specific Gravity , and comparative Wear of Gold. 97 
appear evident, that extreme precision and uniformity in the 
results, can seldom be attained or expected. 
When metals are cast in a mould, they speedily become cold ; 
and, according to the quantity and quality of the metal, the 
figure and position of the mould, and the greater or less rapidity 
of the cooling, metals may vary in texture, and in the relative 
proportion and arrangement of their interstices ; and conse- 
quently the mass, in different parts, may be of unequal degrees 
of density. For, a metal of an uniform quality in other respects, 
generally becomes most dense in the bottom of the mould, 
especially when a long bar of heavy metal is cast in a vertical 
position. 
Those metals which are very ductile, may, by hammering 
and rolling, be brought more nearly to a certain uniform density ; 
for the number and capacity of the interstices, or air-bladders, 
in the interior of the mass, are thus more or less diminished; 
and, although the brittle metals, or semimetals, as they are im- 
properly called, cannot be thus treated, yet, when reduced to 
powder, or into small fragments, they expose a large surface, 
and consequently the error produced by interstices or cavities 
is much reduced.* 
But, neither hammering, rolling, nor pulverization, can be 
applied to those metallic substances, whether simple or mixed, 
• The interstices and cavities here mentioned, are those only which are formed 
during melting and casting ; for the natural grain and texture peculiar to each metal, 
cannot be changed by any of the methods employed to correct the irregularities which 
have been accidentally introduced ; and it appears proper to remark, that without very 
great caution, a new source of error may arise from the reduction of brittle metals into 
small fragments, or powder, which may conceal and retain little bubbles of air, so 
obstinately adherent to the metallic particles, as to require great patience and perse- 
verance before they can be properly and completely disengaged. 
MDCCCIII. O 
