o causes, we have:
Six due to women, four to quarrels, five to other causes, and two
infanticides. Added to the manslaughters previously classified, we have
a total of sixty-two killings, due in twenty cases to quarrels, thirteen
to drink, nine to women, four to disputes over money, one to
race antagonism, five to general causes, three to negligence, two
infanticides, five during the commission of other crimes.

The significant features of this analysis are that about seventy-five
per cent of the killings were due to quarrels over small sums or
other matters, drink and women; over fifty per cent to drink and petty
quarrels; and about thirty per cent to quarrels simply. The trifling
character of the causes of the quarrels themselves is shown by the fact
that in three of these particular cases, tried in a single week, the
total amount involved in the disputes was only eighty-five cents. That
is about twenty-eight and one-half cents a life. Many a murder in a
barroom grows out of an argument over whether a glass of beer has, or
has not, been paid for, or whose turn it is to treat; and more than one
man has been killed in New York City because he was too clumsy to avoid
stepping on somebody's feet or bumping into another man on the sidewalk.

The writer sincerely regrets that his own lack of initiative prevented
his keeping a diary during his seven years's service as a prosecutor. It
is now impossible for him to refresh his memory as to the causes of all
the various homicides which he prosecuted, but where he can do so the
evidence points to a conclusion similar to that deduced from Mr. Nott's
record. The proximate causes were trifling--the underlying cause was
the lack of civilization of the defendant--his brutality and absence of
self-control.

With a view to ascertaining conditions in general throughout the United
States, I asked a clipping agency to send me the first one hundred
notices of actual homicides which should come under its scissors. The
immediate result of this experiment was that I received forty-five
notices supposedly relating to murders and homicides, which on closer
examination proved to be anything but what I wanted for the purpose in
view. With only one or two exceptions they related not to deaths from
violence reported as having occurred on any particular day, but to
notices of convictions, acquittals, indictments, pleas of guilty and
not guilty, rewards offered, sentences, executions, "suspicions"