No.

4B09 December 30, 1961 NATURE 1227

GENERAL NATURE OF THE GENETIC CODE FOR PROTEINS

By DR. F. H. C. CRICK, F.R.S., LESLIE BARNETr, DR. S. BRENNER
and DR. R. J. WAITS-TOBIN
Medical Research _Council Unit for Molecular Biology,
Cavend1sh Laboratory, Cambridge

T HERE is now a mass of indirect evidence which

sug~sts that the amino-acid sequence along the
polypeptide chain of a protein is determined by the
codes of Crick, Griffith and OrgeP. Alternatively,
the correct choice may be made by starting at a fixed
point and working along the sequence of bases three
sequence _of t~e bases along some particular part of (or four, or whatever) at a time. It is this possibility
the nucleic a01d of the genetic material. Since there which we now favour.
are twenty common amino-acids fomtd throughout
Nature, b';lt only four common bases, it has often Experimental Results
?een surmised that the sequence of the four bases is Our genetic experiments have been carried out on
m some way a code for the sequence of the amino- the B cistron of the ru region of the bacteriophage
acids. In this article we report genetic experiments T4, which attacks strains of Escherichia coli. This is
which, to~ether wi~h the work of others, suggest that the system so brilliantly exploited by Benzer&• 7 •
the genetlC code IS of the following general type: The ru region consists of two adjacent genes, or
(a) A group of three bases (or, less likely, a multiple 'cistrons', called cistron A and cistron B. The wild-
of three bases) codes one amino-acid. type phage will grow on both E. coli B (here called
. (b) The code is not of the overlapping type (see
B) and on E. coli Kl2 (:A) (here called K), but a phage
Fig. 1). which has lost the function of either gene will not
(c) The sequence of the bases is read from a :fixed grow on K. Such a phage produces an 1" plaque on B.
starting point. This determines how the long Many point mutation.<> of the genes are known which
sequences of bases are to be correctly read off as behave in this way. Deletions of part of the region
triplets. There are no special 'commas' to show how are also found. Other mutations, known as 'leaky',
t~ select the right triplets. If the starting point is
show partial function; that is, they will grow on K
~sp!aced by one base, then the reading into triplets
but their plaque-type on B is not truly wild. We
IS displaced, and thus becomes incorrect.
report here our work on the mutant P 13 (now
(d) The code is probably 'degenerato'; that is, in re-named FO 0) in the Bl segment of the B cistron.
general, one particular amino-acid can be coded by This mutant was originally produced by the action
one of several triplets of bases. of proflavin•.
We• have previously argued that acridines such
The Reading of the Code as proflavin act as mutagens because they add or
The evidence that the genetic code is not over- delete a base or bases. The most striking evidence in
lapping (see Fig. 1) does not come from our work, favour of this is that mutants produced by acridines
but from that of Wittmann1 and of Tsugita and are seldom 'leaky'; they are almost always completely
Fraenkel-Conrat 2 on the mutants of tobacco mosaic lacking in the function of the gene. Since our note
vi~s produced by nitrous acid. In an overlapping was published, experimental data from two sources
triplet code, an alteration to one base will in general have been added to our previous evidence: (I) we
cha?ge thre.e adjacent amino-acids in the polypeptide have examined a set of 126 ru mutants made with
cham. Their work on the alterations produced in the acridine yellow; of these only 6 are leaky (typically
protein of the virus show that usually only one about half the mutants made with base analogues
amin?·acid a~ a ti:z_ne is changed as a result of treating are leaky); (2) Streisinger 10 has found that whereas
the nbonucleic acid (RNA) of the virus with nitrous mutants of the lysozyme of phage T4 produced by
acid. In the rarer cases where two amino-acids are base-analogues are usually leaky, all lysozyme
a!tered (owing presumably to two separate deamina- mutants produced by profiavin are negative, that is,
tlons by the mtrous acid on one piece of RNA), the the function is completely lacking.
altered amino-acids are not in adjacent positions in If an acridine mutant is produced by, say, adding a.
the polypeptide chain. base, it should revert to 'wild-type' by deleting a base.
Brenner• had previously shown that, if the code Our work on revertants of FO 0 shows that it usually
were universal (that is, the same throughout Nature), ETC.
then all overlapping triplet codes were impossible.
Stortinq point \ __L_,
Moreover, all the abnormal human hmmoglobins
studied in detail• show only single amino-acid changes. • ....-J-. - Overloppinq code
1
The. newer experimental results essentially rule out ,.....--J---1
all simple codes of the overlapping type.
If the code is not overlapping, then there must be
NUCLEIC ACID - I I I I I I I- - -
L...-,--J L--r--' l..-.-r--' - - - -
some arrangement to show how to select the correct ETC.
triplets (~r quadruplets, or whatever it may be) along
the contmuous sequence of bases. One obvious Non·overloppinq code
suggestion is that, say, every fourth base is a 'comma'.
Another idea is that certain triplets make 'sense', Fig. 1. To show the difference between an overlapping code and
a non-overlapping code. The short vertical Jines represent the
whereas others make 'nonsense', as in the comma-free bases of the nucleic acid. The case illustrated Is for a triplet code

