The inner ear in fishes comprises the three semicircular canals and the three otolithic end organs, namely the utricle, the saccule and the lagena [1, 2]. In modern bony fishes (teleosts) the sensory epithelium of each otolithic end organ is overlain by a single calcium carbonate concretion, the otolith . Movement of the fish in a sound field leads to the lagged movement of the denser otolith relative to the fish’s body and the sensory epithelium, which stimulates the sensory hair cells [3, 4]. Fishes using solely this stimulation pathway have limited auditory abilities [4–6]. To improve hearing (that is, to expand frequency detection up to several thousand Hertz and/or to increase sensitivities), species across different taxonomic groups have evolved accessory auditory structures. Inner ears in these species are close to or connected with either (1) intracranial gas cavities (anabantoids, mormyrids) or (2) anterior extensions of the swim bladder (for example, some holocentrids, sciaenids, cichlids, clupeids) or (3) are linked to the swim bladder via a chain of ossicles and ligaments (Weberian apparatus of the otophysans) [7, 8]. The gas bladders, oscillating in a sound field, then transmit energy to the inner ear endolymph, which again results in movement of the otolith relative to the sensory epithelium.
In fishes, it is still difficult to assign a certain vestibular or auditory function to one of the end organs . Although it is clear that the semicircular canals in the inner ear detect only angular accelerations, it is assumed that utricle, saccule, and lagena may serve in both vestibular and auditory functions [4, 9]. It is thus likely that all three otolithic end organs are involved in hearing to different degrees in different species . Nevertheless, physiological and anatomical data indicate that the saccule and, to a yet unknown extent, the lagena are the main auditory end organs. Extirpation experiments by von Frisch and his colleagues demonstrated in Eurasian minnows Phoxinus phoxinus (cypriniforms) that removal of the utricle and semicircular canals did not affect hearing at all, whereas simultaneous removal of the saccule and lagena completely eliminated the detection of sound frequencies above 130 Hz . Dijkgraaf  demonstrated that bilateral extirpation of the saccule and the lagena in the black goby Gobius niger decreased their auditory sensitivity and eliminated sound detection above 400 Hz. The complete removal of both inner ears led to a further decrease in the detectable frequency range. Schuijf  showed that cutting the nerve root of the saccule and lagena of one inner ear in the cod Gadus morhua reduced the ability to localize a sound source. In addition, in most species possessing accessory auditory structures — otophysans (cypriniforms, siluriforms, characiforms and gymnotiforms), anabantoids, mormyrids, and some holocentrids — the swim bladder or gas-filled cavities contact the saccule (for a review see ). None of the so far investigated teleost species, however, possesses a connection between the swim bladder and the lagena.
Ramcharitar et al.  described swim bladder extensions in the silver perch Bairdiella chrysoura that came close to the lagena without, however, directly contacting it or any of the other two end organs. Recently, Schulz-Mirbach et al.  provided the first evidence, based on microtomographic (microCT) scans and dissections, that the swim bladder extensions in the Asian cichlid species Etroplus maculatus approach the lagena; they further showed that E. maculatus has improved hearing abilities compared to other cichlid species lacking such a swim bladder-inner ear connection. From that study, it was not completely clear whether the swim bladder directly contacted the lagena. Moreover, those findings partly contradicted data published by Dehadrai , who illustrated and described the swim bladder-inner ear relationship in Etroplus suratensis without showing or mentioning the lagena. Sparks  noted intracranial bullae of the swim bladder contacting the inner ears, which was not found in other studies [15, 16]. These former studies [15–17] did not use the high-resolutions of microCT imaging or histological analysis we used in our study. Although Dehadrai  performed histological analyses of the swim bladder, he did not describe or illustrate histological sections showing the contact region of swim bladder extensions and inner ears. Nonetheless, histological analysis is necessary in order to determine whether the swim bladder extension abuts the lagena, abuts the bone surrounding the lagena, or is firmly attached to the lagena or to the bone via connective tissue in E. maculatus. Several of the steps during sample preparation for the histological analysis (decalcification, dehydration, embedding, and sectioning procedures), however, may cause artifacts such as shrinkage, distortion or loss of important sections . We, therefore, applied a new methodological approach, a combination of two high-resolution techniques: (1) non-destructive high-resolution microCT imaging by which the fixed and stained sample was investigated prior to decalcification, dehydration, and embedding; and (2) histological serial sectioning of the same specimen. This approach aimed to identify potential artifacts during histological sample preparation. Prior to these morphological investigations, we additionally tested hearing abilities using the auditory evoked potential (AEP) recording technique [19, 20]. In our study, we, therefore, asked (1) whether a direct anatomical connection exists between the swim bladder extensions and inner ears in E. maculatus and, if so, (2) whether the extension is linked only to the lagena or also to the saccule and (3) whether internal bullae are present.
We found a close anatomical link between swim bladder and lagena via a thin bony lamella, indicating that the lagena in E. maculatus plays a role in audition. A saccule attachment to the neurocranium and the close vicinity between the swim bladder extension and the posterior recessus of the upper inner ear also suggest an involvement of the utricle and, thus, of all three end organs in improved auditory sensitivities.