A Scanning Electron Microscopic Study of the Sacculus and Lagena in the Ears of Fifteen Species of Fishes

ARTHUR N. POPPER Department of Zoology and Laboratory of Sensory Sciences, Unioersity of Hawaii, Honolulu, Hawaii 96822 and Kresge Hearing Research Institute, University of Michigan Medical Center, Ann Arbor, Michigan 48109

ABSTRACT The ulstrastructure of the saccular and lagenar maculae were studied in 15 species of teleost fishes, using the scanning electron microscope. Particular attention was paid to hair cell orientation patterns, composition of the ciliary bundles on the hair cells, hair cell distributions, and supporting cell types. The hair cells on both otolithic organs are divided into several groups with all of the hair cells in each group oriented in the same direction. The posterior region of the saccular macula in all species had dorsally oriented hair cells on the dorsal half of the macula and ventrally oriented hair cells on the ventral half. The cells on the anterior end of the macula were oriented anteriorly and poste- riorly, with the posterior group, in most species, being on the dorsal half of the anterior region of the macula. There was considerable inter-specific variation upon this basic pattern. Inter-specific variation on the lagenar macula was considerably less than on the saccular macula. The basic pattern in all of the species includes one dorsal cell group and one ventral cell group. There are four more-or-less discrete ciliary bundles, each varying in the rela- tive size of the kinocilia and . Intermediary forms were also observed, making it difficult to differentiate ciliary bundles in some instances. It was ap- parent, however, that several of the ciliary bundles were found in particular macular regions.

Interest in sound detection by teleost fishes capabilities of the ear (if any), and responses has increased markedly over the past several of the ear to sound source direction. decades (see reviews by Lowenstein, '71; Pop- The teleost ear (fig. 1) consists of three per and Fay, '73; Tavolga, '71) and we now and associated ampullary have a large body of data on detection mecha- regions, and three -containing vesicles. nisms and capabilities, as well as on sound Two of these vesicles, the sacculus and lagena, processing at several levels of the auditory are associated with audition (von Frisch, '36) systems. One area that is now of particular in many species, although their specific con- interest is the role of the in sound tributions are still unknown (see Fay, '75; detection and processing (e.g., Enger, '63, '73; Furukawa and Ishii, '67). Enger et al., '73; Fay, '75; Fay and Popper, '74, Each otolith-containing vesicle has a senso- '75; Furukawa and Ishii, '67; Sand, '74). How- ry hair cell covered epithelium, or macula, in ever, the meaning of inner ear data, in terms close proximity to a single dense calcareous of the function of the ear, is still unclear be- otolith. The macula and otolith are separated cause of a lack of detailed knowledge of the by a thin that may hold morphology and ultrastructure of the audito- the otolith in place on the macula and provide ry regions. These data are needed before a mechanical coupling between the otolith and detailed analysis can be made of responses of sensory hairs. The actual mechanisms for re- the inner ear to sound stimulation, analytic sponse to acoustic stimulation in the ear of

J. MORPH., 153: 397-418. 397 398 ARTHUR N. POPPER

Ahbreoiations

aa, anterior ampulla k, mu, utricular macula ap. posterior ampulla rn, microvilli pl, lagena ca, anterior canal ms. saccular macula s. sacculus cp, posterior canal st. stereocilia

Fig. 1 Drawing (modified from Retzius. 1881, of a medial view of the right ear of Snlrno salur. fishes have not been directly studied, but it generally contains the typical 9 + 2 filament has been suggested that a sound field would pattern found in many metazoan flagella. cause the body of the fish (and the sensory Maximum hair cell depolarization occurs macula) and the considerably denser otolith when shearing is along the cilia towards the to move out of phase with one another, result- kinocilium, and the response is inhibited ing in bending of the cilia on the hair cells when stimulation is reversed 180". Stimula- (e.g., van Bergeijk, '67a; Wever, '69, '71). tion off this axis gives a response whose level The sensory hair cells consist of a cell body is a cosine function of the stimulus direction with a series of cilia on the apical surface. A relative to the axis of maximum response number of workers have provided evidence (Flock, '65, '71). that bending or shearing action on the cilia is Since the hair cells of the inner ear are responsible for the electrical response from functionally polarized, it is likely that this the hair cells (e.g., Bekesy, '60; Lowenstein could significantly affect auditory capabilities and Wersall, '59). Each ciliary (or hair) bun- and mechanisms. In fact, experiments have dle consists of many (30-50 in fishes) stereo- shown that the goldfish (Carassius auratus) cilia which are embedded in a thick cuticle sacculus has hair cells oriented in opposite di- that lies just under the apical cell membrane. rections, resulting in hair cell groups that re- A single, eccentrically placed kinocilium orig- spond to opposite phases (compression and inates inside the cell at the basal body and rarefaction) of a stimulating signal (Furu- penetrates the cell membrane. The kinocilium kawa and Ishii, '67; Piddington, '72). Recent SACCULUS AND LAGENA IN FISHES 399 work with perch (Perca fluviatilid and had- data are available. Furthermore, while we dock (Melanogrammus aeglefinus) has also have a limited idea regarding the functional demonstrated that different regions of the significance of bi-directionality in the saccu- saccular macula have their best responses to lus of the goldfish (Furukawa and Ishii, '671, it stimulation in different directions (Enger et is difficult to extrapolate from this to other al., '73; Sand, '741, again suggesting that in species. The present study is directed towards the inner ear there is a functional polariza- expanding our understanding of the ultra- tion which may be involved with detection of structure of the teleost sacculus and lagena sound source direction. through a scanning electron microscopic Data on hair cell orientation patterns in the (SEMI study of these organs in 15 species. inner ear of fishes show significant correla- While these data are not in any way repre- tions with the physiological data. Furukawa sentative of all (or even most) taxonomic and Ishii ('67), partially on the basis of the groups, they do give a fairly comprehensive biphasic response to sound stimulation in the indication of the variation in ultrastructure eighth nerve, hypothesized the presence of op- and enable us to suggest some general ideas positely polarized hair cell groups in the gold- on the auditory function of the teleost ear. fish sacculus. This hypothesis has been cor- roborated by Hama ('69) and Platt ('771, who MATERIALS AND METHODS have shown that most hair cells on the dorsal Data were obtained on the ultrastructure of half of the saccular macula in goldfish are the and lagenae of 15 species of fish polarized dorsally (kinocilium on the dorsal representing 10 families. The species were side of the cells) while cells on the ventral selected on the basis of availability, although half of the macula are polarized ventrally. Bi- attempts were made to use species that directional orientation has been reported for showed taxonomic and ecological variability five species of catfish (Jenkins, '74), a group in and also to include a number of species for the same superorder (Ostariophysi) as the which sound detection and/or production data goldfish, and for the burbot (Lota vulgaris) by are available. Table 1 lists the species used as Wersall et al. ('65). well as data on number of , sizes, and, The data for the Ostariophysi and Lota where available, citations relating to acoustic closely resemble the bi-directional orientation behavior for these or related genera. pattern found in the saccules of tetrapods Animals were anesthetized with MS-222 (e.g., Lewis and Li, '75; Lindeman, '69; (Sandoz) until they were immobile and then Spoendlin, '65) and the elasmobranch Raja one of two procedures was followed. Animals clauata (Lowenstein et al., '64). However, from 5-20 cm in standard length were placed these data differ somewhat from recent data into a special dissection holder which permit- for a salmonid fish, the lake whitefish, Cor- ted respiration during dissection, thus en- egonus clupeaformis, (Popper, '761, the cod, abling a long dissection time. Specimens out- Gadus morhua, (Dale, '76), and two species of side of this range were decapitated, since they flatfish, Pleuronectes platessa and Limanda did not fit the available holder. In all cases limanda, (Jsrgensen, '76). While the posterior the cranium was opened dorsally, the brain re- in Coregonus and the flatfish have dor- moved by dissection and aspiration, and the sal and ventral hair cell groups, as in other ears exposed. The sacculi in most species were species, the anterior-dorsal saccular quadrant then opened dorso-laterally with a sharp has posteriorly oriented cells and the anteri- blade, and chilled fixative (4%glutaraldehyde or-ventral quadrant has anteriorly oriented in S-collidine buffer for fresh water species; cells. The pattern in Gadus is similar except the same buffer with 4% sucrose for salt water that horizontally oriented cells are also found animals) was perfused into the ears and cra- on the posterior end of the saccular macula. nial cavity for four to eight minutes. Fixation The data now available in the literature always started within three minutes of death indicate that there is considerable inter- in the decapitated animals. Most of the head specific variation in the hair cell orientation tissue was then dissected away (keeping the patterns of the sacculus (and to a lesser ears moist at all times) and a block containing degree the lagena). However, the extent of the ears was placed into a container of fixative this variation cannot yet be evaluated due to for at least 48 hours with one fixative change the small percent of teleost species for which about four to six hours after dissection. 400 ARTHUR N. POPPER

