extraoral taste preferences, Oral taste preferences, food deprivation for food deprivation for

Amino acid (L-isomers) Concentration (M) 1±1.5 hours 17±18 hours 1±1.5 hours 17±18 hours

Threonine 0.1 23Ã 42ÃÃÃ 60 87ÃÃÃ Asparagine 0.1 21 47ÃÃÃ 69ÃÃ 62Ã Glycine 0.1 16 36ÃÃÃ 63Ã 71ÃÃÃ Methionine 0.1 14 41ÃÃÃ 49 56 Glutamic acid 0.01 9 32ÃÃÃ 46 70ÃÃ Control ± 10 15 52 49

Juveniles of Siberian sturgeon, Acipenser baerii (4±6 cm in body length) were tested after being starved for one to one and a half hour or 17±18 hours. Extraoral taste preference is the mean number of snaps of agar pellets containing free amino acids. Oral taste preference is the mean percentage of pellets that were consumed in relation to number of snaps. 50 pellets were offered to 5 juveniles. After 5 min remained pellets were counted and the percentage of consumed pellets per number of snaps was calculated. Pellets containing Cr2O3 were used as Controls. Note the increase in snaps and consumption with duration of starvation time.

Table 15 Palatability and water temperature

Extraoral taste preferences Oral taste preferences

Amino acid (L-isomers) Concentration (M) 12.5 8C208C12.58C208C

Cysteine 0.1 32 35 15 60ÃÃÃ Lysine 0.1 29 28 11 30Ã Serine 0.1 25 28 8 24Ã Threonine 0.1 22 22 5 17 Histidine 0.1 21 27Ã 30 57Ã Asparagine 0.1 20 21 15 21 Arginine 0.1 19 25 9 11 Methionine 0.1 19 20 3 15 Glutamine 0.1 16 23ÃÃ 712 Phenylalanine 0.1 15 18 11 16 Alanine 0.1 14 24ÃÃ 960ÃÃÃ Valine 0.1 10 12Ã 27 12 Glycine 0.1 10 19ÃÃ 14 33ÃÃ Norvaline 0.1 8 11Ã 915 Proline 0.1 3 10ÃÃÃ 515 Aspartic acid 0.01 33 36 11 24 Glutamic acid 0.01 18 28ÃÃ 18 7Ã Isoleucine 0.01 10 14 15 16 Leucine 0.01 10 12 17 15 Tryptophan 0.01 5 13ÃÃÃ 18 15 Tyrosine 0.001 3 9ÃÃÃ 014 Control ± 6 7 6 16

Effect of water temperature on extraoral and oral taste preferences for free amino acids in stellate sturgeon, Acipenser stellatus. Extra- oral taste preference is the mean number of snaps of pellets containing free amino acids. Oral taste preference is the mean percentage of pellets that were consumed in relation to number of snaps. Fifty pellets were offered for 5 ®sh. After 5 min remained pellets were counted and the percentage of consumed pellets per number of snaps was calculated. Pellets controls contained 0.5 mM of the red colour Ponceau 4R (modi®ed from Kasumyan et al. 1993).

324 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

stellate sturgeon did not undergo any appreciable after 3,6 or 24 hours of exposure.The di¡erent heavy modi¢cationat di¡erent temperatures, while the oral metals had di¡erent suppression intensities on taste taste preferences did. Of particular interest is the fact responses at 1 mm. The strongest e¡ect was induced that the rank order of stimulatory e¤ciency di¡ered by HgCl2;CuSO 4 was somewhat less e¡ective than betweentheaminoacidsunderdi¡erenttemperature HgCl2, and the weakest e¡ects were produced by conditions. Amino acids like histidine, asparagine, CdSO4,Pb(NO3)2 and ZnSO4 (Fig. 10). As the HgCl2 phenylalanine, isoleucine, norvaline and proline kept concentration decreased from1 mm to1 nm, the time their rank order at the two temperatures, whereas to observe a change intaste responses increased from glutamic acid, alanine, tryptophan, valine and leu- less than1 hour to 24 or 72 hours, depending on the cine changed their rank drastically (Table 15). palatability of the amino acids used (Fig. 11). Temperature dependence of taste responses to taste The loss of taste responses in ¢sh after the expo- substances of the same type is expressed di¡erently sure to heavy metals is reversible. Some time after in di¡erent animal species (Yamashita 1964;Naka- the cessation of exposure to toxic substances, the mura and Kurihara 1991). We suggest that tempera- taste responses of ¢sh improve to their former func- ture-dependent taste responses might be important tion. In common carp, the taste responses after an for ¢sh and facilitate their seasonal shifts in feeding exposure to 0.1 mm HgCl2 for 3 hours were recovered regime. after 6^12 days at a temperature of 20 8C (Fig. 12), and there appeared to be a direct relationship between the duration of the exposure to a toxic sub- Water pollutants stance and the duration of the recovery period. Fish taste receptors are exposed to the environment and predisposed to the detrimental e¡ects of water LowpH water pollutants. It has been shown that many pollutants, especially heavy metals, a¡ect ¢sh taste reception by European grayling juveniles,6^8 cm in length, were both destroying the taste buds and reducing the sen- used as subjects for studying the e¡ect of low pH in sitivity to the taste stimuli. These events occur after the water on taste preferences in ¢sh. Taste prefer- a short exposure of the ¢sh to polluted water (for ences were estimated for l-isomer amino acids, and review,see Brown et al.1982;Klaprat et al.1992). classical taste substances and the pH of the tank There have been few experiments on the e¡ect of water were lowered by the addition of sulphuric acid. heavy metals on the behavioural taste response in In natural waters, with a pH between 7.6 and 7.8, ¢sh (Kasumyan 1997;Kasumyan and Morsy 1998a). juvenile grayling demonstrated signi¢cant taste pre- The e¡ects of di¡erent heavy metals on taste beha- ferences to pellets containing all the classical taste vioural responses of 9^12-cm long common carp, substances used. In behaviour assays, it was shown after 1-, 3-, 6-, 24-, 48- and 72-hour exposures in that most of the 21 l-isomers of amino acids in the À1 À3 solutions of HgCl2,CuSO4,CdSO4,ZnSO4 and concentration range of 10 to 10 m were either

Pb(NO3)2 were investigated. The amino acids l- ine¡ective or were deterrent stimuli to the grayling. cysteine (0.1 m), l-glutamic acid (0.01 m) and water It was shown that only ¢ve amino acids increased extract of chironomid larvae (75 g LÀ1)wereusedas the consumption of pellets by grayling. When gray- taste stimuli. It was found that in clean water, the ling were exposed to water with a pH of 6.0, the unexposed ¢sh swallowed 99 and 52% of agar^agar pellet consumption decreased in the ¢rst 3^6 hours. pellets containing either l-cysteine or l-glutamic The pellet consumption increased slowly up to acid. The negative e¡ect of heavy metals was evident 24^48 hours, but decreased again after 72-hour after 15-min exposure. The exposure of ¢sh during exposure.When the ¢sh were exposed to a pH of 5.3, 1^3 hours in heavy metal salt solution (1 mm) pro- the pellet consumption decreased to about half of vides a strong and deep suppression on the taste the normal consumption after 3 hours, and to 20% responses of ¢sh. Under these conditions, the ¢sh after 72 hours. The ¢sh mortality was not evident in more often rejected the pellets containing amino waterat pH 6.0 and 5.3 and with a 72-hour exposure. acids with very high palatability than did individuals These results show that short-time exposure in water in a control situation. The exposed ¢sh accepted few with a low pH can induce drastic changes in the abil- pellets, and the retention time of pellets decreased, ity of ¢sh to respond to taste substances (Kasumyan althougth feeding motivation did not change, and and Sidorov 1995b). This ¢nding might explain the all pellets o¡ered were still actively grasped by ¢sh misbalance in biotic communities of acidi¢ed water.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 325 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 10 E¡ect of heavy metals on taste preferences.Taste preferences to l-cysteine (0.1 m)incommoncarp (Cyprinus carpio) exposed to solutions of ¢ve di¡erent heavy metals at1 mm (modi¢ed from Kasumyan and Morsy 1998 a).

Figure 11 E¡ect of mercuryon taste preferences.Taste preferences to l-cysteine (0.1 m)incommoncarp (Cyprinus carpio) exposed to mercury chloride solutions at ¢ve di¡erent concentrations (modi¢ed from Ka su mya n 1997).

pH 3.06, by adding 0.01 ml-aspartic acid, to Taste preferences as a function of pH pH 10.87, gained from the water containing 1.0 ml- No clear relationship was found between the pH of a arginine) were tested on rainbow trout. The most test solution and its palatability, when using amino e¡ective solutions lay predominantly in the range of acids as taste substances (Jones 1989). Eighty-one pH 5.8^6.6, but this predominance was probably solutions of various amino acids (ranged from weighted by virtue of the sheer number of solutions

Figure 12 Recovery of taste preferences. Recoveryof taste preferences to l-cysteine (0.1 m)in common carp (Cyprinus carpio)after exposure to mercury chloride solution (0.1 mm;modi¢ed from Ka su mya n 1997).

326 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

with pH in this range. Aqueous solutions of those sole controlling factor of food palatability for this ¢sh compounds which were palatable, as well as those (Adams et al. 1988). The experiments on juveniles of which were not, ranged in pH from quite acidic to European grayling revealed a weak but signi¢cant quite basic. For instance, the acidic glutamic acid, negative relationship between pH of solutions (0.1 m) the near-neutral proline and the basic arginine were and the palatability of 14 carboxylic acids clearly active. In contrast, neither the more acidic l- (rs ˆÀ0.69; P < 0.01;Ka su mya n 1997). The pa latabi l- aspartic acid, the near-neutral glycine and l-alanine, ity for19 organic acids, mainly carboxylic acids, was nor the basic l-histidine showed any activity. Based determined for three cyprinids: tench, bitterling and on these results, it was concluded that there is no dace. The Spearman's rank correlation coe¤cient direct relationship between pH and palatability, nor between the acceptance ratio of pellets and the pH of were there distinct pH thresholds above or below the solution of taste substances was highly positive which the amino acids became active. Instead, as a for bitterling (rs ˆ 0.91; P < 0.01), slightly positive for class, the amino acids displayed activity over a wide dace (rs ˆ 0.48; P < 0.05) and highly negative for rangeofpH.Thisisnottosaythattheactivityofspeci- tench (rs ˆÀ0.88; P < 0.01;Fig.13;Kasumyan and ¢c amino acids is always pH-independent. The possi- Prokopova 2001;Kasumyan and Fokina, unpub- bility remains that the pH dampens and/or lished). Thus, it seems obvious that the relationship enhances the e¡ects of particular amino acids by between palatability and pH is di¡erent in di¡erent in£uencing their binding kinetics with available gus- ¢sh species. tatory cell receptor sites, and thereby changes their perceived intensity and/or quality.Electrophysiologi- Chemicalnature of stimulantsin ®sh cal recordings do, however, indicate that at least the food organisms magnitude of the gustatory response to amino acids is relatively pH-independent. Except for arginine, the There are two approaches in studying the chemical response of the rainbow trout palatine to highly sti- nature of the taste stimulants. The ¢rst consists of mulatory amino acids is constant at pH above 7 screening of substances that are isolated and identi- (Marui et al. 1983a). In common carp, taste response ¢ed in various ¢sh food organisms byanalysing their magnitudes are constant above pH 6.0 (Marui chemical composition (Konosu et al.1965;Carr et al. et al. 1983b), and the Japanese eel palatine is equally 1977;Miyagawa et al. 1979;Mackie and Mitchell sensitive to glycine at pH 5 and 9 (Yoshii et al. 1979). 1983;I ida et al. 1992;for review, see Carr et al. 1996). At pH below 7, the gustatory response magnitudes This approach is the most common in studying ¢sh of both rainbow trout and common carp increase taste preferences and in the search for ¢sh taste sti- dramatically, but these changes were in parallel to mulants (Mackie 1982;Ikeda et al. 1988a,b;Takaoka the increased response to acidi¢ed natural water et al.1990). alone (Marui et al. 1983a,b). The latter response is Using this approach, it was shown that many probably because of an elevated CO2 tension. The amino acids are highly palatable for ¢sh and are the taste system of ¢sh is highly sensitive to CO2 (Konishi main components responsible for the palatability of et al.1969;Yoshiiet al.1980).Nodi¡erenceswere ¢sh food organisms. For example, a mixture contain- found in rainbow trout tested electrophysiologically ing all the amino acids found in the marine annelid by recording from the palatine nerve and stimulating had the same e¡ect on feeding intake as a complete with free amino acids at two di¡erent pH ranges, the synthetic extract for both Japanese eel and red sea local natural pH of 6.6^6.8 and 7.3^7.8, which were bream. The active constituents in the amino acids achieved by using a limestone pellet ¢lter (Hara et al. fractionwas identi¢edasglycine, l-alanine, l-proline 1999). and l-histidine for Japanese eel (Takeda et al.1984), To measure the e¡ect of pH on the feeding stimula- and glycine, l-alanine, l-lysine, l-valine, l-glutamic tion of ¢sh, a series of substances such as carboxylic acid, and l-arginine for red sea bream (Fuke et al. acids and amino acids (0.01 m) at the same concen- 1981). A m i x t u re o f l-amino acids from the synthetic tration but with di¡erent pH were tested on tilapia. squid extract was stimulatory for European eel and The relationship of pH to palatability was nonlinear. rainbow trout, while the corresponding d-amino Moreover, the results indicate an optimum pH, where acids were ine¡ective taste substances (Mackie and the palatability of food was the highest and which Mitchell1983) or deterrents (Adron and Mackie1978) decreased if the pH was shifted from this value. It for these ¢sh. For rainbow trout, the aromatic and was suggested that acidity might be one but not the basic aminoacids of a squid extract acted as a feeding

