So we rigged up an apparatus which essentially consisted of a glass reaction vessel which was surrounded by a heating element and an oven, and that oven was thermostated at successive temperatures, kept constant... and then this reaction flask was connected by a lead to the mass spectrometer, and at intervals during the reaction we would bleed some of the gas, a minute fraction, into the mass spectrometer which was immediately analysed and give us all the masses which were present and their concentrations by the height of the peaks. So we had all this information and the problem was how to deconvolute it and make it into compounds because we had to, as I say, discern the difference between a B₁₀H and a B₁₁ and so forth. And it turns out that the mathematics for doing this was very similar to the mathematics that was required to deconvolute the Mössbauer spectra, so although the two subjects seem miles apart, the way of reducing the data to something intelligible was fairly similar. And so, as I said, we had Terry Gibbs initially to help us with the mathematics of this, and we found that we could follow the course of the reaction – and again there’s no need to go into detail – but the method worked and we were able to, I think, give a reasonable explanation for how each of the various boron hydrides was formed and having been formed then subsequently reacted to form other boron hydrides... to the total that we could get right through the whole range of the simple boron hydrides, the first ten of them.

So that was a thing that we did with the... with the mass spectrometer and the gas phase reactions, and I was able to talk about this at several of the conferences and we wrote some of it up in full papers as well.