TABLE 1

Species used in studies of ultrastructure of the teleost ear

Standard length Number of References on sound production (PI - Spec LBS Family lin rnm) animals used or sound detection (DJ Salmo salar Salmonidae 240-260 4 Van der Walker, '67 (D) I Saluelinus namaycus Salrnonidae 240-260 5 Van der Walker, '67 (D) ' Ceratoscopelus warmtngi Myctophidae 20.30 2 Marshall, '67 (P,D)I Diuphus brachycephalus Myctophidae 20-30 2 Marshall, '67 (P,D) ' Lanpanyctus sp. Myc tophidae 20-30 2 Marshall, '67 (P,D) I Adioryx xantherythrus 60-70 5 Popper et al., '73 (D) I; Salmon, '67 (PI I; Tavolga and Wodinsky, '63 (D) '; Nelson, '55 (D) I Myripristis berndti Holocentridae 60-70 2 Popper et al., '73 (D);Salmon, '67 (PI; Tavolga and Wodinsky, '63 (D) '; Nelson, '55 (D)' Aequidens pulcher Cichlidae 40-50 3 Myrberg et al., '65 (PI I; Tavolga, '74 (D)I; Fay and Popper, '75 (D) * Tilapia macrocephala Cichlidae 40-50 5 Myrberget al., '65 (PI I; Tavolga, '74 (D);Fay and Popper, '75 (D) Chaetodon miliaris Chaetodontidae 40-50 3 Dascyllus albisella Pomocentridae 40-50 3 Myrberg and Spires, '72 (P) Lagodon rhomboides Spari d a e 40-50 3 Caldwell and Caldwell, '67 (PI, Tavolga, '74 (D) Lepomis macrochirus Centrarchidae 40-50 5 Gerald, '71 (PI Perca flauescens Percidae 40-50 3 Sand, '74 (D) Xenotoca eiseni Goodeidae 20-30 5

1 Data fur species studied with the SEM or a closely related species

After fixation the tissue was washed in membrane could remain on the macula, even Millonig's buffer. The ears were then dis- after careful dissection, interfering with de- sected from the large block of tissue, stained termination of hair cell orientation. In order with 1%osmium tetroxide for one hour, and to eliminate these problems, some of the os- moved to 70% ethanol for storage and further mium stained specimens from each species dissection. Final preparation for the SEM were placed in an ultrasonic cleaner (Convin, involved opening the membranous otolithic '77) for 30 seconds to 2 minutes (the length of capsules and gently pulling the away time varied for each species and was deter- from the sensory maculae. In many cases the mined by trial and error). This treatment re- otolithic membrane came away from the sen- moved the excess otolithic membrane and sory epithelia along with the otolith, thus pro- often "blew" the cilia off the sensory cells viding data on the extent of the membrane in while leaving the microvilli in place on the relation to the otolith and macula. In some support cells. In these cells the bases of the cases the otolithic membrane remained on the stereocilia were visible in the cuticle and macula and then it was either removed by dis- showed up as shallow pits or as white spots section or with a cleaning procedure (see while the kinocilium was represented either below), or left in place for investigation of the by a thicker stub or, in most cases, by a deep relationship between the two structures. hole in the cell membrane where the ki- After dissection the tissue was C0,-critical- nocilium penetrated the cell to the basal body. point-dried, using amyl acetate as the inter- It was also found that sonication for shorter mediary fluid, placed on an aluminum stub, periods of time (10-20 seconds) in many spe- coated with about 150 A thickness of gold, and cies would frequently remove most of the viewed with a JEOL JSM-U3 scanning elec- otolithic membrane but leave the cilia intact, tron microscope. thus permitting studies of cilia structure. In order to study hair cell orientation it was Hair cell orientation patterns were deter- necessary to determine the position of the mined in several specimens of each species kinocilium relative to the stereocilia but this whenever a sufficient number of specimens was sometimes difficult when the cilia were in were available. Low power photographs were place since they could be bent or rotated taken and orientation of hair cells in each during fixation. In addition, excess otolithic region of the macula, as observed on the SEM, SACCULUS AND LAGENA IN FISHES 40 1 were noted on the photos. Cracks and small the macula and it often has a shallow indenta- piece of “debris” on the tissue were used as tion for the macula. In many of the species landmarks in the mapping procedure. (with the exception of the salmonids and pos- sibly the myctophids) there is an anterior RESULTS hook-like lateral projection that swings There were many similarities in inner ear around the anterior edge of the lagenar morphology and ultrastructure in most of the macula into an area containing non-sensory species studied. The features held in common epithelium. will be described first and then followed with The size of the otoliths, relative to the data from each of the species to give varia- volume of the , varies tions from the general pattern. in different species. The lagenar otolith never The sacculus and lagena are found in the fills the membranous chamber, and in most in- cranial cavity at the level of the medulla. The stances it only covers the ventral three- sacculi in the species studied are partially fourths to five-sixths of the macula sensory recessed in the floor of the cranial cavity with region. There is generally a large space lateral the posterior end of the sacculus and all of the to the otolith. The extent to which the sagitta lagena covered by overlying bone. In some spe- fills the saccular chamber is considerably cies, such as the two salmonids and Perca, the more variable. In several species, including membranous labyrinth lies loosely in the bony the cichlids, holocentrids and Chaetodon, the chamber while in other species, such as the otolith almost totally fills the chamber and cichlids, the sacculus and lagena were touches the lateral walls. This is contrasted to difficult to remove since the anterior ends of the condition in salmonids where the otolith the sacculus are partially covered by overly- fills perhaps 75%of the labyrinth volume. The ing bone. most extreme condition observed is in one spe- The most distinct condition was found in cies of myctophid (Lampanyctus sp.) where the holocentrids where the membranous laby- the otolith only fills about 25%of the saccular rinth appeared to be “tied” to the bony walls chamber. of the cranial depression. In both species the A thin otolithic membrane lies between the most anterior tip of the sacculus was also cov- otolith and each macula and is a continuous ered by bone and the membranous chamber sheet over the whole macula, even in regions appeared to be physically connected to the not covered by the otolith. While detailed in- bone at this point. Consequently, it was essen- vestigations of the otolithic membrane were tially impossible to remove the membranous not made, it was clear that membrane fixed labyrinth intact, and it had to be studied in for use in the SEM from several species (e.g., situ before dissection. Adioryr, salmonids, myctophids) have a series The precise position of the lagena varies of “holes.” In several instances a single ciliary with respect to the sacculus, but it is always bundle could be seen in each of these holes found in the dorsal-posterior quadrant of the (fig. 2). saccular chamber. The continuity between the sacculus and lagena vary. In salmonids Hair cell orientation patterns: sacculus the two structures are in a single enlarged In the following discussion the orientation chamber (see fig. 1 and Popper, ’761, while in of the hair cells will be described relative to most other species there is only a small open- the long axis of the appropriate macula rather ing between two discrete chambers. than the axes of the fish. This is necessary The sacculi in all species were oriented at since the otolithic regions are not necessarily an angle to one another with the anterior oriented in the same way in different species edges spread further apart than the posterior and, as a result, discussion of orientation rela- so that the saccules, when viewed from above, tive to the fishes’ axes would be confusing in had a V-like appearance. inter-specific comparisons of orientation pat- terns. However, the general position of each of Otolith and otolithic membrane the macular regions in the ear is similar to A single dense calcareous otolith lies close the same region shown in figure 1. to the sensory macula in both sacculus and In each species the saccular hair cells lagena. The saccular otolith (sagittal has a are divided into four discrete orientation deep medial groove into which the macula fits. “groups” (e.g., fig. 3) with all hair cells in each The lagenar otolith is slightly smaller than group oriented in the same direction. In all 402 ARTHUR N. POPPER