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 327 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Figure 13 Taste preference and pH. Relationship between pH and taste preferences for organic acids at a concentration of 0.1 m intench (Tinca tinca).1 ^ Oxalic acid;2 ^ maleic acid;3 ^ a-ketoglutaric acid;4 ^ tartaric acid;5 ^ malonic acid; 6 ^ citric acid;7^ malic acid;8 ^ glycolic acid;9 ^ fumaric acid;10^ ascorbic acid;11^ succinic acid;12^ adipic acid;13^ butyric acid;14 ^ valeric acid;15^ acetic acid;16 ^ propionic acid;17^ formic acid (from Kasumyan and Prokopova 2001). stimulant, while fractions containing neutral and The assessment of palatability for agar gels £a- acidic amino acids were inactive. The most e¤cient voured by chemical fractions of aqueous extract of stimulants among the amino acids were l-tyrosine, shrimp was made for two salmonid ¢sh, Atlantic sal- l-phenylalanine, l-lysine and l-histidine (Adron and mon and rainbow trout (Mearns et al.1987).The Mackie 1978). High palatability of free amino acids results indicate that various types of substances are was found for various freshwater and marine ¢sh by stimulants for rainbow trout. These are free amino many researchers (Hidaka 1982;Mackie 1982; acids, nucleotides, ammonium bases and some acidic Mearns et al. 1987;Adams et al. 1988;Jones 1989; and neutral compounds, which were not identi¢ed. Kasumyan1997). Bycontrast, theAtlantic salmonappears to be a selec- The free amino acids are not, however, solely tive feeder. Agar gel pellets £avoured by the stock responsible for the high palatability of food organ- solution containing all water-extractable com- isms for ¢sh. For both plaice and dab, the l-amino pounds, but without protein, were only accepted to a acid components and the nonamino acid compo- certain degree, and a high proportion of pellets were nents of synthetic squid mixture acted as feeding sti- rejected. It was supposed that the water-soluble pro- mulants (Mackie1982). ForAtlantic cod, a mixture of teins, which were removed in the ¢rst stage of the both l-amino acids and neutral l-amino acids acted fractionation procedure, are responsible for the palat- as stimulants, but neither were as e¡ective as the syn- abilityof shrimp extract. Further studies arerequired thetic squid mixture. A diet containing 1% of a syn- to investigate this possibility. thetic squid extract was consumed to 100%, In an attempt to ¢nd stimulants for herbivorous nonamino acid fraction was not consumed, all the tilapia, the aqueous extract of Romaine lettuce was amino acids were consumed at 61%, and the neutral fractionated and a bioassay of chemical fraction was amino acids were consumed at 48% (Johnstone and performed. Contrary to Atlantic salmon, agar gel £a- Mackie1990). voured with extract depleted of proteins by precipita- The second approach to the study of the chemical tion with TCA was consumed to the same degree as nature of taste stimulants consists of separation of agar gel £avoured with whole lettuce homogenate, extract of food organisms and analysing the compo- indicating that proteins are not the stimulants. The sition of chemical fractions that show stimulatory lipid-containing ether extract as well as the fractions e¡ect in the bioassay. It should be noted that this containingorganicacidsorsugarsalsohadnostimu- approach is very promising for studies on the chemi- lant activity. The free amino acid fractions only eli- cal nature of taste stimulants, but has seldom been cited positive taste response in tilapia. The ¢ve used (Johnsen and Adams 1986;Takii et al. 1986; amino acids ^ glutamic acid, aspartic acid, serine, Mearns et al.1987). lysine and alanine ^ were found to be stimulants for

328 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

tilapia (Johnsen and Adams 1986). Further studies The oral taste system seems to have a narrower showed that citric acid was one of the major constitu- spectrum of e¡ective amino acids than the extraoral ents of the organic acid fraction, and was found to be taste system. There were only six substances found highly palatable for tilapia when examined as single to be stimulants for the Russian sturgeon, three for taste substances (Adams et al. 1988). the stellate sturgeon and seven for the Siberian stur- Various separation methods were used for identi¢- geon. Alanine had a deterrent e¡ect in this last spe- cation of feeding stimulants for young yellowtail to cies (Table 7). Among the stimulant amino acids, jack mackerel muscle extract (Hidaka et al. 2000). none were common to all the three species, and there The authors found that the stimulatory substances was no signi¢cant correlation between the species had a molecular weight less then 10 000 and that towards responses to amino acids. There was no sig- the substances were absorbed to the stationary phase ni¢cant correlation between sturgeon species con- in an anion-exchange column (DEAE-Sephadex). cerning the oral taste responses. Comparison of the Stepwise elution with NaCl and testing of fractions extraoral and oral taste responses within one species in starch pellets indicated that inosine-50-monopho- revealed a high correlation in the Russian sturgeon sphate and lactic acid are the main feeding stimula- (rs ˆ 0.79; P < 0.001) and lower correlation in the tory substances for yellowtail. Siberian sturgeon (rs ˆ 0.48; P < 0.05) and stellate

sturgeon (rs ˆ 0.44; P < 0.05). It is noteworthy that some substances have an opposite e¡ect on extraoral Relation between oral and extraoral and oral taste receptors. For example, citric acid is a taste preferences strong incitant for sturgeons and increases the fre- There are onlya few studies that permit comparisons quency of pellets caught by ¢sh in comparison with of the oral and extraoral taste systems. They have all blank pellets. But all pellets with citric acid are imme- been done on sturgeons. Sturgeons have numerous diately rejected. taste buds, not only in the mouth, but also on the epi- The extraoral taste system is at least 10 times dermis of both barbels and lips (Pevzner 1981;Devit- more sensitive than the oral system. For example, sina and Kazhlaev 1993). As a result of the external the extraoral taste threshold for citric acid is 5 mm gustatory receptors on the barbels, sturgeon can and the oral taste threshold is 50 mm (Fig. 14; receive information concerning palatability of food Kasumyan and Kazhlaev 1993a). Calculations based items without taking them into the mouth. A rapid on the size of pellets revealed that about 1^2 mg touch of a food item by one barbel is su¤cient to per pellet of taste substance is enough to stimulate assess the food quality.This preliminary assessment appetite in ¢sh. of food, which takes place in a fraction of a second, The extraoral taste system develops more rapidly seems to be essential for inducing feeding behaviour than the oral system during sturgeon ontogenesis. in sturgeon. The ¢nal assessment of food acceptabil- Eleven amino acids were e¡ective for extraoral taste ity is based on the oral sense (Kasumyan and Kazh- buds in Russian sturgeon larvae (2.1^2.5 cm in body laev1993a;Kasumyan1999a). length) compared with 16 amino acids in juveniles In 4^6-cm long Siberian sturgeon,14 of 19 amino (6^7 cm in body length). In the oral taste system, the acids were incitants, in 6^7-cm long Russian stur- di¡erence between larvae and juveniles was much geon,16of 21were incitants and in 6^8-cm long stel- more dramatic, with only one amino acid being e¡ec- late sturgeon,19 of 21amino acids gave a signi¢cant tive for larvae compared with six in juveniles increase in the number of food pellets caught (Table 13). There was no correlation in the oral taste (Table 16).Twelve of these amino acids were common spectra of larvae and juveniles, whereas the corre- to the three sturgeon species, although the correla- spondence in extraoral taste spectra of larvae and tion coe¤cient was high only for Russian sturgeon juveniles was very high (rs ˆ 0.69; P < 0.01;Kasum- versus stellate sturgeon (the Spearman's rank corre- yan et al. 1992). lation coe¤cient rs ˆ 0.84; P < 0.001). The most According to the phylogenetic tree of sturgeons potent incitants for these two species were l-aspartic drawn up on the basis of their karyotypes, the acid and l-glutamic acid and l-cysteine, while l- Russian sturgeon and Siberian sturgeon with 240^ asparagine, l-threonine and l-methionine were the 260 chromosomes evolved more than 80 million most potent substances for Siberian sturgeon. Only years ago (Vasil'ev1985). The stellate sturgeon, with two amino acids, namely l-proline and l-tyrosine, 120 chromosomes, evolved signi¢cantly earlier. were ine¡ective for all three species. Based on these aspects of sturgeon phylogeny, it was

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 329 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Table 16 Extraoral taste preferences to amino acids

Amino acid (L-isomers) Concentration (M) Russian sturgeon Stellate sturgeon Siberian sturgeon

Cysteine 0.1 35ÃÃ 35ÃÃÃ ± Histidine 0.1 31ÃÃÃ 27ÃÃÃ 24Ã Threonine 0.1 30ÃÃÃ 22ÃÃÃ 42ÃÃÃ Arginine 0.1 29ÃÃÃ 25ÃÃÃ 17 Lysine 0.1 29ÃÃÃ 28ÃÃÃ 22 Asparagine 0.1 28ÃÃÃ 21ÃÃÃ 47ÃÃÃ Glutamine 0.1 25ÃÃÃ 23ÃÃÃ 29ÃÃÃ Serine 0.1 24ÃÃÃ 28ÃÃÃ 28ÃÃÃ Phenylalanine 0.1 23ÃÃ 19ÃÃÃ 22Ã Alanine 0.1 21ÃÃ 24ÃÃÃ 30ÃÃÃ Methionine 0.1 16Ã 20ÃÃÃ 41ÃÃÃ Glycine 0.1 14 19ÃÃÃ 36ÃÃÃ Valine 0.1 10 12ÃÃ 30ÃÃÃ Norvaline 0.1 7 11ÃÃ ± Proline 0.1 7 10 15 Aspartic acid 0.01 38ÃÃÃ 36ÃÃÃ 26ÃÃ Glutamic acid 0.01 36ÃÃÃ 28ÃÃÃ 32ÃÃÃ Tryptophan 0.01 23ÃÃÃ 13ÃÃÃ 32ÃÃÃ Leucine 0.01 15Ã 12ÃÃ 30ÃÃÃ Isoleucine 0.01 14 14ÃÃÃ 14 Tyrosine 0.001 14 9 19 Control ± 7 7 15

Three species of Acipenser sturgeons: Russian sturgeon, A. gueldenstaedtii (6±7 cm body length), stellate sturgeon, A. stellatus (6±8 cm body length) and Siberian sturgeon, A. baerii (4±6 cm body length) were exposed to pellets containing amino acids. Extraoral taste pre- ference is the mean number of snaps of a 50 agar pellets by ®ve ®sh during 5-min exposure time. Controls were pellets containing Pon- ceau 4R or 0.3% Cr2O3 (modi®ed from Kasumyan 1999a).

Figure 14 Oral and extraoral taste preferences. Oral (A) and extraoral (B) taste preferences for di¡erent concentrationof citric acid in juveniles of Siberian sturgeon, Acipenserbaerii. Control ^ blank pellets.Vertical columns and T-bars are the mean and the SEM (modi¢ed from Kasumyanand Kazhlaev 1993).

330 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

suggested that extraoral taste sense is evolutionarily were the most e¡ective taste stimuli in electrophysio- more stable than oral taste sense (Kasumyan logical study for Arctic char (Hara et al. 1993), but 1999a,b). were classed as neutral taste substances in beha- vioural assays (Kasumyan and Sidorov 1995d). Hydroxy-l-proline evoked intense neural responses Correlation between bioassay and in rainbow trout palate nerve (Marui et al.1983a), electrophysiological data but was not a palatable taste substance for this ¢sh The input to the brain from any sensory structure (Jones 1990). Similar observations were made for l- can be recorded by electrophysiological means. alanine and betaine, which were ine¡ective taste These methods are powerful techniques as they substances in behavioural assay (Jones1989), but eli- can reveal thresholds or sensitivities, speci¢city, cited strong neural responses in the rainbow trout adaptation and many other functional properties palate (Marui et al.1983a). of these systems. However, the recordings of ner- It has been shown that l-proline is the only amino vous activity does not tell about the animals' reac- acid that evoked responses from the gustatory nerves tion towards a speci¢c chemical;in other words, a in electrophysiological studies in seven Salmonidae discharge of nerve impulses can indicate that the species and in nine strains of the rainbow trout (Hara stimulus was either a stimulant or a deterrent. It et al. 1999). The detection threshold for l-proline was seems important to relate the activity evoked by found to be between 10À7 and10À8 m. These charac- taste substances determined by electrophysiologi- teristic features are consistent with the hypothesis cal methods and those determined in bioassays. proposed by Hara and coworkers (Hara and Marui Such comparisons, however, are available for a few 1984;Ha ra et al.1993) that `Pro'receptor or transduc- species only. tion system is one of the four systems that mediate sti- In the red sea bream, a least square correlation mulation by amino acids in salmonids. However, coe¤cient of 0.78 was found between amino acids proline was an ine¡ective stimulus in the beha- that were e¡ective in initiating feeding behaviour viouralassay provided on di¡erent salmonids ^ chum and those that stimulated gustatory neural activity salmon (Kasumyan and Sidorov 1993b), brown trout (Goh and Tamura 1980a). In the pu¡er, the sub- (Kasumyan and Sidorov1995c), frolich char (Kasum- stances l-alanine, glycine, l-proline, l-serine and yan and Sidorov 1995d) ^ and proline was found to betaine stimulated lip taste receptors in electrophy- be a deterrent for rainbow trout byAdron and Mackie siological recordings. These substances were also (1978), whereas it was found to be a highly palatable e¡ective in stimulating pellet acceptance. The con- substance by Jones (1989). Nucleotides AMP, ADP, centration of amino acids needed to release beha- IMP and UMP have been shown to be e¡ective taste vioural responses was greater by two to three orders stimuli in neural recordings in the pu¡er, but these of magnitude than that required to elicit gustatory compounds did not evoke positive taste responses in neural activity (Hidaka et al.1978;Ohsugi et al.1978). behavioural assays performed on that species In both the electrophysiological studies and in the (Hidaka et al. 1977;Hidaka 1982). Discrepancies behavioural tests, the long-chained aliphatic organic betweenthe sensory input revealed by electrophysio- acids ^butyric, caproic and valeric acids ^werefound logical means and the behavioural output emphasise to be stimulatory substances for Atlantic salmon. that palatability is not synonymous with excitability Recordings from the palatine nerve showed that, of in the ¢sh taste system. the few amino acids tested, all were nonstimulatory, with the exception of l-proline. However, l-proline Correlation between smell and taste was an indi¡erent taste substance for Atlantic sal- preferences mon in bioassays (Sutterlin and Sutterlin1970). There are several examples of discrepancies Olfaction and gustation are the main chemosensory betweenthe sensory input revealed by electrophysio- systems in ¢sh. Both these systems are sensitive to logical means and the behavioural output evoked by chemical substances released by ¢sh food organisms the same substances. In Arctic char, l-glutamic acid and play important roles in ¢sh feeding behaviour was found to be the most palatable substance in beha- (Atema 1980;DÖving 1986;Pavlov and Kasumyan vioural studies (Kasumyan and Sidorov 1995d), but 1990,1998;Maruiand Caprio1992).There is a distinct had no e¡ect in electrophysiological recordings di¡erence in the anatomy of the gustatory and olfac- (Hara et al.1993).Incontrast,l-proline and l-alanine tory organs in ¢sh, and these distinctions are also

