In industry there has to be a lot of component testing and that has to be done in simulations of real environments and then there has to be testing on engine test in essentially an engine environment. But before that you do a lot of material testing. Most of my testing was done on what are called single edge notch bend specimens, the old SENB. And if you imagine something like those wooden bricks that – those games that you can play in the garden where you build towers, what, kind of six inches long and half inch square. Well, if you imagine that in a nickel based alloy or a steel and you imagine there’s a v shaped notch in the top of one side and then I’m going to grow a very sharp crack from that notch and I’m going to do it by sometimes what I call three point bend loading, which means I put a roller underneath that bar directly opposite the notch and two rollers either side of the notch. And then I apply a bending moment. So I bring the top crosshead of my machine down and it applies load through the two rollers on the top of the bar against a fixed crosshead, which is supporting the bottom roller. So I’ve got my – my bar with a notch in the top. And then I apply a sinusoidal wave form to the crosshead of my testing machine so it’s – it’s sinusiodally varying the load and what I’m essentially doing is bending – bending the ends of the bar back and forward. And what that will do, if I’ve got a nice oiled machine, nice sharp notch, if I’ve taken out any compressive residual stresses by – by heat treating the specimen, I’ll be able to grow a nice straight fatigue crack from the notch. And when I’ve grown that to a fixed length, so it’s away from the stress field caused by the notch, I’ve got a specimen on which I can start to do a test to measure the rates of fatigue crack growth. And typically in a metal specimen, I’ll measure the rates of fatigue crack growth by applying a constant current to the specimen. I’ll have a constant current running through the metal specimen. And then either side of the notch on the top, I will spot weld two tiny little wires and I will monitor the potential drop across the crack. And as the crack – as the crack grows with the constant current, the potential drop will – will change. So by – so as I apply my alternating load with my potential drop measurement in place, I can monitor the growth of that crack and I can therefore explore different loading regimes. I can look at different load levels, which give me different stress levels in the bar. And I can look at different crack lengths, ‘cause I’m measuring the crack length with potential drop. So I can explore the stress intensity factor, which is a function of the stress range and the crack length, and I – so I can explore how – the rate of crack growth, the rate of – the change in crack length by the number of fatigue cycles. And that’s – that’s the basis for a lot of our understanding. And then I – I can take that bar. It can be made of different types of nickel based alloy, or I can give that nickel based alloy different thermal treatments. So I did a lot of that. And in the old days when I first started doing that we didn’t have computer control, so if you wanted to make sure your tests ran all through the night, which we had to ‘cause – in order to get enough cycles in, sometimes the test on a single specimen would run for one or two weeks if you were trying to explore behaviour at very slow crack growth rates. And you had to make sure your chart recorder didn’t run out of paper ‘cause that was recording your crack length, it was recording your potential drop, and you had to make sure the machine didn’t trip, then sometimes we spent nights in the lab looking after our tests. But sometimes we – if several of us were testing at the same time, as was quite often, we’d take it in turns to be in to check everybody’s tests and make sure their chart paper hadn’t run out and their pen was still working [laughs] and their machine hadn’t tripped [laughs]. And then when you’d finished that you had a specimen that had either broken in half or you’d stopped the test at a particular point and you then broke the specimen in half to reveal the – the crack you wanted to look at. Then you had this problem in that you had these quite big lumps of extremely strong metal and you had to saw the fractured ends off so you could put them in the scanning electron microscope [laughs].