cases the posterior region of the macula has a anterior side of the lagenar macula are ori- dorsally-oriented hair cell group on the dorsal ented dorsally, while the cells on the posterior half of the macula and a ventrally-oriented side are oriented ventrally. There may be up group on the ventral half of the macula (fig. to 20%variability in intra-group hair cell ori- 3). These cells generally change orientation to entation. This variation is most noticeable at follow macula curvatures. the transition zone between dorsal and Two orientation groups are found on the an- ventral groups where the cells may be ori- terior region of the saccular macula and these ented almost directly towards the opposite show most of the variation in orientation pat- cell group (also see Popper, ’76). terns in the species I have studied. In most The orientation patterns in the three species the cells on the dorsal-anterior quad- lagena types are basically similar (figs. 4-6). rant are oriented posteriorly and in the One type, tentatively called type A (fig. 4)was ventral-anterior quadrant oriented anteriorly previously described for Coregonus (Popper, (e.g., fig. 3; for exceptions see discussion of ’76) and the same pattern is found in the two myctophids and Myripristis). salmonids reported here. The lower two-thirds The transitions between groups of hair cells of this macula is of relatively constant width, vary to some degree, although in all species while the top third (the dorsal “arm”) is some- the change from dorsal to ventral groups or what narrower. The hair cells on the lower from anterior to posterior groups occurs over region are all oriented relative to the fish‘s several (generally two to three “rows” con- dorsal-ventral axis and stay oriented in the taining hair cells from both orientation same direction even as the macula curves. The groups. Transitions from posterior to dorsal cells in the dorsal arm, however, are oriented groups and from anterior to ventral groups parallel to the macula curvature. There is also are somewhat more complex in many species less inter-cell orientation variability than in and may occur over five to 10 hair cell rows in the lower macula region, and transition zone some species (e.g., salmonids) or in one or two cells are oriented in the same direction as the rows in others !myctophids). Where multiple rest of the cells in each group. While the dor- transition rows are found, the cells are often sally oriented cell group is contiguous from oriented at approximately 45” between the the ventral region into the dorsal arm of the orientation patt.erns of the two major groups macula, the ventrally oriented cells terminate that they separate. just below the start of the arm and then “start up” again in the arm. The otolith asso- Lagena hair cell orientation patterns ciated with the type A lagena covers the lower For reasons that are not clear it was con- macula region, although the otolithic mem- siderably more difficult to prepare totally brane continues into the dorsal arm covering satisfactory lagenar maculae of several of the all of the hair cell regions. species than it was to prepare saccular ma- The type B macula (fig. 5) has a wide terial. In these species (e.g., cichlids and ventral area and a narrow dorsal arm. The holocentrids) various fragments of “debris,” cells on the ventral half of the macula turn such as otolithic membrane, remained firmly along with the margins and face almost per- attached to the macula and no amount of dis- pendicular to the cells of the same group on section and/or ultrasonic cleaning could re- the dorsal half of the macula. Furthermore, move the tissue without destroying the senso- the ventral group may fill in below the dorsal ry epithelium. However, in most cases it was cells. Transition between dorsal and ventral possible to get a composite picture of the groups is similar to the transition groups in lagena orientation patterns by studying a type A lagena. The otolith has an anterior number of organs, each having complete hair lateral hook lying over non-sensory epithelial cell patterns in one or more regions. cells. Considerably less inter-specific variation in The type C macula (fig. 6)has cells oriented hair cell orientation pattern is found in the dorsally and ventrally, as in the other la- lagena than the sacculus. I have found three genae, but both cell groups tend to curve as distinct, but similar lagena patterns dis- the macula curves. Transition between cell tributed among the species (except for the one groups is similar throughout the macula with myctophid) for which lagena material was ’ This word IS used nominally and. in fact. hair cells in all of the available. In all species the hair cells on the sensory regions of the teleust ear are not allgned in discrete rows. SACCULUS AND LAGENA IN FISHES 403

Fig. 3 SEM of the whole saccular macula from Leopomis macrochzrus showing the orientation of the hair cells in each of the macula regions. In this and the following illustrations the tips of the arrows indicate the side of the hair cells on which the kinocilium is located in this region. 404 ARTHUR N. POPPER

HAIR CELL TYPES 0 0TYPE I 0 2 LA 3 4

- @ Salmonid 0.2 mm @ Chaetodon milleris Oxm @ Adioryx xontherythrus 0.2- mm Fig. 4 Hair cell orientation patterns and ciliary bundle distribution patterns on the lagenar macula of Salmo and Saluelinus. The arrows point to the direction of orientation of the kinocilia on the hair cells in each region and the dashed lines shows the transition zones between major hair cell orientation groups. The distributions of the various ciliary bundle types are indicated in the figures. The key of ciliary bundle types is for this and all subsequent figures. Fig. 5 Lagenar macula of Chaetodon. Explanation as in figure 4. Fig. 6 Lagenar macula of Adioryx. Explanation as in figure 4.

cells close to the transition zone oriented of the sacculus and lagena and may consist of towards the opposite cell group. The otolith one to ten “rows” which then merge into other covers all but the top one-sixth of the macula cell types (fig. 9). In some instances the F2 and has at the anterior end a laterally-di- bundle is also found on cells making up larger rected “hook” which bends around the anteri- regions, as seen in both cichlids (fig. 8). or edge of the macula and lie? over non-senso- The type F3 ciliary bundle has a long ry epithelial cells. No otolithic membrane was kinocilium, several stereocilia of almost equal associated with the area under the hook. length, and several shorter stereocilia (fig. 10). The distribution of the type F3 ciliary Hair cell ciliary bundles bundle is less well defined than types F1 and Several different types of ciliary bundles F2, but it is often found on the lagena, in were found on the hair cells and each appears areas not covered by the otolith. It IS also the to be dominant in specific macula areas. How- dominant ciliary bundle found on the myc- ever, while these types are fairly discrete, tophid sacculus. there are intermediary forms making it In some instances a type F4 ciliary bundle difficult, at times, to discriminate between has been found. This closely resembles type F1 various bundles. The most common ciliary except that the cilia are all several times bundle, to be called F1 (fig. 71, consists of a longer than those in the F1 bundle. No specific series of short, graded stereocilia, and a distribution pattern was found for the type F4 kinocilium that is 1.1 to 2.0 times longer than cilia and this may be partly due to the fact the longest stereocilia. The cells containing that these closely resemble type F1. F1 ciliary bundles are widely distributed on Supporting cells the saccular macula of most species. Type F1 bundles are also found in the lagenae of sever- The sensory hair cells in all maculae were al species as well as in the central region of surrounded by several microvilli-covered sup- the utricular macula (Popper, unpublished). porting cells. Several different types of sup- The type F2 ciliary bundle has short graded porting cells were found in the various species or non-graded stereocilia and a kinocilium Vanous authors have been using identical systems to designate that is 3 or more time longer than the longest cilia patterns In different species and this could eventually lead to confusion as more patterns are discovered Tentatively I am using the stereocilia (fig. 8). The type F2 bundle is fre- designation F for the patterns found in fishes to eliminate some con quently found on cells located at the borders fusion in later discussions SACCULUS AND LAGENA IN FISHES 405