© 1961 Nature Publishing Group

 1228 NATURE December 30, 1961 voL. 1e2

reverts not by reversing the original mutation but by duced as suppressors of these suppressors. Again all
producing a second mutation at a nearby point on these new suppressors are non-leaky r mutants, and
the genetic map. That is, by a 'suppressor' in the all map within the Bl segment for one site in the
same gene. In one case (or possibly two cases) it B2 segment.
may have reverted back to true wild, but in at least Once again we have repeated the process on two
18 other cases the 'wild type' produced was really a of the new suppressors, with the samo general results,
double mutant with a 'wild' phenotype. Other as shown in Fig. 2, lines i and j.
workers 11 have found a similar phenomenon with All these mutants, except the original FO 0,
ru mutants, and Jinks 12 has made a detailed analysis occurred spontaneously. We have, however, pro-
of suppressors in the hm gene. duced one set (as suppressors of FO 7) using acridine
The genetic map of these 18 suppressors of FO 0 yellow as a. mutagen. The spectrum of suppressors
is shown in Fig. 2, line a. It will be seen that they we get (see Fig. 2, line h) is crudely similar to the
all fall in the Bl segment of the gene, though not all spontaneous spectrum, and all the mutants are
of them are very close to FO 0. They scatter over non-leaky r's. We have also tested a (small) selection
a region about, say, one-tenth the size of the B of all our mutants and shown that their reversion-
cistron. Not all are at different sites. We have rates are increased by acridine yellow.
found eight sites in all, but most of them fall into Thus in all we have about eighty independent r
or near two close clusters of sites. mutants, all suppressors of FO 0, or suppressors of
In all cases the suppressor was a non-leaky r. That suppressors, or suppressors of suppressors of sup·
is, it gave an r plaque on Band would not grow on K. pressors. They all fall within a limited region of
This is the phenotype shown by a complete deletion the gene and they are all non-leaky r mutants.
of the gene, and shows that the function is lacking. The double mutants (which contain a. mutation
The only possible exception was one case where the plus its suppressor) which plate on K have a. variety
suppressor appeared to back-mutate so fast that we of plaque types on B. Some are indistinguishable
could not study it. from wild, some can be distinguished from wild with
Each suppressor, as we have said, fails to grow on difficulty, while others are easily distinguishable and
K. Reversion of each can therefore be studied by produce plaques rather like r.
the same procedure used for FO 0. In a few cases We have checked in a. few cases that the pheno-
these mutants apparently revert to the original wild- menon is quite distinct from 'complementation',
type, but usually they revert by forming a. double since the two mutants which separately are pheno-
mutant. Fig. 2, lines b-g, shows the mutants pro- typically r, and together are wild or pseudo-wild,

..------. r----t

22 101
{ 19v
ALL 10 12 I FCO(+) 15 18
- >II ~
I
9
I I al Z3
II II
.JI !f
I
- - e t

8
~

6
26 I
FCI(-) I
41 bl 40

.Jp
FC6(-) J6 .18
cl JJ J/ .J4 .10
-- ~
32 28 29
I I I
I
........ I I

:!;
FCIO(-) 6Z 6S
dl 64
All 71
+ FCII(-) 72 69

-- I ~68
I I
~I
I
67
I
66

I $6 $7
FC9(-) S/ ss .5J
I II
I
49
I
so
I
$4
I
sz
I
saI
I FC7(-)
48 44
46
~I
gI 4Z
I I 4S
k
102
I
4J
I
47
--
hi 9b 89
JlS91 96
98
I
80 76
78 7S I FC42(+)
77 74
jl

