Fortunately, all that we have learned hitherto from the stars favours the belief that, while a catastrophe of this sort may be possible, it is exceedingly unlikely. We may estimate the probabilities precisely in the same way that an insurance company estimates the chance of a railway accident. Such a company considers the number of accidents which occur among a given number of railway journeys, and from the smallness of the number of accidents compared with the largeness of the number of journeys estimates the safety of railway travelling. Our sun is one among many millions of suns, any one of which (though all but a few thousands are actually invisible) would become visible to the naked eye, if exposed to the same conditions as have affected the suns in flames described in the preceding pages. Seeing, then, that during the last two thousand years or thereabouts, only a few instances of the kind, certainly not so many as twenty, have been recorded, while there is reason to believe that some of these relate to the same star which has blazed out more than once, we may fairly consider the chance exceedingly small that during the next two thousand, or even the next twenty thousand years, our sun will be exposed to a catastrophe of the kind. We might arrive at this conclusion independently of any considerations tending to show that our sun belongs to a safe class of system-rulers, and that all, or nearly all, the great solar catastrophes have occurred among suns of a particular class. There are, however, several considerations of the kind which are worth noting.