Oliver Smithies:

I’m looking at a book that is labeled as 1951, 1. It’s a notebook from my post-doctoral period at the University of Wisconsin in the Department of Chemistry, where I was a post doc with Professor Jack Williams in physical chemistry at the University of Wisconsin in Madison. This notebook is the first of four notebooks which date back to that period and contain notes on various experiments related, in the first book at least, to electrophoresis. At that time there were not very many methods available for carrying out electrophoresis. The method that was used at the University of Wisconsin was the Tiselius method, in which a protein solution was inserted into a container, basically a u-tube, with a high concentration of protein, and then two boundaries were formed, one on the left hand upper part of the u-tube and the other on the right hand part of the u-tube. And then optical machinery as it were, lenses and light sources, etc., looked at the boundary formed, the tops of the two u-tubes between the protein solution and whatever buffer the proteins had been dissolved in, and the recordings were obtained by refractive index differences. The amount of protein was quite considerable, the order of several hundred milligrams per experiment, and the volume of solution was quite appreciable. I don’t remember exactly what, but to the order of maybe ten ccs. And this book describes my introduction to that method by looking at instructions given by the other professor who worked with Jack Williams. There were two of them, Lou Gosting and Alberty. Alberty, I forget his first name for the moment, it will become clear. There is a history, written by me, of those days, that has been published. It was published as a series of images and text under the title “How it all began: a personal history of gel electrophoresis” by myself, published in the journal Methods Mol Biol, volume 869, 1-21. So it’s possible to see a more coherent account of the electrophoresis experiments by going there.

However, let’s go back now to the Book 1, 1951. It says the notes on the papers are from the back, and practical notes from the front. But in fact I don’t find anything at the back. So this really is more or less in chronological order from the front. And it starts with quite simple experiments, with quite simple details, for example the first significant pages dated September 21, 1951: it’s describing my preparing a protein, beta lactoglobulin, which I had been familiar with from my graduate student days, because I’d use that protein for a lot of measurements. They’d also used it in Wisconsin, but I made a new batch of the protein. So it starts off with 10 am, milk from the University Farm was skimmed about 9 am by warming to 35 degrees centigrade, and then I started with 10 gallons. And batch one was 19 ½ liters, to which I added. And it goes on, 39.2 grams of ammonium sulphate per hundred mils, etc. etc., filtered through a coarse paper on plain funnels and onto a Buchner filter. And then it goes on, batch 2 was 15.24 liters, etc. etc. and so it goes on. And everyone can follow what was happening. Several pages later, for example, it says “final LG, October 3”, the LG being lactoglobulin, “decided to unite the recovered lactoglobulin and now recrystallized by me 3 times”. So I had in the period time crystallized it several times in the few pages which are there.

And it talks about Cecil-Ogston modified. Cecil, that was Rupert Cecil, and Ogston was Sandy Ogston, had a modified method for crystallization. And I got a final preparation after one more crystallization. I remember they were rather impressed with my ability to get crystals, because they made large quantities of beta lactoglobulin, but it was an oily material rather like a heavy oil that they would have in a container. And I showed them that all you had to do was scratch the side of the glass container to get a few flakes of glass which acted as a seed, and all the oily beta lactoglobulin turned into crystals. So I got good points for that and they thought it was rather impressive I think. Going on.

There’s another page on lactoglobulin recovery and so forth. And then lactoglobulin checks, checking it in the ultracentrifuge with the help of Sidney Schulman, whom I don’t remember anymore I’m afraid. But on the Svedberg ultracentrifuge and so on. And then on October 8th I did the first electrophoresis experiment with the Tiselius apparatus. Set it up in the cold room, pouring lactoglobulin into the u-tube, no pipettes, etc. etc. and boundaries very sharp, the protein concentration 1.8%, 1.8 grams per 100 milliliters. Let’s see whether it says what the volume was, I don’t see what the volume was there exactly. But it talks about the current was turned on, the time, and the angles of the photograph. Very detailed record of what happened. And then a couple of rather accurate diagrams of what the electrophoresis pattern looks like, and it was rather strange because the ascending and descending boundaries were not all the same, I found, and it’s obviously complicated. The descending boundary gave three peaks or three components, whereas the ascending boundary only gave one. And I later wrote about it, that it was really rather anomalous behavior, because the electrophoresis did not give the same results when you looked at the two boundaries. It was later part of a paper that I wrote about anomalies of different methods of looking at protein purity. So this goes on in this way, detailed calculations of the areas of the trace versus the angle of the knife edge, which was used to make what was called a schlieren boundary image. Here we are October 16th, setting up electrophoresis in the same solution, different angles for the method of illuminating the boundaries to give peaks. Detailed things.

