John Howe sent me copies of his data on seven sheets. Some data were redundant. In Table 1 you may find the original data and in Table 2 - the ones I have used.
A few phalanges were obviously different from the rest by their size. They will be discussed later, after the bulk of the material (ca. 125 middle-sized specimens).

In Fig.1, all Natural Trap Ph3 are compared to all extant Hemiones and Kiangs.

Sorting the anterior Ph3 from the posterior
I’ll discuss elsewhere the general problem of this kind of sorting. For the moment I’ll only say that, unfortunately, two important measurements were not taken : plantar length and plantar circumference. Thus, I tried to sort the Ph3 according to the proportions of the articular surface : in the same individual, the articular width of the anterior Ph3 is larger, but naturally when many individuals are mixed the sorting is uncerain : Fig.2 (all Hemiones and Kiangs), Fig.3 (midle-sized Natural Trap).

Conclusions
Fig.4 compares the middle-sized and the small specimens of Natural Trap to all E. hemionus and E. kiang Ph3. I cannot go farther than saying that they are about the same size.

Fig.5 compares the large specimens of Natural Trap to E. przewalskii Ph3. Fig.6 compares the same to A. occidentalis Ph3.

The specimen 50821 could be a posterior Ph3 of a Caballine or an anterior of A. occidentalis. Others could belong to a very large A. occidentalis (Fig.6).