Deceaber 28, 1952 Dr. Norton Zinder Roekefelier Institute 64th Street and York Ave. New York 2, N.Y, Dear Sorton: The is an answer to yours of the 2lst, whioh only just arrived. The Christyae disorganisation of the mails, and ay owa flood of letters to you (Dec. 14, Dec. 24, and postcard) is Likely to wreck orderly comunl- cation between us. I an answering innediately, hoping you will gst this before you have had to answer amino of the 24th, so that we can go back “ends orderly progression. I promise not to burden you with any aure until TY have hed your reply to this one, Tam sorry if you feel you ars in an awhward position ra Bruce. It alap~ dy Ulustretys the problen of a 3-way collaboration; I could imagine other unfortunate incidents that would be much worse if we felt that matual clearance wae necessery befora A sven telked to 8. The suggestion that Bruce assuas vols authorship was only . trinl bolloon, in the hops that it alyht sispligyy his problem of anthorship. Tine is getting on, aud this paper should heave been in presa wsll before now, Your interest in the sutter is in no danger, and not sven any tentetive action would be taken before consulting you, Gomeone had to speak to someone first. If you feel difforentiy about how this paper should be writtan, you ased havo oo esharrassacnt sbout it. Although we must all have s part in it, the decision (in ay opinion} is preealnently Bruce's. Had we been in closer commnicntion ourselvas, you would have heard about 4% sooner. Still I an personally regretful if you have falc any trouble about this, but an sure there has been no irreparable damage. I still] feel that a 3-ay authorship 11 too complex; $f you want to work cut a dual authorship with Sruce, it ts all right with as. T have ac objection to the use of ny “backcross” data on the ified transdustion if Bruce regards $hes os essential to the present purpose. — To turn to a more scientific question, the "backorceses” did not require additional phages. Fortunately, many of the transductions in this system are still sensitive to the trensducing phaga, so that it was poasible to use PLT22B again, after sensitive orcgeny were discovered is each cosbination, I don't quite follow your reservation about Fla- allelism? Do you mean that tro non-motile stocks each carrying the same Fla~ might each carry modifiers that reetore motility to any other ncnemctilc stock? This wculd be equivalent . to suggesting that asch aon-motile 1s besed on a unique constellation of fastors, the replacement of any ore of which restores motility, I suppose that this is still a formal possibility which could be verfied cnly if wa could select esally for Fle~ —x Flat, For e more complete analgsis, it would be better to work with & group of sutante 211 derived, presumebly by single mutaticnal steps, from the sem wild type strain. However, I do not think this affects the validity of alleliam teats. It might slter the loterpretation of "Fla,-"$ on the simplest veraion, the genetics kgrounds of the different 0 stocks are aore or less the same; to make it sore complicated, Fla,~ aight give flagella in the residual geno- type I, but not in X. Still for A—x B to give a new form (motile), they must carry at least one non-allelio factor, and this fs tacitly named Fla.~» although, as you say, there aight be more than one Fla, factor. Expressed this way, your notion parallels the now rejected hypothesis to explain the alr &, linkage by two alternative, aon-Linked factors. It is fairly likely, from what your letter reports, that your variant differs somwhat from my 22¥. It produces perfectly clear plagues (later some geahularity) on LT-2, I wasn't sure what you meant by recovery of infective centers in the third sentence of thhs paragraph. Do you man that bacteria recover to give colonies, Poissfon referring tp the caloulated faaction of mitiply infected bacteria, or that they give plaques{, (Poisson referring to calculated ratio of infesting partioles to infective centers)? I just oculdn't understand your last two saiéBneeas fn this paragraph at all, can you? fyohinur fam grows more rapidly ther what? Our JV conditions mst haws besn different: YT used undiluted broth lysaaess, and anst have had a sufficiently thick layer to have had considerable skix sHsopption by the broth. Subsequent runs hays given steeper “illing curves dn diluted brotY. The main point is not the absolute desage compariscns, but the survival of TA when phage (plaques) has been greatly reduced, so tht phage could be thus separated artificially from transduction, I have not done the eomplote curve; ospectallg as the distinction of 1~ and 2~hit le orebty delieate. Did you not carry your irradtaticns to higher ‘cess? With the hich dossa, it ts almost impossihle to avoid multiple infection in the essays (calenlated on the bugis of total particlss of phago)}. However, T@ddid got linear dilution rea- ponse, and could not find any difference as between celoulated emltinlichtles of about land 10. 3ti11, the surviving plaques might represent some of the multiply infecte’ canters, though there should have been 2 substantial increase mith 10 as compared to 1. The sain point tg that transductions and plaques could be counted on the dame plate, and the fo-mer were nowlysoganie. A brillfant (2) experiment that didn’t work: UV'd PLT22 does aot protect against 22¥. [I had thought thers sight be recosbination between inactivated PLT22 and active 22¥, protecting against the latter, and inducing lysogenicity. This might than have been a modal for lysovenization tn general. 1 Transduction frequency of 1:50,000 shonld make ft possible to test dual transductions for independence, theugh you way still have to use a selec- tive setup (diauxotrophe). is not entirely uninteresting. If you find the ratio of duals/singles to be 1:200,900 instead of 1:50,000 you coulki argue rather reasonably that the nucleus, not the cell is the unit of transduction, We have the same unpromiaing experience with lwoffing LT22 end L72(22). M666 or 543 Infected with AL¢-228 works much better, but this phage has a low eop and transductive sffictency back on typhimariua. I did understand your delay expertmant this time (hallelndal). the result seoms very neat, but perplexing? I don't think you can correlate the tracks with this: they repreasnt the abortive transductions, and meh of the persie- tence may well be puraly phenotypic. Me don't know anything about delay or segregation in , he initiation of swarns. "hr shonl4 thers He 2 difference in lag? Zean t understand your Xyl reoilt at all. Aren't your Xylets yrowing ? It looks as if-thase had not been transinduead initially at rll, bat why such a high final count? What do you maka of it. I don't envy you the som lexities of your virulence problem. There aust be a number of ways to do the statistical analysis of heterogeneity. One, probably ineffhcient, would be to cut your popplation arbitrarily in half, and compare the sum of the variances of your two halves with the whole. When you say you suspect bimodality, you are suggesting only that the sample is more highly dispersed than you would have expected, but I cannot see any standard variance that you could use to justify your expectation, What you can do, however, is to assume that you have two populations, centered at the two modes, and propor- tional to the square (7) of the modal values. You can then show that you can fit your data to the sum of these two populations with a auch lower variance than to the whole. I don't think there is any way of showing that your samples are taken from an anormally distribyted population, except perhaps by comparing moments of successive orders. I don t know the significance tests for this, but Fisher has worked them out. Have you entirely exhausted the possibilities of an in vitro system? imax How about mouse-serum (! sic) broth, considering serums from normal as well as challenged animals? Your job is to make a testtube a mouse. This letter is written on the plan of your own, but I think needs no P.S., after the deluge of this and mine of the 24th. erely, YA Lh eet Jodnus Lederberg P.S. I did slip. Columbia College Fund has, asked me to be a local "chairman", i.e., to make phone calls to local alumi to remind them of their respoasibi- lities to the Fund. They still have your name lasted locally. I'll have this corrected, but meanwhile will discharge ay task by this senftence. PPS. Do you think it would be appropriate for you to act as a sort of courier din distributing reprints to your colleagues at Rockefeller? I thought it might give you an excuse to communicate with them (or is this not such a problem)! Zt would be a favor to me, but if dropping the papers into sail- boxes or what not would inconvenience you, or seem at all undignified in the local contexts, please say noo This remark does not, of course, apply to your paper.