OLFACTION

IN FISH*

Contents 1. Introduction 2. Morphology of the olfactory system 2.1. Peripheral olfactory organ 2.2, Fine structure of the olfactory epithelium 2.3. Olfactory bulb and tract 3. Sensitivity of the olfactory sense 4. Electrophysiology of the olfactory system 4. I Notes on experimental procedures 4.2. Electrical activity of the olfactory epithelium 4.2.1. Slow potential 4.2.2. Single unit 4.3. Electrical activity of the olfactory bulb 4.3. I. Slow potential 4.3.2. Oscillatory potential 4.3.3. Single unit 4.4. Electrical activity of the olfactory tract and telencephalon 5. Role of olfaction in fish behavior 5. I. Feeding 5.1.1. Olfactory response to food 5.1.2. Survey of active components in foods 5.2. Defense 5.2.1. Fright reactton and alarm substance 5.2.2. Repellent 5.3. Schooling 5.4. Reproduction 5.5. Migration and orientation 5.5. I. Evidence for homing 5.5.2. Olfactory hypothesis 5.5.3. Electrophysiological approaches 5.5.4. Mechanism of learning 5.5.5. Chemical clues 5.5.6. Mechanism of orientation 6. Amino acid as olfactory stimulus 6.1. Structure-activity relationships 6.2. Hypothetical receptor site 6.3. Role of amino acids as chemical signals 7. Summary References

211 272 272 278 283 285 287 287 290 290 292 294 294 296 299 302 303 303 304 305 306 306 309 309 311 312 312 313 315 316 317 319 319 320 324 325 326 327

1. Introduction

Sensory receptors are the immediate detectors of environmental stimuli. These may be chemical (olfaction, taste, etc.),’ physical (photoreception, thermoreception, etc.), or mechanical (position, motion. etc.), the behavioral significance of the signals arriving at the brain through the different channels varies greatly from one species to another. For many animal groups vision is most important. while in others information about the chemical nature of the external environment is more essential. For fishes whose life is entirely restricted to the aqueous environment chemoreception plays a major and sometimes decisive role in behavior such as feeding, defense, spawning, schooling, and orientation and migration. Fish detect chemical stimuli through at least two different channels of chemoreception, olfaction or smell. and gustation or taste. The distinction between these two sensory modalities in fish is not always as clear as in terrestrial organisms, mainly because in fish both olfaction and taste are mediated by dilute * Dedicated ,,‘\

51 A

to the memory

of my late friend Dr. Yasuo 271

Yokoe

aqueous solutions. Practically, in terrestrial air breathers. those receptors which have high sensitivity and specificity, which respond to volatile materials. and which arc distance chemical receptors are distinguished as olfactory. Those receptors of moderate sensitivity, which are stimulated by dilute solutions. and which are usually associated with feeding are considered as gustatory or contact chemical receptors. Howcvcr. in fish differences between olfactory and gustatory senses are not as clear cut. Historically, the existence of the sense of smell in fish has long been controversial. Nagel(1894), for instance, denied the existence of true olfactory sense in aquatic animals. premising that the organ of smell can be stimulated only by gaseous substances. It followed that chemical stimulation in aquatic animals could only be mediated by taste. Since then, many investigators contributed to the study of this problem of physiological differences between the senses of smell and taste in fish (Von Uexkiill, 1895; Herrick, 1908; Parker. 1910. 1911. 1912, 1922; Sheldon, 1909, 1911; Copeland, 1912; Olmsted, 1918; Strieck, 1924; and many others). The most convincing evidence of the existence of the sense of smell in fish was obtained by Strieck (1924). He trained minnows, Pho.~in1r.spho.uinzts, to discriminate pure odorous (coumarin, skatole, and muscone) and taste (glucose, acetic acid. and quinine) substances. These substances were readily detected by intact minnows. However, trained fish were unable to discriminate odorous substances after the forebrain was removed, although they could still perceive taste substances. This clearly demonstrated distinct olfaction and taste receptor function in fish. Much of the general information concerning olfaction in fish has been summarized in the recently published book Olfirctiorl in Fish (Kleerekoper. 1969) and in the chapter on chemoreception (Hara, 1971) of Fish Physiology, Vol. V (Hoar and Randall, 1971). However. during the last couple of years, between publication of the above books and the present time, there have appeared some interesting papers dealing with the fine structure of the olfactory organ, specific stimulatory effectiveness of certain amino acids, and electrophysiological analysis of the olfactory bulbar response. In this review. no attempt has been made to be all-inclusive. Many other topics are covered in reviews and symposia (Teichmann, 1962; Hasler, 1966; Malyukina ct al.. 1969; Hara, 1970).

