meridian; that the definition of the
prime meridian must practically depend upon subsidiary considerations,
no matter where it might be located. In the practical work of
determining longitudes a connection with the prime meridian cannot be
made in each case. What is really determined is the longitude from
some intermediate point, generally in the same country, and in
telegraphic communication with the place whose longitude we wish to
know. This intermediate point would, for the time, be the practical
prime meridian. But the longitude of this point itself must always be
uncertain. Science is continually advancing in accuracy, and we find
that we continually need to correct the longitude of our intermediate
meridian, and hence of all points determined from it. How can this
difficulty of constantly changing longitudes be avoided? He replied
that each system of connected longitudes must rest upon its own basis.
It must be referred to an assumed prime meridian, and the
measurements must be made from that, even if it be found to be
somewhat in error. If some such system had been adopted thirty or
forty years ago, we would have avoided the confusion arising from the
fact that the longitudes given on many maps do not refer at all to any
absolute meridian. All that is known is that the astronomers
determined the longitude of the place, and then the maps had to be
corrected accordingly. The longitude of one place would be determined
from Cambridge, and perhaps in the neighborhood is another place
determined from the observatory at Washington. In either case we know
nothing of the longitude of Cambridge or Washington which the observer
assumed in his calculations.

Generally, in determining longitude, the country adopts the principal
place within its confines as a subsidiary prime meridian, and the
assumed longitude of this place is necessarily selected somewhat
arbitrarily. The longitude, for instance, of Washington was, thirty
years ago, known to be nearly 5 hours 8 minutes and 12 seconds west
from Greenwich. Had we adopted this difference by law, it would have
amounted to choosing for our prime meridian a point 5 hours 8 minutes
and 12 seconds east of Washington, whether we happened to strike the
transit instrument at Greenwich or not. This would have fixed an
assumed longitude for the Cambridge observatory and for all points
within our telegraphic net-work. We should have had a practical
system, which might, however, require