There is an experiment here when trying to measure diffusion. Lou Gosting was I think an assistant professor at the time, but he was one of Jack Williams’ crew, Lou Gosting and Alberty together. And Lou Gosting was very well known for his ability to measure diffusion constants very accurately. And here I am, diffusion measurements with an equivalent of .43% protein. And it shows the tracings, etc., and several pages which are crossed out for some reason. And then Bob, here’s Alberty’s name, Bob Alberty, he had some magnification factors which I took down rather complicated notes of. And then the results of the diffusion are talked about. The diffusion constants of this week, etc. etc. And then I made a comment: the fact that this material has this diffusion constant leaves little doubt that it is lactoglobulin, Bob Alberty suggests that its steady state nature indicates its purity, meaning that it was nicely pure. They go on for many pages of electrophoresis, area checks, electrophoresis buffers. October 31st, boundary formed etc. etc., conductivity measurements, pH, all of them recorded in quite detail. November 2nd, here I’m measuring pH again. November 5th, buffer, protein, etc. Quite a lot of details. Calculations of volume changes, mobilities, descending and ascending boundaries. And so on. It goes on. The results were recorded by photography, and so there’s a page here talking about getting the right exposure during the latter part of the experiment, various exposures were attempted. And I was rather a fan of photography and I liked handling photographic material, and I tried reducing the potassium ferrocyanide to improve the exposures. Or rather the images of the experiments. And here we start with a series of precipitation tests and crystallization, labeled I. In some way I’ve got through the letters of the alphabet and now I’m at a page on the 26th of November labeled I, pH is too high, then J on the 27th of November, then K on the 27th. Further experiments, etc., repeat procedures used in the last set, but run experiments in buffered solutions. Comments of this sort. Nothing very profound. All details of the different experiments, conductivity, quite precise measurements. I’m really always concerned about getting accurate results as precisely as I can. So the pHs and the currents are recorded to four decimals. Time is 11:14, 10 milliamps, 11:14, 11:19, etc. and various other things. This experiment started at 11:14 and I’m still making measurements at 10:00 o’clock that night, I can see. 11:14, 12:05, 2:05, 5:05, 7:05, 10:05, 11:05, 12:05, even 1 o’clock in the morning. I must have been quite energetic on December 4, page M. N, O, P, we’re at P. December 11, last five runs practically indistinguishable, yet very different from the sodium chloride runs, whatever that means. Q, I’m at Q December 15, December 16, here I am January 16. Solubility. I’m using a pencil to write, I don’t know why instead of a pen, probably the pen had quit. Electrophoresis S, again. Blanks. And so on we go for a couple pages. 4 pages with big Xs through everything, I don’t know why. I see, I’m beginning to do some solubility measurements, I hadn’t realized quite what was happening. We’ll stop the recording a moment.

Changing the work that I’m doing. No longer electrophoresis, but I’m beginning to investigate what I could call the precipitation curves, with the protein beta 2 lactoglobulin, which is as I said is very pure, by all preceding methods of measurement, electrophoresis and diffusion and the ultracentrifuge. So this is solubility measurements. And I made up a solution of sodium chloride and sodium acetate and acetic acid to get to a known pH. .04 molar sodium chloride and .01 molar sodium acetate and .01 molar acetic acid. And then I began to add ammonium sulphate to cause precipitation. So this is the beginning of that work. A blank for the dilution factor of adding ammonium sulphate. And the change in extinction with ammonium sulphate buffer.

Some little comments here, I might as well read them. If we have present some substance, for example tyrosine, which is soluble and ammonium sulphate at 100% saturation, then measure the extinction of a solution against the added ammonium sulphate, we can correct for the dilution effect of the ammonium sulphate and for the extinction which it may introduce at this wavelength. And so on. So, June 19th I prepared the solution of tyrosine, so I’m doing it with tyrosine, just to get a sort of blank because it’s not precipitated by ammonium sulphate.