2. Morphology of the Olfactory System The olfactory system may roughly be divided into three parts: (1) a peripheral part represented by the olfactory organ which houses the neuroepithelium and the olfactory bulb; (2) an intermediate part, the anterior olfactory nucleus; and (3) a central part located mainly in the paleocortical area. Very few studies have been done on morphology and function of the central olfactory system either in fish or in other vertebrate forms. The following discussion will consequently be focused on the peripheral olfactory system including the olfactory bulb. 2.1. PERIPHERAL OLFACTDRY

ORGAN

The olfactory organs of fishes are diversely developed. At one extreme they are well developed (macrosmatic forms) such as in elasmobranchs and most eels, and at the other they are poorly developed (microsmatic forms) such as in pike, flying-fish, stickleback, pipe-fish and angler-fish. In sharks and rays, where the chemical senses are highly important, the paired olfactory pits are usually situated on the ventral side of the snout (Fig. 1). The opening of each pit is divided by skin flaps, usually one medially and one laterally attached into anterior inlet and posterior outlet. In some species the posterior outlet opens directly into the mouth. In a holocephalid, Chimaera mnstrosa. for example, a deep furrow runs from each of the two external nares dorsolaterally along the upper lip towards the mouth. The posterior part of this furrow communicates with the anterior part of the mouth cavity, and thus has the function of an internal nare. While the mouth is closed, water passes from the external nare along the naso-oral groove and through the internal nare into the mouth cavity. Since the olfactory chamber communicates dorsally with the naso-oral groove it is always supplied by water (Holl,

273

OLFACTION IN FISH

FIG. 1. Ventral view of head region of the dogfish (Scyliorhinus covering the nasal groove is cut away on the left side. Redrawn ( 1969).

stellaris).

from

The flap Kleerekoper

1973). Wide variations in the location, size and shape of the nostrils have been described in elasmobranchs. In the teleost fishes there are considerable variations in size, structure and degree of development of the olfactory organs. Although there is no extensive review of the nasal anatomy of all species, several studies describe or review numerous species (Burne, 1909; Liermann, 1933; Matthes, 1934; Teichmann, 1954; Holl, 1965; Singh, 1972; Zeiske, 1973, 1974) or consider one species in detail (Laibach, 1937; Eaton, 1956; Johnson and Brown, 1962; Branson, 1963; Pfeiffer, 1963, 1964, 1968, 1969 a; Devitsyna, 1972; Kapoor and Ojha, 1972, 1973 a, b ; Ojha and Kapoor, 1972 a, 1973, 1974). Watling and Hillemann (1964) and John (1972) have dealt with the development of the olfactory apparatus of Arctic grayling (Thymallus arcticus) and cut-throat trout (Salmo clarki) respectively. Kleerekoper (1969) listed numerous citations dealing with the anatomy and development of the olfactory apparatus of several species. Only a few of the typical patterns will be described below. (B)

(A)

Or

an

FIG. 2. Position and internal structure of the nose in the minnow, Pho.xinus phoxinus (A), and eel. Anguilla anguilla (B). an. anterion naris; pn. posterior naris; f, skin flap; or, olfactory rosette. From Teichmann, H., C’mschau Wiss. 7kh.. 62. 588-591. Frankfurt. Germany. 1962.

T. .I. HAKA

FIG. 3. Olfactory organs and brain of Cyclothorzemicrodone. Left, male; right, female. cc. corpus cereberum; d. diencephalon;

eg, eminentia granuralis; fb, forebrain; ob.

olfactory bulb; 00. olfactory organs; ot, optic tectum. From Marshall (1967).

The paired olfactory pits are usually located on the dorsal side of the head (Fig. 2). The eels and morays (Anguilliformes), believed to have the most acute sense of smell, have large and elongate olfactory pits, extending from the tip of the snout to the orbit of the eye (Fig. 2B). In contrast, certain puffers (Tetraodontidae), which are highly visually oriented reef fishes, have completely lost the nasal sacs. There are also great differences in the olfactory organs even between different sexes of the same species. The two most successful groups of bathypelagic fishes, ceratioid anglerfishes and Cyclothone spp., have evolved distinct sexual dimorphism in olfactory organization and size (Marshall, 1967, 1971). The males have large olfactory organs, while these are small or regressed in the females of all species (Fig. 3). The macrosmatic males are assumed to be attracted to the scents of their much larger partner. Marshall (1971) writes: “In the sunless bathypelagic region, pheromones, which can be perceived at greater range than visual signals. seem apt, considering the ‘favourable’ factors of this environment.” In most species of teleost fishes the development of the olfactory organs lies somewhere between these two extremes. It might be expected that the olfactory organs of fishes which hunt by scent would be consistently larger than those which seek food by sight, but this is only partially true. For though the great development of the organ in the eel, conger, and elasmobranchs is accompanied by an acute sense of smell, yet in the rocklings, the loach, and the sole, which also seek their food by scent, the olfactory organs cannot be said to be proportionally more developed than they are in forms which feed by sight, such as plaice and pollack (Bateson, 1890). Each nasal pit generally has two openings, anterior inlet and posterior outlet, which are separated by a nasal Aap or bridge. The patterns of the olfactory flap vary greatly from species to species. In rainbow trout (S&no gairdneri) the flap is concave anteriad, apparently serving to deflect water currents downward into the anterior nostril, a rather general arrangement in bony fish. There is another Hap attached to the rostra1 side of the posterior naris in some species, e.g. whitefish (Covegonus clupeqformis). It is not

