This, it may fairly be said, is very speculative, but the fact remains that celestial bodies appear to be the only places in which the complex elements may be in actual process of formation from their known source--hydrogen. At least we may see what a vast variety of physical conditions these cosmic crucibles afford. At one end of the scale we have the excessively tenuous nebulae, the luminosity of which, mysterious in its origin, resembles the electric glow in our vacuum tubes. Here we can detect only the lightest and simplest of the elements. In the giant stars, also extremely tenuous (the density of Betelgeuse can hardly exceed one-thousandth of an atmosphere) we observe the spectra of iron, manganese, titanium, calcium, chromium, magnesium, vanadium, and sodium, in addition to titanium oxide. The outer part of these bodies, from which light reaches us, must therefore be at a temperature of only a few thousand degrees, but vastly higher temperatures must prevail at their centres. In passing up the temperature curve more and more elements appear, the surface temperature rises, and the internal temperature may reach millions of degrees. At the same time the pressure within must also rise, reaching enormous figures in the last stages of stellar life. Cook has calculated that the pressure at the centre of the earth is between 4,000 and 10,000 tons per square inch, and this must be only a very small fraction of that attained within larger celestial bodies. Jeans has computed the pressure at the centre of two colliding stars as they strike and flatten, and finds it may be of the order of 1,000,000,000 tons per square inch--sufficient, if their diameter be equal to that of the sun--to vaporize them 100,000 times over. Compare these pressures with the highest that can be produced on earth. If the German gun that bombarded Paris were loaded with a solid steel projectile of suitable dimensions, a muzzle velocity of 6,000 feet per second could be reached. Suppose this to be fired into a tapered hole in a great block of steel. The instantaneous pressure, according to Cook, would be about 7,000 tons per square inch, only 1/150000 of that possible through the collision of the largest stars. [Illustration: Fig. 41. Mount San Antonio as seen from Mount Wilson. Michelson is measuring the velocity of light between stations on Mount Wilson and Mount San Antonio. Astronomical observations afford the best means, however, of detecting any possible difference between the velocities of light of different colors. From studies of variable stars in the cluster Messier 5 Shapley concludes that if there is any difference between the velocities of blue and yellow light in free space it cannot exceed two inches in one second, the time in which light travels 186,000 miles.]