Two of the basic requirements of an ear‐model design is that the transfer characteristic of the ear canal and the impedance as measured at the eardrum closely agree with results obtained on real ears. In the case of the electrical analog model, this requires that the number of sections that are used to represent the ear canal provide an upper cutoff well beyond the highest frequency under study. This is particularly true where the effect of ear‐canal taper is of interest. For these purposes, 10‐section analog networks of uniform and of tapered design were employed. For network representation of the eardrum and the adjacent connecting structures, the effectiveness of one‐, two‐, three‐, and four‐branch terminations were explored. Of these, the two‐ and the four‐branch analog networks were found to provide the best overall representations.