are therefore much less convenient to deal with than ordinary letters.
The rate of hand-stamping is therefore only about three-fourths the rate
for ordinary letters.

(_c_) All postcards can be passed through the machine-stamp if
available. The rate is therefore ten times the rate of hand-stamping.

(_d_) A large proportion of halfpenny packets cannot, on account of
their size and shape, be passed through the machine-stamp, and the
figure for the machine-stamp must be considerably reduced for these
packets. The nearest estimate that can be formed for these packets is
7.5 times the rate for hand-stamping.

(_e_) Newspapers cannot be passed through the machine-stamp, but in a
number of cases the wrappers are taken to the post office before the
newspapers are enclosed in them for cancellation of the postage stamps
(in order to secure a prompt despatch when the newspapers are actually
posted). The rate for such stamping is slightly greater than the rate of
hand-stamping for ordinary letters. On the other hand, the rate of
stamping newspaper packets is not more than two-thirds the rate of
hand-stamping ordinary letters. The nearest estimate that can be formed
for all newspapers is that the rate of stamping is four-fifths the rate
of hand-stamping ordinary letters.

* * * * *

NOTE III. _Rates of Sorting._--(_a_) The average rate of sorting for
ordinary letters is taken as the unit.

(_b_) The rate of sorting letters and the rate of sorting postcards may
be taken as identical.

(_c_) Owing to the irregular shape of newspaper packets, and letter
packets over 4 ounces in weight, the average normal rate of sorting must
be taken as considerably less than that for letters.

Both classes are usually sorted at the packet tables and not at the
ordinary letter frames.

(_d_) The letter packets between 1 ounce and 4 ounces in weight present
some difficulty, since they include a considerable number of long
letters, which are sorted at the ordinary letter frames at nearly the
same rate as short letters, while the rest are sorted at the packet
tables at about the same rate as the heavier packets. The figure should
obviously be between (_a_) and (_c_).

(_e_) The halfpenny packets also fall into two classes: (1) those sorted
as short letters, and (2) those sorted at the newspaper frames. A very
large proportion fall into the second class, and the average normal rate
of sorting, as in the case of t