arently strange conduct, had they not observed, in the next instant,
that the arrow, after flying high up, came down again head-foremost and
stuck upright in the back of the turtle!

The turtle dived at once, and all of them expected to see the upright
arrow carried under water. What was their surprise as well as chagrin to
see that it had fallen out, and was floating on the surface! Of course
the wound had only been a slight one, and the turtle would escape, and
be none the worse for it.

But Guapo shared neither their surprise nor chagrin. Guapo felt sure
that the turtle was his, and said nothing; but, jumping into the canoe,
began to paddle himself out to where the creature had been last seen.
What could he be after? thought they.

As they watched him, they saw that he made for the floating arrow. "Oh!"
said they, "he is gone to recover it."

That seemed probable enough, but, to their astonishment, as he
approached the weapon it took a start, and ran away from him! Something
below dragged it along the water. That was clear, and they began to
comprehend the mystery. The _head_ of the arrow was still sticking in
the shell of the turtle. It was only the shaft that floated, and that
was attached to the head by a string! The latter had been but loosely
put on, so that the pressure of the water, as the turtle dived, should
separate it from the shaft, leaving the shaft with its cord to act as a
buoy, and discover the situation of the turtle.

Guapo, in his swift canoe, soon laid hold of the shaft, and after a
little careful manoeuvring, succeeded in landing his turtle high and dry
upon the bank. A splendid prize it proved. It was a "jurara"
tortoise--the "tataruga," or great turtle of the Portuguese, and its
shell was full three feet in diameter.

Guapo's mode of capturing the "jurara" is the same as that generally
practised by the Indians of the Amazon, although strong nets and the
hook are also used. The arrow is always discharged upwards, and the
range calculated with such skill, that it falls vertically on the shell
of the turtle, and penetrates deep enough to stick, and detach itself
from the shaft. This mode of shooting is necessary, else the jurara
could not be killed by an arrow, because it never shows more than the
tip of its snout above water, and any arrow hitting it in a direct
course would glance harmlessly from its shell. A good bowman among the
Indians will rarely miss shooting in this way,--long practi