and Apparatus for producing Water. pr 
ferve to condu& the gazes into the balloon, are made of 
glafs, and cemented into the copper tubes gg, which have 
conical extremities like thofe above defcribed to fix them on 
the cocks. Thefe glafs tubes are raifed a little above 
the edge of the cylinder H, to prevent the water from 
paffing into the balloon, fhould you happen to fill one of the 
glaffes A, and negleét to fhut well the cock C. 
It will be proper to make the glaffes A, pretty large, if 
they can be procured fo, in order that the gazometers may 
contain the more air. It will however be beft to make 
them rather high (28 or 30 inches) than wide, in order 
that the exaétnefs of the fcale 25 may not be diminithed. 
M. Van Marum anfwers an objection which he lays may 
pofiibly be urged againft this apparatus, namely, that the 
combuftion is obliged to be fufpended every time that the 
gazometers need to be replenifhed with either oxygen or 
hydrogen. He fays this is no folid objeétion, as he finds 
that, fince he ufed platina for conveying the eleétric fpark, 
he never fails to kindle the hydrogen the firft moment it 
enters the balloon; and that, befides, the combuftion of a 
cubic foot of hydrogen gaz, in order that the water produced 
may not contain acid, ought to take fix hours. Thole who 
with to make experiments of longer duration, have only to 
attach two other gazometers ; by which means, while the 
one fet is emptying, the other may he replenithed. 
For this purpofe it will be neceffary to have, inftead of 
the cock w, two cocks Z 1, Z 2, (fig. 4) {crewed to a piece 
of copper fixed to the cover of the balloon ; and by which. 
each of thefe cocks has a communication with the bent tube 
In ™, by: means of two holes x, x, which proceed obliquely 
through this piece, and end in the tube 7. The firft gazo- 
meter, which communicates with the balloon by sy cock 
Zi, being almoft emptied, the cock Z2, of the fecond 
gazometer, may be opened, after having made the preffure 
jn the fecond gazometer equal to that's in the firft. It is 
gvident that the preflures of thefe two gazometers being 
equal, 
