424 
NATORE 
[Sepz. 3, 1885 
and thus prevent it from cooling fast enough. We see, 
likewise, the necessity for a calm atmosphere, when we 
reflect that dew can only be deposited by means of the 
body cooling the air around it; now if this air is con- 
stantly renewed, it cannot cool this large body of air to 
any great extent, and hence dew cannot be formed. 
It is very manifest why the body must have a clear 
view of the sky, and why it must bea good radiator in 
order to promote the deposition of dew. Also why it must 
not be of a great mass, for, if it were, the heat from the 
interior might be conducted to the surface, and thus keep 
up the temperature. 
Finally, the substance must be near the earth, for, if 
not, the cooled air will fall down, giving place to warmer 
air. The body will thus have a larger mass of air to cool, 
and it will less easily succeed in bringing this mass below 
the dew point, I shall return to this subject at a later 
stage, when the part played by the aqueous vapour of the 
air is taken into account. Let me here state that there 
are regions in the earth where dew forms an important 
factor in agricultural operations. 
The artificial warming of our rooms is at present 
accomplished very much by radiation. 
of coal or wood acts by this process. The heated car- 
bonic acid gas which is the product of the combustion is 
carried up the chimney and out into the air, so that all 
that remains to heat the room is the light and heat given 
out by the glowing fire. 
It is by no means an economical use of heat, but there 
are other considerations besides those derived from 
economy, and an open fire will always be cherished by 
those nations whose social life is greatly within doors. 
The burning of gas in order to obtain illumination has 
nothing to recommend it. As it is used at present, it 
gives out a great deal of heat compared to its light, as 
well as a quantity of carbonic acid, and other products 
still more deleterious. 
It ought to be replaced by some kind of electric light, 
such as that proposed by Swan, where a thread of carbon 
is kept at a high temperature ina glass vacuum by means 
of an electric current. There the luminous effect is very 
large in comparison with the heat produced, besides which 
there is no foul air or other hurtful product. 
If we regard radiation as a means of increasing our 
An ordinary fire | 
knowledge, apart altogether from its primary and indis- | 
pensable action in rendering us acquainted by means of 
vision with the objects around us, we cannot have a better 
instance than that which is given us in spectrum analysis. 
Here, in the first place, a little reflection will convince us 
that we can gain hardly any knowlege by this means of 
the nature of a luminous solid or liquid body, for all such 
bodies at the same temperature will give out all the 
various rays which are possible to that temperature. 
There is, therefore, no means afforded us by their spectra 
an insular climate like ours, surrounded by sea-water 
which contains chloride of sodium. 
There are three chief points for consideration in the 
study of gaseous spectra :— 
(1) The effect produced by increasing the pressure of 
the gas. 
(2) The effect produced by giving the gas a motion to 
or from the observer. 
(3) The effect produced by increasing the temperature 
of the gas. 
The effect produced by increase of pressure consists in 
a widening of the bright lines. This subject was first 
studied by Frankland and Lockyer, who found that all 
lines are not affected by pressure to nearly the same 
extent. The F line produced by incandescent hydrogen 
was found by them to be peculiarly subject to an increase 
of pressure, widening out in certain cases to a really 
remarkable extent. 
Lockyer, who has since greatly studied this subject, is of 
opinion that it is not Pressure Per se that is influential in 
thickening the lines, but rather the freguency of encounters 
of precisely stmilar molecules. An important application 
of this law of pressure has been made by Lockyer, who 
has for this purpose used the electric arc, placing the slit 
of his spectroscope so as to embrace a section of this arc 
mid-way between its terminals and at right angles to its 
length. Now in the heart or central axis of this arc the 
gaseous particles which give out the light may be sup- 
posed to be somewhat near together, whereas at the 
border or circumference they are comparatively far apart. 
When the spectrum of such a transverse section is taken, 
this is found to consist of a number of bright lines, some 
long and some short. The long lines are those which 
remain visible even when the particles are far apart, while 
the short lines are those which require a greater nearness 
of particles to come out, and are therefore confined to 
the central regions of the arc. 
Suppose now that we take the spectrum of such an arc, 
from terminals composed of absolutely pure iron, and 
that by this means we obtain a number of long and 
short lines, characterising the spectrum of this metal in 
the state of vapour. 
Suppose next that we obtain the spectrum of some 
other metal, such as copper, which is not chemically pure, 
but which, we suspect, contains a little iron. We shall 
obtain, of course, the copper lines well defined and in- 
tense, #/ws an indication of the iron lines; but inasmuch 
as the iron particles are here few and far between, the 
iron lines which make their appearance will be those 
which do not require great nearness of particles in order 
| to come out—in other words, they will be the long iron 
of distinguishing one from another, so that spectrum | 5 ; : 
| of comparison is made much simpler, and we are enabled 
analysis is here impossible. 
It is very different, however, when we come to gases 
which give out spectra consisting of bright lines in | 
a dark background. Here there are various laws which 
combine not only to make spectrum analysis possible, but 
to constitute it an extremely delicate instrument of re- | 
search. J the first place, we have the law that the lines 
given out by any one elementary vapour are different in 
spectral position from those given out by any other. 
Secondly, as a rule such bright lines remain in their places 
throughout a great temperature range. Z/zrdly, an ex- 
ceedingly small amount of the element in question is 
generally sufficient to produce the lines. 
It is stated that by means of the spectroscope the 
presence of less than one two-hundred-millionth part 
. ne : : 
eee) of a grain of sodium may be detected 
Indeed, the difficulty is to get rid of the sodium line in 
lines, and not the short ones. In searching spectro- 
scopically for an impurity it is thus only necessary to 
direct our attention to the long lines of the various metals 
which we suspect to be present. Thus the whole process 
likewise to obtain with comparative ease the true spectra 
of the various elements. 
Let me now say a few words about the effect produced 
by a motion of the radiating gas to or from the observer. 
Suppose that a tram car starts from a station every five 
minutes In a certain direction, and that we are walking 
briskly /owards this station, we shall meet the cars oftener 
than every five minutes. On the other hand, if we are 
walking briskly from the station, they will overtake us 
less frequently than every five minutes. Suppose, again, 
that the whistle of a locomotive engine strikes the air 
I,coo times every second, then if the locomotive be at 
rest, we know from the theory of sound that the one 
blow will have advanced about 13 inches before the next 
is delivered to the air. 
If, however, the locomotive engine be itself travelling 
in this direction, it is evident that the interval between 
the blows will be less, for the engine may have itself 