Fig. 7 SEM from Adioryrshowing type F1 ciliary bundles all oriented in the same direction. Microvilli cln the support- ing cells between the hair cell bundles are typical of those in most of the sensory regions. Fig. 8 SEM from Aeguidens showing a group of type F2 ciliary bundles, found in the ventral portion of the saccular macula of the cichlids (see fig. 19). All cells in this region are oriented anteriorly. Fig. 9 SEM of a border region in the saccular macula of Tilapin showing type F2 ciliary bundle at the macula border. More medially these grade into the type F1 ciliary bundles. Fig. 10 Type F3 ciliary bundle from Lampanyctusshowing opposite polarization of the border between posteriorly and anteriorly oriented hair cells. studied. The most common cell was found on side of the sensory region. It was also found the sensory maculae (X in fig. 11) and con- interspersed with regular supporting cells in tained a relatively even distribution of short the central sensory region on the saccular microvilli (also figs. 7-10). A second type of macula in Myripristis. supporting or perimacular cell (Y in fig. 111, The areas outside of the lagena and saccu- generally found in regions outside of the lar sensory regions also have extensive areas macula, has a relatively featureless surface of non-sensory epithelium which generally except for a ring of microvilli around the cir- have high microvilli density (fig. 11).Many of cumference of the cell. A third type of sup- these cells have a single central cilium which porting cell (Z in fig. 12) has a bulbous appear- was somewhat longer than the other mi- ance (which is conceivably a fixation artifact) crovilli. This type of cilium was not found in and a high microvilli density. This cell is out- supporting cells within the sensory region, 406 ARTHUR N. POPPER

Fig 11 SEM from Crratoscopelus showing two types of supporting cells in a region of the macula just ventral to the sensory region. One cell (XI has an even distribution of microvilli, while another cell (Y)has an almost featureless surface. Fig. 12 A third type of supporting cell from Myipristis having a high microvilli density.

U

- Salmonids 0.2 mm Fig. 13 Saccular macula hair cell orientation pattern and hair cell distribution patterns for both salmonid species. For explanation of figure and cell distribution pattern see figure 4. but similar cells have been reported in the rior tip of the macula. The sacculus and sensory region of the lagena of the goldfish lagena in both species have type F2 ciliary and loach (Saito, '73). bundles on the borders. Type F3 ciliary bun- dles are found on cells located on the dorsal Hair cell Orientation patterns arm of the lagena and there is a small region In the following discussioil the data will where these bundles are found on the anterior supplement the general descriptions already and posterior tips of the sacculus. The sagitta given. covers all but the anterior and posterior tips of the macula. a. Salmo and Salvelinus (fig. 13) The sacculus and lagena patterns in these b. Chaetodon and Dascyllus (figs. 14, 15) salmonids closely resemble the organs in the The saccular hair cell orientation patterns lake whitefish (Popper, '76). The sacculus has are similar in these Hawaiian reef species. four orientation groups with the anterior-ori- The sagitta in both species covers the whole ented group covering the ventral-anterior macula and fills the membranous chamber. quadrant as well as several rows on the ante- The posterior two-thirds of each macula lies SACCULUS AND LAGENA IN FISHES 407 on the vertical plane of the fish, while the an- other species. The posterior end of the sac- terior-dorsal quadrant bends laterally so that culus in Lagodon curves ventrally and its an- it is about 30" from the fish's vertical plane. terior dorsal quadrant bends laterally about The hair cells are organized into four quad- 30%,much as in Chaetodon and Dascyllus. rants, although the most anterior tip of the The lagena in Lagedon is type C while in macula has cells that are oriented posteriorly. Lepomis and Perca it is type B. F2 ciliary bun- Both species have a type B lagena. Chae- dles are found on the marginal cells in all todon, however, has one to three rows of dor- lagenae. Most of the cells in Lagodon have sally oriented hair cells on the ventral and type F1 ciliary bundles except for about ten posterior margin of the macula (fig. 6), even rows on the anterior margin of the macula though the bulk of the cells in these regions which are type F3. Most of the ciliary bundles were oriented ventrally. The dorsal cells end on the lagena in Lepomis are type F3 while about three-fourths of the way up the poste- those on the dorsal arm of Lagodon are F3 rior edge of the macula. These cells were pres- with the rest of the lagena being type F2. ent in lagenae in the three specimens I studied but in no other species. The marginal d. Tilapia and Aequidens (fig. 19) ciliary bundles on the lagena macula (as on The ears in both cichlids are very similar to the sacculus) of both species are type F2 with one another. The saccular maculae lie parallel the bulk of each lagena being covered by type to the vertical plane in both species and they F3 bundles. are completely covered by the sagitta. The saccular hair cells are in the four basic orien- c. Perca, Lepomis, and Lagodon tation groups, although the anterior-oriented (figs. 16-18) cells move into, and fill, a region that would The sacculi in these species are basically normally be occupied by posterior-oriented similar to one another with the maculae cells. The cells just dorsal and ventral to the divided into four quadrants. The anterior 33% transition zone in this region are oriented par- of the macula in Perca (an area not covered by allel to the transition zone. In addition to the otolith) is covered with type F3 ciliary marginal F2 ciliary bundles, there is a small bundles while this area has F1 bundles in the region of type F3 ciliary bundles at the poste-

D LA

- Chaetodon mil iaris 0.2 mm

Dascyl lus albisella 0.2 mm Fig. 14 Saccular macula of Chaetodon. For explanations see figure 4 Fig. 15 Saccular macula of Dascyllus. For explanations see figure 4. 408 ARTHUR N. POPPER D LA

8 Perca f lavescens 0.2 mm

- 0 Lepomis macrochirus 0.1 mm D

Lag adon rhomboides OZm

0 LA

Tilapia~- macrocephola 0.2 mm Flg 16 Saccular macula of Perca For explanations see figure 4. Fig 17 Saccular rnacula of Lepomis. For explanations see figure 4. Fig 18 Saccular macula of Lagodon. For explanations see figure 4. F1g 19 Saccular macula of ’i”tlnpza. For explanations see figure 4. rior tip of the maculae. Marginal F2 cells also oriented at almost a 60” angle to the posterior cover a small region on the ventral-anterior region of the sacculus. region of the maculae (fig. 19). Both species have type B lagenae. Most of e. Xenotoca (fig. 20) the ciliary bundles are type F1 except for a This Mexican molly is the only species series of type F3 in the ventral posterior studied that did not have a noticeable widen- margin of the maculae. The lagena itself is ing of the anterior end of the saccular macula. SACCULUS AND LAGENA IN FISHES 409

- @ Xenotoca eiseni 0,I mm 0 LA

I.

- 0 Adio r y x xa nt her yt hr us 0.2 mm

0 LA

support

-1 -----#’4 -,b--- .. t f

Myripristis murdjans

Fig. 20 Saccular macula of Xenotcca For explanations see figure 4. Fig. 21 Saccular macula of Adioryx. For explanations see figure 4. Fig. 22 Saccular macula of Myripristis. For explanations see figure 4.

