requires that. the descendants of any cell be-immobilized-in the agar, in

. They might be detected by scooping up the entire B)- colony. and plating’
it into medium lacking By and containing the other factors. I have done.
this experimmt a good many times, and have censistently recovered types

' chance of avoiding contamination of prototreph colonies with ‘the parentals*

however, but unlikelgy that the reliability of these data is influenced:

_ YALE UNIVERSITY -
|. QSBORN BOTANICAL LABORATORY __

    
   
     
 
  
 
 
 
 
 
  
  
 
 
 
  
   
 
   
  
 
 
   
     
  
 
  
   
  
 
    
 
  
 

-& NEWHAVEN, CONNECTICUT
. . . ‘ j .

ote Cee
ee ok fe hg ta
| fo Dog on les ta)
! 4 March 22, 1947. o
7 fae. - an CR ate
Dear Mather-- ..- . Fe

fhe interest in our research which your letters indicate is most
gratifying; under the circumstances, one could hardly. consider you
tardy correspondent; I wish again to express again my appreciation ©.»
for the stimukation which these epistolary conversations have provide

 

\Little progress has been made since my last. letter, since neither’
ofthe two projects: crossover-suppressor search, ner 4-strand evidence
has born any fruit. In re the latter, and’ in answer to your question, ;
it would ef ceurse be necessary to recover more than a single product...
from the reduction’ of a. zygote. Theoretically this might. have been pessib
as fellows: B,«'is a Treger’ segregant in the cross B-M-..X T-L-By<5. 0°
to the point there mest the of colonies which appear on B,: supplemented
plates will be By-. Since: the manner in which the crosses are peffprmed |

such B,- colonies one might expect there te be a small-number of cells. --
complementary te-the B,~ segregants which are the ones selected for in thi,
medium. E.g.; one might hope to. find B-M-T-L-,) er other combinations.» 4

other than the B)- constituting the overwhelming majer fraction of the
colony. Unfertunately, these types have censisted exclusively of B-M- .
cells with the same LacV configugstion as the parent, or of B,+ types...
which undoubtedly arise by the (rare) reversion of the By gene. The™
quency of recombination (ca, 10-©) is still so low as to exclude the.’

Since these would be selected for equally as yell or petter than the
hoped fer complementary. segregants, the failure to find. the latter cal
be regarded as any indication of their true status. It may. sometime be
possible to find material in which the major segregant, and both parental
can be selected against, as would be demanded by this. experiment...) get

. As to cressover-buppression,I have been looking for.types.in whig
change between-B;,.and B is inhibited. About 250 isolates of Nitrogen

tard treated material have been fested, but none of these have _shown
failure of interchange.As such, suppressors would not.provide directs.
dence of linear. arrangement; however I‘hoped to ‘have a sufficient icé
lection of them involving ‘xax divers portions of the region BM--TL as
well to be able to map them linearly on this region. But nomluck at “ally.
On the question of the number of linkage groups, I have gone ovexiy tn

possibilities of ‘spurious! linkage, and think the latter is excluded

The crucial data are perhaps the following. (all loci + uhless indicate
(B-M- X T-I-B,-) Be:B+t 11:79 Tao:T+ ° 9237 8 LetLt . 5251. These ‘with
the By sesregation are enough to exclude 'spurious' linkage. It is’ possi

  

  
 

 

 

by the chance of selection against B- types.even on B containing media, et:
4g it stands, however,;:all the data are explained on the basis ofia sing}

group, apropos which.I, might mention that’ the sLacV segregation in 1!
types supports the order B, --B---M--- rather than By cro B my
he P DB ott “4 * :
Fo a,

  
  
 
   
  
 
   
  
 
  
  
 

How far one should go in attempting a ganetic analgsis of this indirect
sort is, it seems to me rather doubtful. I do not have stocks with which:
properly to perform the type off test for linear order which you suggested
Unreasonabke assumptions of the absence of interference have had to be™

made because there is no direct test for it, and I wonder of the subject
has not already been developed to the extent that it should be in the”
present state-of our 'back ground! knowledge. For txample, the question :
of 4-strand crossing over bears directly on the evaluation of linkage.
data. For example,in your analysis ef the Lag, V segregation, you’
assigned the four classes the relative frequencies: :

P9295, serene and p1P2P3+ OE oe

However, 3. and 4- strand double exchanges also ‘yield single exchange’

chromatids which should also be considered among the contributors to. the

| ‘FIFSE three classes. Finally, only a small fraction of triple exchanges -

will give rise to triple-echange chromatids. for the fourth class. 1 have

not attempted te. arithmetize this analysis; perhaps you may have. some

suggestions. ge
In regard: to-the. original analysis, it seems: to me that nas first

three types should be in the ratio p rather’ than PQ ?Poa Teese

on the following basis. On that anatysts.> only single exditene es ware

considered for the first three types. In the absence of interference, . “%&
the probabilities of intébdéfangeé art the class ef. single exthange types 3

should be ‘the same as for all types summed, ise. proportional to the

map distances. One’ can then sskinaks use ng, to deckexminaxkkax estimate.
the absolute distances .~ by. summingx f(p 7 *P9 8 ) cubed, is the probability;
of a triple crossover, of which a determinat fraction are of the type ©

detected: inn What this will all look Like’ on a _four- strand basis
“is: hard to. foresee... : ur a for, A Te eh
AG to oY 3 : Pe

“In the cross B- P- x p- L- “3 we. may have. the. following situations?

 

 

 

Al ; mohs 9,40; FIN the |
I see here the point where my letter was confused, M did not enter

into the cross, and on the basis of the map one should still

expect that B- would be more frequent than B+. The map may be in error,
or my statement that B- is not more frequent than B+ may be incorrect;
I suspect the latter for this reason: the relative frequency of B-

was in this case estimated from the relative number of colonies which
appeared on B supplemented plates. it turns gout that this is not a
reliable procedure : for example, in the experiments in which

B-:B+ was determined for the cross EMM B-li- X T-L-B,-, there were fewer
colonies on B plates than on minimal , although thefe should have been
an additional 12%. This probably stems from the fact that the background
growth of the parental types is greater when 5 is supplied, and this
may crowd out prototroph colonies. At all odds, this point will have
to be looked into, and thanks for bringing it up. +

Thank you also for transmitting the suggestion that we prepare some
ef our material for Heredity. I rather think that we will publish most
of our experimental material in this country for reasons of convenience.
However, 1 have been preparing an analysis of the literature on the
‘genetic! peculaarities of bacteria, of which the re are many which are
perhaps not wideiy known (and authenticated, for that matter). I have
not specifically intended to publish this material, but if the notion
Seems suitable, I shoulc be happy to submit it.

Wy wife, Esther, extends her regards.

Yours singerely,

shua Lederberg.