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 331 Taste preferences Alexander O Kasumyan & Kjell B DÖving

evident when it comes to the functional characteris- 1985). However, palatability and toxicity of prey tics of these chemosensory systems. According to are not necessarily correlated (La Barre et al. 1986; electrophysiological studies, the gustatory system Sammarco et al.1987;Wylie and Paul 1988;Sammarco tends to be more selective than the olfactory system, and Coll 1990;Schulte and Bakus 1992). It has been and the gustatory spectra of e¡ective taste sub- suggested that decrease in palatability might be a stances are more species-speci¢c than the olfactory more important defensive trait for aquatic prey spectra (Goh andTamura1980b;DÖving1986;Caprio organisms than toxicity (Schulte and Bakus1992). 1988;Marui and Caprio1992;Hara1994a). One group of aquatic animals, which are very vul- Bioassays support the notion that there are specia- nerable to ¢sh, are sponges. Nevertheless, some lisations of the olfactory and gustatory systems in sponges are inedible for ¢sh and survive in biotopes ¢sh.The e¡ectiveness of aminoacid solutions in evok- like coral reefs, where ¢sh density and the diversity ing feeding behaviour in common carp via the olfac- of ¢sh are very high. The palatability of sponge tissue tory system (Saglio and Blanc1983;Kruzhalov 1986) and ability of sponge crude extracts to inhibit ¢sh is di¡erent from the palatability of these substances feeding was investigated in a reef lagoon at Yanutha (Kasumyan and Morsy 1996), and demonstrates that Island, Mbengga, Fiji Islands. Sponge tissue was cut the responses of the gustatory and olfactory systems into small pieces (3^5 mm diameter). The reef ¢sh ^ are quite di¡erent. The spectra of palatable amino the sergent-major, the parrot¢sh (Scarus sp.), the acids (Kasumyan and Sidorov1995c) do not coincide damsel¢sh (Dascyllus aruanus) and the green chro- with the spectra of amino acids e¤cient as olfactory mis ^ were fed approximately equal-sized bits of ¢sh stimuli for browntrout (Mearns1986). Glycine,which meat or bread and then alternately o¡ered pieces of evoked obvious bottom food searching behaviour in sponge (Carteriospongia sp. and Dysidea sp.) of similar di¡erent species of sturgeons (Acipenseridae) was an appearance and dimensions. Bread was also soaked oral taste stimulant for the Siberian sturgeon only in100 mL of water into which sponge tissue had been (Kasumyan1994;Kasumyan andTau¢k1994). squeezed. Acceptance or rejection of the food items Of the free amino acids, glycine and l-alanine was recorded along with the subsequent behaviour induce food search behaviour in a number of teleost of the ¢sh (Schulte and Bakus1992). It was found that ¢sh species from di¡erent systematic and ecological reef ¢sh consumed all pieces of the bread or ¢sh meat; groups. Thus, feeding behaviour is elicited by these yet, they never consumed the Carteriospongia sp. or substances in Japanese eel (Hashimoto et al.1968), Dysidea sp. tissue. An exception was one trial in oriental weather¢sh (Harada 1985), Atlantic cod which 50^70% of the bread that had been soaked in (Pawson 1977;Ellingsen and DÖving 1986), winter seawatercontainingaqueous extract of Carteriospon- £ounder (Sutterlin 1975) and goby (Gobiosoma bosci; gia sp. was consumed and only 0^38% of the bread Hoese and Hoese 1967). However, in many ¢sh spe- soaked in Dysidea sp. extract was consumed. These cies, these two amino acids do not possess a high results show that the sponge tissue contained sub- palatability. In several species, they are inert or act stance(s) that have deterrent e¡ect on reef ¢sh and as deterrents;for instance, l-alanine for Siberian prevent sponges from being eaten by ¢sh. Bakus sturgeon and grass carp, and glycine for brown trout et al. (1989) showed that all exposed benthic, soft- and European grayling (see Table 7). In the red sea bodied, macrobiota tested in the Fiji (22 species) bream, a correlation coe¤cient with a probability of and Maldives (17 species) Islands were unpalatable 0.01 was found between the amino acids that were to resident coral reef ¢sh. In a survey of Caribbean e¡ective as taste stimuli in bioassays and those that sponges, crude extract from 44 of 71 species exhib- stimulated olfactory neural activity (Goh and ited properties that deterred ¢sh feeding (Pawlik Tamura1980a,b). et al. 1995). Besides sponges, various other aquatic organisms use deterrents to protect themselves against preda- Naturaldeterrents tors. Addition of live nemertides or polychaetes to There are numerous aquatic plants and animals tank water with Dover sole does not induce any that use chemical defence as protection against obvious response. If, however, the head of a Dover ¢sh. These chemicals might be toxic, venomous or sole happens to come in contact with either of these deterrent compounds. Substances isolated from animals, the ¢sh react with a violent avoidance di¡erent species of marine invertebrates might be response. The Dover sole rapidly swims away, `shak- both deterrent and toxic for ¢sh (Thompson et al. ing'or `nodding' its head. The addition of polychaete

332 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

mucus in the diet as well as saponins (glycosids) iso- wrasse in laboratory assays so long as the extract lated from blue star¢sh also reduces the food con- had the same concentration as in the natural state. sumed (Mackie and Mitchell 1982b). Various corals Stevensine also deterred feeding in a wide assem- (Mackie1987;Fenical and Pawlik1991;Kerr and Paul blage of ¢sh in assays performed on reefs where 1995;E pi fa n io et al. 1999b), green algae and cyano- A. corrugata is found (Wilson et al. 1999). Furanoses- bacteria (Paul et al. 1993;Thacker et al.1997),nudi- terpene variabilin, an ichthyodeterrent, has been iso- branchs (Rogers and Paul 1991), nemertean (Heine lated from the sponge Ircinia strobilina (Epifanio et al. et al.1991), commonoceanic holoplankton organisms 1999a). A medium ring haloether, brasilene, has (McClintock et al. 1996), echinoderms (Bryan et al. been isolated from the opisthobranch, and this com- 1997), gastropods (McClintock and Janssen 1990), pound acts as a feeding deterrent against several spe- aquatic bugs, mites and beetles (Kerfoot et al.1980; cies of ¢sh (Kinnel et al. 1979). Concentrations of Gerhart et al.1991;Lokensgard et al.1993), polychaete secondary metabolites in prey tissue can be quite worms (Prezant 1980;Yoshiyama and Darling 1982), high, for example, scalaradial constituted 2.4% of prey ¢sh (Winn and Bardach 1959;Gladstone 1987), the total dry mass in the Paci¢c sponge (Rogers and eggs, hatchlings and tadpoles of anurans (Voris Paul 1991). Concentration of stevensine in the Carib- and Bacon 1966;Kruse and Francis 1977;Kats bean sponge A. corrugata, ranged from 12 to et al. 1988;Komak and Crossland 2000;Crossland 30 mg mL À1 sponge tissue (mean 19.0 mg mLÀ1; 2001)and othergroups of aquatic organisms use che- Wilson et al.1999). mical deterrent to make themselves inedible for ¢sh. Awell-known neurotoxin found in the skin of puf- Du¡y and Paul (1992) suggested that chemical fer ¢sh species, tetrodotoxin, also has deterrent prop- defences areless e¡ective in high- than in low-quality erties. Behavioural studies demonstrated that some foods. ¢sh species reject toxic pu¡er tissues and arti¢cial The existence of chemical defence mechanisms food pellets containing tetrodotoxin (Yamamori et al. in food organisms is a widely recognised phenom- 1980, cited in Saito et al.1984;Yamamoriet al.1988). enon. Unfortunately, few of the deterrent sub- Electrophysiological experiments in rainbow trout stances have been identi¢ed as the research on and Arctic char demonstrated that tetrodotoxin was such chemical defences has only recently begun. a highly e¡ective taste stimulus (Yamamori et al. Many of the unusual secondary metabolites that 1988). Aquarium observations in which Dusky have been isolated from soft-bodied marine inverte- grouper was kept in starved condition showed that brates and marine weeds are proposed to function theyswallowed live mice, but promptly rejected small as agents that deter predation (for review, see Paul pu¡er ¢sh (Freitas et al.1992). The fertilised eggs and 1992;Pawlik 1993). Some deterrents have been newly hatched larvae of a tetraodontid, such as identi¢ed as saponins (Lucas et al. 1979), p olyket ide sharpnose pu¡er ¢sh, are unpalatable to potential sulphates (Rideout et al. 1979), terpenoids, aceto- egg predators. In experiment, both eggs and larvae genins, unusual fatty acids, phlorotannins, malyn- were mouthed, then immediately rejected by wrasse gogamides (Hay 1991;Thacker et al.1997)and (Thalassoma lunare), and damsel¢sh (Dascyllus reticu- steroids (pregnanes;Gerhart et al. 1991;Lokensgard lates). It was suggested that tetrodotoxin is present et al.1993). in skin, liver, intestine and ovaries of tetraodontids, A new polymeric pyridinium alkaloid, named thus making the eggs and larvae of these ¢sh unpala- amphitoxin, has been isolated as the chemical table.The nonpalatabilityof their fertilised eggs prob- defence of the Caribbean sponge Amphimedon com- ably explains the absence of parental care in these pressa (Albriz io et al.1995). Two compounds, an alka- domersal spawning species (Gladstone1987). loid called oroidin and its carboxypyrrole hydrolysis Chemical defence among prey organisms is espe- product, 4,5-dibromopyrrol-2-carboxylic acid, have cially important in tropical coral reefs, where ¢sh been identi¢ed as primary defensive agents of Carib- may forage the bottom using in excess of bean sponges of the genus Agelas against the reef ¢sh, 150 000 bites mÀ2 dayÀ1, and where either ¢sh or bluehead wrasse (Channas et al. 1996). Recently, sea-urchins alone are capable of removing nearly dibrominated alkaloid stevensine, closely related to 100% of the daily production (Carpenter 1986). The the compound oroidin, has been isolated from the dominance of certain coral reef species may result in Caribbean spongeAxinella corrugata. A crude extract part from their unpalatability to ¢sh. Fish-feeding of the sponge, which includes abutanol-soluble parti- experiments on coral reefs showed that all exposed tion of the extract, deterred feeding of the bluehead soft-bodied benthic macrobiota animals were unpa-

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 333 Taste preferences Alexander O Kasumyan & Kjell B DÖving

latable and rejected by a variety of predators (Bakus re¢nements that use substances that attract the ¢sh et al.1989). More than half of the common coral reef from a distance and substances that have the correct invertebrates are toxic to ¢sh (Bakus 1981). It was palatability for the ¢sh species that preferably should found that ¢sh of the temperate zone might be more be caught.The arti¢cial baits have beguntheirera in a¡ected by seaweed deterrents then the tropical sport¢shing,andthetimeisripefortheiruseincom- herbivorous ¢sh (Cronin et al. 1997). It was also mercial ¢sheries. shown that the combination of secondary meta- Knowledge of feeding behaviour of ¢sh in aquacul- bolites and calcium carbonate in the aragonite ture is important so that feeds and feeding techni- form proved a more e¡ective deterrent than either ques can be designed to encourage consumption secondary metabolites or aragonite alone (Meyer and hence survival and growth whilst minimising and Paul 1995). metabolic energy expenditure in feeding (Jobling It seems to be a general feature that ¢sh stimulants 1994). The eventual incorporation of feeding stimu- are species-speci¢c, thus the spectra of stimulants lants into diets consistingof cheap and normally non- have di¡erent composition for di¡erent species. This palatable protein sources may be of practical ¢nding is in conspicuous contrast to the ¢nding that signi¢cance to the ¢sh farming industry (Jobling deterrents, like secondary metabolites, are not spe- et al.2001). Furthermore, wastage must be minimised cies-speci¢c, but are highly e¤cient fora range of ¢sh because of high feed costs and the potentially dele- species. For example, a crude extract of sponge (Axi- terious e¡ects of waste food on water quality. Non- nella corrugata) at the same concentration as in the acceptance of arti¢cial food, which can cause large sponge tissue signi¢cantly deterred feeding in a vari- ¢sh mortality, is a serious problem in aquaculture, ety of ¢sh species present on the reef where the especially in the rearing of marine ¢sh larvae (de la sponge inhabited. They were wrasses (Thalassoma Higuera 2001). Large amounts of free amino acids, spp.)and Halichoeres spp.,snapper,parrot¢sh(Scarus which have high solubility in water, were lost during spp.) and Sparisoma spp., grunts, tile¢sh, porgy and mastication by ¢sh. Forcommoncarp, the percentage angel¢sh (Pomacanthus spp.) and Holacanthus spp. loss of amino acids within 15 min after feeding (Wilson et al. 1999). Moreover, crude extracts of 15 of reaches more then 70% for arginine, leucine, iso- 17various coral reef macrobiota in the Fiji and Mald- leucine, lysine, methionine, valine, alanine, glutamic ives Islands including sponges, algae, corals, asci- acid, glycine, proline, serineand some others (Yamada dians and sea anemones were unpalatable to andYone1986). gold¢sh (Bakus et al. 1989) that have never been There are several publications that show the exposed to tropical reefs. Pure secondary metabolites importance of introducing taste stimulants into of di¡erent species of marine demosponges from San modern ¢sh cultivation technologies. It was shown Diego, California inhibited feeding of gold¢sh as well that start-feeding of European glass eels with a trout as feeding of several species inhabiting the same area fry crumble can be improved signi¢cantly by adding where demosponges were collected (Thompson et al. a mixture of the amino acids l-alanine, glycine, l- 1985). histidine and l-proline to the feed (Kamstra and Heinsbroek1991).The mixture used as attractant was based on the mixture applied by Takii et al.(1984). Application in aquaculture and ®sheries In this mixture, the authors also used a nucleotide, Information about the palatability of feeding sub- uridine-50-monophosphate (UMP), which had a stances provides a basic knowledge in ¢sh physiol- small synergistic e¡ect. In the experiments of ogy, but such information might be of signi¢cant Kamstra and Heinsbroek 1991), UMP was omitted value when applied to cultivation and ¢sheries. The from the mixture as it was so costly that it would pre- growing use of ¢sh in aquaculture demands sources vent practical application. Addition of amino acid for less expensive food for the ¢sh. The use of addi- mixture signi¢cantly increased the ad libitum meta- tives that improve the palatability of arti¢cial food bolizable energy of the diets from 45 kJ kgÀ0.8 dayÀ1 for ¢sh is a probable application. The possible to 50-60 kJ kgÀ0.8 dayÀ1, which resulted in an destruction of ¢sh biotopes by trawling could pro- increase in growth rate from 1.3 g kgÀ0.8 dayÀ1 to mote and enforce less damaging ¢shing methods, 2.4 g kgÀ0.8 dayÀ1 (Heinsbroek and Kreuger 1992). such as passive ¢shing gears like traps and long lines. Using a mixture that stimulates feeding, which These ¢shing methods have been used for centuries contained l-alanine, l-serine, inosine-50-mono- and are not outdated. Their use, however, requires phosphate and betaine in plant-based diets, improved

