So we had... we knew of an enzyme that was exclusively associated with mitochondria; from our own work we knew of an enzyme that was almost, certainly exclusively associated with microsomes – glucose 6-phosphatase. And so the idea came to me to use those enzymes as markers, and that really became a key... a key step or a key notion in our approach to these problems. The use of marker enzymes to tell us where the particle... the host particles of the enzymes were. So using cytochrome oxidase we knew about mitochondria; assaying glucose 6-phosphatase we would know about microsomes. And so we did experiments in which we used the standard four-fraction scheme and compared the distribution of cytochrome oxidase, acid phosphatase and glucose 6-phosphatase. There was nothing wrong with our technique. Cytochrome oxidase behaved exactly as we expected and so the extra amount of acid phosphatase in the microsomal fraction could not be ascribed to mitochondria that had somehow moved from one fraction to the other. So the technique was okay; so now we were left with two explanations. Dual localisation or single localisation? Now, I think already I had at that time... I had a sort of prejudice in favour of single localisation because we had two examples: cytochrome oxidase – single localisation in mitochondria; glucose 6-phosphatase – single localisation in microsomes. But nevertheless, of course, we were open minded, and what we did was to start varying techniques to find out whether we could, by changing the centrifugation protocol, establish more clearly... or distinguish more clearly between the two... the two explanations. But here again, you know, in this story... I forget to tell you, but serendipity, chance, comes in all the time. I mean, if I can have some credit it is for making use of it, but without serendipity we would not have made the discoveries we made. And one of the... one of the serendipitous happenings was the centrifuge breaking down, and there was a young medical student Françoise Appelmans, who was working with us at that time, and she was... her job was to prepare the mitochondrial fractions by centrifugation that were used for all these latency experiments... studying how enzymes would be released from the particles. And so she came back to... she was crestfallen and said, ‘Centrifuge has broken down; I can't do the work.’ And so I said, ‘Well, never mind, I think we can get enough material by using a lower power... lower power centrifuge; we won't get all but we'll get... we'll get enough to do the experiments. And, well, we didn't. We got a nice sediment but there was almost no acid phosphatase or glucuronidase in the sediment because the centrifuge that we had used wasn't powerful enough to bring down those smaller particles. So this was the first hint that the enzyme may be in a smaller particle, and then we changed the scheme, the centrifugation scheme, and eventually hit upon a five-fraction scheme. Almost the same as the four-fraction scheme except that between the main mitochondrial fraction and the main microsomal fraction, which contained the bulk of cytochrome oxidase on one hand, the bulk of glucose 6-phosphatase on the other... when I say bulk, I mean 80%. We isolated a small fraction in between – an intermediary fraction we called L at that time – for light mitochondria. Contained only about 5% of the total nitrogen, the total protein, but it contained 50% of acid phosphatase. So now we had three peaks: one for cytochrome oxidase; one for acid phosphatase; and one for glucose 6-phosphatase. So the evidence was clear, acid phosphatase belonged neither to the mitochondria nor to the microsomes but it belongs to some unknown particle that has centrifugation... sedimentation properties such that it sediments at somewhat higher speeds, mostly, than the mitochondria, but lower speeds than the microsomes.