Rear Professor Klee~=<

I was quite delighted to receive your note, and look forward to studying
the report you mentioned. fhe prospect of opening up this kind of communica-
tion is of course the main purpose of semmmax publishing notes like thet in
the PNAS and I & hope to see it flourééh. It may be rather informal, but it
is convenient to pass on some more background on what we are doing in the
form of the enclosed copies pf other material,

At this moment I am just in the course of final revision of part 2 of the
DENDRAL system, and do not know whether I Will be able to assimilate your new
information into it, but will surely make every effort to do so in writing up the
later parte. I hope you will give me the privilege ofquoting it as a forthcoming
publication.

Having assured ourselves to a practical certainty that the graphs through
18 vertices either have Hamiltonian circuits, or ere unions of graphs of
lower order, we can therefore feel somewhat comfortable that we can anticipate
all the graphs of practical consequence in organic chemistry, and in this sense
the problem is solved for our immediate purposes. However, I am sure there sre
deepir generalizations than Iknow how to deal with, and which might simplify
many aspects of programming the hypothesis-generator, and also the problam of
effictent search techniques for matching substructures (ff which I have just not
attempted to deal with-- perhaps you know of Seeeengt Sussenguth's work on this)
I also have some immediate problems in dealing with symmetries of the various
graphs so as to produce an mmman economical generator for the canonical forms
without undertaking the intolerable labor of retrospectively checking the iso-
morphism of various mappings.

Unleas this is already citedin your report, could you send me Bouwkamp's
initials abd full address, as IW would indeed like to follow up your reference
to his compilation, I feel we have the trivalent graphs fairly well én hand
now even up to 18 vertices, but he may well have a much more attractive to
our approach to quadrivalent vertices- (Igwill have to send you a draft of
part 2/dendral which has progressed somewhat over the pnas note.)

aes SMncerely,
. /
fo Joshue Laderberg
aoe . °