ed to warn me of approaching
danger. It is not that sound which I have elsewhere described as the
alarm-sound, and which is used only in case of imminent and awful
danger; but this sound is used in case of remote danger or in announcing
something unusual. As nearly as I can represent the sound by letters, it
would be "e-c-g-k," and with this word I have been warned by these
little friends many times since I first heard it from Nellie.

[Sidenote: NELLIE'S ACTIONS ALMOST HUMAN]

In the following experiment this sound was used with great effect.
Nellie's cage occupied a place in my study near my desk. She would stay
awake at night as long as the light was kept burning, and as I have
always kept late hours, I did not violate the rule of my life in order
to give her a good night's rest. About two o'clock one morning, when I
was about to retire, I found Nellie wide awake. I drew my chair up to
her cage, and sat watching her pranks as she tried to entertain me with
bells and toys. I tied a long thread to a glove, which I placed in a
corner of the room at a distance of several feet from me, but without
letting her see it. I held one end of the string in my hand, I drew the
glove obliquely across the floor towards the cage. When I first
tightened the string, which I had drawn across one knee and under the
other, the glove moved very slightly, and this her quick eye caught at
the very first motion. Standing almost on tip-toe, her mouth half open,
she would peep cautiously at the glove, and then in a low whisper would
say "e-c-g-k"! And every second or so would repeat it, at the same time
watching me, to see whether I was aware of the approach of this goblin.
Her actions were almost human, while her movements were as stealthy as
those of a cat. As the glove came closer and closer she became more and
more demonstrative, and when at last she saw the monster climbing up the
leg of my trousers, she uttered the sound aloud and very rapidly, and
tried to get to the object, which she evidently thought was some living
thing. She detected the thread with which I drew the glove across the
floor, but seemed in doubt as to what part it played in this act. I saw
her eyes several times follow the thread from my knee to the glove, but
I do not think she discovered what caused the glove to move. Having done
this for a few times, however, with about the same result each time, I
relieved her anxiety and fright by allowing her to examine the glove,