March 9, 1871 | 

NATURE 
SA 

Fourth, the same when the blowing was finished, but 
before the spiegeleisen was added. 
Fifth, the finished steel when poured into the ingot 
moulds, 
I 2 3 4 5 
Combined carbon t'o00 §=63'040.-—«1'640. 0190-0370 
Graphite . . .2°570 trace trace trace trace 
Silicon . . . .2°260 01955 0470 trace _ trace 
Sulphur. . . .@107 o091 01098 0'093 o'090 
Phosphorus . .0°073 0070 0070 0'070 0059 
Manganese. . . 0410 trace trace trace o'540 
We shall now be able to understand the changes I have 
described as occurring in the flame. Before the full com- 
bustion of the carbon can commence, there is about 24 
per cent. of silicon to be converted into silicic acid. Ina 
charge of 6 tons this amounts to 3 cwt. For the complete 
combustion of this, nearly 3} cwt. of oxygen, or about 
14 cwt. of atmospheric air, is necessary. My explanation 
of the smaller and less brilliant flame that at first roars 
from the mouth of the converter is that it is mainly a silicon 
flame, mingled, however, with a small proportion of carbon 
flame; that the amount of silicon combustion goes on 
diminishing, and in a proportionate degree the carbon 
combustion increases, as the demand of the silicon upon 
the oxygen of the blast diminishes in consequence of its 
less abundant diffusion among the melted iron. 
I shall not be surprised if this explanation is contro- 
verted, as in offering it I fly in the face of the spectro- 
scope, which has made such glorious conquests that 
modern philosophers are disposed to trust it as implicitly 
as successful soldiers rely upon the general who has led 
them continually to victory; but without failing in due 
deference to those who are more skilful than I am in the 
use of this instrument, I am satisfied that in this and 
other cases where the question has been to determine the 
presence or absence of the mefadlozds, the negative replies 
of the spectroscope have been too hastily accepted. 
Prof. Roscoe, who devoted a considerable amount of 
time and labour to the spectroscopic examination of the 
Bessemer flame, says, ‘‘ Those who are practically en- 
gaged in working this process would like spectrum analysis 
todo a great deal more; they would like to be told 
whether there is any sulphur, phosphorus, or silicon in 
their steel ; questions which, unfortunately, at present 
spectrum analysis cannot answer, for this very good reason, 
that these substances do not appear at all as gases in the 
flame, but that they either remain unvolatilised in the 
molten metal, or swim on its surface in the slag of the ore ; 
and, consequently, the lines of these bodies are not seen 
in the spectrum of the flame.” Dr. Watts’s observations 
and conclusions accord with those of Prof. Roscoe. 
If by the above Prof. Roscoe is to be understood as 
asserting that no portion of the Bessemer flame at any 
period of its existence is due to the combustion of silicon, 
or that silicon is not present in the Bessemer flame, I must 
very decidedly affirm that such conclusion is erroneous. 
I do not for a moment question the accuracy of the 
observations of both Prof. Roscoe and Dr. Watts. I merely 
maintain that the absence of “the lines” of silicon in the 
spectrum of the flame does not prove its absence as a 
constituent in producing such flame, and for the following 
reasons :— 
We know that that the silicon existed in the pig-iron in 
the proportion already stated, and that although a very 
small quantity of that which ordinary analysis detects 
may have existed as entangled silicate in the pig, and 
another small portion is of course oxidised in the cupola, 
the bulk of it enters the converter as unoxidised silicon, and 
that it is oxidised and converted into silicic acid during 
the blow. We also know that silicon when heated in air 
or oxygen burns brilliantly, and that the product of such 
combustion when heated with a blast such as that which 
supplies its oxygen in the Bessemer converter, is sufficiently 

volatile to form concretions in the throats of furnaces 
which have been compared to natural chalcedony. Besides 
this a large quantity of solid matter is mechanically forced 
into the flame, and is seen above as a red smoke, which, 
without the slightest indication of unburnt carbon, is often 
sufficiently dense to hide the mid-day sun. The greater 
the quantity of silicon in the pig the more dense is this 
red smoke, which appears to consist of silicate and peroxide 
ofiron, Imaintain, therefore, that silicon is there, and that 
it must contribute to the luminosity of the flame, though 
it shows no characteristic “lines” in the spectrum. 
Under such circumstances, we have no good a priori 
reasons for looking for the silicon “es; a continuous 
spectrum being that which we are theoretically justified 
in anticipating as the result of such combustion of silicon, 
and this is exactly what the spectroscope reveals. The 
spectrum of the Bessemer flame at the commencement 
and early stages of the blow is of a most uncommunica- 
tive continuous character: occasional flickerings and 
vanishing ghosts of lines and bands come and go with 
perplexing irregularity ; and even the brilliant and ever 
obtrusive sodium lines do not appear at this stage, but 
commence with spasmodic flashings across the spectrum 
at about the period when the elongation and brightening 
of the flames which I have described is most decidedly 
taking place. When the flame has reached its maximum 
of extension and brilliancy, the sodium lines cease their 
intermittent flashings, and become a steady stream of 
light, the lithium band appears (though not in every blow), 
and the whole spectrum becomes striped, but the con- 
tinuous spectrum still remains as the permanent back- 
ground, 
My general reason for questioning the negative conclu- 
sions of the spectroscope in reference to silicon and the 
other metalloids is, that these bodies usually give a con- 
tinuous spectrum when, as elements, they combine freely 
with oxygen, as in direct unrestrained combustion under 
ordinary pressure in the open air. It appears to me that 
there is thus presented a broad distinction between the 
spectra of the metals and of the non-metallic elements, 
which is of great practical importance, and which has not 
been sufficiently considered, when conclusions have been 
based on negative spectroscopic results. I have already, in 
chapter 13 of “ The Fuel of the Sun,” referred to the worth- 
lessness of the negative evidence of the spectroscope in 
reference to the non-existence of the metalloids in the sun, 
and maintain that “they may all be there though the 
spectroscope should not detect one of them.” The mere 
fact that nothing but metals (I include hydrogen with 
these) should have been discovered in the sun is very 
suggestive. W. MattTieu WILLIAMS 


SOCIETIES AND ACADEMIES 
LONDON 
Royal Society, February 23.—‘‘On the Thermo-electric 
action of Metals and Liquids,” by George Gore, F.R.S. Itis well 
known that the degree of rapidity with which a metal immersed in 
an acid, alkaline, or saline liquid is corroded varies con- 
siderably with the temperature, and that the speed of corro- 
sion usually increases with the heat ; alsoa few experiments have 
been published (Gmelin’s ‘‘ Handbook of Chemistry,” vol. i. 
p- 375) showing that changes of electrical state occur in metals 
under such circumstances ; but a further examination of the rela- 
tions of the temperature and chemical change to the electrical 
state has not, that I am aware, yet been made. 
In an investigation on the development of electric currents by 
unequally heated metals in liquids (PA:/. Mag. 1857, vol. xiii, 
p- 1), I found that hot platinum was electro-negative to cold 
platinum in liquids of acid reaction, and positive to it in alkaline 
ones, provided in all cases chemical action was completely or 
sufficiently excluded. In the present experiments I have 
endeavoured to ascertain what electrical changes are produced in 
cases where chemical action more freely occurs, and I have there- 