ALL

r~ --
81
I
8Z
I
I
j I
I
88
84
8J
I
87
FC47C+)
I
I 82
Bla segment B1bt Bib2.
J.'ig. 2. A tentative map---<mly very roughly to scale-of the left-hand end of the B cistron, showing the position of the FO family of
mutants. The order of sites within the regions covered by brackets (at the top of the llgure) Is not known. Mutants in italics have
only been located approximately. Each line represents the suppressors picked up from one mutant, namely, that marked on the line
in bold llgures

© 1961 Nature Publishing Group

 No. 4B09 December 30, 1961 NATURE 1229
must be put together in the same piece of genetic Double Mutants
material. A simultaneous infection of K by the
two mutants in separate viruses will not do. We can now ask: What is the character of any double
mutant we like to form by putting together in the
The Explanation in Outline same gene any pair of mutants from our set of about
eighty? Obviously, in some cases we already know
011r explanation of all these facts is based on tho the answer, since some combinations of a + with a -
theory set out at the beginning of this article. were formed in order to isolate the mutants. But, by
Although we havo no direct evidence that the B definition, no pair consisting of one + with another +
cistron produces a polypeptide chain (probably has been obtained in this way, and there are many
through an RNA intermediate), in what follows we combinations of + with - not so far tested.
shall assume this to be so. To fix ideas, we imagine Now our theory clearly predicts that all combina-
that the string of nucleotide bases is read, triplet by tions of the type + with + (or - with - ) should
triplet, from a. starting point on the left of the B give an r phenotype and not plate on K. We have
cistron. We now suppose that, for example, the put together 14 such pairs of mutants in the cases
mutant FO 0 was produced by the insertion of an listed in Table 1 and found this prediction confirmed.
additional base in the wild-type sequence. Then this
addition of a base at the FO 0 site will mean that the Table 1. DOUBLE MUTANTS HAVINO THE r PHENOTYPE
reading of all the triplets to the right of FO 0 will
be shifted along one base , and will therefore be incor-
- With - + With +
FC (1 + 21) FC (0 + 58) FC (40 + 67)
rect. Thus the amino-acid sequence of the protein FO (23 + 21) PC (0 + 38) FC (40 + 58)
which the B cistron is presumed to produce will be FC (1 + 23) PC (0 + 40) PC (40 + 55)
PC (1 + 9) FC (0 + 55) FC (40 + 1\4)
completely altered from that point onwards. This FC (0 + 54) FC (40 + 38)
explains why the function of the gene is lacking. To
simplify the explanation, we now postulate that a At first sight one would expect that all combinations
suppressor of FO 0 (for example, FO 1) is formed by of the type ( + with -) would be wild or pseudo-wild,
deleting a base. Thus when the FO l mutation is but the situation is a little more intricate than that,
present by itself, all triplets to the right of FO 1 and must be considered more closely. This springs
will be read incorrectly and thus the function will be from the obvious fact that if the code is made of
absent. However, when both mutations are present triplets, any long sequence of bases can be read
in the same piece of DNA, as in the pseudo-wild correctly in one way, but incorrectly (by starting
double mutant FO (0 + 1), then although the at the wrong point) in two different ways, depending
reading of triplets between FO 0 and FO 1 will be whether the 'reading frame' is shifted one place to
altered, the original reading will be restored to the the right or one place to the left.
rest of the gene. This could explain why such double If we symbolize a shift, by one place, of the reading
mutants do not always have a true wild phenotype frame in one direction by - and in the opposite
but are often pseudo-wild, since on our theory a direction by +-, th@n we can establish the convention
small length of their amino-acid sequence is different that our + is always at the head of the arrow, and
from that of the wild-type. our - at the tail. This is illustrated in Fig. 3.
For convenience we have designated our original We must now ask: Why do our suppressors not
mutant FO 0 by the symbol + (this choice is a pure extend over the whole of the gene ? The simplest
convention at thifl stage) which we have so far con- postulate to make is that the shift of the reading
sidered as the addition of a single base. The suppres- frame produces some triplets the reading of which is
sors of FO 0 have therefore been designated -. The 'unacceptable'; for example, they 10ay be 'nonsense',
suppressors of these suppressors have in the same way or stand for 'end the chain', or be unacceptable in
been labelled as + , and the suppressors of these last some other way due to the complications of protein
sets have again been labelled - (see Fig. 2). structure. This means that a suppressor of, say,
FO 0 must be within a region such
r----r----r----r----r----r----r----r----r- that no 'unacceptable' triplet is pro·
•A B C•A B C•A B CIA B C•A B (lAB C•A B (lAB Cl duced by the shift in the reading
~~~~_.~~~--~~_.~~~--~_. ____ ~._~~_.~~--
frame between FO 0 and its sup-
pressor. But, clearly, since for any
sequence there are two possible mis-
readings, we might expect that the
r-·--r----r----r----r----r----r----r----r-
+ 'unacceptable' triplets produced by
a _., shift would occur in dif-
'A B C'A B (IB C A'B C A'B C A'B C A'A B C A B C' 1
ferent places on the map from those
.,.
~~~~--._--~~_._.~~~~~~--~~~--~~~~-
.,. produced by a· +-shift .
deletoon odd it ion Examination of the spectra of
suppressors (in each case putting
t in the arrows ->- or +-) suggests
r - - - - r - - - - r ~ - - - r - - - - r - -
•w - r - - - - r - - - - r - - - - r - that while the ~ shift is acceptable
1 A B ('A B ( ( A B'C A B'C A B'C A B A BC A B Cl__ anywhere within our region (though
1 1 1