And the final entry in this book is electrophoresis. Entry S. Another experiment started but the plate cannot be found for this run, presumably it was abandoned for some reason. And so we come to the end of this book with the pipette calibrated and found to be accurate to plus or minus less than 1/10 of a percent and reproducible plus or minus 1/10 of a percent. I was trying to make sure that my pipettes were accurate. These were relatively small micro pipettes and the slope was very good. And I’m recording to rather a large number of decimal places. 5.3103 plus or minus .9, so it should have been 5.31 plus or minus .9 to be more precise. Almost at the end of this book. It’s saying that a very rough test with .1% protein shows that precipitation begins at between 25 and 30% saturation. No difference by naked eye once the precipitation began. Maybe a very narrow range. So there are several more pages again with rather too many decimal places. February 17th, now I’m starting to scale up the experiments. I set up 16 tubes with a range of 32-47 grams per hundred milliliters of ammonium sulphate. And only the last three had a precipitate which came out at room temp but not at 5 degrees, etc. etc. “Results promising”, however, I was evidently rather pleased with it. And that ends this book except for a couple of references to Bob Alberty’s paper on reversible boundary spreading and pH spreading by Longsworth, papers from 1948 and 1947. And we’ll have a look at Book 2 and see how the story matures.

So here we are looking at Book 2, which is still in the Department of Chemistry at the University of Wisconsin, 1952. And I say “If found, please return to O. Smithies, Department of Chemistry, University of Wisconsin, Madison, Wisconsin, USA, or Balliol College, Oxford, England.” So I still am a little bit homesick, I imagine. It begins with solubility measurements, so taking up the work that had already been begun in Book 1. The pH of the solutions were determined roughly and tubes were set up Thursday, February 21. Measurement on Saturday, February 23. And I make a note “the last few tubes were dirty with filter paper threads. All stock solutions have to be filtered through glass.” Filter paper threads were not acceptable. And the series of measurements of the transmission of the solution over the starting concentration. PH 4.5, set up another experiment on Saturday, etc. These tubes were centrifuged Monday, but determination was not possible until Tuesday. Some tubes showed very slight cloudiness. So evidently something had come out of solution after they had already been centrifuging. So I made a note: normally therefore centrifuge only before measurements. A large page full of data. PH 4.8, set up another series of tubes. Thursday, March 6, the a.m. centrifuge of the tubes, measured them in the evening. Several were cloudish, so I have to decant immediately, the centrification is over. So it’s obvious that something was happening that I didn’t understand there, but there are pages of measurements there, all with very precise measurements to four decimal places of 31.24, 32.35, and the transmission, etc. etc. Saturday March 8: set up an experiment with pH 4.5 again. Centrifuged 12 of the tubes, 6 were left. Sunday, working weekends, March 9. Pages of the experiments. Keeping going. March 16, more measurements. Then back to making lactoglobulin again, or doing something to it, because here it is: preparation of lactoglobulin for future work, Thursday the 20th of March. I recovered the old lactoglobulin which had been made January, it had come down as an oil and later crystallized, as I had mentioned. And notes on preparing lactoglobulin for another experiment. It goes on several pages. Back to exposure experiments, some more exposure experiments on the Schlieren optical system. Trying to get to different films to work.

Now here is the start of another set of experiments which I did. So far we’ve looked at the centrifuge, we’ve looked mainly at the electrophoresis, we’ve looked somewhere at the centrifuge measurements, which I haven’t noticed yet, but now we’re going to look at the protein by an antibody method. So there’s a change in topic here: rabbit antibody production, May 20. To make 50 milliliters of nearly 2% beta lactoglobulin in .15 molar sodium chloride that’s isotonic and a little bit of merthiolate to keep it sterile. It talks about making that. So eventually I end up with .97% beta lactoglobulin and .15 molar sodium chloride. Friday, May 22, two rabbits were injected subcutaneously, with 20 milligrams for each rabbit. And Monday, again, injection subcutaneously of 10 milligrams. And Wednesday and Friday. So three or four total doses. And then I make a note June 2 that Rabbit 2 has produced a litter. Gave 1 its fifth injection and number 3, a new rabbit, its first of 20 milligrams. And so on. It records Wednesday, Friday, Monday, Wednesday, etc., over that period of time. I was having help here, and I don’t have a note of who the person was helping. Except I have a note here “I decided to try Bob’s BSA, anti-bovine serum, with 15 milliliters in centrifuge cups and using .250 molar acetic acid to dissolve the precipitate to try to learn how to quantitate the precipitate that one got with the antibody.”

