Experiments and Discoveries. 73 
with every plate for that purpose. This table comprises a se- 
ries of attractions obtained by the plate, comprehending all pos- 
sible limits for every class of vessels, and amongst them, of 
course, will be found that of any vessel in question ; correspond- 
ing to which, are given two numbers, the one being the distance 
of the centre of the plate below the pivot of the needle, and the 
other its distance from the centre line of the pedestal ; and at 
this depth and distance in the line of no attraction, already men- 
tioned, the plate is ultimately to be fixed. 
This being determined, the plate may now be placed either 
fore or aft of the compass ; in the former case, doubling the ef- 
fect of the vessel, and in the other neutralising it. When the 
former method is adopted, the plate is not a fixture, but is mere- 
ly applied, whenever it is thought necessary to correct the course, 
or to take the variation of the compass by azimuth, or ampli- 
tude observations ; in the other, it remains fixed in its place, du- 
ring the voyage, and the needle thereby left free to obey only 
the magnetic power of the earth. The former is better suited 
for southern voyages, in which the local attraction is not great, 
and the latter, in northern voyages, where this disturbance is 
very considerable ; it is, in fact, essentially necessary in the lat- 
ter case, to preserve the due action of the needle, which, as we 
shall see, is enabled, by this means, to continue its action with 
the plate attached, after it would cease altogether to act without 
it ; so that the experiment not only preserves the due direction 
of the needle, but it also renders it active after it would other- 
wise cease to traverse on its pivot. With respect to the plate, it 
has hitherto been made double ; viz. of two plates screwed to- 
gether, in such a manner, as to combine any strong irregular 
power of one with a like weak point in the other ; by which 
means a more uniform attraction is obtained. Mr Barlow, how- 
ever, seems to think that this precaution would not be necessary, 
if iron, weighing about 6 lb. per square foot, were employed ; 
but with thinner plate-iron, viz. of about 3 lb. per foot, it is re- 
quisite, not only for the purpose above stated, but also to pre- 
vent any accidental bending, by a fall, or otherwise. The plates 
may vary from 12 to 16 inches in diameter, according to the 
power of the vessel. They have a hole in their centre, through 
which is passed a brass socket, with a broad head, and with an 
