nt ourselves with what offers the fewest disagreements. An obvious
method, if we could depend on Columbus's dead reckoning, would be to see
for what island the actual distance from the Canaries would be nearest
to his computed run; but currents and errors of the eye necessarily
throw this sort of computation out of the question, and Captain G. A.
Fox, who has tried it, finds that Cat Island is three hundred and
seventeen, the Grand Turk six hundred and twenty-four nautical miles,
and the other supposable points at intermediate distances out of the way
as compared with his computation of the distance run by Columbus, three
thousand four hundred and fifty-eight of such miles.

The reader will remember the Bahama group as a range of islands, islets,
and rocks, said to be some three thousand in number, running southeast
from a point part way up the Florida coast, and approaching at the other
end the coast of Hispaniola. In the latitude of the lower point of
Florida, and five degrees east of it, is the island of San Salvador or
Cat Island, which is the most northerly of those claimed to have been
the landfall of Columbus. Proceeding down the group, we encounter
Watling's, Samana, Acklin (with the Plana Cays), Mariguana, and the
Grand Turk,--all of which have their advocates. The three methods of
identification which have been followed are, first, by plotting the
outward track; second, by plotting the track between the landfall and
Cuba, both forward and backward; third, by applying the descriptions,
particularly Columbus's, of the island first seen. In this last test,
Harrisse prefers to apply the description of Las Casas, which is
borrowed in part from that of the _Historie_, and he reconciles
Columbus's apparent discrepancy when he says in one place that the
island was "pretty large," and in another "small," by supposing that he
may have applied these opposite terms, the lesser to the Plana Cays, as
first seen, and the other to the Crooked Group, or Acklin Island, lying
just westerly, on which he may have landed. Harrisse is the only one who
makes this identification; and he finds some confirmation in later maps,
which show thereabout an island, Triango or Triangulo, a name said by
Las Casas to have been applied to Guanahani at a later day. There is no
known map earlier than 1540 bearing this alternative name of Triango.

San Salvador seems to have been the island selected by the earliest of
modern inquirers in the sevente