not outside it) the shift +-, starting

~ from points near FO 0, is acceptable
t
Stortonq
deletion
over only a more limited stretch.
This is shown in Fig. 4. Some-
poont where in the left part of our region,
Fig. 3. To show that our convention for arrows is consistent. The letters .A, B and c between FO 0 or FO 9 and the
each represent a different base of tho nucleic acid. For simplicitv a repeating sequence FO 1 group, there must be one or
of bases, ABC, ill shown. (This would code for a polypeptide for which evef\' amino-acid
was the same.) A triplet code is asaumed. The dotted lines repre•ent the imaginary more unacceptable triplets when
'reading frame' Implying that the sequence is read in sets of three starting on the left a -<c-. shift is made; similarly for

© 1961 Nature Publishing Group

 1230 NATURE December 30, 1961 veL :gz

the region to the right of the FO 21 cluster. _Thus FCO

we predict that a combina~ion. ~f _a + w1th a
j.

- will be wild or pseudo-wild 1f 1t mvolves a -+ FC I I - t•- -- -- ' FC 21

shift, but that such pairs involving a -<- shift will be £li I t FC40
phenotypically r if the arrow crosses one or more of
FCb • t FC 36
the forbidden places, since then an unacceptable
triplet will be produced. FCIO. - - - - - - - - - - - - - - - - - - - - - : - FC bJ
- tFCbb
Table 2. DOUBLE MUTANTS OF THE TYPE ( + WITH -) ~·---------------------------.
FC'I-; ---------------_.t FOB
~ FC 41 FC 0 FC 40 FC 42 FC 58 • FC 63 FC 38
FCl lV W W W W fC4bi <-------;- FC7 ;- FC47
FC86 W W W W W
FC74--;' t FC42
F09 w w w w w 1-----------------------T FC47
FC 82 w w w w FCBb

FC 21 w w w Fig. 4. A simplified version of the genetic map of Fig. 2. Each

line corresi>onds to the suppressor from one mutant, here under-
FC 88 w w lined The arrows show the range over which suppressors have so
far been found, the extreme mutants being named on the map.
FC87 w Arrows to the right are shown solid, arrows to the left dotted