334 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

feed acceptability and growth performance of cul- et al. 1995). The palatability of each of the 19 compo- tured striped bass, fed low-cost, plant-based diets nents that make up a feed for 9^12-cm long stellate (Papatryphon and Soares 2000). Improvements sturgeon juveniles was tested. The components were in feed intake, weight gain and feed e¤ciency as a ¢sh meal, krill meal, squid meal, meat and bone meal, result of feeding stimulant supplementation have dry whey, ¢sh protein concentrate, Na-caseinate, been previously observed in gilthead sea bream yeast hydrolysate, PVF yeast, PVF enzymated yeast, (Tandler et al.1982) and European sea bass (Dias et al. Eprin yeast, coarsely cut sun£ower seed, coarsely 1997). cut soyabean, wheat £our, wheat bran, fukus grist, Anotherexample showing the signi¢cance of taste PM-2 premix, PF-2V premix and FinnStim (Finnish qualities in arti¢cial ¢sh feed has been shown in the Sugar Co. Ltd). It was found that water extract results obtained by Stradmeyer (1989). It was found of 13 of the components, mainly the ¢sh protein that just after neutralising the salmon feed (Frippak concentrate, PVF enzymated yeast and premixes Feeds Ltd.;pH 3^4) pellet ingestion by Atlantic sal- included in agar-agar pellets provoked the increase mon fry increased by a factor of two in comparison of grasp activity in ¢sh. Extracts of PF-2V premix, to non-neutralised food. It has been shown that krill meal, ¢shmeal concentrate, Eprin yeast and withinthe ¢rst minute of feeding, ¢shate more than coarsely cut soyabeans signi¢cantly increased the 50% of the total feed consumed during the total per- pellet consumption. The Japanese pilchard ¢shmeal iod (4 min;Oikawa and March 1997). Thus, there is stabilised by various antioxidants such as santo- probably an advantage in applying feeding stimu- chin, anfelan-1, anfelan-5 and ionol released di¡er- lants to the surface of feed pellets, thus minimising ent kinds of behavioural taste responses in stellate the amount required and facilitating its detection by sturgeon. The FinnStim was ine¡ective as both oral the ¢sh. and extraoral taste stimuli. Based on the results It has been reported that ¢sh fed a diet of poor obtained, it was possible to correct the feed formu- palatability were satiated sooner than ¢sh fed a more las so as to improve their palatability for ¢sh acceptable diet. If ¢sh are fed to satiation with a diet (Kasumyan et al. 1995). of poor palatability and then given a more palatable High protein content and a balanced amino acid diet, they will begin eating again until they reach a pro¢le make the earthworm, one of the most culti- new satiation level (Ishiwata 1968). It has been vated species in vermiculture, a useful alternative shown that arti¢cial food with good chemosensory protein source for ¢sh. However, this invertebrate characteristics is digested by ¢sh more e¡ectively contains unpalatable compounds in the coelomic (Takeda andTakii1992). £uid. The removal of these compounds resulted in Another possible application of feeding stimulants an increase in feed intake in rainbow trout ¢nger- may be to mask deterrent ingredients that may lower lings, and the use of squid extract as a feeding stimu- the palatability of diets. Unpalatable plant proteins lant further improved feed intake of ¢sh fed are used as substitutes for ¢shmeal protein in com- formulations containing earthworm meal (Carde- mercial diets. Certain antibiotics have been shown to nete et al.1993). reduce the palatability of feed (Schreck and Mo¤tt 1987;Poe and Wilson 1989;Robinson et al.1990; Epilogue Hustvedt et al. 1991;Robinson and Tucker 1992). Supplement of feeding stimulants might mask In this review, we have presented the status for taste deterrent ingredients and improve feed intake (Toften preferences in ¢sh, but evidently, what is done so far et al.1995;Dias et al. 1997;Papatryphon and Soares is only a beginning of an extended series of experi- 2000). ments that will provide a sound scienti¢c basis for Using special additives might increase palatability our understanding of the subject. In the present of arti¢cial food, but it might also be raised by remov- review,we have stressed the speci¢cityof taste prefer- ing the components with bad taste or by replacing ences in ¢sh. We also point out the great variability those components with equivalents, which do not among individuals. Further, it seems that the ¢sh is provoke negative taste responses. To this end, one indi¡erent to previous taste experience. This prop- must estimate the palatability for each of the compo- erty is probably unexpected among ¢shers and aqua- nents that make up the feed. In one special study, it culturists. It should be noted that the example of was shown that components of arti¢cial food have patroclinous inheritance of taste preferences does strongly di¡erent taste qualities for ¢sh (Kasumyan not need to be the same for all species.We hope that

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 335 Taste preferences Alexander O Kasumyan & Kjell B DÖving

we have made it clear that ¢sh taste is much more References than the ¢ve modalities which dominate mammalian Adamek, Z., Fasaic, K. and Debeljak, L. (1990) Lower tem- taste. perature limits of plant food intake in young grass carp Taste preferences are under genetic control, and (Ctenopharyngodon idellaVa l.). Ichthyologia 22,1^8. hence, readily inheritable. The fact that speci¢c Adams, M.A., Johnsen, P.B. and Hong-Qi, Z. (1988) Chemical classes of substances, such as the amino acids and enhancement of feeding for the herbivorous ¢sh Tilapia nucleotides, remain as feeding stimulants across a zillii. Aquaculture 72,95^107. broad taxonomic expanse of ¢sh orders suggests that Adron, J.W.and Mackie, A.M. (1978) Studies on the chemical the genetic controlof taste preference in ¢sh has been nature of feeding stimulants for rainbow trout, Salmo phylogenetically conservative. Moreover, there is a gairdneri Richardson. Journal of Fish Biology12,303^310. strong conservatism of taste preference within spe- Albrizio, S., Ciminiello, P., Fattorusso, E., Magno, S. andPaw- cies;even those now diverged into di¡erent strains lik, J.R. (1995) Amphitoxin, a new high molecular weight antifeedant pyridinium salt from the Caribbean sponge and yet there is almost no genetic conservation of Amphimedon compressa. Journal of National Products 58, taste preference between even closely related species 647^652. withinthe samegenus having similardiets and occu- Andriashev,A.P.(1944) Role of sense organs in searching for pying similar ecological niches. A possible reason food in Gaidropsarus mediterraneus. Journal of General for this apparent contradiction is that we areworking Biology (Zhurnal Obshei Biologii) 5,123^127(inRussian). with substances that are onlyof marginal interest to Appelbaum, S. (1980) Versuche zur Geschmacksperzeption the ¢sh taste receptors. Therefore, it seems a useful einiger Susswasser¢sche im larvalen und adulten Sta- aspect for the future to uncover the unknown sub- dium. Archiv fu« r Fischereiwissenschaft 31,105^114. stances that evidently take action in ¢sh taste prefer- Appelbaum, S., Adron, J.W., George, S.G., Mackie, A.M. and ences. This is needed because there are certainly Pirie, B.J.S. (1983) On the development of the olfactory taste substances that are much more potent than and the gustatory organs of the Dover sole, Solea solea, during metamorphosis. Journal of Marine BiologicalAsso- those commonly in use. ciation, UK 63,97^108. A major breakthrough has been the uncovering of Atema, J. (1971) Structures and functions of the sense of the molecular basis for taste receptors. Future studies taste in the cat¢sh (Ictalurus natalis). Brain, Behaviour will undoubtedly uncover the properties of these and Evolution 4,273^294. receptors and which substances the ¢sh can detect. Atema, J. (1980) Chemical senses, chemical signals and Finally, we would like to add that experiments on feeding behaviour in ¢shes. In: Fish Behaviour and its ¢sh taste preferences can be made with simple Use in the Capture and Culture of Fishes (eds J.E. Bardach, means and a modest budget, and can from someone's J.J. Magnuson, R.C. May and J.M. Reinhart). International viewpoint look like children's play.However, through Center for Living Aquatic Resources Management, careful planning and systematic data collection, and Manila, pp.57^101. by scrutinizing the results, it is possible to obtain Atema, J., Holland, K. and Ikehara, W. (1980) Olfactory responses of yellow¢n tuna (Thunnus albacares)toprey important information for science, ¢sheries and odors: chemical search image. Journal of Chemical aquaculture. Ecology 6,457^465. Bailey,M. and Sandford, G. (1999) The Ultimate Encyclopedia Acknowledgements of Aquarium Fish and Fish Care. Lorenz Book, London. Bakus, G.J. (1981) Chemical defence mechanisms on the We wish to thank Sergey Sidorov, Abdurahman Great Barrier Reef, Australia. Science 211,497^499. Kazhlaev, Nikolaj Pashchenko, Amal M. H. Morsy, Bakus, G.J., Stebbins,T.D., Lloyd, K. et al. (1989) Unpalatabil- Ekaterina Nikolaeva, Olga Prokopova and Elena ity and dominance on coral reefs in the Fiji and Maldive Fokina for their valuable contributions to many of Islands. Journal of Zoology and Aquatic Biology1,34^53. the experiments cited in this review.We are grateful Bannister, L.H. (1974) Possible functions of mucus at gusta- to George Alexander, Paul Hart, Kurt Kotrschal tory and olfactory systems. In:Transduction Mechanisms in Chemoreception (ed.T.M. Poynder). Information Retrie- and Kathryn Mearns for helpful comments on an val, London, pp.39^48. earlier version of our manuscript and for help Bardach, J.E. and Atema, J. (1971)The sense of taste in ¢shes. with the English. This study was supported by the In: Handbook of Sensory Physiology. Vol. 4 (ed. L.M. Bei- Russian Foundation for the Basic Research, the der). Springer-Verlag, Berlin, pp.293^336. International Science Foundation, the European Bardach, J.E., Todd, J.H. and Crickmer, R.K. (1967) Orienta- Science Foundation and the Norwegian Research tion by taste in ¢sh of genus Ictalurus. Science 155, Council. 276^1278.

336 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Bardach, J.E., Winn, H.E. and Menzel, D.W. (1959) The role Caprio, J. (1984) Olfaction and taste in ¢sh. In: Comparative of the senses in the feeding of the nocturnal reef preda- Physiology of Sensory Systems. Olfaction and Taste in Fish tors Gymnothorax moringa and G. vicinus. Copeia 2, (eds L. Bolis, R.D. Keynes and S.H.P.Maddrell). Cambridge 133^139. University Press, Cambridge, pp.257^283. Barlow, L.A. and Northcutt, R.G. (1995) Embryonic origin Caprio, J. (1988) Peripheral ¢lters and chemoreceptorcells in of amphibian taste buds. Developmental Biology 169, ¢shes. In: Sensory Biology of Aquatic Animals.(edsJ. 273^285. Atema, R.R. Fay, A.N. Popper and W.N. Tawolga). Beauchamp, G.K., Cowart, B.J. and Mennella, J.A. (1994) Springer-Verlag, Berlin, pp.313^338. Infant salt taste: developmental, methodological, and Cardenete, G., Garzo¨ n, A., Moyano, F. and de la Higuera, M. contextual factors. Developmental Psychology 27, (1993) Nutritive utilization of earthworm protein by ¢n- 353^365. gerling rainbow trout (Oncorthynchus mykiss). In: Fish Beauchamp, G.K., Cowart, B.J. and Moran, M. (1986) Devel- NutritioninPractice. (eds S.J. Kaushik and P. Luquet). opmental changes in salt acceptability in human infants. INRA Editions, Paris, pp.923^926. Developmental Psychology19,75^83. Carpenter, R.C. (1986) Partitioning herbivory and its e¡ects Bernache, L. (2001)The evolutionary history of brown trout on coral reef algal communities. Ecological Monographs (Salmo trutta L.) inferred from phylogeographic, nested 56,345^365. clade, and mismatch analyses of mitochondrial DNAvar- Carr,W.E.S. (1976) Chemoreception and feeding behaviour iation. Evolution 55,351^379. inthe pig¢sh, Orthopristis chrysopterus: characterization Boikova, O.S. (1986) Feeding of ¢sh in Lake Glubokoe. Hydro- and identi¢cation of stimulatory substances in a shrimp biologia141, 95 ^111. extract. Comparative Biochemistry and Physiology 55A, Brand, J.G. and Bruch, R.C. (1992) Molecular mechanisms of 153^157. chemosensory transduction: gustation and olfaction. Carr, W.E.S. (1982) Chemical stimulation of feeding beha- In: Fish Chemoreception (ed. T.J. Hara). Chapman & Hall, viour. In: Chemoreception in Fishes.(ed.T.J.Hara).Elsevier London, pp.126^149. Scienti¢c Publishing Co., Amsterdam, pp.259^273. Branson, B.A. (1966) Histological observations on the Carr, W.E.S., Blumenthal, K.M. and Netherton, J.C. III sturgeon chub, Hybopsis gelida (Cyprinidae). Copeia 4, (1977) Chemoreception in the pig¢sh, Orthopristis 872^876. chrysopterus: the contribution of amino acids and Bridcut, E.E. and Giller, P.S. (1995) Diet variability and fora- betaine to stimulation of feeding behaviour by various ging strategies in brown trout (Salmo trutta): an analysis extracts. Comparative Biochemistry and Physiology 58A, from subpopulations to individuals. Canadian Journal of 69^73. Fisheries and Aquatic Sciences 52,2543^2552. Carr,W.E.S. and Chaney,T.B. (1976) Chemical stimulation of Brown, S.B., Evans, R.E.,Thompson, B.E. and Hara,T.J. (1982) feeding behaviour in the pin¢sh, Lagodon rhomboides: Chemoreception and aquatic pollutants. In: Chemorecep- characterization and identi¢cation of stimulatory sub- tion in Fishes (ed. T.J. Hara). Elsevier, Amsterdam, pp. stances extracted from shrimp. Comparative Biochemis- 363^393. try and Physiology 54A,437^441. Bryan, J.E. and Larkin, P.A. (1972) Food specialization by Carr,W.E.S., Netherton, J.C. III, Gleeson, R.A. and Derby,C.D. individual trout. Journal of the Fisheries Research Board of (1996) Stimulants of feeding behavior in ¢sh: analyses of Canada 29,1615^1624. tissues of diverse marine organisms. Biological Bulletin Bryan, P.J., McClintock, J.B. and Hopkins, T.S. (1997) Sruc- 190,149^160. tural and chemical defenses of echinoderms from the Cembella, A.D.,Therriault, J.-C. and Beland, P.(1988)Toxicity nothern Gulf of Mexico. Journal of Experimental Marine of cultured isolates and natural populations of Protogo- Biology and Ecology 210, 173^186. nyaulax tamarensis from the St. Lawrence estuary. Jour- Cantoni, G.L. and Anderson, D.G. (1956) Enzymatic cleavage nalofShell¢shRessearch7,611^621. of dimethylpropiothetin by Polysiphonia lanosa. Journal Challenger, F.and Simpson, M.I. (1948) Studies on biological of Biological Chemistry 222,171^177. methylation. Part XII. A precursor of the dimethylnsul- Caprio, J. (1975) High sensitivity of cat¢sh taste receptors to phide evolved by Polysiphonia fastigata. Dimethyl-2-car- amino acids. Comparative Biochemistry and Physiology boxyethylsulphonium hydroxide and its salts. Journal of 52A,247^251. Chemical Society 30,1591^1597. Caprio, J. (1978) Olfactionand taste inthe channel cat¢sh: an Channas,B.,Pawlik,J.R.,Lindel,T.andFenical,W.(1996) electrophysiological study of the responses to amino Chemical defense of the Carribbean sponge Agelas cla- acids and derivates. Journal of Comparative Physiology throdes (Schmidt). Journal of Experimental MarineBiology 123, 357^371. and Ecology 208,185^196. Caprio, J. (1982) High sensitivity and speci¢city of olfactory Colgan, P.W., Brown, J.A. and Orsatti, S.D. (1986) Role of diet and gustatory receptors of cat¢shtoaminoacids. In: Che- and experience in the development of feeding behaviour moreception in Fishes.(ed.T.J.Hara).ElsevierScienti¢c in largemouth bass, Micropterussalmoides. Journal ofFish Publishing Co., Amsterdam, pp.109^134. Biology 28,161^170.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 337 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Coughlin, D.J. and Strickler, J.R. (1990) Zooplankton capture Farr, J.A. and Herrnkind,W.F. (1974) A quantitative analysis by a coral reef ¢sh: an adaptive response to evasive prey. of social interaction of the guppy, Poecilia reticulata Environmental Biology of Fishes 29,35^42. (Pisces: Poeciliidae) as function of population density. Cronin, G., Paul, V.J., Hay, M.E. and Fenical, W. (1997) Are Animal Behaviour 22,582^591. tropical herbivores more resistant than temperate herbi- Fenical, W. and Pawlik, J.R. (1991) Defensive properties of vores to seaweed chemical defenses? Ditepenoid metabo- secondary metabolites from the Carribbean gorgonian lites from Dictyota acutiloba as feeding deterrents for coral Erythropodium caribaeorum. Marine Ecology Pro- tropical versus temperate ¢shes and urchins. Journal of gress in Series 75,1^8. Chemical Ecology 23, 289^302. Finger,T.E. (1976) Gustatory pathways in the bullhead cat- Crossland, M.R. (2001)Abilityof predatory nativeAustralian ¢sh. Part I. Connections of the anterior ganglion. Journal ¢shes to learn to avoid toxic larvae of the intro- of Comparative Neurology165, 513^526. duced toad, Bufo marinus. Journal of Fish Biology 59, Finger, T.E., Drake, S.K., Kotrschal, K., Womble, M. and 319^329. Dockstader, K.C. (1991) Postlarval growth of the peri- Dabrowski, K. and Rusiecki, M. (1983) Content of total and pheral gustatory system in the channel cat¢sh, Icta- free amino acids in zooplanktonic food of ¢sh larvae. lurus punctatus. Journal of Comparative Neurology 31 4, Aquaculture 30,31^42. 55^66. Davenport, C.J. and Caprio, J. (1982) Taste and tactile record- Finger, T.E. and Morita, Y. (1985) Two gustatory systems: ing from the ramus recurrens facialis innervating £ank facial and vagal gustatory nuclei have di¡erent brain- taste buds in the cat¢sh. Journal of Comparative Physiol- stem connections. Science 227, 776^778. ogy147A,217^229. Fischer, Z. (1973) The elements of energy balance in grass De la Noue, J. and Choubert, G. (1985) Apparent digestibility carp (Ctenopharyngodon idellaVal.). Part IV.Consumption of invertebrate biomasses by rainbow trout. Aquaculture rate of grass carp fed on di¡erent types of food. Polish 50,103^112. Archives of Hydrobiology 20, 309^318. Devitsina, G.V. and Kazhlaev, A.A. (1993) Development of Flerova, G.I. and Gdovskii, P.A. (1976) The rate of excitement chemosensory organs in Siberian sturgeon, Acipenser propagation in olfactory nerve ¢bres and responses in baeri, and stellate sturgeon, A. stellatus. Journal of the smell brain of ¢shes under varied temperature condi- Ichthyology 33,9^19. tions.Voprosi Ikhtiologii16,119^125 (in Russian). Devitsina, G.V. and Kazhlaev, A.A. (1995) Chemosensory Formaker, B.K. and Hill, D.L. (1990) Alterations of salt taste systems and their heterohronic morphogenesis in perception in the developing rats. Behavioral Neuro- sturgeons. Bio¢zika 40,146^150. science104,356^364. Dias, J., Gomes, E.F.and Kaushik, S.J. (1997) Improvement of Fowler, K.L., Pauley,G.B. andThomas, G.L. (1978) The e¡ects feed intake through supplementation with an attractant of the food preferences and stocking rates of grass carp mix in European seabass fed plant-protein rich diets. (Ctenopharyngodon idella Val.) on mixed plant commu- Aquatic Living Resources10, 385^389. nities. Aqautic Botany 5,261^276. DÖving, K.B. (1986) Functional properties of the ¢sh olfac- Franco, J., Johnstone, A.D.F.and Mackie, A.M. (1991) Studies tory system. In: Progress in Sensory Physiology, Vol. 6. of bait preference in the cod Gadus morhua L. Characteri- SpringerVerlag, Berlin e.a, pp.39^104. sation of feeding stimulants using an operant condition- Du¡y, J.E. and Paul, V.J. (1992) Prey nutritional quality and ing technique. Fisheries Research10,229^242. the e¡ectiveness of chemical defenses against tropical Freitas, J.C., Rantin, F.T., Nonaka, K.O. and Almeida, F.E.X. reef. Oecologia 90, 333^339. (1992) Bradycardia in groupers (Epinephelus guaza) Dussault, G.V. and Kramer, D.L. (1981) Food and feeding induced by taste stimuli from pu¡er ¢sh (Sphaeroides behavior of the guppy, Poecilia reticulata (Pisces: Poecilii- spengleri) skin secretion. In: RecentAdvancesinToxicology dae). CanadianJournal of Zoology 59,684^701. Research. Bradycardia in Groupers (Epinephelus Guaza) Ellingsen, O.F. and DÖving, K.B. (1986) Chemical fractiona- Induced byTaste Stimuli from Pu¡er Fish (Sphaeroides Spen- tion of shrimp extracts inducing bottom food search gleri) Skin Secretion,Vol. 2 (eds P.Gopalakrishnakone and behavior in cod (Gadus morhua L.). Journal of Chemical C.K. Tan). National University of Singapore, Singapore, Ecology12,155^168. pp.465^475. Epifanio, R.D.A., Gabriel, R., Martins, D.L. and Muricy, G. Fryer, G. and Iles, T.D. (1972) The Cichlid Fishes of the Great (1999a) The sesterpene variabilin as a ¢sh-predation Lakes of Africa: the Biologyand Evolution. Oliver and Boyd, deterrent in the Western Atlantic sponge Ircinia strobi- Edinburgh. lina. JournalofChemicalEcology25, 2247^2254. Fuke, S., Konosu, S. and Ina, K. (1981) Identi¢cation of feed- Epifanio, R.D.A., Martins, D.L., Villaca, R. and Gabriel, R. ing stimulants for red sea bream in the extract of marine (1999b) Chemical defences against ¢sh predation in worm Perineries brevicirrus. Bulletin of the Japanese three Brazilian octacorals: 11b,12b-epoxypukalide as Society of Scienti¢c Fisheries 47, 1631^1635. a feeding deterrent in Phyllogorgia dilatuta. Journal of Gerhart, D.J., Bondura, M.E. and Commito, J.A. (1991)Inhibi- Chemical Ecology 25,2255^2265. tion of sun¢sh feeding by defensive steroids from aquatic