June 17, I’m now making serum, four hours at 37 degrees overnight in the cold room, etc. Required for the precipitation curves. Precipitin curves, that’s the antibody precipitation. A standard beta lactoglobulin solution. More accurate precipitin tests. Three pages on this. A note that the corks give absorption in the ultraviolet when extracted with ascetic acid. So stop using corks. A comment: “There is no doubt that this serum has a low antibody content, the precipitates small in test also tested in a gelatin gel with an antigen above for zonal precipitation. Precipitate very light. Same test with Bob’s potent serum gives a very heavy precipitate.” So I was getting rather poor antibodies. Tried to do some light scattering to measure the precipitant. September 23, I’m beginning to be a little bit more quantitative, because I’ve got 16 tubes set up roughly with the old 1/50 beta lactoglobulin in 1:1:1:2  dilutions, a rough test for the anti-serum. Quite reasonable looking precipitation curve.  There is an optimum ratio of antibody to solution to get precipitation, which of course is well known. A comment then on solubility, electrophoresis, purification, etc. repeat. And get the ten milliliter test tubes clean and dry. Two good 5 milliliter pipettes, etc., getting everything very clean. A comment: get new rabbits. I know at that time, I don’t know whether I have a note of it, but I had an accident because in trying to clean my tubes for solubility measurements with the ammonium sulphate, I used to use a cleaning solution which was made with concentrated nitric acid plus a little bit of alcohol. And it would sit there quietly for a few seconds and then would start to give a catalytic reaction and produce large quantities of nitrogen oxide and NO2 brown fumes, and the boiling would clean anything. It’s very dangerous. And there wasn’t a hood in the Physical Chemistry Department, so I used to do this reaction outside. And then one day by mistake, I used fuming nitric acid instead of nitric acid, which meant there was no latent period, and when I added the alcohol to the fuming nitric acid, it immediately gave this very powerful reaction and sprayed my face and my arms. Fortunately I had glasses on, but the glass frames were pitted and I got severe burns over quite a lot of my face and couldn’t get to the water very quickly. Had to get through a door and down a flight of steps before I could get to water. And managed to get most of my face and arms free of the acid, except for one place where I still have a scar. But my friend there, who shared the same room as me in the lab, Dick Goldberg, was very concerned and took me to the hospital immediately. And very annoyed that the nurses didn’t realize I was in sort of shock and it was really very serious. And he was, I remember, very upset about it. But I recovered. I never used that nitric acid method again. When I tried to publish about it later on, the journal wouldn’t let me describe the reaction, quite wisely, because it should never have been used. Incidentally, I learned it from my chemistry teacher in high school who used to use it to demonstrate this reaction. It was a reaction I learned as a high school student. Dangerous things, nobody would do anything like that these days.

I’m turning the pages here. The serum and dilution tubes. So I’m back to not antibody experiments, but to precipitation experiments. I could see precipitates were a (?) pretty hopeless cell design, I have a note, can’t use the vertical mask, must try horizontal one. I don’t know what that refers to. Trying to measure optical density. Trying to get light scattering to measure the turbidity. I think I made a mistake here, these are still antibody experiments. Saturday, October 9, trying to use light scattering, not very satisfactory. And a comment here: “The cells are too dirty. Clean and try again.” New cells. The light scattering. Just three bottles, I was trying light scattering. I don’t remember ever publishing on this so I may never have got any useful results on it. Rough test of scattering, about four times more, etc. Then another change, plan for future work, solubility. Centrifuge tests, 15,000 rpm. 2 milliliters solution. I did some ultracentrifuge tests. More light scattering, November 6, light scattering, nothing very profound. For reasons that I don’t remember at all, they’re using allantoic fluid and anti-flu serum. I think I must have been trying to help my girlfriend at that time who later I married, Lois, later Smithies, so she was doing antibodies and I was testing her allantoic fluid with anti-flu serum to try to get a method of measuring antibodies in allantoic fluid. Some measurements contaminating (?). Back to lactoglobulin November 12, test of lactoglobulin concentration for solubility work. Unfortunately merthiolate absorbs strongly at 278 millimicrons. Cannot use it or the solutions prepared with it. November: precipitation almost immediately in all but the first two tubes. PH 8 5.4. December 1, working on December 1 I see, set up five lots of 16 cubes. That’s 80 cubes. December 3, December 5, and here are some notes on various papers now. So the idea that the back of the book would be notes on papers in this case is true. That I’m getting into coming backwards from various papers. I’m going to go to the back and start at the back. Virus work with Lois K. Kitsy, with LKK, and it talks about a paper by Oster in Science 1946. A paper by Gerald Oster Rockefeller. This is going backwards in this book. Notes on various papers. Going backwards again from the end of the book. Zinn 1948, Zinn 1948, Zinn 1948, and there it has reached the forward notes on the experiments, so that’s it. End of Book 2, 1952.

