20 ORGANIC BEHAVIOUR 
constitution of nature, neither more nor less mysterious than, 
let us say, crystallization or chemical combination. But 
associated and closely interwoven with all that is distinctively 
organic there is much which can to some extent be interpreted 
in terms of physics and chemistry. 
The animal * has sometimes been likened to a steam-engine, 
in which the food is the fuel which enters into combustion 
with the oxygen taken in through the lungs. It may be worth 
while to modify and modernize this analogy—always remember- 
ing, however, that such an analogy must not be pushed too far. 
In the ordinary steam-engine the fuel is placed in the fire- 
box, to which the oxygen of the air gains access; the heat 
produced by the combustion converts the water in the boiler 
into steam, which is made to act upon the piston, and thus set 
the machinery in motion. But there is another kindof engine, 
now extensively used, which works on a different principle. 
In the gas-engine the fuel is gaseous, and it can thus be 
introduced in a state of intimate mixture with the oxygen with 
which it is to unite in combustion. This isa great advantage. 
The two can unite rapidly and explosively. In gunpowder the 
same end is effected by mixing the carbon and sulphur with 
nitre, which contains the oxygen necessary for their explosive 
combustion. And this is carried still further in dynamite and 
gun-cotton, where the elements necessary for explosive com- 
bustion are not merely mechanically mixed, but are chemically 
combined in a highly unstable compound. 
But in the gas-engine, not only are the fuel and the oxygen 
thus intimately mixed, but the controlled explosions are cacsed 
to act directly on the piston, and not through the intervention 
of water in a boiler. Whereas, therefore, in the steam-engine 
the combustion is to some extent external to the working of 
the machine, in the gas-engine it is to a large extent internal 
and direct. 
Now, instead of likening the animal as a whole to a steam- 
engine, it is more satisfactory to liken each cell to an automatic 
* The following paragraphs are taken with some slight changes from 
* Animal Life and Intelligence,” pp. 30-39. 
