January 6, 1950.

Mr. Gordon Allen,
155 Corona Avenue,
Pelham 65, N.Y.

Dear Gordoh:

There have been very few published accounts of recombination by
other workers, and such as there are, may mention the experiments
rather casually. However, there can be cited:

Haas, et al PNAS 34: 229, 1948
Cavalli and Heslot, Nature (current— I haven't seen it yet)
Newcombe Rec. Gen, Soc. Amer. 1949.

In addition, as you must know, there is unpublished work at Columbia
(Tom Nelson), Chicago (Novick and Szflard), Indiana (iss Kahn) and
Cold Spring Harbor (Demerec), and New York (Allen).

The “Hfr" atrain is from K-12 (58-161). Its present status is doubt-
ful,as the effects gay be due to syntrophic efficiency rather than sexual
potency. I am inclined to accept, however, that it may be 100 - 1000 x
as active as standard stocks, under optimum conditions. He wrote last
month that in the cross S” x Ag™ (4.e., using inhibitor selection} that
an augaentation of ma 100x was found, and this should be more or less
definite. I think the question is still open whether the Hfr effect is
oppositional -- i.e., in the direction of heterothallisn.

As to outcrossing, Cavallis strain "123", which I have put down as
W-1258, very definitely crosses with K-12. However, except for a rather
complex nutrition (3 amino acids +? 1 vitamin, not yet run down), W-1258
does not greatly differ from K-12. I have an additional strain, +1113
isolated from chicks, amxotroph mutants from which appear to cross with
K-12 mtants. This strain differe in many respects from K-12, but the
yields on crossing are very low. In addition, Norton Zinder has one
experiment in which "Salmonella coli" (i.e., Vi antigen) has crossed with
K-12. If all these are correct, the score is about 4/ 16 + . I would not
completely dismiss the 30 far negative results, however, Zinder also has
promising, but still not yet conclusive,evidence of recombination in
several other Salmonella species. That is, occasional prototrophs, but
not yet satisfactory recombination of unselected mabkers.
Concerning 4-strand crossing-over:
i ba vd

On EMS Lac or Maly a small percentage of prototrophs are found which
are obviously sectored (mmmkix diametrically), containing a Lac+ and a Lac-
component. They occur too frequently, on dilute plates, I think to be entirely
explainable as coigcident colonies. The problem is to decide whether they¢
are the issue of the same zygote! Since so many cells are obviously multinucleate,
it would seem to be possible for a single fusion to result in a binucleate
diploid cell. Alternatively, the zygote nucleus might exceptionally undergo
one mitosis—- we know that about .1% of the prototrophs may be diploid even
in ncn-Het crosses.

In some runs, I think there has been a significant correlation between
the components ofa "duplex" prototroph, suggesting interference in crossing-
over, and supporting their origin from the same segregating nucleus. (And in
turn, of course, 4-strands). However, these studies involvedconly Mal and
Lac, and, as you know, I am deeply suspicious of anything that depends on
Mal. The problem should be gone into more ddeply, using a larger array of
characters. Aother point that kas to be controlled is the possiblity off
reverse mutations on the plates. However, I think this is unlikel. ,» but
as far as possible, stable allels should be used. For example, (Exp. 636;10/25/49):
58-161 x N~677 on EMS Mal. 6 duplex colonies/ "seweral hundred" (heavy

background)
The twins are reported in order, resp.
Mal Lac Xyl MtL T5
Ll +e ++ ee — ss The first 4 pairs are indubitably
2 e+. -- +- += ss correlated re Lac ahd T5. I also have
3 to t+ eee -- ss rather more extensive data of the
A t+- -~+ --= - 9 ss same kind, but without Xyl and Mtl.
5 + + - + + ss
6 +- - ~ += ss EMS Mal has been used here, because
the largsst number of duplex prototrophs
containing is found on this medium. With Mal,

as many as 20% of the + segregants may ba duplex. I've just found the
earlier data (45 twins) The} are distributed (Lac and Mal):

M-l- M-L+
M+L— 19 5 This would speak for rather intense
M+L+ 4 1? coincidence of double crossing over, but you can

see why reversion has to be eliminated, although I don't
think it is at work here.

If you would be interested to take up this problem, I would be very pleased >
to send you any additionallstocks that might be useful (stable Lac-; multiple marker,
etc.) I have been thinking of dabbling in it further myself, but I think it might
be just the thing to divert you, and provide a distinct fact-grubbing problem (vs.
shot in the dark) to push concurrently with the meningococci. It ought especially
to be done with the Lac+t— twins which occur at ca. 1% on EMS Lac Bi

There is some other evidence for 4—strands from the diploids. Not infrequently,
the diploids recevered show "double-reduction", e.g., are homozygous Lac- where the
parents were different. This indicates a) crossing-over from a 4~strand system in
the formation and recovery of the "diploids" — i.e., they are not merely the
persistent zygotes, or b) a complex compression of a serkes of fusion and reduction
cycles. In either case, there is very little precedent! I have to look at the
"spontaneous" diploids to see whether they may be doubly reduced.