I’m looking now at Book 3 which is 1952-1953. Still O. Smithies lab notes, and still at the University of Wisconsin Chemistry Department. Inserted at the beginning of this book is a letter from Bob and I’m not sure which Bob this is at the present time, I might remember. But he’s writing from back in England, the Queen’s Road Estate, Beeston, Nottinghamshire, November 27, 1952. “Dear Oliver, just a line to let you know how things are progressing in a ‘saner land.’” Evidently feeling back in England that it’s more sane than the US. He’d been at Nottingham for four months, so far haven’t managed to start any work. Then he goes on, talks about things. “Still I must say, life is very pleasant and we are living fairly comfortably. Food isn’t too bad, the only thing in short supply are good eggs at the moment, butter and cheese. We have as much meat as we want. Steak, stew, and roast every week, and we aren’t short the other items.” He does talk about work a bit. He said “while I am in the scrounging mode, I will mention another item which I am looking for. I require a small amount of the number of plasma protein fractions, but unfortunately they are unobtainable to my knowledge in England. Most people seem to beg them from the almighty Cohn at Harvard or from secondary sources.” He’s going to write to Jack Williams and to Harold Deutsch, who was also at Wisconsin, “but this may take quite a time, if you could ask around to see if anyone has any plasma fractions I would be very grateful.” And he says what he’s wanting: human serum albumin and gamma 2 globulin, etc. etc. “Sorry to be such a b. nuisance,” b. standing for bloody, “I saw Arthur about three weeks ago, he’s now settled in Birmingham, and took back to the happy routine. But I must close now Oliver to earn some more money for the income tax office. I can hardly live at the moment, supposing a wife and baby to be. Please remember me to all at Madison. Hope you’re enjoying life and that you’re not working too hard. Best wishes for Christmas and the New Year.” A very nice memory.

Here we go back to experiments. Precipitation experiments. Pages of notes of that sort. Double the protein concentration and see what happens. December 23, summary of results to date. Tube number 46-58. 5.7, December 13, January 2, so on it goes. Pages of experiments. Cell giving trouble on January 8. Now January 12, which is 6 days, I began to see what was happening in these precipitation experiments, and really what was happening was that they were crystallizing, the protein was crystallizing. So I have here on the left hand side, apparently some amorphous, and that means amorphous precipitate. And then too clear, but it looked like a couple of little triangular crystals, I thought at first they might be glass, but the diamonds and amorphous material in tube 3. Tube 4: remains of the diamond crystals and no needle crystals. Diamonds – many imperfected by some what I call hedgehogs, the little criss-cross crystals of needles, and I’m beginning to realize that there are different forms of crystals present in these solutions. The protein is pure enough. That when it’s precipitated with ammonium sulphate, it forms an amorphous precipitate at first, but as it’s kept, they turn to crystals. So tube 7, some diamonds, messy, with all sizes of diamonds and funny needle like crystals, but I’m clearly beginning to understand something that’s going on now because the next few pages are pages of little diagrams of the forms of the crystals that are there. June 14, I set up a pH 5.1 experiment with a large number of tubes. 9 days later, and then on the 15th, tube 1 slight amorphous. One battered crystal, diamond shaped crystal, to many fine diamonds and some clusters. Very little amount of amorphous material. Three diamonds and some very small single bead like structures. So I’m beginning to realize that what does that mean as I say, a change in crystal form, which became part of the paper that I later published on them.