775

OLFACTION IN FISH

known whether the posterior flap has any role in controlling the flow of water out of the posterior nostril. A current of water enters the anterior and leaves through the posterior nostril either passively through the locomotion of the fish in water, or actively by ciliary action within the pits or by the pumping action of the nasal sac generally with the aid of an accessory nasal sac (a posteriorly lobed diverticulum of the nasal sac proper). In sticklebacks, blennies and some other species the olfactory sacs possess only one nostril (Pipping, 1926; Liermann. 1933). Parker (1910) observed that water flowed through the olfactory sac of catfish in g-10 sec. According to Teichmann (19.59). 22lOsec was required for the passage of the water through the nasal sac of eels. The rate of flow is considerably accelerated by forward movement of fish. Johnson and Brown (1962) observed that movement of water through the olfactory chamber of black rockfish, Sehastodes mrlunops. was brought about by the alternate compression and expansion of the accessory sacs during normal respiration of the fish. In centrachid fishes, almost no water entered through the anterior naris during ordinary resting respiration when examined by placing Indian ink in water near the nares (Eaton. 1956). In the garfish. Lepisosteus OSSPUS,flow of water through the nasal capsule is produced by action of cilia. In excised lamellae, the velocity of a water stream across the lamellar face varied between 1.5 and 2.9 mmjsec (average, 2.2 mmjsec) (Bashor et ul., 1974).

(A)

(B)

FIG. 4. The olfactory rosettes’of minnow. Phozims pho\-imrs (A). and anguilla (B). Redrawn from Teichmann (1954).

ccl. A~y~ri//u

The nasal sac is lined with the olfactory epithelium. which is generally raised from the floor of the organ into a complicated series of folds or lamellae to make a rosette (Fig. 4). The arrangement, shape and degree of development of the lamellae in the olfactory rosette of teleosts vary considerably from species to species. Normally, a central ridge or raphe is formed rostrocaudally on the bottom of the olfactory chamber. From this first ridge, a varying number of transverse lamellae radiate. The development of the lamellae begins caudad and progresses rostrad. so that the oldest and largest lamella is located most posteriorly. In a freshwater, air-breathing murrel, Cha~ta punctatus, the lamellae in the olfactory rosette are arranged with their long axis parallel to the anteroposterior axis of the body. New lamellae are added at its two lateral ends (Kapoor and Ojha, 1973; also see Shibuya, 1960). Thus, the number of lamellae within a species depends on size of the individual. In Lota lota, however, the number of lamellae is relatively constant in fishes from 280 to 41.5 mm in length (Pfeiffer, 1965). Figure 5 is a schematical representation of the typical arrangements of the olfactory lamellae

T. J. HAKA

FIG. 5. Arrangement of the olfactory lamellae in ten species of fish. From Teichmann (1954).

in a number of species studied by Teichmann (1954). In Table 1 some of the representative data on the number of the olfactory lamellae studied by various authors are listed. As already mentioned, the number increases, to some extent, with the body length of the fish. In addition to the formation of new lamellae, each lamellae increases in size. Thus, the area of the olfactory epithelium of the individual fish is considerably increased by the formation of new lamellae and by the growth of those already present. Burne (1909) distinguished three types of olfactory rosettes; oval (in most fishes), round (in Esox) and elongate (in Aquilla). Fish with round rosettes normally have only a few lamellae and usually show little or no behavioral responses to olfactory stimulation (microsmatic). Species with oval rosettes are most common and intermediate between the other two. Secondary folding of the olfactory lamella was first found in rainbow trout by Teichmann (1954). However, he thought the secondary folding he observed might be an artefact caused by fixation, since he could not confirm this finding in fresh material. Pfeiffer (1963) later confirmed the presence of the secondary lamellae in Pacific salmon, Oncorhynchus spp., as well as in rainbow trout. There were between 5 and 10 secondary

277

OLFACTIONIS FISH TABLE 1. NUMBEROF LAMELLAEOBSERVEDIN SOME SPECIES

Number Of

Species Sticklehack (Gasrrrostrus ucuhtus) Pike (EX1.Y IIICiLIS) Arctic grayling

Reference

lamellae Wunder

(1957)

Wunder

(1957)

Watling

and Hillemann

I l-19

Wunder

(1957)

12-14

Hara.

Law and Hobden

(1973)

IZ- I6

Hara.

