Maroh 6, 1949, Dear Max, I have the cultures you sent, and your very impressive pedi- gree of 2/26 A. Your numbering scheme is very ingenious, and will make it wuch more painless to discuss what oomes up. To be sure that I've gotten it streight, I'm sending back the exeapnded line pedigree, together with eertain inferences. First the pedigree. The critical cell divisions seem, of course, to te tgose of 3 and 5. 5 seems to have segrecated at its division to Maxm give what may be inferred to be a haploid segregant (11) which thereafier breeas true (47-46; 93-102}, and « diploid hetero- zygote which divided only equationally (12 -- 25,26 -- 51; 219-222) for the next 4 generations. I've lookea at the 6 segrezants a little further_and was a little surprised to find that they were all Xyl+ V\". I probably will have more to say about ASL, but Te seems to be unusual in being Xyl-varlagated (-v), but with + predominant to -. My past experience with H-72 nis bé=n thot - predominated over + about 13:1 among the segregants, end I have also never seen a ohange of type before in the diploids. This mey be on ex- ceptional example of crossing-over without reduction, but I'll have to look at it some more. If th™s Farvene) betore 5, it may explain why all the 6 segregants are the crossover type Lac-Xyl+, Otherwise,this type of behavior might call for two-strand crossing pver ( or else that we had a rarey four-strand double). I would have expected, on a 4-strand basis,that a oell like il, if crossing-over ocourred, might give one crossover, 23, Lac-Xyl+, and one non- Grossover, 24, Lao-Xyl-. But it 18 unsefe to try to gencralize on just one observation, and 5 may have changed type from I!-72. How a cell, 5, can segregate to one haploic and one diploid product may not be easy to explain, but perhaps we can appeal to the probable aultinucleate condition of coli cells. (Cytological work on H-7e2 has been started. If the "Robinow bodies" are nuclei, this presumption is justified). Then, we can imagine that in a binu- eleate cell mignk, one nucleus might segregate, the other remain diploid. Of the segregants, one pair carries a kkksa lethal, which shows up in the 6 descent; the other does not segregate from the diploid until 5 divides. From this point of view, the ternary fissions that you mentioned would be especially interesting. But this kind of explanation can be made to suit almost any segregation pattern. I hope that some simpler patterns will turn up. M.R.Zelle- 3/6/49- 2 I think your procedure of recovering microcolonies at a very early stage is very well advised, and removes one of the more in- portant anxieties I had about this program. As to classification: Certainly, any cell which produces mosaacs must be a heterozygote. Likewise, eny oell which is homo- geneous over 100 - 1000 colonies is very unlikely to be hetero- zygous, although conceivably it could have been 2 heterozygote which segregated uniformly by the time you recovered it. The cultures which have no mosaics, but mixtures of + and -, pose another problem, especially if there ere nearly equal im numbers. Such cultures could represent hetexozygotes which have segregated completely. They also might represent the frst reduc- tion civision of a heterozygote, as could be yerified by finding only twe combinations of factors (including V*, Xyl and nutrition) in the whole nopulation. -pe “léthal” celle are most interesting, as I have had te postulate them to account for the deviation of the Lac segregation ratio from Bf/ 1+ %31-, to kKxxrxkx 1 : 7.5 for H-72. Enclosed is a draft of a manusrript that has been sent to PNAS going over this point. When maintsined on EMS, most Lac+ prototrophs are hetero- zygous. Only a few %, at most, will be prototrophic segregants. I stijl haven't checked directly on the nutrition of H-72 Segre jants, but would infer thet at least M, T, and L are hetero- zygous. I would also put biotin and #M thiamin into the testing media. Thess are Giftficult to score for, but chances should not be taken against the cultures being B- or B,- . In testing the uutrition, simply aad a drop of a dilute suBpension of freshly grown cells (most conveniently scraped from mutrient agar) to 10 sc of miniaal liquid medium supplemehtdd as follows: a. Bh, MTL (+) Lack of growth after 24-36 h. in b. BB, MT (-12) -X denotes a reyiirement for X, co. " MD (=f) provided there is adequate growth d. " TL (=M) in the + culture. You will, of course, have to use well-cleaned glassware, and dilute inocula. Controls with known parents are desirable. Reverse-mutaution certainly does occur, ana is pushed by seleotive pressure on lactose-EMB. heZederberg is studying this system. (Abstract reprint enclosed). It should cause no trouble except in cultures repeatedly traksferred on EMB-Lac. Reversion has nothing to do with the segregation phenomenon. I can see no pressing advantage to sending the original oultures, provided care is taken to include a complete sample (i.e., no fresh single-colony isolation). For now, I would appreciate getting the mosaic single“cell-isolates as well as the segregants, but this should npt be necessary later, “MeReZs 3/6/49 3 You have the best estimate of the frequency of segregation in your pedigree. Out of 8 successful fissions of a heterozygote, there was one segregation. This seems quite reasonable from the sectored appearance of the colonies on EMB. Assuming that diploids and hap- loids grow at the same rate, this means, on a Yrotigh average, that the proportion of heterozygotes will diminish by 1/8 each fission oyole; tes, Rakxthret?3 cd enxmamnxfxemckxoxkaxtcirenxakewtrrtx epauaknkiomex that a heterozygote is lost every third generation. This means that the diploids will increase 7-fold each tnree mgt's, the haploids will increase 8-fold, plus one for each diploid. If h is the number of heterozygotes initially, and s the segregants, ah = ij ads = 68s +h, where the time unit is three at” Ct generations. I ran across these differential equations once before in connection with some yeast work, and one rets: dh= 7 bh where n=h +s = total populetion,. on 3 on and, for the boundary condition of a populétion starting from a single h cell, h/n=1/n . That is, the prpportion of h will be 10% when n = 108, which seems to be substantially correct. Having just gone through the algetra, it is upparciit that the frequency of sesregetion ts the reciprecal of the log population size for which the proportion of heterozygotes is xkmimakx dininished by the factor 1/1C. This enalysis excludes the lethals, but more informétion about then will be needed before = mathematical model can be set un. H-7e2 is & prototroph from the cross mentioned in my letter Nov. 15. Succinate is put in BMS-lhec so thet Tec~ emguergsorkyxeam be detectec as white ecclcentes. You / prototrophs probabiy don't neec to use it for your rurperes. I'm sending E. coli K-12, "58" which xeytires tiotin. Bernard D. Davis, in New York, also has some such. Sincercly, Joshuc Ledexherg Feo. baybe 1tts about time we got together. Any suggestions? Enc: kis reprint.