To perform laryngeal research involving long-term survival surgery, a permanent tracheostomy is often necessary. For experiments using a long-term induced canine phonation model, we required a tracheostomy that was placed as low as possible, to maximize the subglottic space superior to the stoma. The ideal experimental tracheostomy would also be safe and easy to perform, require no tracheostomy tube, and be low maintenance, requiring minimal cleaning or suctioning. Tracheostomies were performed in 37 dogs based on previously published methods. If the stoma was placed below the twelfth tracheal ring, the perioperative mortality rate was 57% because of kinking of the trachea and subsequent airway obstruction. When the tracheostomy was performed above this level, the mortality rate was reduced to 3%. A number of significant modifications in technique were made to achieve this improvement and resulted in the last 12 dogs having no complications. Several of the tracheostomies were maintained for more than 18 months. The method derived meets the above criteria for the ideal experimental tracheostomy and also meets our needs for a long-term induced phonation model.