Law and Hobden

(1973)

ICIX

Wundcr

(1957)

30 32

Wunder

(1957)

6&90

Wunder

(1957)

XC90 230

Shibuya Pfeiffer

2 9- 18 II

15

(1964)

(Th~wdlrt.swticus) Minnow (Ph0.uiuu.s pho\-irncs) Brook trout (Salcdimta,fontidi.s) Whitefish (Car-q011frs clupruformis) Rainbow trout (Sul~no gairdrxvi) Burhot (L&I /ora) Eel (Anguilltr am&l/a) Challrla LIrp.5 Haplopagrus quutltheri

(1960)

(I 964)

foldings per lamella in 4@50cm salmon and trout (Fig. 6), but no secondary folding observed in 15 cm long salmon. Secondary foldings have also been observed in grayling (Watling and Hillemann. 1964), Baltic sea trout, Salmo trutta trutta (Holl, 1965; Bertmar, 1972), Atlantic salmon, S. salar (Sutterlin and Sutterlin, 1971), Lota lota (Devitsyna, 1972), brook trout, Salurlinus firhalis and lake whitefish (Hara rt al., 1973) and garfish (Lrpisosteus) (Bashor rt The secondary is also in sharks 1934) and dipnoi (Neoceratodus, Protopterus) (Pfeiffer, The secondary is a of size. is not related to marine phase the life of Pacific (Pfeiffer. 1963) Baltic sea (Bertmar. 1972). lamellae are folded when sea trout leaves the New primary continue to formed, and secondary folding these occurs on in sea. The folding process the primary results in increase in surface area the olfactory However, all most of increase can referred to indifferent non-sensory (Pfeiffer, 1963; 1972; Hara al., 1973; et al., Even so, of the increases effective of the in the chamber. which. turn. seems result in eventual increase total olfactory Attempts have made to the total of the epithelia in species to particular olfactory (Teichmann, 1954; and Kapoor. 1973; Kapoor Ojha, 1973). (1954) investigated surface area the olfactory and of retina of species of teleosts. Species round rosettes the smallest of olfactory (Eso.u, about of the body surface; 0.4?3. The olfactory epithelium found in with oval (Gohio, 3.6%. Phosinus. 1.9yJ. In hgdla and Lota, believed to be macrosmatic, the area was 1.4 and 1.3”/, respectively. On the basis of these measurements, Teichmann classified these fishes into three groups: (1) species in which the eye and nose are well developed (Pho~inus and Go&o), (2) species in which the eye is better developed than the nose (Esox and Gasterostrus), and (3) species in which the nose is well developed compared with the eye (Anguilla and Lota). However. it is doubtful whether such simple relation exists between the surface area of the olfactory epithelium and sensitivity to odors, because there is no simple relation between the surface area of the olfactory epithelium and the number of receptors it contains. In no fish is the sensory epithelium uniformly distributed on the surface of the olfactory was

T. .1. HAI~A

FIG. 6. Photomicrographs of an olfactory Iamclla of sockeye salmon, O,rc,o~,I~~rl~hrrs w&a (A). and its cross section to dcmonstratc the secondary folding (B). pf, primarl lamella; sf. secondary lamella. From Pfeiffcr (1063 c). Reproduced by permIssion of the National Research (‘ouncil of Canada.

lamellae, but it generally occurs in isolated sensory areas separated by columnar ciliated cell areas (indifferent epithelium). According to Ho11 (1965). there are three types of arrangements of the sensory epithelium in the lamellae: (1) continuous except for the dorsal parts of the lamellae (Zctalurus, Anguillu, Prrca, Sulmo, etc.), (2) separated in large areas between the lamellae (Esox), and (3) dispersed in small islets (Phoxinus, Cyprinus, Curassius, etc.). Teichmann (1954) estimated the number of sensory cells present in the olfactory epithelium of the freshwater teleosts mentioned above. These figures are comparable with. 40,000 sensory cells per square millimeter of human olfactory epithelium (Ehrensvtird, 1942) 120,000 cells in the rabbit (Allison and Warwick, 1949) and 16,000 cells in the dog (Neuhaus and Miiller. 1954). 2.2. FINE STRUCTURE OF THE OLFACTORY EPITHELIUM

The microscopical study of the vertebrate olfactory organ began as early as in 1850. However, the olfactory epithelium of fish has not received as much attention as that of other vertebrates. The historical development of the histological studies on all cell types and their fiber connections in the olfactory epithelium has been described elsewhere

OLFACTION

IU!L

IN

FISH

279

,

FIG. 7. Diagram of fine structure of the olfactory epithelium of the eel. Anguilla I. receptor cells; 2, supporting cells: 3. ciliated cells; 4. basal cells; 5, goblet cells; 6. club-shaped secretory cells: 7, olfactory knob with cilia. From Ho11 (1965). anguilla.