338 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

beetles: structure-activity relationships. Journal of Che- Hay, M.E. (1991) Fish^seaweed interactions on coral reefs: mical Ecology17, 1363^1370. e¡ects of herbivorous ¢shes and adaptations of their prey. Gladstone,W. (1987) The eggs and larvae of the sharpnose In: The Ecology of Fishes on Coral Reefs.(ed.P.F.Sale). pu¡er¢sh Canthigaster valentini (Pisces,Tetraodontidae) Academic Press, San Diego, pp.96^119. are unpalatable to other reef ¢shes. Copeia1,227^230. Hayama,T. and Caprio, J. (1989) Lobule structure and soma- Glaser, D. (1966) Untersuchungen u« ber die absoluten Gesch- totopic organization of the medullary facial lobe in the macksschwellen von Fischen. Zeitschrift fu« rVergleichende channel cat¢sh, Ictalurus punctatus. Journal of Compara- Physiologie 52, 1^25. tive Neurology 285,9^17. Goh,Y. and Tamura,T. (1980a) E¡ect of amino acids on the Hayama,T. and Caprio, J. (1990) Somatotopical organization feeding behaviour in red sea bream. Comparative Bio- of the intermediate nucleus of the facial lobe in the chan- chemistry and Physiology 66C,225^229. nel cat¢sh, Ictalurus punctatus. Brain Research 516,137^ Goh, Y. and Tamura, T. (1980b) Olfactory and gustatory 140. responses to amino acids in two marine teleosts ^ red Heine, J.N., McClintock, J.B., Slattery,M. andWeston, J. (1991) seabreamand mullet. Comparative BiochemistryandPhy- Energetic composition, biomass and chemical defense siology 66C, 217^224. in the common antarctic nemertean Parborlasia corruga- Gomahr, A., Palzenberger, M. and Kotrschal, K. (1992) Den- tus McIntosh. Journal of Experimental Marine Biology and sity and distribution of external taste buds in cyprinids. Ecology153,15^25. Environmental Biology of Fishes 33,125^134. Heinsbroek, L.T.N. and Kreuger, J.G. (1992) Feeding and Gosselin, S., Fortier, L. and Gagne, J.A. (1989) Vulnerability growth of glass eels, Anguilla anguilla L. The e¡ect of of marine ¢sh larvae to the toxic dini£agellate Protogo- feeding stimulants on feed intake, energy metabolism nyaulax tamarensis. Marine Ecology Progress in Series 57, and growth. Aquaculture and Fisheries Management 23, 1^10. 327^336. Grobecker, D.B. and Pietsch,T.W. (1979) High-speed cinema- Herrick, C.J. (1901) The cranial nerves and cutaneous sense tographic evidence for ultrafast feeding in Atennariid organs of the North American silurid ¢shes. Journal of angler¢shes. Science 205,1161^1162. Comparative Neurology and Psychology11,177^249. Hansen, A., Reutter,K. and Zeiske, E. (2002) Taste bud devel- Hidaka, I. (1982) Taste receptor stimulation and feeding opment in the zebra¢sh, Danio rerio. Developmental behavior in the pu¡er. In: Chemoreception in Fishes.(ed. Dynamics 223,483^496. T.J.Hara). Elsevier Scienti¢c Publishing Co.,Amsterdam, Hara, T.J. (1994a) The diversity of chemical stimulation of pp. 243^257. ¢sh olfaction and gustation. Reviews in Fish Biology and Hidaka, I. and Ishida,Y.(1985) Gustatory response inthe Shi- Fisheries 4,1^35. maisaki (tiger¢sh) Therapon oxyrhynchus. Bulletin of the Hara,T.J. (1994b) Olfaction and gustation: an overview. Acta Japanese Society of Scienti¢c Fisheries 51,387^391. Physiologica Scandinavica152,207^217. Hidaka, I., Kohbara, J., Araki,T., Morishita,T., Miyajima,T., Hara,T.J., Carolsfeld, J. and Kitamura, S. (1999) The variabil- Shimizu, S. and Kuriyama, I. (2000) Identi¢cation of ity of the gustatory sensibility in salmonids, with special feeding stimulants from a jack mackerel muscle extract reference to strain di¡erences in rainbow trout, Oncor- for young yellowtail Seriola quinqueradiata. Aquaculture hynchus mykiss. CanadianJournal of Fisheries and Aquatic 181,115^126. Sciences 56,13^24. Hidaka, I., Kyiohara, S. and Oda, S. (1977) Gustatory Hara,T.J. and Marui,T. (1984) Multiplicity of taste receptors response in the pu¡er. Part III. Stimulatory e¡ectiveness for amino acids in rainbow trout: evidence from cross- of nucleotides and their derivatives. Bulletin of the Japa- adaptation and kinetic analysis. Chemical Senses 8,250. nese Society of Scienti¢c Fisheries 43, 423^428. Hara,T.J., Sveinsson,T., Evans, R.E. and Klaprat, D.A. (1993) Hidaka, I., Nyu, N. and Kiyohara, S. (1976) Gustatory Morphological and functional characteristics of the response in the pu¡er. Part IV. E¡ects of mixtures of olfactory and gustatory organs of three Salvelinus spe- amino acids and betaine. Bulletin of the Faculty of Fish- cies. CanadianJournal of Zoology 71,414^423. eries, Mie University 3,17^28. Harada, K. (1985) Feeding attraction activities of amino Hidaka, I., Ohsugi,T.and Kubomatsu,T.(1978) Tastereceptor acids and nitrogeneous bases for oriental weather¢sh. stimulation and feeding behaviour in the pu¡er Fugu Bulletin of the Japanese Society of Scienti¢c Fisheries 51, pardalis. Part I. E¡ect of single chemicals. Chemical Sense 461^466. and Flavor 3, 341^354. Harborne, J.B. (1993) Introduction to Ecological Biochemistry, de la Higuera, M. (2001) E¡ects of nutritional factors and 4th edn. Academic Press, London. feed characteristics on feed intake. In: Food Intake inFish. Hashimoto, Y., Konosu, S., Fusetani, N. and Nose, T. (1968) (eds D. Houlihan, T. Boujard and M. Jobling). Blackwell Attractants for eels in the extracts of short-necked clam. Science, Oxford, pp.250^268. Part I. Survey of constituents eliciting feeding behaviour Hoagland, H. (1933) Speci¢c nerve impulses from gustatory by the omission test. Bulletin of the Japanese Society of and tactile receptors in cat¢sh. Journal of GeneralPhysiol- Scienti¢c Fisheries 34,78^83. ogy16,685^693.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 339 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Hoese, H.D. and Hoese, D. (1967) Studies onthe biologyof the amount. Bulletin of the Japanese Society of Scienti¢c Fish- feeding reaction in Gobiosoma bosci. Tulane Studies in eries 34,785^791. Zoology14,55^62. Ivlev,V.S. (1961) Experimental Ecology of the Feeding of Fishes. Holm, J.C. andWalter,B. (1988) Free amino acids in live fresh- Yale University Press, New Haven. water zooplankton and dry feed: possible importance for Jakubowski, M. (1983) New details of the ultrastructure ¢rst feeding in Atlantic salmon fry (Salmosalar). Aquacul- (TEM, SEM) of taste buds in ¢shes. Zeitschrift fu« rMikros- ture 71, 341^354. kopish-Anatomische Forschung 97,849^862. Hoogerhoud, R.J.C. (1989) Second International Symposium Jakubowski, M. and Whitear, M. (1990) Comparative mor- Vertebr. Morphol. Stuttgart, pp.19^21. phology and cytology of taste buds in teleosts. Zeitschrift Horppila, J., Ruuhijarvi, J., Rask, M., Karppinen, C., Nyberg, fu« r Mikroskopish-Anatomische Forschung104,529^560. K. and Olin, M. (2000) Seasonal changes in the diets and Jobling, M. (1994) Fish Bioenergetics. Chapman & Hall, relative abundances of perch and roach in the littoral London. and pelagic zones of a large lake. Journal of Fish Biology Jobling, M., Arnesen, A.M., Baardvik, B.M., Christiansen, 56, 51^72. J.S. and JÖrgensen, E.H. (1995) Monitoring feeding beha- Hughes, S.G. (1991)Response of ¢rst-feeding spring chinook viour and food intake: methods and applications. Aqua- salmon to four potential chemical modi¢ers of feed culture Nutrition1, 131^143. intake. Progressive Fish-Culturist 53,15^17. Jobling, M., Gomes, E. and Dias, J. (2001) Feed types, manu- Hughes, S.G. (1993) Single-feeding response of chinook sal- facture and ingredients. In: Food Intake in Fish.(edsD. mon fry to potential feed intake modi¢ers. Progressive Houlihan,T. Boujard and M. Jobling). Blackwell Science, Fish-Culturist 55,40^42. Oxford, pp.25^48. Hunter, J.R. (1980) The feeding behaviour and ecology Johnsen, P.B. and Adams, M.F. (1986) Chemical feeding sti- of marine ¢sh larvae. In: Fish behaviour and its use in mulants for the herbivorous ¢sh, Tilapia zillii. Compara- the capture and calture of ¢shes. International Center for tive Biochemistry and Physiology 83A,109^112. Living Aquatic Resources Management, Manila, pp. Johnstone, A.D.F.and Mackie, A.M. (1990) Laboratory inves- 287^330. tigations of bait acceptance by the cod Gadus morhua L. Hustvedt, S.O., Storebakken,T. and Salte, R. (1991) Does oral Identi¢cation of feeding stimulants. Fisheries Research 9, administration of oxolinic acid or tetracycline a¡ect feed 219^230. intake of rainbow trout? Aquaculture 92,109^113. Jones, K.A. (1980) Temperature dependent attraction by Iida, H., Nakamura, K. andTokunaga,T.(1985) Dimethyl sul- gold¢sh to a chemical feeding cue presented alone and ¢de and dimethyl-b-propiothetin in sea algae. Bulletin of in combination with heated water. Physiology and Beha- the Japanese Society of Scienti¢c Fisheries 51,1145^1150 vior 33,509^515. (in Japanese). Jones, K.A. (1983) E¡ects of temperature on the activity and Iida, H., Nakamura, K. and Tokunaga,T. (1992) Non-protein orientation of gold¢sh (Carassius auratus). Comparative nitrogenous compounds in muscle extract of oceanic Biochemistry and Physiology 74A, 777^780. cephalopods. Nippon Suisan Gakkaishi 58,2383^2390. Jones, K.A. (1989) The palatabilityof amino acids and related Iida, H. and Tokunaga, T. (1986) Dimethyl sul¢de and compounds to rainbow trout, Salmo gairdneri Richard- dimethyl-b-propiothetin in shell¢sh. Bulletin of the Japa- son. Journal of Fish Biology 34,149^160. nese Society of Scienti¢c Fisheries 52, 557^563. Jones, K.A. (1990) Chemical requirements of feeding in Ikeda, I., Hosokawa, H., Shimeno, S. and Takeda, M. (1988a) rainbow trout Oncorhynchus mykiss (Walbaum): palat- Identi¢cation of feeding stimulant for jack mackerel in ability studies on amino acids, amides, amines, alcohols, its muscle extract. Bulletin of theJapanese Society of Scien- aldehydes, saccharides, and other compounds. Journal of ti¢c Fisheries 54,229^233. Fish Biology 37,413^423. Ikeda,I.,Hosokawa,H.,Shimeno,S.andTakeda,M.(1988b) Kaku,T.,Tsumagari, M., Kiyohara, S. and Yamashita (1980) Identi¢cation of feeding stimulant in the krill extract for Gustatory responses inthe minnow, Pseudorasboraparva. jack mackerel. Bulletin of theJapanese Society of Scienti¢c Physiology and Behavior 25,99^105. Fisheries 54,235^238. Kamstra, A. and Heinsbroek, L.T.N. (1991) E¡ects of attrac- Ikeda, I., Hosokawa, H., Shimeno, S. and Takeda, M. (1991) tants on start of feeding of glass eel, Anguilla anguilla L. Feeding stimulant activity of nucleotides, tryptophan Aquacultureand Fisheries Management 22, 47^56. and their related compounds for jack mackerel. Nippon Kanwal, J.S. and Caprio, J. (1983) An electriphysiological Suisan Gakkashi 57, 1539^1542. investigation of the oropharyngeal (IX^X) taste system Ishida,Y. and Hidaka, I. (1987) Gustatory response pro¢les in the channel cat¢sh, Ictalurus punctatus. Journal of for amino acids, glycinebetaine and nucliotides in sev- Comparative Physiology150 A, 345^357. eral marine teleosts. Bulletin of the Japanese Society of Kanwal, J.S. and Caprio, J. (1988) Overlapping taste and tac- Scienti¢c Fisheries 53,1391^1398. tile maps of the oropharynx inthe vagal lobe of the chan- Ishiwata, N. (1968) Ecological studies on the feeding of nel cat¢sh, Ictalurus punctatus. Journal of Neurobiology ¢shes. Part VI. External factors a¡ecting satiation 19, 211^222.