W, wild or pseudo-wild phenotype; ll', wild or pseudo-wlld com- across the 'unacceptable' regions for the -<- shifts,
bination used to Isolate the suppressor; r, r phenotype.
• Double mutants formed with FC 58 (or with l!'C 34) give sharp but these we can avoid by a proper choice of mutants.
plaques on K. We have so far examined the six cases listed in
Table 3 and in all cases the triples are wild or pseudo-
We have tested this prediction in the 28 cases wild.
shown in Table 2. We expected 19 of these to be The rather striking nature of this result can be
wild, or pseudo-wild, and 9 of them to have the r seen by considering one of them, for example, the
phenotype. In all cases our prediction was correct. triple (FO 0 with FO 40 with FO 38). These three
We regard this as a striking confirmation of our mutants are, by themselves, all oflike type ( + ). We
theory. It may be of interest that the theory was can say this not merely from the way in which they
constructed before these particular experimental were obtained, but because each of them, when
results were obtained. combined with our mutant FO 9 (-),gives the wild,
or pseudo-wild phenotype. However, either singly
Rigorous Statement of the Theory or together in pairs they have an r phenotype, and
So far we have spoken as if the evidence supported will not grow on K. That is, the function of the
a triplet code, but this was simply for illustration. gene is absent. Nevertheless, the combination of all
Exactly the same results would be obtained if the three in the same gene partly restores the function
code operated with groups of, say, 5 bases. Moreover, and produces a pseudo-wild phage which grows on K.
our symbols + and - must not be taken to mean This is exactly what one would expect, in favourable
literally the addition or subtraction of a. single base. cases, if the coding ratio were 3 or a multiple of 3.
It is easy to see that our symbolism is more exactly Our ability to find the coding ratio thus depends
as follows: on the fact that, in at least one of our composite
mutants which are 'wild', at least one amino-acid
+ represents +m, modulo n must have been added to or deleted from the poly-
- represents -m, modulo n peptide chain without disturbing the function of the
gene-product too greatly.
where n (a positive integer) is the coding ra.tio (that This is a very fortunate situation. The fact that we
is, the number of bases which code one amino-acid) can make these changes and can study so large a
and m is any integral number of bases, positive or region probably comes about because _this part
negative. of the protein is not essential for its functwn. That
It can also be seen that our choice of reading this is so has already been suggested by Champe
direction is arbitrary, and that the same results (to and Benzer 18 in their work on complementation in the
a first approximation) would be obtained in whichever ru region. By a special test (combined infection on
direction the genetic material was read, that is, K, followed by plating on B) it is possible to examine
whether the starting point is on the right or the left the function of the A cistron and the B cistron
of the gene, as conventionally drawn. separately. A particular deletion, 1589 (see Fig. 5)
covers the right-hand end of the A cistron and part
Triple Mutants and the Coding Ratio of the left-hand end of the B cistron. Although
1589 abolishes the A function, they showed that it
The somewhat abstract description given above is allows the B function to be expressed to a considerable
necessary for generality, but fortunately we have extent. The region of the B cistron deleted by 1589
convincing evidence that the coding ratio is in fact is that into which all our FO mutants fall.
3 or a multiple of 3.
This we have obtained by constructing triple Joining two Genes Together
mutants of the form ( + with + with +)or (- with -
with - ). One must be careful not to make shifts We have used this deletion to re-inforce our idea
that the sequence is read in groups from a fixed
Table 3. TRIPLE MUTANTS HAVING A WILD OR PSEUDO-WU.D PHENO· starting point. Normally, an alteration confined
TYPE to the A cistron (be it a deletion, an acridine mutant,
FO (0 + 40 + 38) or any other mutant) does not prevent the _expre~si~n
FC (0 + 40 + 68)
FC (0 + 40 + 67) of the B cistron. Conversely, no alterat10n w1thm
FC (0 + 40 + 64) the B cistron prevents the function of the A cistron.
FC (0 + 40 + 65)
FC (1 + 21 + 23) This implies that there may be a region between the

© 1961 Nature Publishing Group

 No. 4Boe December 30, 1961 NATURE 1231
two cistrons which separates them and allows their P6J
functions to be expressed individually. WT3