However, there is a slight “bulge” on the dor- region. The bulk of the ciliary bundles in the sal-anterior region of the macula - a condi- ventral cell group, as well as bundles in both tion seen in no other species reported here. groups located in the dorsal arm, are type F3 The saccular otolith covers the posterior 80% while the rest of the dorsal group are type F1. of the macula. Transition from posterior to dorsal cell groups occurs in the bulge region, f. Adioryx (fig. 21) with transition cells oriented 45’ between the This species of squirrelfish (subfamily Holo- two cell groups. centrinae) will be treated separately from the The type C lagena has marginal F2 ciliary related Myripristis since there are significant bundles, except in the posterior ventral differences in their sacculus patterns. How- 410 ARTHUR N. POPPER ever, lagena patterns are very similar in both two hair cell groups then become equal in size species. further anteriorly. Type F3 ciliary bundles The surface of the saccular macula in start about half-way from the posterior end of Adioryxlies parallel to the vertical axis of the the macula and gradually replace the F1 bun- fish and the macula is at a 30’ angle to the dles further anteriorly. In this same region horizontal plane. Consequently, the anterior bulbous supporting cells (fig. 11) become end is higher than the posterior. The otolith interspersed among the more typical support- covers all but the dorsal-anterior region of the ing cells within the sensory area. macula. The hair cells are arranged in the Another striking feature of this sacculus is four basic orientation patterns found in most the anterior region which has anterior-ori- other species. The anterior group fills the ented cells facing posterior cells and another ventral anterior region and anterior tip of group of anterior cells at the most anterior the macula. The transition zone between an- region of the macula. The more posterior teriorly and posteriorly oriented hair cells group of anterior-oriented cell is contiguous changes direction, much as the ventral mar- with the ventrally-oriented cells. gin of the macula curves. Consequently, some The enlarged dorsal perimacular region of t.he hair cells in the anterior and posterior contains supporting cells typical of those in cell groups are oriented directly towards one the sensory regions with some bulbous cells another. interspersed, especially in regions closer to The bulk of the saccular ciliary bundles are the ventral hair cell regions. type F1 except for marginal F2 cilia. Gen- erally there are one to four rows of F2 cells. h. Myctophids (figs. 23-25) The lagena in Adioryx is type B with the The sacculi in the three species of meso- otolith covering all but the dorsal tip of the pelagic lanternfishes studied are similar to macula. The most striking feature of the one another. The sagitta in Ceratoscopelus and macula in this species, and in Myripristis, is Diaphus almost fills the membranous cham- that the cilia are longer than in the other spe- ber but does not cover the anterior 25%)of the cies studied. The marginal ciliary bundles are macula. The sagitta in Lampanyctus is as deep similar to type F2 bundles, while the rest of as that in the other species, but it does not the macula is covered with what appear to be cover the anterior 50%of the macula. There is cells intermediate between types F3 and F4. some indication that a cupula-shaped struc- ture that lightly stains with osmium lies over g. iwyripristis (fig. 22) the otolithic membrane anterior to the otolith This squirrelfish is a member of the holo- in all of the species. However, this could not centrid subfamily Myripristinae. While the be verified due to the limited number of speci- lagenar macula closely resembles that of mens available. Adioryx, the saccular macula is significantly The saccular macula of each species has the different from that in other species studied. four typical orientation groups. The posterior The macula is about twice the length of that sacculus has dorsally and ventrally oriented in Adioryx, although the specimens used were cell groups, while the anterior and posterior approximately the same standard length. The groups occupy somewhat different regions long axis of the macula is at a 45’ angle to the than in other species. The posterior orienta- long axis of the fish’s body and its surface lies tion group covers the whole anterior end of on the ’s vertical plane. The otolith is the macula. The transition zone between an- triangular shaped, as opposed to the ovoid terior and ventral orientation groups is clear- shape in most other species studied. There is a ly seen in all three species due to the presence groove on the otolith not only for the macula of a demarcation line that only contains sup- outline, but also for the large “support” area porting cells (fig. 27). The demarcation line extending from the macula. narrows as it extends anteriorly and it is The posterior portion of the macula has finally obliterated by hair cells at its anterior typical type F1 ciliary bundles all the way to end. Preliminary observations indicate that the border. The macula becomes thinner this line marks the separation between two about one-third of the way from the posterior branches of the saccular portion of the eighth end and the dorsal cell region becomes only nerve, one innervating the region of the one or two cells deep for a short distance. The macula containing anterior and posterior hair SACCULUS AND LAGENA IN FISHES 411

- Diaphus brachycephalus 0.2 mm

- Ceratoscopelus warmingi 0.2 mm

D

Lampanyctus -sp. Fig. 23 Saccular macula of Diuphus. For explanations see figure 4. Fig. 24 Saccular macula of Ceratoscopehs. For explanations see figure 4. Fig. 25 Saccular macula of Lampanyctus. For explanations see figure 4. cell groups and one innervating the dorsal and the maculae). Each of the cells has a single ventral cell group region. long central cilium and an indeterminate Due to the small size of the lagena in each number of shorter cilia-like structures. No of these species it was difficult to prepare such cells were found in other species studied. specimens for study successfully, except in Lampanyctus (fig. 26). Approximately 85% of DISCUSSION the hair cells in this species are oriented The data presented here extend earlier ventrally with only a single row of dorsal cells work (Dale, '76; Jsrgensen, '76; Popper, '76) on the margin and a small group in the demonstrating inter-specific variation in the ventral region. All of the ciliary bundles are ultrastructure of the sacculus and lagena of type F2 except for some dorsally-oriented cells teleost fishes. While the variations in hair cell in the ventral-medial region of the macula orientation patterns are found in both sac- which are F3. culus and lagena, they are particularly evi- In addition to normal hair cells, each of the dent in the relationship between the orienta- myctophids has a group of large, square, cells tion groups on the anterior half of the saccu- just ventral to the posterior-ventral margin of lar macula. The posterior half of the saccular the saccular macula (fig. 28). There were 13 macula contains hair cells oriented dorsally cells in Lampanyctus and 20 to 30 in the other and ventrally and this pattern resembles, to a species (perhaps reflecting size differences in considerable degree, the saccular pattern in 412 ARTHUR N. POPPER ostariophysine fishes (e.g., Hama, '69; Jenk- in individual species, as well as the taxonomic ins, '74; Platt, '77) and many tetrapods (e.g., and functional significance of the inter- Lewis and Li, '73; Lindeman, '69; Spoendlin, specific variation in the morphology and ul- '65). The anterior macula region in the 15 trastructure of the teleost ear, is still not non-ostariophysines, as well as in a cod (Dale, known. However, while more extensive ultra- '76) and several flatfish (Jorgensen, '761, have structural and physiological data are neces- horizontally oriented hair cells, a pattern sary to enable us to make specific statements unlike that found in any other vertebrates. on the evolution of the teleost ear and on cor- The significance of these orientation patterns relations between structure and function, sev- eral suggestions may be made relating to specific patterns of the inner ears in several species as well as to certain very general aspects of inner ear function. One aspect of inner ear structure and teleost taxonomy that can be considered is the relationship between ear structure and hair cell orientation patterns in the ostariophysine and non-ostariophysine fishes. Each of the non-ostariophysines that has been studied (with the exception of Lotu vulgaris, Wersall et al., '65) has horizontally and vertically ori- ented hair cells and the sacculus is consid- erably larger than the lagena in all non- ostariophysines species (e.g., von Frisch, '36; Retzius, 1881; Werner, '60). However, there are only vertically oriented hair cells in six ostariophysines (five catfish and one carp) (Hama, '69; Jenkins, '74; Platt, '77) and the lagena is considerably larger than the sac- - culus in all Ostariophysi. While more data are @ Lampanyctus -sp. 0.05mm needed, especially for the ostariophysines, we might tentatively suggest that the marked Fig. 26 Lagenar macula from Lampanyctus. Note that there is only a small region with dorsally oriented hair differences in inner ear structure between the cells. For explanations see figure 4. ostariophysine and non-ostariophysine fishes

