June 6, 1956

Dr, Norman Horowitz
Galtech
Pasadena 4, Calif,

Dear Norman:

Thank you for your letter of the 4th. It is not hard to see why the
penicillin-sensitive strains are gone, but I hoped you might have fept
the collection of temperature-mitants that you and Leupold har amassed,
in which some of the p.-s. strains are included. I am not sure I made
this clear; the strains are cited in Rowley's paper. I will write to
Rowley if you think he is beck in Lomion.

As to your counter-comments, I wish we could do this in extenso, in
person, because we could spend Mays exchang&ng letters. I am sure you will
compensate for the circumstanges under which my comment was prepared; the
written version follows the transcript rather closely, and did not have the
benefit of the transcripts of the principal's remarks. For that reason,
the "critijue" that I felt was expected of me had to be directed at what I
undereteod as a general background of discussion one the 1:1 theory, plis
whatever specific remarks of your own I could assimilate on the spot. I am
aorry if I have misattributed views to you that you do not share (ani am
pleased to see the very large agea that wea agree on.) On rereaiing my account,
it does no’ seem that any specific statements were attribute! to you, and ay
criticisms were certainly ingended to be directed a& what I consider a rather
generalised erroneous formulation, and not to any personalities. The idea of
a gene making an enzyme is stated fairly explicitly in a number of older papers
(Beadle-Chem Rev 45; Tatum & Beadle Jann Mo Bot Gard 467; PNAS 1941;) and while

Emerson

these may no longer be representative cf your own views, you have to take account
of the usual cultural lag. The main point I want to stress is that the real
answers on mechanisms of gens-enzpme relationships are not likely tc come from
genetic experiments.

I am afraid that I am at fault for evoking part of your criticism by overlooking
a typographical error until just now. At p. 167, line 6 up, my ms. had read
fexperimentally indefeasible", but I missed the error in proof, so it's zy own
fault. I don't know whether the correct version is more congenial to you; it
should be more intelligible. I was not terribly clear about the different levels
on which the theory can be used. As a purely empirical matter, one can ask whether
there are any apparent enzyme-pleiotropisms; there are quite 2 few in ™. coli, and
you have contributed some yourself, but the axperimants you and Yanofsky cited
sean to be the first concertec afforts to Bind specific examples in Neurospora.
But at a deeper level, the theory 1s indefeasible because you could always
explain away any exceptions by considering them to be secondary effects: inhibitors:
quantitative levels, ete. Here we are in agreement, that the only plausible way
to do genetic experiments ia to assume a single prdmary effect: in fact why don't
just go ahead and postulate an ultimate unit of function which we can

call a "phystological gene" regardless of its behavior in recombinatdional
and mutational analysis. It would be impossible to disprove such a postulate.
I do not consider that the evikience favors identifdcation of the "physiological"
gene with the units of crossing-over or of mutation, and it is misleading

to promulgate the theory in such a form as to encourag& the expectation

that mutations with a given physlologicsl effect mst be allelic, or that
mutations with manifold effects must be separable into physiological units
by ercssing-over. In view of your own couments, I am obviously beating a
dead horse, execpt that there are some biochemist who still doen't know

that yet. Af not by developmental analysis.

Sincerely,

Joshua Lederberg