Let's take his experiment. He has a test tube in which he has the building blocks for nucleic acids, in which he has the right pH conditions, ionic strength conditions, everything are right, and the isolated enzyme. Now all he has to do is to add a few strengths of nucleic acids which then can be replicated by the enzyme using the energy-rich building blocks. Now if he would let this experiment go, you would have soon saturated your test-tube with nucleic acid strengths, so what he is doing, he lets it grow exponentially, let's say by a certain factor, let's say he lets it grow ten times, so that he gives... he adds a certain number of strengths and they replicate now so that he has ten times as many strengths. Then he stops the experiment and takes out one-tenth of that solution, so comes back to the... and puts it into a new glass with such a reaction mixture. Lets it grow a factor 10, stops it, takes out one-tenth. You can go on indefinitely with that. Because you always keep the concentration at the same level, it grows a factor 10 and you take only a tenth out of it. But evolution goes on. During that time you can adapt it to a new purpose, and Spiegelman was the first to do such experiments.

First of all it was a great thing to show that all you need is the enzyme, the replicase, and the template ‒ the nucleic acid ‒ and the building blocks, and can do that in the test tube then.

[Q] But you mentioned he didn't know about the error threshold.

That's right.

[Q] When you did the evolutionary experiments in your laboratory you know of the error threshold. Wasn't it Charles Weissmann who was, so to speak, the...?

Let me... I come back to that. But let me go on with Spiegelman. What he did is really the first to show you can do that with single molecules in the test tube and you get evolution. The first experiment he did was a complete phage Q-beta with his four thousand and some nucleotides, and what happened... the molecule became shorter and shorter and shorter, up to five hundred, so it went down to about a tenth of the length. Why?

[Q] With the same efficiency?

Yes, even with a larger efficiency. Only... you can evolve only the better adapted ones. So if the one... the shorter one grows out of the solution then it must be better than the wild type, otherwise the wild type would have won the competition. Now what happens? This phage is replicated in a test tube and you give the energy-rich building blocks, so the phage doesn't have to infect a cell any more, a coli cell.  So that means it throws away all the information which it would need to infect cells, and it doesn't need... it gets its enzyme, it gets its food, its building blocks, so it throws away everything until it has a shorter nucleic acid which still is recognised by the enzyme and...  If you would make it even shorter than that, then the enzyme... then the efficiency of recognition would be less. But here he got something which is a tenth of the original length and therefore takes much less time to do one, so the speed goes up, and even the speed per nucleotide went up by a factor of three. That was the first experiment... very interesting.

Then he did another experiment. He took a short nucleic acid, which I mentioned... the 216 nucleotides which he called a midi variant... and he added to his solution a selection pressure, in form of a substance which inhibits the replication. In other words there is an interchelating substance which sits in between the bases and hinders the reaction. Now what can evolution do? You can evolve strands which do not get hindered by this substance, so you favour those strands which can cope with the selection pressure, and sure enough he got out a strand which was much better than the wild type, with this... in the presence of this inhibitor. And he even got out a strand which was addicted to ethidium bromide, in other words which worked better in presence of this substance than in absence.

These were quantitative evolution experiments. The mutants he got out were not very efficient, in that the mutants were for error mutants, and that's due to the fact that he was too precise... the short end with that given error rate did not give him efficient way of evolution.