which occurred
three weeks after the wreck, was also observed as a holiday; and despite
our forlorn and rather precarious situation, we contrived to make a
fairly jolly day of it, the only discordant element being the boy
Julius, who early became sulky for some unaccountable reason, and spent
the entire day upon the topgallant forecastle with a rifle, shooting at
sea-birds and wasting some two hundred rounds of ball cartridge. I felt
strongly inclined several times to take the rifle forcibly from him, but
the mere hint of such a thing seemed to distress his mother so keenly
that I did not refer to it a second time. Yet I must confess that I
bitterly begrudged the utterly useless expenditure of so many good and,
in our case, valuable cartridges.

Now, it must not be supposed that, in our anxiety to recover as much
wreckage as possible, we forgot to keep a diligent look-out for passing
ships, for we did not. Nor did we neglect to dispatch a copy of our
appeal for help, securely sealed up in a bottle, regularly every day.
But thus far the horizon had remained blank while daylight lasted;
therefore if perchance any ships had passed us, they must have done so
during the night. Up to this we had all been working so hard that we
had deemed it hardly worth while to sacrifice our hours of rest for the
very doubtful advantage of maintaining a night watch; but with the
conclusion of what we considered our heaviest task, so far as actual
labour was concerned, we decided that it might be of advantage to keep a
look-out at night time, at least during the moonlit nights. We should
then be able to see a passing ship at such a distance as would enable us
to attract her attention by means of a flare. Accordingly it was
arranged that four of us, namely, Julius, the two stewardesses, and I,
should each take one watch in succession.

In that latitude, which was only a few degrees north of the Line, day
and night were approximately of equal length, and for all practical
purposes the night might be reckoned as beginning at six p.m., and
ending at six a.m. Therefore if each of us kept a watch of three hours,
we should cover the twelve hours between us. But by this arrangement
the same person would keep one particular watch every night, and, of
course, the least arduous of the watches would be that from six o'clock
to nine o'clock p.m.; I therefore decided to split this watch into two
dog-watches of one and a half hours each, by which