January 17th, a Saturday, a note on the left says “very satisfactory pH 5.1”. A new lot of 5.1 experiments. January 21st, found in November, ammonium sulphate solutions and glass bottles now have been contaminated with the mold. Decided to throw out all but those made since January 14. So I made new solutions up. On January 30th, I made a note the instrument is unstable. Measuring up to the density. The blank compared with this morning was about 99.8 and I was worried about it. Changed the distilled water in the blank cell and the value came up to 101.3. And I made a note the distilled water contains something absorbing, in the future use only double distilled water. February 14th, 13 days, crystals appearing in various forms. The next page is very detailed. Recording of numbers. But it says not good enough, repeat both the new tubes and leave for at least 7 days. Then I use new lactoglobulin and new ammonium sulphate, made up new lactoglobulin on Monday, February 9. That was a 20 day experiment. And making more lactoglobulin, a note on Saturday February 14th: the precipitate looks now to contain crystals. Perfect crystals in the top part, on the low power. Little rectangular crystals somewhere. Little bit of parallelogram crystals. The beta lactoglobulin was crystalizing ok. Next time used degassed water it says. Saturday, February 21st, 30 days and looking at crystals again in the tubes. I was repeating these experiments many times. 6 days, February 26th, 6 days. A pH of 5.7. A page here, a summary of the two times pH 5.1, and crossed out through all the observations, “useless, too high a concentration”. I don’t know what that means. Probably everything precipitated. Crossing out pages. A comment Thursday, March 5: “Stopped, the instrument is hopeless. Stored solutions with parafilm seals, remade the battery connections.” So something was wrong. Trying to get material that might be an impurity to look in the ultracentrifuge. PH 4.8 and so on and so forth. Preparing more lactoglobulin. I didn’t at this time know what was later found to be the case, that beta lactoglobulin had two genetic forms. And I was confused about this and was having one form that I called P and one form that I called S. It’s rather difficult to reconstruct at this point. Details of precipitation. Some notes on papers. So there were again some notes on papers starting from the back of the book. Starts with Stanley and Anderson, 1941, and Zinn, 1948, and Doty. Coming backwards now from the back. Chemical abstracts looking for light scattering. Going backwards again. Notes on the thesis by E.P. Geiduschek. Doty and Edsall. Some talks on multicomponent systems. So going backwards again. Falconer and Taylor, Biochemical Journal, 1946. There is a page there which rings a bell. It’s talking about the shape of precipitation curves. And the one in the top right hand corner, a constant solvent, is special case a variable solvent, I don’t know exactly what that means. The diagram shows the precipitation curve if there are two components that are differentially precipitated as salt concentration increases. And it’s the form of a precipitation curve, which I used in my publication of this work. Whether I had used it before then I’m not sure. And that meets then the forward drawing preparation of lactoglobulin, so that’s the end of Book 3.

So here we are in Book 4, the last book from my postdoctoral period, 1953. The usual thing at the front. Where to return it to if found. So it’s starting with a Saturday experiment. The weekends never seemed to make any difference throughout my scientific life. PH 4.8, number 2, etc. We’re getting the usual sort of results with pH 4.8. No great, if any, advantage over pH 5.1 is the comment made there on trying an experiment at pH 4.8. Always making lactoglobulin: here is March 30, on a preparation of beta 2 lactoglobulin. It was commenced on the 24th of March. I’ve crystallized lactoglobulin. No, crystallized lactoglobulin was taken from the first dialysis sac etc. etc. Comments on it. Just making lactoglobulin again. This business they call the P fraction. Not really clear anymore what that was, except that there was a P fraction and a D fraction and an S fraction. It was realized that it wasn’t a single component. Wednesday April 1, there’s a comment set up P and S anyway. There’s at least 2 components, P and S. 2 lactoglobulin Thursday, April 2, more lactoglobulin. Here I have a comment on Monday, April 6, “very good, 3 components. Component 1, .8%; component 2, about 20%; and component 3, about 80%. P number 2 and number 3 are therefore important.” That’s the 20% and the 80% components. So there was obviously two components there. I think the comment here is the comment of gelation. And that corresponds to amorphous precipitate, rather than waiting long enough for them to crystallize. So you could get an amorphous precipitate, or the precipitation with ammonium sulphate. Amorphous precipitate would occur at a quite high concentration of ammonium sulphate but if you waited long enough then crystallization would occur, and solubility would decrease with the crystals present, so the precipitation curve would move to the left, and that was part of what I eventually published. Rather confusing at the time until I understood it.