340 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Kanwal, J.S. and Finger, T.E. (1992) Central representa- Kasumyan, A.O. and Nikolaeva, E.V.(1997) Taste preferences tion and projections of gustatory systems. In: Fish of Poecilia reticulata (Cyprinidontiformes). Journal of Chemoreception. (ed. T.J. Hara). Chapman & Hall, London, Ichthyology 37,662^669. pp.79^102. Kasumyan, A.O. and Nikolaeva, E.V.(2002) The comparative Kapoor, B.G., Evans, H.E. and Pevzner, R.A. (1975) The analysis of taste preferences in ¢shwith di¡erent ecology gustatory system in ¢sh. Advances in Marine Biology 13, and feeding. Journal of Ichthyology 42, 203^214. 53^108. Kasumyan, A.O. and Ponomarev, V.Y. (1986) Study of the Kasahara, K. and Nishibori, K. (1987) E¡ect of heating tem- behaviour of zebra¢sh, Brachidanio rerio,inresponseto perature on volatiles of roasted laver. Nippon Suisan Gak- natural chemical food stimuli. Journal of Ichthyology 26, kaishi 53,673^676 (in Japanese). 96^105. Kassil,V.G. (1972)Taste. In: Physiology of Sensory System,Vo l. Kasumyan, A.O. and Ponomarev,V.Y. (1989) Role of Chemore- 2.Nauka,Leningrad,pp.562^606(inRussian). ception in Feeding Behavior of Cyprinids. Chemosensitivity Kasumyan, A.O. (1992) In: Chemical Signals inVertebratesVI and Chemocommunication in Fishes. Nauka, Moscow, pp. (eds R.L. Doty and D. Muller-Schwarze). Plenum Press, 123^131(in Russian). NewYork, Philadelphia, pp.135^139. Kasumyan, A.O. and Prokopova, O.M. (2001) Taste prefer- Kasumyan, A.O. (1994) Olfactory sensitivity of the sturgeon ences and the dynamics of behavioral taste response in to free amino acids. Biophysics 39, 519^522. the tench,Tinca tinca (Cyprinidae). Journal of Ichthyology Kasumyan, A.O. (1997) Gustatory reception and feeding 41,640^653. behavior in ¢sh. Journal of Ichthyology 37,72^86. Kasumyan, A.O. and Sidorov, S.S. (1993a) Behavioral Kasumyan, A.O. (1999a) Olfaction and taste in sturgeon responses of caspian brown trout parr, Salmo trutta behaviour. Journal of Applied Ichthyology15,228^232. caspius Kessler, to the main types of taste substances. Kasumyan, A.O. (1999b) Olfaction, extraoral and oral taste Moscow University Biological Sciences Bulletin 48, senses have di¡erent level of stability in ¢sh phylogeny. 48^53. Chemical Senses 24,530. Kasumyan, A.O. and Sidorov,S.S. (1993b) Taste sensitivityof Kasumyan, A.O. (2000) Individual variabilityof taste prefer- chum salmon, Oncorhynchus keta, to types of taste sti- ences and of behavioral response to taste stimuli in carp muli and amino acids. Sensory Systems 6, 222^225. Cyprinus carpio. Journal of Ichthyology 40,661^669. Kasumyan, A.O. and Sidorov, S.S. (1994) Comparison of Kasumyan, A.O. and Devitsina, G.V.(1997) The e¡ect of olfac- intra- and extra-oral taste responses to free amino acids tory deprivation on chemosensory sensitivity and the in three sturgeon species of the genus Acipenser.Biophy- state of taste receptors of Acipenserids. Journal of sics 39, 523^526. Ichthyology 39, 786^798. Kasumyan, A.O. and Sidorov, S.S. (1995a) A comparative Kasumyan, A.O. and Kazhlaev, A.A. (1993a) Behavioral analysis of the taste responses of young salmon trout, responses of early juveniles of Siberian sturgeon, Acipen- Salmo trutta trutta, from populations of the Baltic and ser baeri, and stellate sturgeon A. stellatus (Acipenseri- White seas. Doklady Biological Sciences 343,289^391. dae) to gustatory stimulating substances. Journal of Kasumyan, A.O. and Sidorov, S.S. (1995b) Environmental Ichthyology 33,85^97. Science and Vulnerable Ecosystems. 5th SETAC-Europe Kasumyan, A.O. and Kazhlaev, A.A. (1993b) Formation of Congress, Copenhagen,284 pp. searching behavioral reaction and olfactory sensitivity Kasumyan, A.O. and Sidorov, S.S. (1995c) Gustatory proper- to food chemical signals during ontogeny of sturgeons ties of free amino acids in Caspian trout, Salmo trutta cas- (Acipenseridae). Journal of Ichthyology 33,51^65. pius. Journal of Ichthyology 35,8^20. Kasumyan, A.O. and Morsy, A.M.H. (1996) Taste sensitivity Kasumyan, A.O. and Sidorov, S.S. (1995d) The palatability of of common carp Cyprinus carpio to free amino acids and free amino acids and classical taste substances to frolich classical taste substances. Journal of Ichthyology 36, charr, Salvelinus alpinus erythrinus (Georgi). Nordic Jour- 391^ 403. nal of Freshwater Research 71,320^323. Kasumyan, A.O. and Morsy, A.M.H. (1997) Taste prefer- Kasumyan,A.O.andSidorov,S.S.(1998)Tastepreferencesin ence for classic taste substances in juveniles of the young Arctic char, Salvelinus alpinus, brook trout, S. fon- grass carp Ctenopharyngodon idella (Cyprinidae, Pisces) tinalis, and lake char, S. namaycush.In:Biology andEvolu- reared on various diets. Doklady Biological Sciences 357, tion of Chars of the Nothern Hemisphere,18pp. 284^286. Kasumyan, A.O., Sidorov, S.S., Kazhlaev, A.A. and Pash- Kasumyan, A.O. and Morsy, A.M.H. (1998a) E¡ect of heavy chenko, N.I. (1995) In: 2nd International Symposium on metals on the feeding activityand taste responses of carp Sturgeons (eds A.D. Ershanovich and T.I.J. Smith).VNIRO Cyprinus carpio. Copper, cadmium, zinc and lead. Journal Publishing, Moscow,pp.278^288. of Ichthyology 38,393^409. Kasumyan, A.O., Sidorov, S.S. and Pashchenko, N.I. (1993) Kasumyan, A.O. and Morsy,A.M.H. (1998b) VIII InternSym- E¡ect of water temperatures on taste sensitivity of fry of posium on Fish Physiology. Uppsala University, Uppsala, the stellate sturgeon Acipenser stellatus to free amino Sweden,101pp. acids. Doklady Biological Sciences 331,265^267.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 341 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Kasumyan, A.O., Sidorov, S.S., Pashchenko, N.I. and Nem- rus. Bulletin of the Japanese Society of Scienti¢c Fisheries chinov, A.V. (1992) Extra- and intraoral sensitivity to 32,881^886. amino acids in young sturgeon, Acipenser gueldenstaedti. Konosu, S., Fujimoto, K., Takashima, Y., Matsushita, T. and Doklady Biological Sciences 322, 82^84. Hashimoto, Y. (1965) Constituents of the extracts and Kasumyan, A.O. and Tau¢k, L.R. (1994) Behavioral reaction amino acids composition of the protein of short-necked of juvenile sturgeons (Acipenseridae) to amino acids. clam. Bulletin of theJapanese Society of Scienti¢c Fisheries Journal of Ichthyology 34,90^103. 31, 680^686. Kats, L.B., Petranka, J.W. and Sih, A. (1988) Antipredator Konosu, S. and Hayashi,T.(1975) Determinationof b-alanine defenses and the persistence of amphibian larvae with betaine and glycine betaine in some marine inverte- ¢shes. Ecology 69,1865^1870. brates. Bulletin of theJapanese Society of Scienti¢cFisheries Kawamura, G. and Washiyama, N. (1989) Ontogenetic 41,743^746. changes in behavior and sense organ morphogenesis in Kotrschal, K. (1991) Solitary chemosensory cells ^ taste, largemouth bass and Tilapia nilotica. Transactions of the common chemical sense or what? Review in Fish Biology American Fisheries Society118,203^213. and Fisheries1, 3^22. Kerfoot,W.C., Kellog, D.L. Jr and Strickler, J.R. (1980) Visual Kotrschal, K. (1992) Quantitative electron microscopy of observations of live zooplankters: evasion, escape, and solitary chemoreceptor cells in cyprinids and other tele- chemical defenses. In: Evolution and Ecology of Zooplnak- osts. Environmental Biology of Fishes 35,273^282. ton Communities. (ed. W.C. Kerfoot). University Press of Kotrschal, K. (1995) Ecomorphology of solitary chemosen- New England, New Hampshire, pp.0^27. sory cell systems in ¢sh: a review. Environmental Biology Kerr, J.N.Q. and Paul,V.J. (1995) Animal^plant defense asso- of Fishes 44, 143^155. ciation: the soft coral Sinularia sp. (Chidaria,Alcyonacea) Kotrschal, K. (1996) Solitarychemosensorycells: whydo pri- protects Halimeda spp. from herbivory. Journal of mary aquatic vertebrates need another taste system? Experimental Marine Biology and Ecology186,183^205. Trends in Ecology and Evolution11, 110 ^114. Kinnamon, J.C. (1987) Organization and innervation of taste Kotrschal, K. (2000) Taste(s) and olfaction(s) in ¢sh: a buds. In: Neurobiology of Taste and Smell. (eds T.E. Finger review of spezialized sub-systems and central integra- andW.L.Silver).Wiley,NewYork,pp.277^297. tion. P£u« gerArchiv ^ European Journal of Physiology 439 Kinnel, R.C., Dieter, R.K., Meinwald, J. et al. (1979) Brasile- (Suppl.), R178^R180. nyne and cys-dihydrorhopophytin: antofeedant med- Kotrschal, K. and Finger,T.E. (1996) Secondary connection ium-ring haloethers from a sea-hare (Aplysia brasiliana). of the dorsal and ventral facial lobes in a teleost ¢sh, the Proceedings of the National Academy of Science USA 76, rockling (Ciliata mustela). Journal of Comparative Neurol- 3576 ^3579. ogy 370, 415^ 426. Kiyohara, S.,Yamashita, S. and Harada, S. (1981) High sensi- Kotrschal, K., Van Staaden, M.J. and Huber, R. (1998) Fish tivity of minnow gustatory receptors to amino acids. brains: evolution and environmental relationships. Physiology and Behavior 26, 1103 ^1108. Review in Fish Biology and Fisheries 8,1^36. Klaprat, D.A., Evans, R.E. and Hara,T.J. (1992) Environmen- Kotrschal, K.,Whitear, M. and Finger,T.E. (1993) Spinal and tal contaminants and chemoreception in ¢shes. In: Fish facial innervation of the skin in the gadid ¢sh, Ciliata Chemoreception. (ed. T.J. Hara). Chapman & Hall, London, mustela (Teleostei). Journal of Comparative Neurology 331, pp.321^341. 407^417. Knights, B. (1985) Feeding behaviour and ¢sh culture. Krinner, M. (1934) Uber die Geschmacksemp¢ndlichkeit In: Nutrition and Feeding in Fish. (eds C.B. Cowey, der Ellritzen. Zeitschrift fu« rVergleichende Physiologie 21, A.M. Mackie and J.G. Bell). Academic Press, London, 317^242. pp. 223^241. Kruse, K.C. and Francis, M.G. (1977) A predation deterrent in Komak, S. and Crossland, M.R. (2000) An assessment of the larvae of the bullfrog, Rana catesbeiana. Transections of introduced mosquito ¢sh (Gambusia a¤nis holbrooki)as theAmerican Fisheries Society106,248^252. a predator of eggs, hatchlings and tadpoles of native and Kruzhalov, N.B. (1986) Behavioral responses of carp ¢nger- non-native anurans.Wildlife Research 27,185^189. lings Cyprinus carpio L. to amino acids.VoprosiiIkhtiologii Konishi, J., Hidaka, I., Toyota, M. and Matsuda, H. (1969) 26, 1016^1022 (in Russian). High sensitivity of the palatal chemoreceptors of carp Kubitza, F., Lovshin, L.L. and Lovell, R.T. (1997) Identi¢ca- to carbon dioxide. Japanese Journal of Physiology 19, tion of feed enhancers for juvenile largemouth bass 327^341. Micropterus salmoides. Aquaculture148, 191^200. Konishi, J. and Zotterman, Y. (1961) Taste functions in the La Barre, S.C., Coll, J.C. and Sammarco, P.W. (1986) Competi- carp: an electrophysiological study on gustatory ¢bres. tive strategies of soft corals (Coelenterata: Octocorallia). Acta Physiologica Scandinavica 52, 150 ^161. Part III. Spacing and agressive interactions between Konosu, S., Cheng,Y.-N. and Hashimoto,Y. (1966) Constitu- alcyonaceans. Marine Ecology Progress in Series 28, ents of the extracts of a marine worm, Perinereis brevicir- 147^156.

