In 2007 the field of nuclear astrophysics celebrated its 50th birthday.

I discuss especially my summer with Willy Fowler at Kellogg Radiation Laboratory in 1951, where I did my ‘triple alpha’ work. I also go back even earlier to Arthur Eddington and Hans Bethe. The 1953 summer school in Ann Arbor only gets a mention.

Primitive meteorites and interplanetary dust particles contain small grains that originated in stellar outflows and supernova explosions. These μm- and sub-μm-sized presolar grains can be isolated and studied for their isotopic compositions in the laboratory. They are recognised as stardust by their isotopic compositions, which are completely different from those of the Solar System. They provide new information on stellar evolution, nucleosynthesis, mixing processes in asymptotic giant branch (AGB) stars and supernovae, and Galactic chemical evolution. Red giants, AGB stars, Type II supernovae and possibly novae have been identified as stellar sources of the grains. Of the eight nuclear processes proposed by Burbidge et al. (1957), signatures of all except the r-process can be found in presolar dust grains.

Starting from a sketch of the s-process concept formulated 50 years ago, the nuclear physics data for s-process calculations are briefly reviewed with emphasis on the status of neutron capture cross sections and beta decay rates. Accurate and comprehensive experimental data are mandatory as direct input for s-process calculations as well as for improving the complementary information from nuclear theory. The current challenges of the field are discussed in the light of new or optimized methods and state-of-the-art facilities, indicating the potential for accurate measurements and the possibility to study cross sections of radioactive isotopes. These opportunities will be considerably enriched by the enormous improvements provided by new facilities.

In this talk I shall start by describing how we set about and carried out the work that led to the publication of Burbidge et al. (1957, hereafter B2FH). I then shall try and relate this work and the circumstances that surrounded it to the larger problem of the origin and formation of the universe. Here it is necessary to look back at the way that ideas developed and how, in many situations, astronomers went astray. Of course this is a personal view, though I very strongly believe that if he were still here, it is the approach that Fred Hoyle would take.

I start by describing the problems originally encountered by Gamow and his associates in trying to decide where the helium was made. This leads me to a modern discussion of the origin of 2D, 3He, 4He and 7Li, originally described by B2FH as due to the x-process. While it is generally argued, following Gamow, Alpher, and Herman, that these isotopes were synthesised in a big bang I shall show that it is equally likely that these isotopes were made in active galactic nuclei, as was the cosmic microwave background (CMB), in a cyclic universe model. The key piece of observational evidence is that the amount of energy released in the conversion of hydrogen to helium in the universe is very close to the energy carried by the CMB, namely ∼4.5 × 10−13 erg cm−3.

An overview of the sources for heavy elements in the early Galaxy is given. It is shown that observations of abundances in metal-poor stars can be used along with a basic understanding of stellar models to guide the search for the source of the heavy r-process nuclei (r-nuclei). Observations show that this source produces very little of the elements from C through Zn, including Fe. This strongly suggests that O–Ne–Mg core-collapse supernovae (SNe) from progenitors of ∼8–11 M⊙ are the source for the heavy r-nuclei. It is shown that a two-component model based on the abundances of Fe (from Fe core-collapse SNe) and Eu (from O–Ne–Mg core-collapse SNe) gives very good quantitative predictions for the abundances of all the other elements in metal-poor stars.

I discuss the development and resolution of the solar neutrino problem, as well as opportunities now open to us to extend our knowledge of main-sequence stellar evolution and neutrino astrophysics.