To evaluate the ability to assess laryngeal function and to diagnose unilateral laryngeal paralysis (uLP) via airway endoscopy and carbon dioxide (CO2 ) stimulation. Experimental study. Six healthy, adult beagles. Dogs were anesthetized with sevoflurane and dexmedetomidine. Laryngeal activity was observed via endoscopy placed through a laryngeal mask airway (LMA). The absolute and normalized glottic gap areas (AGGA and NGGA, respectively) and the glottic length (GL) were measured at inspiration and before and after surgically induced uLP. Measurements were obtained at eupnea and during hypercapnic hyperpnea produced by the administration of CO2 . Values for each hemilarynx were also measured. Video recordings were observed by 2 surgeons who scored function as normal or uLP. The AGGA and NGGA increased similarly during CO2 administration in intact dogs and in dogs with uLP; the GL increased in dogs with uLP but not in intact dogs. The AGGA and NGGA of the intact hemilarynx increased more than those of the affected hemilarynx in dogs with uLP. uLP was correctly identified more frequently by observers at hypercapnic hyperpnea than during eupnea. The increase in AGGA and NGGA at peak inspiration during CO2 administration was not limited by uLP, but asymmetry in hemilarynx AGGA and NGGA was observed in dogs with uLP. CO2 administration facilitated the identification of uLP. Laryngeal endoscopy through an LMA coupled with administration of CO2 in anesthetized dogs facilitates the observation of arytenoid function and may improve the diagnosis of naturally occurring mild laryngeal paralysis.