indicate that this takes place more frequently between 10 and 12 miles.
From this point onward the performance depends on the efficiency of the
steam production, which does not appear to be able to utilize more than
16% of the weight on the drivers. The diagrams presented by Mr. Dennis
in his paper on virtual grades, and by John A. Fulton, M. Am. Soc.
C. E., in his discussion of that paper, indicate that similar results
would be shown were they extended to include the distance named.

From this it would appear that a locomotive is capable of hauling a
larger train on grades less than 10 miles in length than on longer
grades, and that, even when unexpectedly stopped, it is capable of
starting again as soon as the steam pressure is sufficiently built up.
Conversely, it should be practicable to use a higher rate of ascent on
shorter grades on any given line without decreasing the load which can
be hauled over it. In other words, what is known as the "ruling grade"
is a function, strictly speaking, of the length as well as the rate of
grade.

In any discussions of the practicability of using a higher rate on the
short grades, which the writer has seen, the most valid objection has
appeared to be the danger of stalling and consequent delay. As far as
momentum is relied on, this objection is valid. Within the limits of the
load which can be handled by the steam, it has small value, as it is
only a question of waiting a few minutes until the pressure can be built
up to the point at which the load can be handled. As this need only be
an occasional occurrence, it is not to be balanced against any material
saving in cost of construction.

The writer does not know of any experiments which will throw much light
on the value of heat storage as separated from momentum, though the