(cf. Kleerekoper, 1969). Since the first description by Schultze (1856), it has been recognized that the olfactory epithelium in vertebrates consists of three cell types: receptor cells, supporting cells and basal cells (Fig. 7). In some species (Anguilla, Myxocephalus, etc.), large flasklike mucus cells are interspersed among supporting cells (Holl, 1965). The individual olfactory receptor cells of fish are similar to those of other vertebrates in general appearance, though there is a great variation in details even within a particular olfactory organ. The fine structure of the epithelium has been studied in a number of species of fish (Trujillo-Cenoz, 1961; Bannister, 1965; Bronshtein and Ivanov, 1965; Vinnikov, 1966; Wilson and Westerman, 1967; Thornhill. 1967; Gemne and Diiving, 1969; Schulte and Holl, 1971; Schulte, 1972; Lawry. 1973 ; Breipohl rt al., 1973 a, b). The receptor cell, which is a bipolar primary sense cell, sends a slender cylindrical dendrite toward the surface of the epithelium and is directly connected with the olfactory bulb by its axon. The dendrite terminates in a minute swelling (olfactory knob) which, in the majority of cases, bears a variable number of cilia (Figs. 7 and 8). Unlike those of higher vertebrates, the olfactory knob in fishes is merely an apical swelling of the dendrite and does not project beyond the epithelial surface. The knob has an approximate diameter of 2-3 p, the dendrite of l-2 ,u. and the cell body of 5-8 cc. The dendrite is characterized by the presence of thin tubules, about 3008, wide. aligned parallel to its length. The cytoplasm contains numerous mitochondria. The perikaryon of the sensory cell is mostly occupied by the nucleus. In Phoxinzrs pho.~inus Bannister (1965) found three distinct types of receptor cells: (1) those bearing cilia, (2) those bearing microvilli, and (3) those bearing neither cilia nor microvilli, but rising as a simple rod about 4 p into the mucus. These three morphologically different receptor cell types have also recently been observed in the eel (Schulte, 1972). Main features of the fine structure of the epithelium of the African bichir, Culumoichthys coluhuricus, are generally the same as those described in other fish species (Schulte and Holl, 1971). In the Australian lungfish, Neocerutodus forstrri, however, the receptor cells bear only microvilli (Theisen. 1972). which have been also reported in the elasmobranchs (Ginglyonzostonlu cirraturn and Rhirzohuttrs lrntiginosus (Reese and Brightman, 1970). It is not known which part of the receptor cell is involved in the initial events of olfactory stimulation. If we view the surface of the olfactory epithelium through the

FIG. 8. Electron mtcrograph illustrating the lint structure of the distal segment of the olfactory dendrites of F,rxoyiu liw~lt~~. Long cilia arise from the olfactory knob (O.K.) and project into the lumen of the olfactory chamber. B.B., basal hodieb; I>. desmosome; M. mitochondria; SC.. supporting cells. From Trujillo-Cenciz (I 961).

scanning electron microscope we are struck by the dense carpet of the epithelial cilia (Fig. 9). It is reasonable to assume that the receptor site lies on the membranes of such olfactory cilia, since these are the first points of contact with odorous molecules. In fishes, moreover, the olfactory knob does not protrude beyond the limiting surface of the epithelium as in other vertebrates. If this is the point of excitation, number, length and motility of the cilia would become significant in enlarging the active receptor area and increasing the chance of contact between cilia and odorous molecules. The fine structure of the cilia is identical with that of common kinocilia; nine pairs of peripheral fibrils are arranged in a circle around two central fibrils. In the European carp, Curassius carassius. in addition to typical cilia. aggregations of cilia have been observed; the fibrils are grouped together in clusters instead of forming individual cilia and enveloped by a single limiting membrane (Wilson and Westerman. 1967). Similar

FIG. 9. Scanning electron micrographs of the sensor) (A) and cillatcd indifferent epithelia (B) of sea trout. SU/MI truttrr ttxttu. cn. ciliated nonsensorv cells; h. hole: m. microvilli: mb, mucus balls; nc. nonsensory cilia: rc. receptor cilia From Bertmar (1972 a).

cilia ry aggregations

have also been reported to occur on the olfactory epit helium of ary/~fir~u.s (Lowe, 1974). The fine structl ire of 1;he WY .egations is constant, consisting of 9 + 2 complexes enclosed by a single IIlembra. ne, that lgh the number of complexes varies within any ciliary aggregation. It is suggeslted that the ciliary aggregations are motile structures which are not directly in volved in olfac:tion, but may be associated with the circulation of fuid between the lame11ae. Fou r to 20 cilia per receptor ending were counted in Lota, Am@la and Calamwwichtl lYS Gad US morhua and Melanogramrnus

