On 18 August 1868 a total solar eclipse crossed India, and Jules Janssen, observing from Guntur, turned a spectroscope on the prominences at the Sun's edge. Among the expected lines he recorded a bright yellow one at a wavelength close to, but not the same as, the familiar sodium pair. In October of the same year Norman Lockyer in London, using a method that let him observe prominences without an eclipse, saw the same line. Lockyer and Edward Frankland concluded that no known element produced it and proposed a new one, named for Helios.
The claim was not well received. Announcing an element observed only in the atmosphere of the Sun, with no sample, no weight and no chemistry, was a considerable liberty, and for a quarter of a century helium was a line in a spectrum and an argument about it.
The matter was settled in 1895. William Ramsay, treating the uranium mineral cleveite with acid, collected a gas showing exactly Lockyer's yellow line; Per Teodor Cleve and Nils Abraham Langlet in Uppsala reached the same result independently and almost simultaneously. Helium was real, it was on Earth, and it had been sitting inside uranium ore the whole time.
That last detail is the explanation. An alpha particle is a helium nucleus. Every alpha decay in the crust makes another helium atom, and the gas accumulates in the same porous rock traps that hold natural gas. Terrestrial helium is therefore not primordial but a by-product of radioactive decay, manufactured slowly over billions of years and, once released into the air, lost to space permanently. The world's supply comes from a small number of gas fields, and it is finite in a way that most resources are not.