Fig. 27 SEM of the distinct demarcation line found in the three myctophids between the anteriorly oriented hair cells (to the left of the line) and ventrally oriented hair cells (to the right of the line). This line is shown in figures 23 thru 25. Fig. 28 SEM of series of large cells found in the three species of myctophid. These cells are found just below the ventral-posterior quadrant of the macula. SACCULUS AND LAGENA IN FISHES 413 indicate significantly separate evolutionary to the holocentrid sub-family Holocentrinae, a paths for the sacculus and lagena in the two group without direct contact between the an- groups. It is possible that the differences are terior end of the swim bladder and the audito- related to the presence of direct acoustic cou- ry bulla. However, the anterior region of the pling, via the Weberian , to the swim swim bladder in species of is bladder in the ostariophysines; but without closer to the auditory bulla than in Adioryx more data on ear structure in species closely (Nelson, '55) and from the behavioral data one related to ostariophysines, this suggestion can conclude that this anterior projection cannot be tested. enhances sound detection sensitivity, as has Due to lack of functional data related to been suggested for other species (Lowenstein, hair cell orientation patterns we cannot yet '71; Popper and Fay, '73; Tavolga, '71). Data discuss the significance of most inter-specific on sensitivity are not available for Myripristis differences of the sensory maculae. However, but this species is a member of the sub-family several points should be made about the sac- Myripristinae, a group with swim bladder an- cular macula patterns in the holocentrids terior projections that are in contact with the since there are significant differences in the auditory bulla (see Nelson, '55).We might ex- sacculus within this family. trapolate from Holocentrus and Adioryx and Several aspects of the sacculus in Myr- speculatively suggest that sensitivity in Myri- ipristis are of interest. The saccular oto- pristis is enhanced as compared with Adioryx. lith is triangular-shaped as compared to It is then possible that the sacculus pattern, the ovoid shape in most of the other species and general morphology of this otolithic (including Adioryx) I have looked at (al- structure in Myripristis, is in some way corre- though triangular shape appears typical of lated with the differences in swim bladder- other members of the same sub-family, inner ear relationships or, perhaps, the gen- Frizzel and Lamber, '61). The shape of the eral role of the swim bladder as a participant macula itself also differs from that found in in sound detection in some species and not in any other species of vertebrate with which I others (see Fay and Popper, '74, "75). am familiar. Finally, the positions of the various orientation groups differ considerably Ciliary bundles from those in other species. In several macula A number of different ciliary bundles have regions the sensory cells are interspersed with been found on the hair cells in . It is large, atypical, supporting cells, in addition to necessary to point out, however, that the four the more typical cells found in the sensory types reported here are essentially only points regions. These enlarged supporting cells are on a continuum and that intermediary forms also found in special regions which lie in a were often observed. Similar observations on dorsal depression on the otolith surface. gradation between six types of ciliary bundles The adaptive basis for the different saccular were reported in several amphibians (Lewis structures in the two holocentrids is not clear and Li, '75; Lewis and Neminac, '72); and and they are especially interesting since the Dale ('76) has observed various types in species live sympatrically on Hawaiian reefs Gadus but has chosen to consider all of these a and produce similar sounds for communica- continuum of a single type of ciliary bundle. tions (Salmon, '67). One suggestion is that the In fishes, as in amphibians, and lampreys two species have different sound detection ca- (Hoshino, '75; Lewis and Li, '75) the different pabilities or possibly detect or analyze sounds ciliary bundles are found in relatively discrete in different ways. The only data on sound macular regions and these regions may in- detection capabilities for holocentrids, how- clude cells from several orientation groups. ever, are for species different than those stud- The presence of different types of ciliary bun- ied here, so only indirect comparisons can be dles, and their location in specific areas, sug- made. Tavolga and Wodinsky ('63) found that gests a potential functional correlate for the Holocentrus ascensionis and Adioryx vexillarus different bundle types. (formerly H. uexillarus) have best pure tone Comparisons of hair cell ciliary bundle auditory sensitivity at 600-800 Hz, although types can only be made with one other teleost the highest frequency detected by HoZocentrus species, Gadus rnorhua (Dale, '76). Dale has is 2,800 Hz while it is about 1,200 Hz for reported ciliary bundles that resemble the Adioryx. Sensitivity was 10 db better in Holo- types F1 and F2 found in the species I have centrus than in Adioryx. Both species belong studied, and he also reports lagena ciliary 414 ARTHUR N. POPPER bundles that more nearly resemble the bun- nisms or sensitivity. Considering differences dles found on the lagena in both holocentrids, in morphogenesis, Lewis and Li ('75) point out rather than the cells found in other species. that their type A bundle, which somewhat re- Some comparisons can also be made with data sembles the teleost F2 ciliary bundles, is often for the Japanese lamprey, Entosphenus japon- found on the edges of the saccular macula, an icus, which has two types of ciliary bundles area that they have identified as one of (Hoshino, '75) as does Lampetra fluviatilis growth for that organ (Lewis and Li, '73; also (Lowenstein et al., '68). Hoshino ('75) found see Corwin, '77). The type F2 ciliary bundles, one bundle, designated type A, that resembles are also found peripherally and while nothing the teleost F2 bundle. The lamprey type B is yet known about morphogenesis of the bundle has a long kinocilium and shorter, non- teleost otolithic organs, F2 cells "grade" into graded, stereocilia. While the general F2 form type F1 cells in the more central areas of the in teleosts has short but graded stereocilia, I maculae (fig. 9). did, at times, see stereocilia that appeared to be equal in height in some salmonids. While Functional significance of inner ear preliminary, these data indicate that similar structure ciliary bundles are to be found throughout the The complex structure and orientation pat- aquatic anamniotes. terns of the teleost sacculus and lagena lead The most extensively described ciliary bun- to the conclusion that these organs may be dles have been the six types found in the bull- highly adapted for signal detection and proc- frog, Rana catesbeiana (Lewis and Li, '75). essing and that these functions may vary in Some of the types in Rana resemble those in different species. However, due to the paucity teleosts. Most notably, Lewis and Li found one of data it is still essentially impossible to bundle (their type F) which closely resembles speculate on the functional significance of the the teleost F3 bundle, while another (type B) variation observed in the 15 species reported is similar to the teleost type F2. While no frog here. Considering the morphology and ultra- ciliary bundle fully resembles the teleost F1 structure of the sacculus and lagena in gen- bundles, the type E has short graded eral, however, we can suggest that the ear stereocilia with a slightly longer kinocilium. may be adapted for providing some complex However, the kinocilium on this frog cell has types of signal analysis Le., frequency or a bulbous ending. Similar endings are found intensity discrimination) or in some species, in other frog hair cells, as well as in several for providing the animals with information reptile cells (Bagger-Sjoback, '74) but have for determining the direction of a sound yet to be reported for fishes (see also Dale, '76; source in the space around the animal. Platt, '77). Several workers have suggested that the Several generalizations about distribution otolith and macula may move in different pat- of the various ciliary bundles in fishes may be terns relative to one another in response to made, but these must be qualified since there signals that differ in frequency or intensity is considerable variation in the different spe- (van Bergeijk, '67b; Popper and Clarke, '76; cies studied. In general, the type F1 bundle Sand, '74). As a result, different hair cell makes up the dominant cell on the sacculus in regions would respond to different signals, all species except the myctophids where the thus providing for some degree of signal anal- dominant bundle is F3. The type F2 bundle is ysis at the level of the inner ear. The move- most often found on macula margins. The ment patterns of the otoliths and maculae type F3 ciliary bundle is the pattern most would possibly be "enhanced" by the complex often found in macula areas not directly shapes of these structures, and the inter- covered by the otolith, but this is not always specific variation in the otoliths and the the case, as seen in myctophids where the maculae would then be responsible for whole saccular macula has F3 bundles. differences in detection capabilities or signal The functional or structural significance analysis by different species. While it is not of each of the ciliary patterns remains to be yet possible to predict the specific nature of determined. Lewis and Li ('75) suggest that the movements, a possible basis for variation structure could reflect differences in stages of in response on the saccular macula in many morphogenesis, or in relationships to overly- species is likely to be coupling of different hair ing structures (otolith and/or otolithic mem- cell groups to the otolith. In many species brane), or differences in stimulation mecha- reported here some portions of the saccular SACCULUS AND LAGENA IN FISHES 415 macula are directly covered by the sagitta fashion analogous to that proposed for mam- while other regions are only overlain by the malian localization by van