110 The N.Z. Journal of Science and Technology. [Mar. 
would have been done on each sun, so the rise in temperature would be 
only one-quarter of that mentioned in the last example. 
But, on the other hand, most stars have considerable proper motions, 
and in nearly all cases of collision these initial velocities will increase the 
relative velocity of the impact. 
Again, the attraction between suns is proportional to the product of 
their masses; so in considering a collision between a pair of exactly equal 
and similar suns we may say that, other things being equal, the force that 
draws them together is proportional to the square of their common density. 
Doubtless many stars are denser than our Sun, and if twice as dense they 
would meet with twice the velocity. 
A collision of two stars, then, is capable of transforming kinetic energy 
into heat in such enormous quantity for each unit of mass involved 
that the resulting temperature may be hundreds of millions of degrees 
centigrade. 
As far, then, as the requisite energy is concerned, the impact of two 
stars is well able to account for the phenomenon. But some astronomers, 
while granting that this is true, consider that the chance is very small 
of impacts between stars occurring sufficiently frequently to account for 
the numbers of observed novae, and point out that impacts between other 
orders of celestial bodies are more likely. 
The probability of stellar impacts will be discussed later. Meanwhile, 
let us consider briefly the last objection. 
In our solar system, in addition to the Sun and its attendant planets 
with their satellites, there are numerous comets and countless meteorites. 
In our Milky Way, besides the stars which probably have their own 
planetary systems, there are star clusters, planetary nebulae, and the great 
irregular gaseous nebulae. Outside the Milky Way there are between 
half a million and a million spiral nebulae within the range of the largest 
existing telescopes. Each one of these is probably a cosmic system, perhaps 
resembling our own Milky Way. 
Now, everything in the Universe, as far as we know, is in rapid motion, 
and these motions, though doubtless organized, appear to us almost as 
indiscriminate as those of molecules in a gas. And just as the molecules 
of a gas encounter one another in their rapid flight, so cosmical bodies and 
systems of every order appear liable to collide. We should therefore study 
the probable effects of every possible variety of celestial encounter and see 
whether any of them agree with the facts of observation. 
It is often stated, for example, that, gaseous nebulae being extremely 
large, and stars by comparison insignificantly small, the impact of a star 
with a nebula is much more probable than of one star with another. This 
is reasonable enough ; but to suggest that a typical nova is the result of 
such a collision is a very different matter. 
Gaseous nebulae, though of enormous volume, have very small masses, 
and must be of exceedingly low density. That of the Great Nebula of 
Orion has been estimated not to exceed one ten-thousand-millionth part of 
the density of air at normal atmospheric pressure,* and the density of air 
is less than one thousandth part of that of the Sun. 
On approaching such a nebula a star would not have its velocity appreci¬ 
ably increased, and it would enter the nebula with about its original proper 
motion. Supposing this to be 30 miles a second, or one-tenth of a relative 
velocity that is common in the case of stellar impacts, let us consider the 
* Clarke, Problems in Astrophysics, p. 534. 
