按图索骥 is formed by the sequential combination of four characters: 按 (meaning 'press' or 'follow'), 图 (meaning 'picture'), 索 (meaning 'seek'), and 骥 (meaning 'thoroughbred horse'). It originates from an anecdote where a person literally used a diagram to identify a steed, leading to a literal meaning of 'following a picture to search for a horse.' The characters are arranged to depict this specific action, creating the idiom's structure.