Woods Hole September 3, 1964
Dear Dr. Pauling :
I realized that, if I could at all help it, I should not touch the shores of France before the Rutgers paper is written.
Once I arrive over there, I couldn’t concentrate on it sufficiently. Dr. Vogel wants the manuscript soon ,and you are going
to leave for Australia. So I have remained in Woods Hole, a very good place, off season, for writing a paper, and I shall
be here until October 10, at least. There is indeed that much work left ot be done in relation to this paper !
The paper is going to be long. But the editors may accept a long paper, and from people’s reactions to chemical paleogenetics
I gathered, at Rutgers, that much needed to be said.
Thank you so very much for the cheque, the most enjoyable mail item that has so far reached me here !
The Rutgers meeting was excellent. A number of papers had only a tenuous relation to the topic of the conference,
but nearly all were interesting. Some significant disagreements, of which you will find a reflection in the enclosed fragment
of my new draft, arose between participants.
To-day would be a late day for sending you the complete draft of the manuscript, yet I am even worse off in that
I can send you only a part of it. I have no hopes for getting the remaining parts to you before the 9th. Would it be possible
for me to send the last section or sections of the draft to Australia for your O.K.? There is no “détente” in my struggle
with dead-lines and lack of sufficient time. In my experience, an accurate definition of the dead line is a line over which
one nearly drops dead.
The enclosed fragment of the manuscript is interrupted in the middle of the write up of the material contained in
your letter of September 12. I find it aggravating to be a mathematical moron, but I can’t get around this fact. I don’t grasp
how you get for τ = 1, P. = 0.90 (p. 2 of your letter). You refrain from having the number of differences between hemoglobin
chains that are being compared (p. 3 of your letter). Should the figure 21.5 not have been halved and the value for τ = 1
have been derived from the halved figure ? I probably get it all wrong. I am sending you back a Xerox copy of your letter
for the case you want to consult it.
I wonder whether you will feel like giving further consideration to the question of back mutations. According to
a proposal you will find in the enclosed section of the draft, the basic rate of evolutionary effective amino acid substitutions
is essentially determined by a to-and-fro movement between a few types of residues that are nearly equally “good” from the
point of view of natural selection. If this is true, evolutionarily effective back mutations should be quite frequent. The
postulated shuttle effect will however not occur at an equal rate for all types of amino acid residues. Some residues which
have no functional equivalent among other residues will not take part in the shuttle effect. In their case evolutionarily
effective back mutations will be very rare. The rare occurrence of these residues is a further reason why this should be so.
I think one may select, as significantly involved in back mutations,
asp, glu, asn, glm, lys, arg, gly, ala, val, leu, ser, thr.
A priori, the most frequent reversible transitions may be expected to be :
Number of types
asp→glu, asp→asn 2
glu→asp, glu→glm 2
asn→asp, asn→glm 2
glm→asn, glm→glu 2
ala→gly, ala→val 2
val→ala, val→leu 2
On the average, any of the twelve residues listed are expected to shift to 1.5 other residues from which a reverse
shift may easily occur. The figure of 1.5 is likely to be an underestimate, because other shifts that are not listed may also
be frequent and easily reversible. Perhaps 2.0 would be a better estimate.
The sum of the number of residues of the twelve listed amino acids represent the following proportions in different
hemoglobin chains :
human alpha 110/141 = 0.78
“ beta 114/146 = 0.78
horse alpha 112/141 = 0.79
“ beta 117/146 = 0.80
cattle alpha 115/141 = 0.81
human gamma 115/146 = 0.79
Thus the total number of molecular sites should perhaps not been taken into account for evaluating back mutations,
but, in the case of mammalian hemoglobins, only 79 % thereof. At some of the sites comprised in the 79 % evolutionarily effective
mutations may be so rare that evolutionarily effective back mutations occur with negligible frequency. This effect may to
some extent be compensated by back mutations at sites that are not counted in the 79 %.
One may postulate that, among the possible one step transitions allowed by the genetic code, those that are most
frequently evolutionarily effective will also be those after which evolutionarily back mutations most frequently occur. Indeed
the most frequently successful mutations in either sense are expected to be those that lead to the smallest functional changes.
I have not made a thorough study of the different transitions that occur at the different sites. At glycine sites,
the situation is as follows. If we compare all known globin chains, we find at present a total of 33 glycine sites (sites
at which glycine is found to occur at least in one type of chains). In the group of mammalian chains listed in the little
table above there are 23 glycine sites, one of which is invariant. At 10 out of the 22 variable glycine sites, alanine is
occasionally found. No other transition is as frequent, in accordance with expectation. (However serine is a second very frequent
transition). The number of glycine sites at which other amino acids are found in the mammalian chains under consideration
is as follows :
ala 10 (= at 10 different glycine sites ala is occasionally found)
The number of observed transitions per site varies from 0 to 3 in this mammalian group, with a mean of 1.8.
Perhaps we may expect that the mean number of observed transitions at all sites that are considered for back mutations
is close to 2. This figure checks with the one obtained above. Does this agreement suggest that, at the 79 % of the globin
sites, we have an equally good chance for a back mutation after any transition occurred, the chances for an evolutionarily
effective back mutation and an evolutionarily effective forward mutation are equal.
Do these pieces of information allow you to take back mutations in account in your treatment?
At Rutgers, I met Dr. Jennings, the head of the Biochemistry Division of the Office of Naval Research. To my surprise
he asked me questions relating to your research grants. He wondered whether you wanted financial support to be continued,
and whether you had exhausted the funds put at your disposal in the past. He said, “If Dr. Pauling had wanted us to continue
support, he would have let us know it by now”. I said you continued your association with Cal Tech and that, as far as I knew,
you would continue to require money for research projects and that the funds now at your disposal were pretty nearly exhausted.
I hope it was the right thing for me to say. I believe my answer won’t do any harm, whether you decide to ask the ONR for
further support or not.
Warmest wishes for you and Mrs. Pauling!