c=: 5 1956

A personal note on the technique of discovery.

Last Thursday I was preoccupied with an insight into an approach to chemotherapeutic
screening. It my be almost commonplace, and its parts are not novel, butin toto
the time may be ripe for it. (See proposal dated 11/30/56 to Bristol Co.-—-J Lein)

I had an opportunity to examine the technique of scientific generalization, and to
realize that this jas been my own principal method of discovery.

The chain of thought here started from the specific questions: what would be the

most likely derivatives of DAP to serve as pdtential anutagonists, and how should
they be isolated. C-chlorinated and -alkylated products suggested themselves. But

if they had to be synthesized from separate precursors, therw would be a tedious
chemical job. Mumiux How about making them from DAP itself which is commercially
available? But this is rather expensive, why not use DAP isolated from culture
filtrates or from cell walls? But actually, is it necessary to &a&olate the DAP
first, when not use the walls directly? Then generalization: would not this apprdach
cover more difficult materials (e.g/ the heptoses and the lipids and the amAnosugars )?
In fact, even unknown constituents, wall or not.

Further generalization, have crude natural products of various origin even been used
as empricdl starting materials for specifically screenable activities?

It then occurred to m that this had bsen a general pattern in my scientific work.
Steps:

&) Consider a specific, sometimes a minor technical problem. Because it is restricted,
it can be analyzed in its essentials, or else intuitively solved.

b) What are the eseentials, expressed in the mosf¢ general form? [Example: replica
plating is equivabaht to milti-tube selection].

and

s) “hat problems of general import can be solved by this technique? Example:
selection for specific genobypes , as used in Neurospora reversion studies, is also
applicable to selection for rare recombinants. Example: one cycle of enrichment
in replica plating can be reiterated for effective indirect selection. Example:
pneumococeus transformation & Salmonella transduction are aspects of same pheno-
menon; lysogenic conversion is correlated at even a furtier level. Example: hypertonic
sucrose might protect against lysis by other agents besides lysozyme.

Not all probhems and discoveries are technical. In many cases one has, rather, to
frame specific hypotheses and find critical mxperiments to decide between them.
Abstraction is of course a necessary precedent to imagination here too. But I am
not sure this has been my strongest technique.

Moral: Solve specific problems, narrow enough to be grappled with. The stimulus of
this accomplishmant, the abstraction of the technique, aad its broader application
will then usually suggest other soluble problems and approaches.

I think that this mntal cast has somethhgg to do with two ways of science (as I think
Seymour quoted from Luria): (1) to start with a methodology and find phenomena; (2)

to start with a phenomenon and find a rule. Not many people do both well: the biophysi-
gxixt cists have generally prospered with (2), which is also safer fop graduate stud-
ents; my own actual emphasis has been on (1). If this is a reflection of talent,

it might be better if I did not stick too,1 ong with the study of the various
systems I uncover, which I have pergapstione or thought + had done.