April 15, 1949.

Dear Max,

Jaum enclosing 10 g. d-xylose which may be able to tide you over until
your own order is filed. We just got hold of this ourselves; our old supply
had just become depleted.

I've looked at the cultures you sent this last time, and find all of them
mixtures of Lac-/# and Xyl~/# when expected. 2-160 is almost entirely Xyl-
but there were a very few "papillae" of Xyl# indicating thet it had been
mixed. On Lae these aould not have been told from reverse—-amtation, but
Xyl- is exceedingly stuble, 2nd I have never seen «= reversion. However, I
saw very few mosiic colonies, but for most of the cultures, this has no
pressing, immediate imoobtance.

The family 2: 199-202 was looked at most carefully. 199 is of course
Lac-Xyl-. From its sisters and cousins I tried to reisolate the mosaic or
heterozygote, but so far have failed. I was especiaily interested to verify
whether 2:200 was realiy still heterozygous, but so far have boen able to
recover only # and ~. I haven't lonked yet to see whether 211 the recombination
types are represented in the culture, i.e., Lac#Xyl-; Lac#Xyl#; Lac-Xyl# and
Lac~Xyl~. I think that u distinction should be made between cells from which
segregating heterosygotes can be recovered, and those whgre heterozygosis is
inferred from a mixture of # and -. You state that 2-200 was mosaic, and I
was able to verify a similar situation (demonstrable heterozygotes as sibs
of segregants) in the last batch you sent, so I don't doubt it. I haven't
looked yet at 2:200. The reuson that these two kinds of heterozygotes should

possibly be differentiated is in hopes of picking up the first reduction-division
ef meiosis. But from the looks of it already, the nuclear situation is rather
complicated, and may hopelessly obscure the situation.

I hadn't thought that H-72 was any less unstable than H-1 or H-62. Nor
is H-168.

You my be interested in some segregation data on H-168. Mosaic colonies
were individuslly streaked out on the indicated medium, 1 — and 1 # picked from
each and tested on the several sugars.

Of 10U pulred selections on xylose HMB:

Xylose# Xylose- fhe rare * types were
checked to see if there was
Luc Gal -7 “9g NS oe any reciprocity in the streak
~ Z 7 1+ 5, from which they came. None
_ # g 0 x 0 was found. Only occasionally
_ # " a 0 was another representative of
# _ # i 9 the * found in the streak, so
# _ . 5 1 * each mosddc repreeents a large
# # # 0 Ll % mumbsr cif indepenedent segre-
# # s 0 12 ganse, although not an infinite
population suitabke for statas-
— — tical analysis.
100 100

Of 100 or ériginal paired selections on lactcse (in a few streaks, one or the other
glass could not be cleanly isolated)

Lactose # Lac~

Gal Mtl = Xyl

# - - 93 9 Agsin there were no reciprocal
7 - ~ 0 9B correluthoass.
# # # 1 0
- # # ® 1
~ # - 0 1
others 0 en ee
92 97

These data give the best estimate of the true mean segregation frequency of Lac, Xyl,
etc. The fact that both Lac- and # give a low proportion of Xyl #
(ca 1%) shows that the segregation of Lac in the Xyl- selections gives an unbiassed
estimate, uninfluenced by fluctuations from mosédc to mogaic, of 56% Lac-. This is
quite different from what is found in H~-72, I don't know why. Lac and Gal are ahmost
completely linked, as are Xyi and Mtl. In prototroph segregants from transient zygotes,
there is much more crossing over between Lae and Gal. You will also notice that
in the first set above, there is much more between Xyl and Lac among the Xeké« Xyl-
than the Xyl#. This presuaably has something to do with the aberrant region,
Finally, recippocujs of the rare types could not be find. This fits your observation
that segreguksts are nob paired with segregants; the opposite type presumably gives
an inviable nucleus.

Hope to have a long talk with you in Cinncinnati. I expect to give a 10 min.
paper on this stuff. Let me know your plans; maybe we can work out some way to bring
our stuff together.