Until recently the only reported action of the ratite tongue during feeding was retraction during swallowing  and depression of the mouth floor to allow for the effective intra-oral transfer of food to the proximal esophagus using the ‘catch and throw’ method . The ratite tongue has been classified as rudimentary and of little functional significance during feeding when compared to the tongues of neognaths . However, recent morphological studies have revealed a number of diverse functions for this organ in ratites [12, 13, 17]. These include swallowing, cleaning of the palate , lubrication, mechanoreception, taste and mechanical and immune protection  in D. novaehollandiae, and cleaning of the choana in R. americana. Furthermore, as demonstrated here, the ratite tongue potentially plays a far more pivotal role during feeding by closing off the weakest part of the adducted glottis during swallowing and thus preventing the inhalation of food and water (choking). This appears to be achieved, in S. camelus, by folding of the flat tongue root which allows the lingual pocket to encase the adducted glottis, and in D. novaehollandiae, by the specialised structure of the tongue root which slides over the rostral aspect of the adducted glottis. We have termed this proposed mechanism the linguo-laryngeal apparatus.
Despite the lingual pocket in S. camelus[2, 4, 6, 8–11] and triangular tongue root in D. novaehollandiae[2, 12] having been described, why has such an elegant mechanism eluded discovery for so long? Possible explanations for the prolonged obscurity of this mechanism are, firstly, most morphological studies are conducted on preserved material. Regardless of the fixative used (commonly formalin or alcohol) the tissues become hardened and the manipulations performed here on fresh specimens are impossible in preserved tissues. Secondly, functional studies [4, 5] using diagnostic imaging techniques do not detect the soft tissue and cartilage adequately or provide an intra-oral view thus making it impossible to demonstrate or interpret the movements we have described. Additionally, and only compounding the problem, it was generally accepted that the small ratite tongue was rudimentary, and therefore of little functional significance.
The proposed functioning of the linguo-laryngeal apparatus would rely on muscle action decreasing the distance between the hyobranchial apparatus (and thus tongue) and the larynx. The detailed study on the musculature of the hyobranchial apparatus of S. camelus and D. novaehollandiae provides supportive evidence in this regard. The muscle groups responsible for decreasing the distance between the two components are the hyolaryngeal and extrinsic hyolingual retractor muscles. The M. cricohyoideus, a hyolaryngeal muscle, appears to be a powerful muscle in S. camelus and is partially located in the tongue body which is a situation unknown in other taxa . The M. ceratocricoideus is unique to paleaognaths and is considered an unusual hyolaryngeal muscle . The hyolaryngeal muscles contract to decrease the distance between the basiurohyal (and thus the tongue) and the larynx during retraction of the tongue . This muscular action would account for the proposed functioning of the linguo-laryngeal apparatus. Additionally, the extrinsic hyolingual retractors also show some unique features in ratite species. In S. camelus the M. serpihyoideus inserts directly on the cricoid cartilage, again a situation unique in known avian taxa . However, in D. novaehollandiae this muscle acts directly on the ceratobranchials and urohyal . The M. hyomandibularis acts directly on the ceratobranchials in S. camelus but in D. novaehollandiae shows a unique configuration not reported in other avian taxa and is divided into an M. hyomandibularis lateralis (inserting on the mid-ceratobranchial) and M. hyomandibularis medialis (inserting on the urohyal) .
In S. camelus it would appear that the main muscle pulling the tongue, and thus the lingual pocket, over the adducted glottis would be the M. cricohyoideus, which in S. camelus has a unique conformation. In D. novaehollandiae the main contributors to pulling the tongue root over the adducted glottis would appear to be those muscles attaching to the urohyal, namely the M. serpihyoideus and M. hyomandibularis medialis of which the latter is again unique to D. novaehollandiae. The existence, conformation and positioning of such unique muscles, coupled with the synergy of the precisely formed anatomical structures reported in this study, further support the proposed functioning of the linguo-laryngeal apparatus.
There is currently no consensus as to the presence of a structure that functionally replaces the epiglottis in birds. It was originally suggested  that the posterior border of the tongue of S. camelus functioned like an epiglottis; however, this was later refuted . To further support this observation  in S. camelus, as well as our own conclusions, it was noted that a fold in the base of the tongue of Apteryx australis covers the glottis when the tongue is retracted . Other suggestions as to the function of the lingual pocket in S. camelus add further support to our findings on the linguo-laryngeal apparatus. The lingual pocket, observed by high-speed cineradiography, is reported to change shape during intraoral transport of food and is closed during tongue retraction . However, the pocket may have appeared ‘closed’ due to it being filled by the rostral aspect of the laryngeal mound as the tongue is pulled caudally (see Figures 1c2b3b). The muscular action of the M. ceratoglossus on the paraglossalia, ‘closing’ the pocket during swallowing , would undoubtedly aid in anchoring the pocket over the rostral aspect of the laryngeal mound. A secondary fold is present in the lingual pocket, which would provide an increase in surface area for mucus-producing glands, enhancing mucous production and secretion for the ingestion of dry food . This additional mucous would facilitate a smooth sliding motion of the lingual pocket over the laryngeal mound. In D. novaehollandiae it was originally proposed that the tongue root functioned like an epiglottis  but, as in S. camelus, this was subsequently refuted . However, this role for the tongue root of D. novaehollandiae was again proposed  and has now been tentatively demonstrated. It may be possible that neognathous birds (the jay and flamingo  and the chicken and domestic birds ) possess similar mechanisms (although less specialised), as has been previously suggested. This mechanism consists of a transverse, semi-lunar fold at the entrance to the glottis  or folds opposite the base of the tongue  that can function as a rudimentary epiglottis. However, the action of these folds was never functionally demonstrated which is most likely why their proposed role as an “epiglottis” has not been accepted and recognised. In the chicken, this mechanism does seem possible where, in the fresh state, the flat, smooth tongue root (which is relatively long) forms a semi-circular fold when the tongue is retracted and which covers the rostral part of the glottis (personal observation). Thus it has been suggested and debated, but never conclusively stated, that the tongue root in birds may function, albeit partially, as a form of epiglottis.