uthority and all
command over the general staff of the army, and the latter course
was adopted. The ancient controversy was up to 1888 no nearer
settlement than it was in 1869, though in General Sheridan's time
some progress had been made in the persistent efforts to deprive
the general-in-chief of the little authority which had been left
to General Sherman. General Sheridan had, with his usual gallantry
and confidence, renewed the contest, but had been worsted in his
first encounter with the Secretary, and then gave up the struggle.

Upon my assignment to the "command of the army" in 1888, I determined
to profit so far as possible by the unsatisfactory experience of
Generals Scott, Grant, Sherman, and Sheridan--at least so far as
to avoid further attempts to accomplish the impossible, which
attempts have usually the result of accomplishing little or nothing.
In fact, long study of the subject, at the instance of Generals
Grant and Sherman, earnest efforts to champion their views, and
knowledge of the causes of their failure, had led me to the conclusion
heretofore suggested, namely, that under the government of the
United States an actual military commander of the army is not
possible, unless in an extreme emergency like that which led to
the assignment of Lieutenant-General Grant in 1864; and that the
general-in-chief, or nominal commanding general, can at most be
only a "chief of staff,"--that or nothing,--whatever may be the
mere title under which he may be assigned to duty by the President.

A NEW POLICY AT ARMY HEADQUARTERS

As the first step in the experimental course decided upon, I sent
an order in writing to the adjutant-general, directing him never,
under any circumstances, to issue an order dictated by me, or in
my name, without first laying it before the Secretary of War; and
I made it known to all the staff that I disclaimed the right to
issue any order to the army without the knowledge of the President
or the Secretary. I also forbade the issuing of any order in my
name without my knowledge. The first rule was easy, the latter
very difficult, to enforce. I found, with no little surprise, that
the office of the "commanding general" usually learned for the
first time of routine orders issued in his name by seeing them
published in the New York papers the next day; and it was quite
difficult at first to make it distinctly understood that such a
practice coul