I made an automatic pipette with a mercury seal made to deliver .2 mls per stroke. Eventually I made a rather nice automatic pipette for pipetting 1 ml solution. It was quite accurate. It was the beginning of what would be later on the Eppendorf micropipetters, but many years before. I never published it so I never made any money out of it. But I’m beginning to make an automatic pipette then, in April of 1953. Here then begin my attempts to make an automatic pipette again. There’s a .2 ml automatic pipette I’m trying to make in different ways. To pipette .2 mls. And one to pipette 1 ml. Actually I still have the 1 ml automatic pipette that I made. It worked but the precision of tubing was inadequate in those days, and there weren’t any O-rings to seal, and the use of Teflon was rather problematic as one would say these days. But it actually was quite accurate in the end. But I didn’t use it for the experiments which were published, so I never published my automatic pipette. But here is more tests of the automatic pipette to see how reproducible it was. Saturday, May 9th, for example, .2 automatic pipette plus one ml — I can’t read that word. So I have the comment that I remember very well, that thus the pipettes are good to better than a quarter of a percent. I’m weighing these solutions afterwards. The mean was 25.98 plus or minus 0.1. That’s a quarter of a percent. So did make an automatic pipette that was repeatable to a quarter of a percent. Remains to calibrate and set one milliliter pipette and repeat the check once again. I rather enjoyed the thought of an automatic pipette in those days. And I have Wednesday, May the 13th set up 8 tubes with a calibrated auto-pipette, 0.2 ml used and 1 ml used. The first 8 were set up rapidly without excessive care and using a fairly wide tip on the 1 milliliter pipette. And I have notes on the new construction, where it was a Teflon compressible washer which was oiled. The other way around, the .2 ml one, where there was a plunger going what would be nowadays the same as an Eppendorf, a highly polished central rod with a Teflon washer around it for the .2 ml which would be like an O-ring, and the other one was the opposite way around where the tube itself was highly polished for the .1 ml one and it was like a Teflon compressible washer ceiling. I have a comment here that “dirt on the cells follows the blank, but not very good anyway, check again, re-setup with narrow tips.” So the tips were disposable, the tips were Pasteur pipettes, glass Pasteur pipettes. So in a sense it had disposable tips as well. It really was the equivalent of a modern wand but way before the present ones. I didn’t need it again so I never did publish it. Pages of the same sort of thing. Had to make special adapters to spin these solutions at a high rpm, so the adapters had to be very carefully made otherwise the tubes would break under the centrifugal force. So I have a note here, June 1, “3 breakages!” A tube broken. The new large tube. Getting ready June 2 for an ultracentrifuge run. And then June 2 the ultracentrifuge run itself. .4 grams of pure wet crystals made up to 11 mls with buffer, etc. etc., dialyzed, and test the d and mu when runs are finished. I don’t even remember what d and mu are, anymore. Optical density and concentration, perhaps. So the extinction coefficient of a .1% solution, 278 millimicrons for this betaglobulin solution is 0.92 plus or minus .01. And then with a new cell, this is the ultracentrifuge run. June 4, ultracentrifuge run, washed the cell out with solution and filled from the bottom. Watched for water boundary. Exposure time slipped, more details of the ultracentrifuge and then back to electrophoresis again, boundary electrophoresis. Interestingly enough I’m just using a pencil for all those notes. My June 5 experiment says “as near perfectly clean run as ever made. Only 21 showed any trouble and 76.8x is the maximum reading obtained, probably when the precipitate was out of the field.” I’m not sure what that means but I was pleased with the run. But not so pleased June 8 where there’s a big cross. And I’m thinking about publication, because I’m saying curves to be used for paper one on the phase change — that’s the formation of different crystals. PH 4.1, 5.1, 5.4, 5.7. It’s a comment on where these were March 8, January 16, December 3 and 4, and March, where these different experiments were done. PH 4.8, March 6. Two days with a “good” under it. Continuing then — not very good, but essentially as before. PH 5.1, very satisfactory, but recalculate excluding tube 40. PH 5.1, measurements, pH 5.1, and so on. And then here’s one summarizing 1 day, 3 days, 13 days, and 30 days what the precipitation curves were looking like. So then I’m also talking about paper 2, materials on solubility for paper 2. P versus S, the two fractions. But I do see at the bottom a comment that “no doubt that the auto pipettes have made the method good. See CPII.” And that comes to the end of this book where there are still about ten pages of blank. So that’s the end of my period of recorded results from the University of Wisconsin at Madison.  I’m not quite sure when the last day was, the last recorded date in this book is March 18, 1952, but I’m sure that there some days thereafter where the date is not recorded. But that’s the end of this postdoctoral period.