2x2

T. .I. HAKA

(Gemne and Diiving, 1969; Schulte and Holl, 1971; Schulte, 1972). The cilia project to the basal portion of the olfactory knob, each cilium forming an angle approximately 30’ with the vertical cell axis. Estimates of length and diameter of cilium are in the ranges of I CL30 p and 0.2-0.4 p, respectively. Motility of the olfactory cilia was found by Hopkins (1926) in some amphibians but not in teleosts. However, it is not surprising that the olfactory cilia are motile, since, as mentioned above, they have all the characteristics of kinocilia. In fact, Bronshtein (1964) observed that the olfactory cilia in C’q~prinus,Esox, and Laml~tra are motile ir? vitro. The cilia in these three species displayed non-coordinated whiplike movements which were asynchronous and irregular even within the same cell. The microvillous receptor cells of the eel bear from 30 to 60 microvilli. each of which is 0.1 ,U in diameter and up to 5 11 in length. Each microvillus contains tubules of 160 t%in diameter. The microvillous receptor cells have been observed also in Pho.~inus and Curassius curassius. A single rod-shape appendage of a third receptor type observed in the eels is 0.8 ,u in diameter and projects up to 4~ above the epithelial surface (Schulte, 1972). The functional differentiation between ciliated and microvillous receptor cells is not known. In addition to the receptor cells described, a new type of cellular element-the secondary neurone or spindle-shaped cell-has been reported in the olfactory epithelium of Cl~r~u punctatu.~ (Kapoor and Ojha, 1974). The secondary neurone lies at a level lower than that of the primary neurones and the dendritic ends of the former synapse with the axonal ends of the latter. The axons of the secondary neurones pass out of the basement and contribute to the formation of bundles of olfactory fibers. In fishes and all other vertebrates, as will be discussed later. the axon of the olfactory receptor has been believed to synapse with the dendrite of a mitral cell only after it reaches the olfactory bulb. The supporting cells are polygonal columnar epithelial cells which occupy the interstices between the receptor cells. In many cases, they bear a small number of either cilia or microvilli, and have prominent oval nuclei situated basally. The cytoplasm immediately beneath the limiting surface is relatively scant with electron-dense vesicles. In the deeper parts of the cells as much endoplasmic reticulum and mitochondria as in the receptor cells are present. The significance of the supporting cells in olfactory perception is not understood. Recent ultrastructural studies by Breipohl et al. (1974) have shown that in goldfish the supporting cells do not merely fill the spaces between the sensory receptor cells but their apical cytoplasm envelops the dendrites of the receptor cells. A small dendrite is sometimes enclosed spirally by a supporting cell. This enclosure is believed by these authors to be due to an active penetration by the dendrites. The basal cells occupy the lower third of the epithelium. Their nuclei are almost identical with those of the receptor cells. Frequent mitosis in these cells suggest a function as stem cells for the replacement of the neurones (Graziadei. 1971). The turnover of cells within the olfactory epithelium has been also suggested in the lamprey, Lumpctra jiuviatilis (Thornhill, 1970) and sea trout, Sulmo truttu (Bertmar, 1972 c). Besides the cell types described. Bertmar (1972 d, 1973) found a labyrinth cell, a cell type unique to vertebrates, in the olfactory epithelium of Baltic sea trout. He suggested that these cells probably help to maintain an optimum ion balance. which is of great ecological importance to this migrating species. In addition to the primary olfactory neurones, free nerve endings of the trigeminal nerve occur in the olfactory epithelium. These were observed in the teleosts, Suluelinus alpinus, Coregonus worthmarmi and Sulmo truttu by Aichel (1895). The functional significance of the trigeminal endings is not understood, but they have been shown to be odor-sensitive in the tortoise. Gophrrus polyphrnus, and in rabbits (Tucker, 1963). Stone et al. (1968) suggested a central regulatory control over olfactory afferent inputs in rabbits. While scattered accounts on the histochemistry of the mammalian olfactory mucosa have been published (Bourne, 1948; Baradi and Bourne, 1951, 1953; Bronshtein, 1965; Shantha and Nakajima, 1970). very little is known about the distribution of enzymes

2x3

OLFACTICININ FISH

and other substances and their functional significance in the olfactory epithelium of fish. Ojha and Kapoor (1972 b) demonstrated histochemically the activity of acid and alkaline phosphatase, lipids, phospholipids, glycogen, and acid mucopolysaccharides in the olfactory epithelium of the teleost Channa punctatus. According to Bronshtein (196.5), these chemically active substances show the mosaic nature of the spatial distribution in higher vertebrates, but it is less distinct in fishes.

ob

’ (A)’

(B)

FIG. 10.

Dorsal view of the brains and the olfactory organs of rainbow trout. Saltno gairdneri (A), and catfish, Parasilurus asotus (B). ce, cerebellum; mc. mesencephalon; ob. olfactory bulb; 01, olfactory lobe; on, olfactory nerve; or. olfactory rosette; ot. olfactory

tract.