342 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Lamb, C. and Finger, T.E. (1995) Gustatory control of feed- Marui,T. and Caprio, J. (1992) Teleost gustation. In: Fish Che- ing behavior in gold¢sh. Physiology and Behaviour 57, moreception. (ed. T.J. Hara). Chapman & Hall, London, 483^488. pp. 171^198. Larher, T., Hamelin, J. and Stewart, G.R. (1977) L'acide Marui, T., Caprio, J., Kiyohara, S. and Kasahara, Y. (1988) dimethylsulfonium-3 propanoque de Spartina anglica. Topographical organization of taste and tactile neurons Phytochemistry16, 2019^2020. inthe facial lobe of the sea cat¢sh, Plotosuslineatus.Brain Lemm, C.A. (1983) Growth and survival of Atlantic salmon Research 446,178^182. fry on semimoist or dry starter diets. Progressive Fish- Marui,T., Evans, R.E., Zielinski, B. and Hara,T.J. (1983a) Gus- Culturist 45,72^74. tatory responses of the rainbow trout (Salmo gairdneri) Leyd ig, F. (1851) U« ber die aussere Haut einiger Su« sswasser- palate to amino acids and derivates. Journal of Compara- ¢sche. Zeitschrift fu« rWissenschaftliche Zoologie 3,1^12. tive Physiology153 A,423^433. Liem, K.F.(1973) Evolutionary strategies and morphological Marui,T., Harada, S. and Kasahara,Y.(1983b) Gustatory spe- innovations: cichlid pharyngeal jaws. Systematic Zoology ci¢ty foraminoacids inthe facial taste systemof the carp, 22,425^441. Cyprinus carpio L. Journal of Comparative Physiology Lindstedt, K.J. (1971) Chemical control of feeding behaviour. 153 A,299^308. Comparative Biochemistry and Physiology 39A, 553^581. Marui,T.and Kiyohara, S. (1987) Structure-activity relation- Lokensgard, J., Smith, R.L., Eisner,T. and Meinwald, J. (1993) ships and response features for amino acids in ¢sh taste. Pregnanes from defensive glands of a belostomatid bug. Chemical Senses12, 265^275. Experientia 49,175^176. McBride, J.R., Idler, D.R., Jones, R.E. and Tomlinson, N. Lucas, J.S., Hart, R.J., Howden, M.E. and Salathe, R. (1979) (1962) Olfactory perception in juvenile salmon. Part I. Saponins in eggs and larvae of Acanthaster planci (L.) Observation on response of juvenile sockeye to extracts (Asterioidea) as chemical defenses against planktivorous of food. Journal of the Fisheries Research Board of Canada ¢sh. Journal of Experimental Marine Biology and Ecology 19,327^334. 49,155^165. McClintock, J.B. andJanssen, J.(1990) Pteropod abductionas Mackie, A.M. (1982) Identi¢cation of the gustatory feed- a chemical defense in a pelagic Antarctic amphipod. Nat- ing stimulants. In: Chemoreception in Fishes.(ed.T.J. ure 346,462^464. Hara). Elsevier Scienti¢c Publishing Co., Amsterdam, McClintock, J.B., Swenson, D.P., Steinberg, D.K. and pp. 275^291. Michaels, A.A. (1996) Feeding-deterrent properties of Mackie, A.M. (1987) Prelimenary studies on the chemical common oceanic holoplankton from Bermudian waters. defenses of the British octocorals, Alcyonium digitatum, Limnology and Oceanography 41,798^801. and Pennatula phosphorea. Comparative Biochemistry and Mearns, K.J. (1986) Sensitivity of brown trout (Salmo trutta Physiology 86A,629^632. L.)and Atlantic salmon (Salmosalar L.)fry to amino acids Mackie, A.M. and Adron, J.W.(1978) Identi¢cation of inosine at the start of exogenous feeding. Aquaculture 55, and inosine-50-monophosphate as the gustatory feeding 191^200. stimulants for the turbot, Scophthalmus maximus. Com- Mearns, K.J., Ellingsen, O.F., DÖving, K.B. and Helmer, S. parative Biochemistry and Physiology 60A,79^83. (1987) Feeding behaviour in adult rainbow trout and Mackie,A.M.,Adron,J.W.andGrant,P.T.(1980)Chemical atlantic salmon parr, elicited by chemical fractions and nature of feeding stimulants for the juvenile Dover sole, mixtures of compounds identi¢ted in shrimp extract. Solea solea (L.). Journal of Fish Biology16,701^708. Aquaculture 64,47^63. Mackie, A.M. and Mitchell, A.I. (1982a) Chemical ecology Metcalfe, N.B., Huntingford, F.A. andThorpe, J.E. (1986) Sea- and chemoreception in the marine environment. Publi- sonal changes in feeding motivation of juvenile Atlantic cations Du Centre national pour l'exploitation des oce¨ans, salmon (Salmo salar). Canadian Journal of Zoology 64, Actes de colloques14,11^24. 2439^2446. Mackie, A.M. and Mitchell, A.I. (1982b) Further studies on Meyer, A. (1988) In£uence of age and size on the response to the chemical control of feeding behaviour in the Dover novel prey by ¢sh of the cichlid ¢sh Cichlasoma mana- sole, Soleasolea. Comparative Biochemistry andPhysiology guense (Pisces: Cichlidae). Ethology 78,199^208. 73A,89^93. Meyer,K.D. and Paul,V.J. (1995) Variation in secondary meta- Mackie, A.M. and Mitchell, A.I. (1983) Studies on the chemi- bolite and aragonite concentrations in the tropical green cal nature of feeding stimulants for the juvenile Eur- seaweed Neomeris annulata: e¡ects on herbivory by opean eel, Anguilla anguilla (L). Journal of Fish Biology ¢shes. Marine Biology122,537^545. 22,425^430. Michel, W. and Caprio, J. (1991) Responses of single facial Mackie, A.M. and Mitchell, A.I. (1985) Identi¢cation of gus- taste ¢bers in the sea cat¢sh, Arius felis, to amino acids. tatory feeding stimulants for ¢sh applications in aqua- Journal of Neurophysiology 66,247^260. culture. In: NutritionandFeedinginFish. (eds C.B. Cowey, Millkin, M.R. (1982) Qualitative and quantitative nutrient A.M. Mackie and J.G. Bell). Academic Press, London, requirements of ¢shes: a review. Fisheries Bulletin 80, pp. 177^189. 655^685.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 343 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Mistretta, C.M. (1991) Developmental neurobiology of the Osinov, A.G. and Bernache, L. (1996) `Atlantic'and `Danube' taste system. In: Smell and Taste in Health and Disease. phylogenetic groupings of brown trout, Salmo trutta, (eds T.V. Getchell, L.M. Bartoshuk, R.L. Doty and J.B. complex: genetic divergency, evolution and conserva- Snow). Raven Press, New-York, pp.35^64. tion. Journal of Ichthyology 36, 762^786. Miyagawa, M., Nakamoto, S.,Yamane, K. et al. (1979) Studies Osse, J.W.M., Sibbing, F.A. and Van Den Boogaart, J.G.M. on organic constituents of the snow crab (Chionoecetes (1997) Intra-oral food manipulation of carp and other opillio). Part II. Free amino acids in extracts. Bulletin of cyprinids: adaptationa and limitations. ActaPhysiologica theJapanese Society of Scienti¢c Fisheries 45,115^120. Scandinavica161,47^57. Morita, Y. and Finger,T.E. (1985) Topographic and laminar Papatryphon, E. and Soares, J.H.J. (2000) The e¡ect of diet- organization of the vagal gustatory system in the gold- ary feeding stimulants on growth performance of striped ¢sh, Carassius auratus. Journal of Comparative Neurology bass, Morone saxatilis, fed a plant feed stu¡-based diet. 238, 187^201. Aquaculture185,329^338. Motta, P.J. (1982) Functional morphology of the head of the Paul, V.J. (1992) Chemical defenses of benthic marine intertial suction feeding butter£y ¢sh, Chaetodonmiliaris invertebrates. In: Ecological Roles of Marine Natural (Perciformes, Chaetodontidae). Journal of Morphology Products. (ed. V.J. Paul). Comstock, Ithaca, New York, 174,283^312. pp. 164^188. Motta, P.J. (1988) Functional morphology of the feeding Paul,V.J., Cronan, J.M. Jr and Cardellina, J.H.I. (1993)Isolation apparatus of ten species of Paci¢c butter£y ¢shes of new brominated sequiterpene feeding deterrents from (Perciformes, Chaetodontidae): an ecomorphological tropical green alga, Neomeris annulata (Dasycladaceae: apporoach. Environmental Biology of Fishes 22,39^67. Chlorophyta). Journal of Chemical Ecology19,1847^1860. Murofushi, S. and Ina, K. (1981) Survey of feeding stimu- Pausan,K.(1984)Aquarium Fishes. Country Life Books,Mid- lants for the sea bream present in the dried pupae of dlesex. silkworms. Agricultural and Biological Chemistry 45, Pavlov, D.S. and Kasumyan, A.O. (1990) Sensory principles 1501^1504. of the feeding behaviour of ¢shes. Journal of Ichthyology Nakajima, K. (1991) E¡ects of diet-supplemented dimethyl- 30,77^93. b-propiothetin on growth and trust power of gold¢sh, Pavlov, D.S. and Kasumyan, A.O. (1998) The structure of the carp and red sea bream. Nippon Suisan Gakkashi 57, feeding behaviour of ¢shes. Journal of Ichthyology 38, 673^679. 116^128. Nakajima, K., Uchida, A. and Ishida,Y. (1989a) E¡ect of sup- Pawlik, J.R. (1993) Marine invertebrate chemical defenses. plemental dietary feeding attractant, dimethyl-b-pro- Chemical Review 93, 1911^1922. piothetin, on growth of gold¢sh. Nippon Suisan Gakkashi Pawlik, J.R., Chanas, B.,Toonen, R.J. and Fenical,W. (1995) 55,1291. Defenses of Carribbean sponges against predatory reef Nakajima, K., Uchida, A. and Ishida,Y. (1989b) A new feed- ¢sh. Part I. Chemical deterrency. Marine EcologyProgress ing attractant, dimethyl-b-propiothetin, for freshwater Series127,183^194. ¢sh. Nippon Suisan Gakkashi 55,689^695. Pawson, M.G. (1977) The responses of cod, Gadus morhua Nakajima, K., Uchida, A. and Ishida,Y.(1990) E¡ect of a feed- (L), to chemical attractants in moving water. Journal ing attractant, dimethyl-b-propiothetin, on growth of du conseil International pour l'exploration de la mer 37, marine ¢sh. Nippon Suisan Gakkashi 56,1151^1154. 316 ^318. Nakamura, M. and Kurihara, K. (1991) Di¡erential tempera- Pevzner, R.A. (1981) Ultrastructure of taste receptors in ture dependence of taste nerve responses to various taste Chondrichthyes. Part I. Adult sturgeon. Tsitologija 23, stimuli in dogs and rats. American Journal of Physiology 760^766 (in Russian). 261, R1402^R1408. Pevzner, R.A. (1985) Ultrastructure of taste receptors in Nikolaeva, E.V.and Kasumyan, A.O. (2000) Comparative ana- Chondrichthyes. Part III. Sturgeon larvae in yolk-feeding lysis of the taste preferences and behavioral responses period.Tsitologija 27,1240^1246 (in Russian). to gastatory stimuli in females and males of the guppy, Poe, W.E. and Wilson, R.P. (1989) Palatability of diets con- Poecilia reticulata. Journal of Ichthyology 40, 479^484. taining sulfadimetoxine, ormetoprim, and Romet 30 to Nikolski,G.V.(1963)The Ecology of Fishes. Academic Press, channel vat¢sh ¢ngerlings. Progressive Fish-Culturist London. 51,226^228. Ohsugi,T., Hidaka, I. and Ikeda, M. (1978) Taste receptor sti- Prescott, J. and Stevenson, R.J. (1995) E¡ects of oral mulation and feeding behaviour in the pu¡er, Fugu par- chemical irritation on tastes and £avors in frequent dallis. Part II. E¡ects produced by mixtures of and infrequent users of Chilli. Physiology and Behavior constituents of clam extract. Chemical Senses and Flavor 58, 117^1127. 3,355^368. Prezant, R.S. (1980) Anantipredator mechanism of the poly- Oikawa, C.K. and March, B.E. (1997) A method for assess- chaete, Phyllodoce mucose, with notes on similar ment of the e¤cacyof feed attractants for ¢sh.Progressive mechanisms in other potential prey. US FisheriesBulletin Fish-Culturist 59,213^217. 77,605^616.