Bergeijk ('621, thin otolithic membrane. Consequently, cou- Hall ('651, Erulkar ('72) and others. Further, pling of otolith movement to the cells in con- by adding a third receptor (or group of recep- tact with the otolith, and those only covered tors) oriented 90" from the first two, an ani- by the otolithic membrane, would be sig- mal could tell sound direction in three dimen- nificantly different. This would result in sions (van Bergeijk, '641. Thus, if information different hair cell stimulation across the sac- in afferent fibers from different orientation cular macula. Evidence for such regional groups are kept separate to a level of some variation due to direct and indirect coupling comparator, then a single teleost sacculus to the otolith is not yet available for fishes. would provide sufficient information in two However, an analogous system does occur in dimensions. Three-dimensional information the basilar papilla of the alligator lizard (Ger- would be derived from two ears oriented in rhontous multicuinatud where different physi- different directions as well as by the cur- ological responses have been recorded from vature of the maculae (both sacculus and neurons innervating different regions of the lagena), since this produces cells oriented in basilar papilla (Weiss et al., '76). These different directions even within the four workers suggest that the response differences major orientation groups. are related to anatomical findings that hair Finally, it is necessary to note that the hy- cells in one portion of the basilar papilla are in pothesis proposed here for determination of contact with the , while sound source direction has several limita- cells in another region of the basilar papilla tions. One limitation is that this system are not in contact with the membrane. would not permit discrimination of signals The structural adaptations of the inner ear that are 180" apart. Consequently, fishes may also have evolved to function in deter- would theoretically have difficulty telling mining the direction of a sound source as a "front from back" (Schuijf, '75; Schuijf and prelude to actually approaching the source Buwalda, '75). A second limitation is that the (sound localization). A number of workers proposed hypothesis may only be applicable to have recently provided evidence that the species not using the swim bladder for sound inner ear may be involved in sound orienta- detection or, in certain instances where the tion in some teleost species (Enger et al., '73; swim bladder is used but where it is not di- Moulton and Dixon, '67; Sand, '74; Schuijf, rectly coupled to the ear. Species, such as the '75; Schuijf and Buwalda, '75). Particularly Ostariophysi, having both ears stimulated by important are data showing that the levels of a single swim bladder, lose any directional microphonic response from different regions information about the impinging signals. of the saccular macula vary as a function of However, in non-ostariophysines the swim stimulus direction (Enger et al., '73; Sand, bladder could respond directionally to dis- '74). These directional responses are closely placement signals and then set up a sound correlated with data on orientation patterns field which would be used to provide di- of hair cells in various macular regions (also rectional stimulation to the ear. Consequent- see Dale, '76), suggesting that the different ly, it is possible that stimulation by bone con- microphonic responses represent stimulation duction, or with the swim bladder, could re- of hair cells that are oriented in different sult in detection of sound direction. directions. The amount of directional information ACKNOWLEDGMENTS available to fishes is potentially more exten- This work was done while I was in residence sive than just from four discrete orientation at the Kresge Hearing Research Institute and groups. Of particular importance are the the Division of Biological Sciences, University graded responses for the hair cells (Flock, '65, of Michigan. I would like to express my sin- '71; Harris and Van Bergeijk, '62; Schwartz, cerest thanks to Doctor J. E. Hawkins, Jr., for '67). Schwartz ('67) has proposed that two providing me with laboratory facilities and hair cells, oriented at 90" to one another, can for valued discussions during the course of provide precise two-dimensional directional this work. I would also like to express my ap- information (also see van Bergeijk, '64) if the preciation to R. Preston for invaluable help in level of response of the two cells can be com- all phases of this investigation. Doctor L. R. pared at some level in the CNS, perhaps in a Aronson and J. Taylor kindly supplied me 416 ARTHUR N. POPPER with several species used in this study. Addi- 1973 Masking of auditory responses in the me- dulla oblongata of goldfish, J. Exp. Biol., 59: 415-424. tional animals were provided by the Great Enger, P. S., A. D. Hawkins, 0. Sand and C. J. Chapman Lakes Fisheries Laboratory of the U. S. Fish 1973 Directional sensitivity of saccular microphonic and Wildlife Service and my particular potentials in the haddock. J. Exp. Biol., 59: 425-433. thanks to Doctors R. Reinert and D. Rottiers Erulkar, S. D. 1972 Comparative aspects of spatial localization of sound. Physiol. Rev., 52: 237.360. for their assistance. Ms. N. Clarke served gal- Fay, R. R. 1975 Sound reception and processing in the lantly in collecting the deep sea species and carp: saccular potentials. Comp. Biochem. Physiol., 49A: preparing them for study while on a rolling 29-42. ship. The staff of the University of Michigan Fay, R. R., and A. N. Popper 1974 Acoustic stimulation of the ear of the goldfish. J. Exp. Biol., 62: 243-260. Scanning Electron Microscopy-Electron Mi- 1975 Modes of stimulation of the teleost ear. J. croprobe Analysis Laboratory provided inval- Exp. Biol., 62: 379-388. uable help in all phases of the SEM work. I Flock, A. 1965 Electron microscope and electrophysio- would also like to thank J. Corwin for suggest- logical studies of the lateral-line canal organ. Acta Oto- Laryngol. (Suppl.), 199: 1-90. ing the use of the ultrasonic cleaner, J. Bruce 1971 Sensory transduction in hair cells. in: for doing the many illustrations, and the staff Handbook of Sensory Physiology. Vol. 1. W. R. Lowen- of the Kresge Hearing Research Institute stein, ed. Springer-Verlag. Berlin, Heidelberg, New photographic shop for considerable assistance. York, pp. 396-411. Frisch, K. von 1936 her den Gehorsinn der Fische. Doctors W. C. Stebbins and W. N. Tavolga Biol. Rev., 11: 210-246. kindly read and commented upon the Frizzell, D. L., and C. K. Lamber 1961 New genera and spe- manuscript. cies of myripristid fishes in the gulf coast Cenozoic, Work was supported by Public Health Ser- known form otoliths (Pisces, ). Bull. U. Missouri School Mines and Metallurgy, No. 100,pp. 3-25. vice grants from the National Institute of Furukawa, T., and Y. Ishii 1967 Neurophysiological Neurological and Communicative Disorders studies on hearing in goldfish. J. Neurophysiol., 30: 1377- and Stroke, NS-09374 to A. N. Popper and NS- 1403. 05065 and NS-05679 to J. E. Hawkins, Jr.; by Gerald, W. 1971 Sound production during courtship in six species of sunfish (Centrachidae). Evol., 25: 75-87. Program Project Grant NS-05785 to M. Law- Hall, J. L. 1965 Binaural interaction in the accessory rence, and by a biomedical science support superior olivary nucleus of the cat. J. Acoust. SOC.Amer., grant from the University of Hawaii to A. N. 37: 814-823. Popper. Hama, K. 1969 A study on the fine structure of the sac- cular macula of the goldfish. Z. Zellforsch., 94; 155.171. LITERATURE CITED Harris, G. G., and W. A. van Bergeijk 1962 Evidence that the lateral organ responds to near-field displacements of Bagger-Sjoback, D. 1974 The sensory hairs and their sound sources in water. J. Acoust. Soc. Amer., 34: 1831- attachments in the lizard basilar papilla. Brain, Behav. 1841. and Evol., 20: 88-94. Hoshino, T. 1975 An electron microscopic study of the Bekkesy, G. von 1960 Experiments in Hearing. McGraw- otolithic maculae of the lamprey (Entosphenus japoni- Hill, New York, 745 pp. cus). Acta Oto-Laryngol., 80: 45-53. Bereijk, W. A. von 1962 Variations on a theme of Jenkins, D. B. 1974 A preliminary report on anatomical Bekesy: a model of binaural interaction. J. Acoust. Soc. studies of the inner ear in catfishes (Siluriformesl. Anat. Amer., 34: 1431-1437. Rec., 78: 382. 1964 Directional and non-directional hearing in Jergensen, J. M. 1976 Hair cell polarization in the fish. In: Marine Bio-Acoustics. W. N. Tavolga, ed. flatfish inner ear. Acta Zool. (Stockholm), 57: 37-39. Pergamon Press, Oxford, pp. 281-300. Lewis, E. R.,and C. W. Li 1973 Evidence concerning the 1967a The evolution of vertebrate hearing. In: morphogenesis of saccular receptors in the bullfrog Contributions to Sensory Physiology. W. D. Neff, ed. (Ram catesbeianai. J. Morph., 139: 351-362. Academic Press, New York, pp. 1-49. - 1976 Hair cell types and distribution in the 1967b Discussion. In: Marine Bio-Acoustics. otolithic and auditory organ of the bullfrog. Brain Res., Vol. 2. W. N. Tavolga. ed. Pergamon Press, New York, p. 83: 35-50. 244 Lewis, E. R.,and P. Nemanic 1972 Scanning electron mi- Caldwell. D. K., and M. C. Caldwell 1963 Underwater croscope observations of saccular ultrastructure in the sounds associated with aggressive behavior in defense of mudpuppy (Necturus maculosus). 2. Zellforsch., 123: territory by the pinfish, Lagodon rhornboides. Bull. So. 441-457. Calif. Acad. Sci., 66: 69-75. Lindeman, H. H. 1969 Regional differences in structure Corwin, J. T. 1977 Morphology of the macula neglecta in of the vestibular sensory regions. J. Laryngol. Otol., 83: sharks of the genus Careharhinus J. Morph.. 152: 341- 1-17. 362. Lowenstein, 0. 1971 The labyrinth. In: Fish Physiology. Dale, T. 1976 The labyrinthine mechanoreceptor organs Vol. 5. W. S. Hoar and D. J. Randell, eds. Academic Press, of the cod, Gadus rnorhua L. (Teleostei: Gadidae) Norw. New York, pp. 207-240. J. Zool., 24: 85.128. Lowenstein, O., M. P. Osborne and R. A. Thornhill 1968 Enger, P. S. 1963 Single unit activity in the peripheral The ultrastructure of the labyrinth of the of a teleost fish. Acta. Physiol. Seand. lamprey (tampetra fluvklilis L.1. he.Roy. Sac Lond. (Suppl.), 59: 210. B., 170: 113-134. SACCULUS AND L,AGENA IN FISHES 417