2.3. OLFACTORY BULB AND TRACY The olfactory nerve fibers. fila olfactoria, which are centrally directed axons of the receptor cells, arise in the nasal mucosa and terminate in the olfactory bulb, where they make a special synaptic contact with the bulbar neurones in the glomerulus (spherical masses of dense nervous tissue). In the carp, the olfactory nerve consists of two main bundles, medial and lateral. The medial bundle is derived from the more rostra1 lamellae, while the lateral from the more caudal. The fibers of the two bundles often cross before they reach the bulb so that fibers from each reach all parts of the bulb (Sheldon, 1912). The fibers themselves do not branch until they terminate. These fibers are non-myelinated and extremely small. In Lota lota, for example. Gemne and Diiving (1969) estimated a mean diameter of 0.13 ,LL,the number of fibers per ,u2 to be 29 and a total number of 10’ axons in the olfactory nerve bundle. The length of the olfactory nerves varies greatly depending upon relative positions of the olfactory bulb in different species (Fig. 10). Three types of positions can be recognized: (1) the bulbs are located close to noses, the olfactory nerves are very short and a long olfactory tract is present (pedunculated). This type is found in all elasmobranchs, in Curassius, Ictalurus and other cyprinids, cobitids. silurids, etc; (2) the bulbs are close to the hemispheres of the forebrains so that long olfactory nerves are present (sessile). Examples are Anguillu, Esox, Sulm and the majority of teleosts; (3) the position

,M

7. J.

HAM

of the bulbs is intermediate between nose and forebrain. This has been found in ~url;crp.~ raninus (D6ving. 1967), and in Gyrnnothorur kid&o and Coryphaena hipptms (Uchihashi, 1953). In the young Carassius, the olfactory bulbs are closely located to the telencephalic hemispheres. As a fish grows. elongation of the skull results in an increase in the olfactory tract (Uchihashi, 1953; Schnitzlein, 1964). Thus. in a gar, Lrpi,so.stcw.s OSSYUS.measuring 60 cm, the olfactory nerves reach approximately 15 cm. It is generally recognized that the neurones do not increase in number by mitosis in the postembryonal stage. However. there has accumulated much histological evidence indicating regeneration after degeneration of the olfactory epithelium or nerve (for detailed discussion, see Takagi. 1971). In goldfish, histological and functional regeneration takes place after dissection of the olfactory nerve fibers (Zippel and Westerman, 1970; Zippel rt al.. 1970). After completion of regeneration. no difference in training and learning behavior between fish that has been operated on and control fish can be observed. The microscopic structure of the olfactory bulb is essentially similar in all vertebrates. In fishes the olfactory bulb is poorly differentiated and lamination is not as distinct as in higher vertebrates. Nieuwenhuys (1967) distinguishes five layers. from the outside toward the inside: (1) a layer of primary olfactory fibers. (2) a glomerular layer, (3) an external cell layer which often contains besides large mitral cells also smaller elements, (4) a layer of secondary olfactory fibers. and (5) an internal cell layer. In many forms there is a gradual transition between layers 3 and 4, and 4 and 5. The dominant feature of the bulb is the synaptic contact between the primary olfactory nerve fibers and dendrites of secondary neurones, mitral and stellate cells. Single mitral cells of fishes ’ have from one to five thick dendrites, each of which ends in one or more glomeruli (Allison, 1953; Nieuwenhuys, 1967). This is remarkably different from the mammalian mitral cells in which only a single main dendrite ends in each glomerulus. It is significant that the axon of a receptor cell does not terminate in more than one glomerulus, and that each glomerulus receives impulses only from a limited group of several olfactory receptor cells. Therefore. secondary neurones in fish would be in direct synaptic contact with a larger number of receptors than in mammals. Smaller cells in the bulb include stellate, spindle-shaped and granule cells. The long dendrites of the stellate cells extend fanwise toward the periphery. where their branches end in the glomerular zone. The axons of the stellate and spindle cells run also with the olfactory tracts to the hemispheres. Information from the olfactory bulbs is conveyed through the olfactory tracts to the basal telencephalic areas. In all vertebrates the olfactory tract consists of two main bundles, lateral and medial. Both bundles are further subdivided into several small bundles. The lateral olfactory tract is composed entirely of centripetal fibers. arising largely from mitral cells of the lateral part of each bulb. Most of the fibers of the medial olfactory tract originate from mitral cells far rostrally in the bulb. Some fibers run directly to the bpothalamus, while some cross in the anterior commissure. Centrifugal nerve fibers running to the olfactory bulbs have also been described (Sheldon, 19 12). The majority of the nerve fibers in the olfactory tract are myelinated and their diameters are less than 6.5 p in Lota lota. Most of the fibers in the lateral portion of the medial tract were smaller than 0.5 ~1 (Diiving and Gemne. 1965). In goldfish smaller diameters (0.2-1.8 p) were reported (Westerman and Wilson, 1968). The number of receptor cells in the epithelium is far greater than the number of secondary neurones in the olfactory bulb. In the burbot. the number of receptors and olfactory tract fibers was estimated to be 1I x IO“ and 10’ respectively (Diiving and Gemne. 1965; C&-me and Dbving, 1969). Therefore the ratio between receptor cells and the myelinated fibers in the tract is 1000: 1. This is the same order of the convergence ratio between receptor cells and mitral cells in the rabbit calculated by Allison and Warwick (1949). Secondary olfactory fibers from the olfactory bulbs extend to all parts of the forebrain except to the olfactosomatic area. The medial olfactory tract runs to the dorsal and medial olfactory areas, and some fibers run directly to the hypothalamus. The lateral olfactory tract runs to the dorsal and lateral olfactory areas. and some fibers may

2x5

OLI ACTION oh FISH

extend directly to the habenular nuclei. Most of the forebrain areas are inter-connected by tertiary olfactory fibers (Sheldon. 1912; Aronson, 1963).

