MR. ROBERT MALLET ON VOLCANIC ENERGY. 
203 
Trachytes. 
A. 
B. 
c. 
Silica .... 
77-92 
76-67 
61-03 
Alumina . . . 
12-01) 
14-23 
m-21 
Oxide iron . . 
1-32/ 
\ 4-84 
Lime .... 
0-76 
1-44 
1-43 
Magnesia . 
0-13 
0-28 
2-07 
Potash . . . 
3-27 
3-20 
7-16 
Soda .... 
4-59 
4-18 
4-64 
A and C are the extremes of five analyses, B Bunsen’s normal trachyte of Iceland. 
Porphyry, Pitchstones, Obsidian, &c. give closely allied results. 
The preceding analyses of rocks are all taken from Blum’s fine work, 4 Lithologie oder 
Gesteinlehre ’ of 1860. 
166. The writer has been enabled to obtain, from measurements made daily and habi- 
tually in the progress of British plate-glass manufacture, certain returns, from which the 
coefficient of contraction in that material, between a temperature not far below perfect 
fusion and that of the atmosphere, may be obtained, though not with perfect accuracy, 
yet with an approximation to truth as great as, or possibly greater than, by means of any 
small number of direct experiments. 
167. In the manufacture of plate-glass the molten glass is suddenly poured out from the 
melting-pot upon the surface of a large cast-iron horizontal table. In breadth the mass 
is restrained from spreading laterally beyond two parallel strips of iron, fixed upon the 
table, the thickness of which determines that of the glass plate itself, which is produced 
by rapidly rolling over the heap of viscous glass a very heavy iron roller, by the passage 
of which the heap is evenly spread out, and a nearly rectangular plate produced, two of 
the sides being quite rectilinear and parallel, while the two other sides or ends are a 
little irregular. 
The moment after the roller has thus acted, a superintendent applies to the sheet a 
pair of graduated beam-callipers (specially prepared), and measures from a mean point 
chosen in the upper and lower ends of the plate (which he marks at once on the glass ) 
the mean length of the plate (the breadth being always the same), and the glass plate 
is as soon as possible slid off the table and removed to the “ leer” to be slowly cooled. 
After it has become quite cold and is removed from the “leer” for storage, it is again 
measured by the same callipers applied to the same marked places ; the width of the 
plate (known while hot by the space between the parallel rulers) is now also measured. 
The surface of the plate is then calculated from those two dimensions, and the dimen- 
sion and surface registered in a book, and so of every plate and of every day’s work in 
the year. 
168. Whether this be the universal practice of the trade or not, it has been so at the 
mdccclxxiii. 2 E 
