Dear Luca:

Your letter of the 10th (and remains of the ms,) arrived just today. I wish it
could have come a week ago, so that I could have given it long and thoughtful considera-
tion. I am gfd¥d7 greatly pressed to finish a mass of paper work by the end of the
week so that we can leave for our vacation, I ope you will forgive the paucity of
this reply. I tried to put your letter asideg but was so distracted by thinking about
it that I could not proceed mith my other work. As you will see, I wrote you just a
few days ago to send some pictures.

Thahk you for the cultures. iim583 arrived broken, but may be still recoverable,

If you do not hear further, please assume it was CK. We will have time only to grow

it out before leaving. I am pleased that the m8, also is in a good status. There is

no ohjection at all {Y¢ to the ms. for the Congress; TEX MH AMAT RRR RE ERY it doesn't
matter how you handle it. The furthest thing from my mind is the possibility of
giving you any other paner.

Your conceptions of the self-elimination of the Hfr, and its possible relationship
to centromeres are in (the usual J!) astonishing agreement with our own. ‘le have not
been able to fit all the data together so nicely, however. Esther has been working
more closely on the genetics of the Gal mutations: there is a Cluster of at least four
very closely linked loci, very close to the point of elimination. For, although they
are very closely linked to each other, some are regularly hemizygous in the diploids,
some are heterozygous. As you know, Lp agrees with the former. However, we do have
some diploids heterozygous for the region, and we had already planned to try to
cross Hfr into these, but there are technical difft¢ulties that will take some time
to iron out.

There is this difficulty that has hindered me from framing any unitary theory
of elimination: Gal and Mal mark two distinct eliminated regions, usually corresponding
to the F+ side, There are, however exceptions which appear to be independent of each
other. But perhaps the Mal exceptions (i.e. cases where the F+ contribution persists)
represent crossing over with the defective centromere, while the corresponding Gal
exceptions are another progess, perhaps failure of elimination. But if this is the
Case, one should find among the diploids both hemi- and heterozygotes for a given
Gal locus, and this has not yet been found.

It is not crucial which Gal locus is represented in W-945, as all of the Gal loci
seem to be very closely linked. (That is all of the full negatives, not such as repre-
sented in 1-677, which is messy). 4-945 may be Gal,, or possiblg a new locus.

W-583 carries what we call Gal,~. In usual diploids, Gal, is not eliminated; Gal, is

eliminated, and from some of my older data so is Galo. W-1294 is not supposed to
be galactose-negative: it may carry some slow or suppressor mutations as well as Gal,-.
Curiously, while Gal, is digenic in diploids from Het crosses, it is (in few cases
available) only haplogenic in other non-disjuntional types.

The édd/ easiest explanation of Mal and S is that they are terminal. This would
require only a single break to account for their loss. I think the extent of elimination
in this region must be quite regular; otherwise one might expect occasional discrepancies
between these two loci, such as are never found.

The basic datum in Fried's work is that in crosses where M and Lac were unselected,
and crosses were based on selection either of $ x TL or P (linked to Lac) x TL, there
Was no significant correlation of M and Lac, although there were the now usual deviations
from independent segregation of either: this does not prove that M and Lac are unlinked;
it fails to prove that they are, and invalidates the previous evidence (Lac ratio in

prototrophs) of the lihkage. Certainly one could set up special schemes that would give
the same result, where the linkage might be obscured by other forces.