Lowenstein, O., M. P. Oshorne and J. Wersall 1964 Struc- Retzius, G. 1881 Das Gehororgan der Wirbelthiere. Vol. ture and innervation of the sensory epithelia of the 1. Samson and Wallin, Stockholm. labyrinth in the thornback ray (Rqja clavata). Proc. Roy. Saito, K. 1973 Fine structure of the macula of lagena in SOC.Lond. B, 160: 1-12. the teleost inner ear. Acta Anat. Nipponica, 48: 1-18. Lowenstein, O., and J. Wersall 1959 A functional inter- Salmon, M. 1967 Acoustical behavior of the menpachi, pretation of the electron-microscopic structure of the Myripristm berndti, in Hawaii. Pacific Science, 21: 364- sensory hairs in the cristae of the elasmobranch Raja 381. clauata in terms of directional sensitivity. Nature, 184: Sand, 0. 1974 Directional sensitivity of microphonic 1807-1808. potentials from the perch ear. J. Exp. Biol., 60: 881-899. Marshall, N. B. 1967 Sound-producing mechanisms and Schuijf, A. 1975 Directional hearing of cod (Gadus the biology of deep-sea fishes. In: Marine Bio-Acoustics. morhua) under approximate free field conditions. J. Vol. 2. W. N. Tavolga, ed. Pergamon Press, New York, pp. Comp. Physiol., 98: 307-332. 123-134. Schuijf, A., and R. J. A. Buwalda 1975 On the mechanism Masterton, R. B. 1974 Adaptation for sound localization of directional hearing in the cod (Gadus morhua L.). J. in the ear and brainstem of mammals. Fed. Proc., 33: Comp. Physiol., 98: 333-343. 1904-1907. Schwartz, E. 1967 Analysis of surface wave perception Miller, M. 1974 Scanning electron microscopy of the in some teleosts. In: Lateral-Line Detectors. P. H. Cahn, lizard papilla basilaris. Brain, Behav. and Evol., 10: ed. Indiana University Press, Bloomington, pp. 123-134. 95-112. Spoendlin, H. 1965 Ultrastructural studies of the Moulton, J. M., and R. H. Dixon 1967 Directional hearing labyrinth in squirrel monkeys. In: The Role of the in fishes. In: Marine Bio-Acoustics. Vol. 2. W. N. Tavolga, Vestibular Organ in the Exploration of Space. NASA Sp., ed. Pergamon Press, pp. 187-232. 77: 7-22. Myrberg, R., Jr., E. Kramer and P. Heinecke 1965 Sound Tavolga, W. N. 1971 Sound production and detection. In: production by cichlid fishes. Science, 149: 555-558. Fish Physiology. Vol. 5. W. S. Hoar and D. J. Randell, eds. Myrberg, R., Jr., and J. Y. Spires 1972 Sound discrimina- Academic Press, New York, pp. 135-205. tion by the bicolor damselfish, Eupornacentrus partitus. J. 1974 Signalhoise ratio and the critical band in Exp. Biol., 57: 727-735. fishes. J. Acoust. SOC.,55: 1323-1333. Nelson, E. M. 1955 The morphology of the swim bladder Tavolga, W. N., and J. Wodinsky 1963 Auditory capacities and auditory bulla in the Holocentridae. Fieldiana: Zool., in fishes: pure tone thresholds in nine species of marine 37: 121-130. teleosts. Bull. Amer. Mus. Nat. Hist., 126: 177-240. Piddington, R. W. 1972 Auditory discrimination be. Response of salmon to low fre- tween compressions and rarefactions by goldfish. J. Exp. Van der Walker, J. 1967 Biol., 56: 403-419. quency sound. In: Marine Bioacoustics. Vol. 2. W. N. Platt, C. 1977 Hair cell distribution and orientation in Tavolga, ed. Pergamon Press, Oxford, pp. 45-58. goldfish otolith organs. J. Comp. Neur., 172: 283-298. Weiss, T. F., M. J. Mulroy, R. G.Turner andC. A. Pike 1976 Popper, A. N. 1976 Ultrastructure of the auditory Tuning of single fibers in the cochlear nerve of the alliga- regions in the ear of the lake whitefish. Science, 192: tor lizard: Relation to receptor morphology. Brain Res., 1020-1022. 115: 71-90. Popper, A. N., and N. L. Clarke 1976 The auditory system Werner, C1. F. 1960 Das Gehororgan der Wirbeltiere und of the goldfish (Carassius auratus): effects of intense des Menschen. G. Thieme, Leipzig. acoustic stimulation. Comp. Biochem. Physiol., 53A: Wersall, J. A,, A. Flock and P. G. Lundquist 1965 Struc- 11-18. tural basis for directional sensitivity in cochlear and Popper, A. N., and R. R. Fay 1973 Sound detection and vestibular receptors. Cold Spring Harbor Symp. Quant. orocessine bv fish: A critical review. J. Acoust. SOC. Biol., 30: 115-132. kmer., 55 1515-1529. Wever, E. G. 1969 Cochlear stimulation and Lempert’s Popper,.~ A. N., M. Salmon and A. Parvulescu 1973 Sound mobilization theory. Arch. Otolaryngol., 90: 68-73. localization by the Hawaiian squirrelfishes Myrzpristis 1971 The mechanics of hair-cell stimulation. berndti and M. argyromus. Anim. Behav., 21: 86-97. Ann. Otol. Rhinol. Laryngol., 80: 786-804.