-
WT2 WTI WT5
We argued that the deletion 1589 will have lost
this separating region and that therefore the two W l6
·"'
(partly damaged) cistrons should have been joined
together. Experiments show this to be the case, Fig. 6. Genetic map of P 83 and its suppressors, W'l' l, etc.
for now a.n alteration to the left-hand end of the The region falls within segment B9a near the right-hand end of
the B cistron. It is not yet known which way round the map i,;
A cistron, if combined with deletion 1589, can prevent in relation to the other figures
the B function from appearing. This is shown in
Fig. 5. Either the mutant P43 or Xl42 (both of close to the FO 9 site. By a suitable choice of
which revert strongly with acridines) will prevent the partners, we have been able to show that two are +
B function when the two cistrons are joined, although and two are -. Secondly, we have two mutants
both of these mutants are in the A cistron. This is (X146 and X225), produced by hydrazine10 , which
also true of X142 Sl, a suppressor of Xl42 (Fig. 5, fall on or near the site FO 30. These we have been
case b). However, the double mutant (X142 with able to show are both of type - .
X142 81), of the type ( + with -),which by itself is Thus unless both acridines and hydrazine usually
pseudo-wild, still has the B function when combined delete (or add) an even number of bases, this evidence
with 1589 (Fig. 5, case c). We have also tested in this supports a coding ratio of 3. However, as tho action
way the 10 deletions listed by Benzer', which fall of these mutagens is not understood in detail, we
wholely to the left of 1589. Of these, three (386, 168 cannot be certain that the coding ratio is not 6,
and 221) prevent the B function (Fig. 5, case j), although 3 seems more likely.
whereas the other seven show it (Fig. 15, case e). We We have preliminary results which show that other
surmise that each of these seven has lost a number of acridine mutants often revert by means of close
bases which is a multiple of 3. There are theoretical suppressors, but it is too sketchy to report here.
reasons for expecting that deletions may not be A tentative map of some suppressors of P 83, a
random in length, but will more often have lost a mutant at the other end of the B cistron, in segment
number of bases equal to an integral multiple of the B 9a, is shown in Fig. 6. They occur within a shorter
coding ratio. region than the suppressors of FO 0, covering a
It would not surprise us if it were eventually shown distance of about one-twentieth of the B cistron.
that deletion 1589 produces a protein which consists The double mutant WT (2 + 5) has the r phenotype,
of part of the protein from the A cistron and part, as expected.
of that from the B cistron, joined together in the same
polypeptide chain, and having to some extent the Is the Code Degenerate l
function of the undamaged B protein. If the code is a triplet code, there are 64 (4 x 4 x 4)
possible triplets. Our results suggest that it is unlikely
Is the Coding Ratio 3 or 6 ? that only 20 of these represent the 20 amino-acids
It remains to show t.hat the coding ratio is pro- and that the remaining 44 are nonsense. If this
bably 3, rather than a multiple of 3. Previous rather were the case, the region over which suppressors of
rough extimatesl0• 14 of the coding ratio (which are the FO 0 family occur (perhaps a quarter of the B
admittedly very unreliable) might suggest that the cistron) should be very much smaller than we observe,
coding ratio is not far from 6. This would imply, on since a shift of frame should then, by chance, pro-
our theory, that the alteration inFO 0 was not to one duce a nonsense reading at a much closer distance.
base, but to two bases (or, more correctly, to an even This argument depends on the size of the protein
number of bases). which we have assumed the B cistron to produce.
We have some additional evidence which suggests We do not know this, but the length of the cistron
that this is unlikely. First, in our set of 126 mutants suggests that the protein may contain about 200
produced by acridine yellow (referred to earlier) amino-acids. Thus the code is probably 'degenerate',
we have four independent mutants which fall at or that is, in general more than one triplet codes for
each amino-acid. It is well known that if this were
F'unction so, one could also account for the major dilemma of
o!B the coding problem, namely, that while the base
~A ciatron~ ~Bdstron~ cistron
composition of the DNA can be very different in
J
different micro-organisms, the amino-acid composi-
I tion of their proteins only changes by a moderate
amounta. However, exactly how many triplets
I code amino-acids and how many have other functions
we are unable to say.
)(
@J @x
Future Developments
)fl(
@l @x
Our theory leads to one very clear prediction.
(i)J CDx Suppose one could examine the amino-acid sequence
\\\'\ of the 'pseudo-wild' protein produced by one of our
double mutants of the ( + with -) type. Conven-
1'
deletion
1589
tional theory suggests that since the gene is only
Fig. 5. Summary of the results with deletion 1589. The first two
lines show that without 1589 a mutation or a deletion in the A altered in two places, only two amino-acids would
cistron does not prevent the B cistron from functioning. Deletion be changed. Our theory, on the other hand, predicts
1589 (line 3) also allows the B cistron to function. The other that a string of amino-acids would be altered, covering
cases, in some of which an alteration in the A cistron prevents
the function of the B cistron (when 1589 is also present), are the region of the polypeptide chain corresponding
discussed in the text. They have been labelled (a.), (b). etc., for to the region on the gene between the two mutants.
convenience of reference, although cases (a) and (d) are not
discu&!ed in this paper. ./implies function; x implies no function A good protein on which to test this hypothesis is