3. Sensitivity of the Olfactory

Sense

All the available evidence points to a great acuity of the olfactory sense in many fish species both in the capability of detecting and discriminating natural and artificial odors. However. no systematic study has been attempted to establish threshold concentrations of various odorous substances in various fish species. As mentioned earlier, the existence of both olfactory and gustatory senses in fish had long been a controversial issue. This occurred because olfaction was believed to be mediated only by air and hence could not exist in fish. Nevertheless. no special attention seems to have been paid to this point in the past. This is well represented in the fact that various investigators have utilized as olfactory stimuli for fish the same kinds of chemicals which are odorous to humans. Moreover. most of the attempts to establish threshold concentrations of various odorous substances in fish has been done primarily through conditioning or electrophysiological technique. The data thus obtained are therefore not always consistent. Let us first consider some of the topics in the earlier studies. As early as 1930. Bull was able to condition blennies. BI~li.s guttortrgine and B. pholis, to respond towards seawater extracts of natural foods. such as h’cwis and Mytilus. The threshold for perception was between concentrations slightly less than 0.000375 and 0.00075’jj,, of the weight of living food substances in sea-water. However, these fishes were unable to respond to an artificial olfactory stimulus such as musk at the final concentration of about 5 x 10e4 mg/l. Although Bull himself regarded the seawater extracts of food substances as gustatory stimulants. he was probably the first to attempt to determine the limits for olfactory perception of biologically significant substances. Most of the attempts to establish threshold concentrations of various odorous substances in fish are of relatively recent date. Some of the published data are listed in Table 2. Although these data are somewhat scattering, one of the most extensive study is probably that by Teichmann (1959). who determined threshold concentrations for various odorous substances in eels through conditioning experiments. As shown in Table 2, the threshold concentrations obtained by Teichmann for these highly odorous substances are extremely low. 3 x lo- ” for /J-phenethyl alcohol, for example, is equivalent to 1.77 x lo3 molecules/cm3. If one can assume that the space in the nasal cavity of the small eels employed by Teichmann is about 1 mm3. and that the olfactory response can be induced with a single puff of stimulant solution into the nasal cavity. no more than a few molecules of the stimulus substance could have been in contact with the receptor sites before any appreciable response was induced. This threshold concentration is comparable to that of the best studied pheromone of the moth. Bomh~s tori. in which males can detect their females sexual attractant. bombykol, at a concentration of 3.1 x lo4 molecules/cm3 (Kaissling, 1971). Using beef-heart extracts. Pfeiffer (1969 a) studied the relative sensitivity of the olfactory organs of two species of tropical African polypterid fishes. Calunzoichthys could detect beef-heart extracts at concentrations as low as lO_ “’ to IO- ’ I g:l. while the thresholds for Pol~~ptrrus were IO- ” to IO I4 g, I. Thcsc threshold values were compared with those for Phoxir~~rs that detected the extracts at concentrations of IO-” to IO-‘g/l. Pfeiffer claims that the olfactory acuity of Po/~~pte~zr.ssurpasses that of P~OS~HUS by several factors of 10, indicating the Polypteridae being macrosmatic. Recent electrophysiological studies have revealed that a number of chemical substances which are strongly odorous to humans are non-stimulatory to fish, and that a large number of substances that are regarded as odorless are highly stimulatory to fish (Sutterlin and Sutterlin. 1971; Suzuki and Tucker, 1971 ; Hara. 1972; Hara et al., 1973). Compounds that induced little or no stimulatory responses in the olfactory bulb when the nares of rainbow trout were infused at the concentration of 10p4~ are listed

2X6

T. J. HAKA TAHL~: 2.THRESHOLDVALL.LS OF VAKKWS SUISIAWLSI-ORSEMI.FISHSPWILS Species

Substance

Pho.uinus phovinus

Eugcnol b-Phenethyl alcohol

Hyhorhynchus

notatus

Phenol p-Chlorophenol

Oncorhynchus

kisutch

Morpholine Amino acids p-Phenethyl alcohol I-Menthol Citral Terpineol Eugenol p-Ionone X-Ionone b-Phenethyl alcohol Amino acids Benzene Mononitrobenzene 1.3-Dinitrobenzene 1,3,5-Trinitrobenzene Phenol Resorcinol Phloroglucinol Acetic acid Butyl alcohol Morpholine Eugenol Amino acids Amino acids Amino acids Amino acids

Anguilla anguilla

Salmo gairdneri Leuciscus

Ictalurus

rutilus

catus

and I. melas Oncorhynchus nerka Salnw salar Ictalurus catus Salvelinus fontinalis Coregonus clupeaformis

Threshold 1.1.7 X IO’

Ncurath (1949)

I :2.? x IO’ I :6.7 X 10: