January 26, 1955

Dear Bruce:

0
I

  HelLlo!  I hope you are going to be patient with me about this manu-
script. It is no exaggeration that 1 have dropped everything else
abruptly ix&m in order to ooneentrate &n it, anl try to resolve my
ambiguous thoughts about the problem.

      Let m9 say to s&art that my own experiments, with -x ST666 almost
    exalusively, papallel your results quite closely. I have not been so
    fortunate as you, however) in being able to trsce the descendancy of
@    'W oells over many generations, aa you have. This must be due to the      /
      rather lower inoidenoe of ftEtB cells in this system, and to what seems
    a rather lower number of motile progeny, on &he average. mat data I
    do have (I've already sent you some) do indeed support the unequal
    distribution of such progeny among sibs, but some of the numbers are
    on the s&aky side of 10. iAt least for now, do let me speak of a

    ; 51    polycatenate = E, oligocatenate for the cells or clones wlhth fewer
    \/    motile9 arxi (uni)-catenate for the strict %nilinear~i case]. Aiiost of
           my earlier experiments were devoted to studying the life expectancy of
          single pckafr chains, for example one was followed to the 59th (sic)
           generation after its fipt;f initiation, whioh was certainly not less than
                                                                           --
           45 generations after its separation as a single chain. That is, I was
utl '-l sp
     ti       primarily concerned in getting as rigorous proof as possible of the
      articulate and non-pmproduotive character of the tQ&i.l.ity-c_onferring-~articlelt;
    this may be taken as amply settled on both sides of the Atlantic. In my experience,
    however, I have almost never seen further increase in the number of chains after
    about the 13-lf+th generation, which ia simply to say I had never caught the poly-

15)    catenate cell among the numerous chains already produced.
      There are just tpn, questions I am not altogether happy about: is there actually
Q ,,. an j&crease in the numbers of the mcpla,
  3, and secondary chains,            in accbrd with your hierarchy of primary
               or is iL still possible that all the mp's are already formed
    in a polycatanate cell, and are then distributed albeit non-randomly at successive
    divisions; and, how certain is the correspondence between polycatenate and trail-

16)

forming cells? As to the first, I am not greatly bothered by the dispruportions;
owing to the dmsl.1 numbers involved there is no good evidence that the apportionment

0 from oliPocatenates is random either (how are your data on this?; my own offer
  i examples like 6:1:0:0 ; 7:l ; 7:l:l:O ; 7:4:0:0, which should be none too frequent
    on a random basis.) And, indeed, I am rather more sympathetic than you to Bissetts
    notion; some time ago I had done dome ex?eriments on TZ-labelled cells which suggested
    that they regularly grew from the pole opposite the TZ granule, which accounts for
       (and div&d)   the subpolar position being maintained. In any
171   @nt, it is too uncertain that the mclg's are instantly flagella for our obeervations
      to be deuisive.   Concerning the trails, I would not yet reject the role of ac-
    dcidental factors, as you state it at the bottom of 4g - 4h, and 1 think it will be,
    necessary to get more direct proof that unicatenates are not, and polycatenates are,
    able to form trails. In your experiment, is the Poisson distribution applicable?
    That is, was the number of motile cells per drop uniform or normally distributed,
   6 c. 'Jhat was the number?  1 have been transferring drops with just one cell each
    and have about a 50-608 successful recovery, but as told you before, virtually no
18)     trails.

19)


(Index Nos: for
conv. reference)

9)




log

12)



13)

Bruce, I would like you to bear with me on two possible courses, which I
should like to try simultaneously. One is a terser account, to be considered
preliminary, and designed say for the Proe Nat &zad Yci US, (where it can be pub-
lished within 10 - 15 weeks) where w4 can summarize our accordant data, and
emphasize especially the unicatenate end of the story (which is the most
intesesting to ny ti, perhaps because least speculative). You can then follow
this up at your own convenience, and feel no hindrance from my part. Alternatively,
I will try to continue some 1Iy)re experiments, including a look at your own
0-x SW541 naterdal, for example, to try and convince myself more definitely
of your choice among tha alternative interpretations, one way or the other.
I have in mind particularly to look more closely at trails and the chemotactic
and physical factors that might be inrolled. Since an accidental experience
with phenol some tims ago, as I may have written, I have no doubt of the
importance of tactic behavior, be it "apobatic" or %trophic".

One other question: one way of correlating trails, swarma, and x-catenates
would be by coneddering the frequency of swarms as a common denominator. I think
you will agree that every cell that engenders a motile clone should be deteoted
as a amarm. Have you detailed data on the ratio of trails: awarms on one hand,
and poly-: oligo48atenate : motile clones on the other? I have to collect tqy own
scattered pedigrees on this point, but offhand I would judge that they do agree
with your notion. I think I should like to take another look at -X SW67, which
forms such beautiful trails.

Query? Have you any mOre recent quantitative data on incidence of swarms and
trails per phags? In your
of better than 16    "report* you estimated, for TM2 -x SW541 an efficiency
         ! I don't recall any direct comparison; this must have been
what impelled mb to n&e SW-665 (a SW541 Xyl-), transduction to which proved to
be only the usual, about 10-6, for Xyl

   Do you account dirsc tly for the small incidence of polycatenates among
ti initial isolations? Do you think the others are al.1 oligocatenate a?
                                  -----------I *

Just to be the devil's advocate, let m propose an alternative version(s).
1) in re trails, that owing to the lack of a chemotactic impulse, most motile
cells are content to swim and grow on the surf ace. Very few start a trail, and
those that do only after all of the motile cells are unicatenate, Ans: if
trail-foropers are not distinctive polycatenate cells, it should be possible to
modify the incidence of trails either by manipulating the medium, or with chemdtac
tic pressure. Will do.

   2) in re polycatenates. *"let us not multiply particles without necessity". We
\   can readily presume that, in transduction, some fragments implant; otherr do
  not and are lost. Meanwhile, even in the latter (which might be expected to be
  most frequent) the gene has left its product. The product is not particulate (ne-
  cessarily) but is sooner or later assimilated into pp particles (flagella). The
  product is, however, not soluble and is disproportionately distributed. [*en
  simpler would be the accumulation of flagella themselves, but it may be awkward
  to think of acowating as many as lOO+ flagella. What do you think? what is the
  maximum number of ckains you have dobserved? 1.  If you like, the intermediate pro-
  duct (essentially equivalent to your Y3 par$iolet*) might be an enzyme which could
  function only when present above a certain threshold amount,

1) and 2) are not directly dependent on one another.     If I may state a
general outlook on the problem, it would be possible to postulate any number of
elements in the oath from gene to flagella or motility; the genetic literature is


   full of hypothetical hierarchies,and you can take your choice whether the inter-
=)  mediate elements do or do not Wselver replicate (cf. Sewall Wrightis reviews,
   Amer. Natural., 791289; Ann, Rev. Physiol. 7275; Physiol, Rev. 21:487, and
   Spiegelman `a fantasies in CSH 1946 at ~~2'71). The present case is unusually simple
   in some respects, but we have an unknown numberfi of parameters in the way thet
   the transduced fragment might function. I want to distinguish, if we cans between
a)   what is reasonable, and what is reasonably certain.

   If I can find the time, on top of the two "goursesW supra, I will try to set up
22)   another tape by way of verb& coaanentarg on the ms. Perhaps one reason for hesitancy
   in using ths tape is that I did not know whether you could conveniently listen in
   private, as some comarsnte are 1ikUy to be designed as more intimate conversation*

Pc@r t
,.j.+ 4`4 j.-
boshua Lederberg