344 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Raderman-Little, R. (1979) The e¡ect of temperature on the Sammarco, P.W., La Barre, C.L. and Coll, J.C. (1987) Defensive turnover of taste bud cells in cat¢sh. Cell and Tissue strategies of soft corals (Coelenterata: Octocorallia) of Kinetics12,269^280. the Great Barrier Reef. Part III.The relationship between Reutter, K. (1986) Chemoreceptors. In: Biology of the ichthyotoxicity and morphology. Oecologia (Berlin) 74, Integument.Vertebrates Vol. 2 (eds J. Bereiter-Hahn, A.G. 93^101. Matoltsy and K.S. Richards). Springer-Verlag, Berlin, Saxena, S.C. (1959) Adhesive apparatus of a hill-stream pp. 586^604. cyprinid ¢sh, Garra mullya (Sykes). Proceedings of the Reutter, K. (1987) Specialized receptor villi and basal National Institute of Sciences of India B25, 205^214. cells within the taste bud of the European silurid ¢sh, Scharrer, E., Smith, S.W. and Palay, S.L. (1947) Chemical Silurus glanis (Teleostei). In: Olfaction and Taste. Part IX. sense and taste in the ¢shes, Prionotus andTrichogaster. (eds D. Roper and J. Atema). Annals NY Acad. Sci, NY, Journal of Comparative Neurology 68, 183^198. pp.570^573. Schildknecht, H. (1971) Evolutionary peaks in the defensive Reutter, K. (1992) Structure of the peripheral gustatory chemistry of insects. Endeavour 30,136^141. organ, represented by the siluroid ¢sh Plotosus lineatus Schreck, J.A. and Mo¤tt, C.M. (1987) Palatabilityof feed con- (Thunberg). In: Fish Chemoreception.(ed.T.J.Hara). taining di¡erent concentrations of erythromycin thio- Chapman & Hall, London, pp.60^78. cyanate to chinook salmon. Progressive Fish-Culturist Reutter, K., Breipohl, W. and Bijvank, G.J. (1974) Taste bud 49, 241^247. types in ¢shes. Part II. Scanning electron microscopical Schulte, B.A. and Bakus, G.J. (1992) Predation deterrance in investigations on Xiphophorus helleri Heckel (Poeciliidae, marine sponges: laboratory versus ¢eld studies. Bulletin Cyprinodontiformes, Teleostei). Cell and Tissue Research of Marine Science 50,205^211. 153,151^165. Sheldon, R.E. (1909) The reaction of dog¢sh to chemical sti- Reutter, K. and Witt, M. (1993) Morphology of vertebrate muli. Journal of Comparative Neurology19,273^311. taste organs and their nerve supply. In: Mechanisms of Sibbing, F.A. (1988) Specializations and limitations in the Taste Transduction (eds S.A. Simon and S.D. Roper). CRC utilization of food resources by the carp, Cyprinus carpio: Press, Boca Raton, pp.29^82. a study of oral food processing. Environmental Biology of Rideout, J.A., Smith, N.B. and Sutherland, M.D. (1979) Fishes 22, 161^178. Chemical defense of crinoids by polyketide sulphates. Sibbing, F.A. (1991) Food capture and oral processsing. In: Experientia 35,1273^1274. Cyprinid Fishes: Systematics, Biology and Exploitation Ringler, N.H. (1985) Individual and temporal variation in (eds I.J.Win¢eld and J.S. Nelson). Chapman & Hall, Lon- prey switching by brown trout Salmo trutta. Copeia 4, don, pp.377^412. 918^926. Sibbing, F.A., Osse, J.W.M. and Terlouw,A. (1986) Food hand- Robinson, E.H., Brent, J.R., Crabtree, J.T. and Tucker, C.S. ling in the carp (Cyprinus carpio): its movement patterns, (1990) Improved palatabilityof channel cat¢sh feeds con- mechanisms and limitations. Journal of Zoology (London) taining Romet-301. Journal of Aquatic Animal Health 2, 210A,161^203. 43^48. Silver,W.L. (1987) The common chemical sense. In: Neuro- Robinson, E.H. and Tucker, C.S. (1992) Palatability of sara- biology of Taste and Smell. The Common Chemical Sense £oxin HCl-medicated feed to channel cat¢sh. Journal of (eds T.E. Finger and W.L. Silver).Wiley, NewYork, pp. 65^ Applied Aquaculture1,81^87. 87. Rogers, S.D. and Paul,V.J. (1991) Chemical defenses of three Silver,W.L. and Finger,T.E. (1991) The trigeminal system. In: Glossodoris nudibranchs and their dietary Hyrtyos Smell and Taste in Health and Desease. (eds T.V. Getchell, sponges. Marine Ecology Progress Series 77,221^232. L.M. Bartoshuk, R.L. Doty and J.B. Snow). Raven Press, Saglio, P. and Blanc, J.M. (1983) Etude en olfactometre de la NewYork. pp.97^108. reponse comportementale aux l-acides amines chez le Specziar, A.,Tolg, L. and Biro, P.(1997) Feeding strategy and juvenile de carpe, Cyprinus carpio L. Biology of Behaviour growth of cyprinids in the littoral zone of Lake Balaton. 8,359^370. Journal of Fish Biology 51, 1109 ^1124. Saito,T., Noguchi,T., Harada,T., Murata, O. and Hashimoto, Stevens, D.A. (1990) Personality variables in the perce- K. (1984) Tetrodoxin as a biological defence agent for puf- ption of oral irritation and £avor. In: Chemical Senses, fers. Bulletin of the Japanese Society of Scienti¢c Fisheries Irritation, Vol. 2 (ed. B.G. Green) edn New York, 51,1175^1180. pp.217^228. Sakata, K. (1989) Feeding attractants and stimulants for Storebakken, T. and Austreng, E. (1988) Feed intake mea- marine gastropods. Bioorganic Marine Chemistry 3,115^ surements in ¢sh using radioactive isotopes. Part II. 129. Experiments with Atlantic salmon and rainbow trout in Sammarco, P.W.and Coll, J.C. (1990) Lack of predictability in sea-pens. Aquaculture 70, 277^288. terpenoid function: multiple roles and integration with Storebakken, T., Austreng, E. and Steenberg, K. (1981) A related adaptations in soft corals. Journal of Chemical method for determination of feed intake in salmonids Ecology16, 273^289. using radioactive isotopes. Aquaculture 24,133^142.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 345 Taste preferences Alexander O Kasumyan & Kjell B DÖving

Stradmeyer, L. (1989) A behavioural method to test feeding metabolites on feeding by juvenile rabbit¢sh and parrot- responses of ¢sh to pelleted diets. Aquaculture 79,303^ ¢sh. Marine Ecology Progress Series147,21^29. 310. Thompson, J.E.,Walker,R.P.and Faulkner,D.J. (1985) Screen- Stradmeyer,L., Metcalfe, N.B. andThorpe, J.E. (1988) E¡ect of ing and bioassays for biologically-active substances from food pellet shape and texture on the feeding response of forty marine sponge species from SanDiego, California, juvenile Atlantic salmon. Aquaculture 73,217^228. USA. Marine Biology 88, 11^21. Stradmeyer, L. and Thorpe, J.E. (1987) The responses of Toften, H., Jorgensen, E.H. andJobling, M. (1995) The studyof hatcheryreared Atlantic salmon, Salmo salar L. parr to feedingpreferencesusingradiography:oxytetracyclineas pelletted and wild prey. Aquaculture and Fisheries Man- afeedingdeterrentandsquidextractasastimulantindiets agement18,51^61. forAtlantic salmon. Aquaculture Nutrition1,145^149. Strieck, F. (1924) Untersuchungen uber den Geruchs- und Trudel, P.J. (1929) Untersuchungen u« ber Geschmacksreak- Geschmackssinn der Elritze (Phoxinus laevis). Zeitschrift tion der Fische auf `su« sse' Sto¡e. Zeitschrift fu« r Vergle- fu« rVergleichende Physiologie 2,122^154. ichende Physiologie10,367^409. Stuart, J. (1996) How to Keep Tropical Aquarium Fish. Tiger Tucker, D. (1983) Fish chemoreception: periferal anatomy Books Intern, London. and physiology. In: Fish Neurobiology: Brainstem and Sutterlin, A.M. (1975) Chemical attraction of some marine SensoryOrgans,Vol.1 (eds R.G. Northcutt and R.E.Davis). ¢sh in their natural habitat. Journal of the Fisheries University of Michgan Press, Ann Arbor, pp.311^349. Research Board of Canada 32,729^738. von Uexku« ll, J. (1895) Vergleichend-sinnesphysiologische Sutterlin, A.M. and Sutterlin, N. (1970) Taste responses in Untersuchungen. I. U« ber die Nahrungsaufnahme des Atlantic salmon (Salmo salar)parr.Journal of the Fisheries Katzenhais. Zeitschriftfu« r Biologie 32,548^566. Research Board of Canada 27,1927^1942. Van Damme, R., Bauwens, D.,Vanderstighelen, D. and Ver- Takaoka, O.,Takii, K., Nakamura, M., Kumai, H. andTakeda, heyen, R.F.(1990) Responses of the lizard Lacertavivipara M. (1990) Identi¢cation of feeding stimulants for marbled to predator chemical cues: the e¡ects of temperature. rock¢sh. Bulletin of the Japanese Society of Scienti¢c Fish- Animal Behaviour 40,298^305. eries 56,345^351. Van Diggelin, J., Rozema, J., Dickson, D.M.J. and Broekman, Takeda, M. and Takii, K. (1992) Gustation and nutrition R. (1986) b-3-Dimethylsulphoniopropionate, proline and in ¢shes: application to aquaculture. In: Fish Chemo- quaternary ammonium compounds in Spartice anglica reception (ed. T.J. Hara). Chapman & Hall, London, in relation to sodium chloride, nitrogen and sulphur. pp.271^287. New Phytologist103, 573^586. Takeda, M.,Takii, K. and Matsui, K. (1984) Identi¢cation of Vasil'ev, V.P. (1985) Evolutionary Karyology of Fish. Nauka, feeding stimulants for juvenile eel. Bulletin of theJapanese Moscow. Society of Scienti¢c Fisheries 50,645^651. Voris, H.K. and Bacon, J.P. (1966) Di¡erential predation on Takii, K., Shimeno, S.,Takeda, M. and Kamekawa, S. (1986) tadpoles. Copeia 3, 594^598. The e¡ect of feeding stimulants in diet on digestive Weber, E.H. (1827) U« ber das Geschmacksorgane des Karpfen enzyme activities of eel. Bulletin of theJapanese Society of and den Ursprung seiner Nerven. Archiv fu« rAnatomie Scienti¢c Fisheries 52,1449^1454. and Physiologie,309^315. Takii, K., Takeda, M. and Nakao,Y. (1984) E¡ect of supple- Welsch, U. and Storch,V.(1969) Die Feinstruktur der Gesch- ment of feeding stimulants to formulated feeds on feed- macksknospen von Welsen [Clarias batrachus (L.) and ing activity and growth of juvenile eel. Bulletin of the Kryptopterus bicirrhis (Cuvier et Valenciennes)]. Zeits- Japanese Society of Scienti¢c Fisheries 50,1039^1043. chrift fu« r Zellforschung und MikroscopischeAnatomie 100, Talbot, C. (1985) Laboratory methods in ¢sh feeding and 552^559. nutritional studies. In: Fish Energetics: New Perspectives Whitear, M. (1971) Cell specialization and sensory func- (eds P. Tytler and P. Calow). Croom-Helm, London, pp. tion in ¢sh epidermis. Journal of Zoology (London) 163, 125^154. 237^264. Talbot, C. and Higgins, P.J. (1983) A radiographic method for Whitear, M. (1992) Solitarychemosensorycells. In: Fish Che- feeding studies on ¢sh using metallic iron powder as a moreception (ed. T.J. Hara). Chapman & Hall, London, pp. marker. Journal of Fish Biology 23, 211^220. 103^125. Tandler, A., Berg, B.A., Kissil, G.W. and Mackie, A.M. (1982) Whitear, M. and Kotrschal, K. (1988) The chemosensory E¡ect of food attractants on appetite and growth rate of anterior dorsal ¢n in rockling (Gaidropsarus and Ciliata, gilthead bream, Sparus aurata L. Journal of Fish Biology Teleostei, Gadidae): activity, ¢ne structure and innerva- 20,673^681. tion. Journal of Zoology (London) 216,339^366. Tester, A.L. (1963) Olfaction, gustation and the common Whitear, M. and Moate, R.M. (1994) Chemosensory cells in chemicalsenseinsharks.In:Sharks and Survival.(ed. the oral epithelium of Raja clavata (Chondrichthyes). P.W.Gilbert).D.C.Heath,Boston,pp.255^282. Journal of Zoology (London) 232,295^312. Thacker, R.W., Nagle, D.G. and Paul, V.J. (1997) E¡ects of Wilson, D.M., Puyana, M., Fenical,W.and Pawlik, J.R. (1999) repeated exposure to marine cyanobacterial secondary Chemical defense of the Caribbean reef sponge Axilella

346 # 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 Taste preferences Alexander O Kasumyan & Kjell B DÖving

corrugata against predatory ¢shes. Journal of Chemical in ¢sh: a possible mechanism for avoiding marine Ecology 25,2811^2824. toxins. CanadianJournal of Fisheries and Aquatic Sciences Winn, H.E. and Bardach, J.E. (1959) Di¡erential food selec- 45, 2182^2186. tion by moray eels and a possible role of the mucous Yamashita, S. (1964) Chemoreceptor response in frog as envelope of parrot ¢shes in reduction of predation. Ecol- modi¢ed by temperature change. JapaneseJournal of Phy- ogy 40,296^298. siology14,488^504. Witt, M. and Reutter, K. (1990) Electron microscopic demon- Yoshii, K., Kamo, N., Kurihara, K. and Kobatake, Y. (1979) strationof lectin binding sites inthe taste buds of theEur- Gustatory responses of eel palatine receptors to amino opean cat¢sh Silurus glanis (Teleostei). Histochemistry acids and carboxylic acids. Journal of General Physiology 94, 617^628. 74,301^317. Wootton, R.J. (1998) Ecology of Teleost Fishes. Kluwer Aca- Yoshii, K., Kashiwayanagi, M., Kurihara, K. and Kabataki,Y. demic Publishers, Dordrecht. (1980) High sensitivityof eel palatine receptors to carbon Wylie, C.R. and Paul,V.J. (1988) Feeding preference of the sur- dioxide. Comparative Biochemistry and Physiology 66A, geon ¢sh Zebrasoma £avescens in relation to chemical 327^330. defenses of tropical algae. Marine Ecology Progress Series Yoshiyama, R.M. and Darling, J.D.S. (1982) Grazing by 45,23^32. the intertidal ¢sh Anoplarchus purpurescens upon a dis- Yamada, S. and Yone, Y. (1986) Loss of dietary amino acids tasteful polychate worm. Environmental Biology of Fishes during masticationbycarp. BulletinoftheJapaneseSociety 7,39^45. of Scienti¢c Fisheries 52,673^676. Zimmer-Faust, R.K. (1987) Crustacean chemical perception: Yamamori, K., Nakamura, M., Matsui, T. and Hara, T.J. towards a theory on optimal chemoreception. Biological (1988) Gustatory responses to tetrodoxin and saxitoxin Bulletin172,10^29.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4,289^347 347