© 1961 Nature Publishing Group

 1232 NATURE December 30, l%1 VOL. 192

the lysozyme of the phage, at present being studied are particularly grateful to Prof. C. F. A. Pantin for
chemically by Dreyer 17 and genetically by Streisin- allowing us to use a room in the Zoological Museum,
gerto. Cambridge, in which the bulk of this work was done.
At the recent Biochemical Congress at Moscow, the
1
audience of Symposium I was startled by the Wittman, H. G., Symp. 1, Fifth Intern. Cong. Biochem., 1961 for
refs. (in the press). '
announcement of Nirenberg that he and Matthaeits • Tsuglta, A., and Fraenkel-Conrat, H., Proc. U.S. Nat. Acad. Sci., 48,
had produced polyphenylalanine (that is, a poly- 636 (1960); J. Mol. Bioi. (In the press) .
peptide all the residues of which are phenylalanine) 'Brenner, S., Proc. U.S. Nat. Acad. Sci., 43, 687 (1957).
• For refs. see Watson, H. C., and Kendrew, J. C., Nature,l80 670
by adding polyuridylic acid (that is, an RNA the bases (1961). •
of which are all uracil) to a cell-free system which can 1
Crick, F. H. C., Griffith , J. 8., and Orgel, L. E., Proc. U.S. Nat.
synthesize protein. This implies that a sequence of Acad. Sci., 43, 416 (1957).
macils codes for phenylalanine, and om work suggests
1
Benzer, S., Proc. U.S. Nat. Acad. Sci., 45, 1607 (1959), for refs. to
earlier papers.
that it is probably a triplet of uracils. ' Benzer, 8., Proc. U.S. Nat. A cad. Sci. , 47, 403 (1961); see his Fig. 3.
It is possible by various devices, either chemical or 1 Brenner, S., Benzer, S., and Barnett, L., Nature, 182, 983 (1958).

enzymatic, to synthesize polyribonucleotides with • Brenner, 8., Barnett, L., Crick, F. H. C., and Orgel, A., J. Mol.
Bioi., 3, 121 (1961).
defined or partly defined sequences. If these, too, "Streislnger, G. (personal communication and in the press).
will produce specific polypeptides. the coding problem 11 Feynman, R. P.; lBenzer, 8.; Freese, E. (all personal communi·
is wide open for experimental attack, and in fact cations).
many laboratories, including our own, are already "Jinks, J. L., Hereditv, 18, 153, 241 (1961).
"Champe, S., and Benzer, S. (personal communication and in pre·
working on the problem. If the coding ratio is indeed paration).
3, as our results suggest, and if the code is the same "Jacob, F., and Wollman, E. L., Sea:ualitv and the Genetic~! of Bacteria
throughout Nature, then the genetic code may well (Academic Press, New York, 1961). Levlntbal, C. (personal
communication).
be solved within a year. " Orgel, A., and Brenner, S. (in preparation).
We thank Dr. Alice Orgel for certain mutants and "Sueoka, N. Cold Spring Barb. Svmp. Quam. Bioi. (in the press).
for the use of data from her thesis, Dr. Leslie Orgel "Dreyer, W. J., Symp. 1, Fifth Intern. Cong. Biochem., 1961 (In the
press).
for many useful discussions, and Dr. Seymour 11 Nirenberg, M. W., and Matthael, J. H., Proc. U.S. Nat. Acad. Sci.,
Benzer for supplying us with certain deletions. We 47, 1588 (1961).

SCIENCE AND WORLD AFFAIRS

activities arose independently from different working
T HE Seventh Pugwash Conference on Science
and World Affairs was held at Stowe, Vermont,
during September 5-9. Forty-one scientists from
groups, and this is reflected in several places in the
statement. This is a welcome indication of the
twelve countries attended the Conference*. essential unity in science. The discussions were
This Conference had as its theme "International carried on in a. spirit of friendly co-operation, and
Co-operation in Pure and Applied Science". The full agreement was reached by the entire Conference
previous conferences have been chiefly concerned on the suggestions enumerated here.
with ways of preventing the misuse of science in the
wholesale destruction of mankind. In this Conference (I} Co-operation in the Earth Sciences
at Stowe, constructive international co-operation in
science was discussed, because it is a way to create The planet Earth is the common abode of all
trust between nations, a. trust which develops from humans. They have a. common interest, both intel-
common interests and from experience in working lectual and practical, in increasing the knowledge of
together. the structure and dynamics of the Earth.
Science misused by nations to foster their com- The following proposals were made by the Confer-
petitive interests as world powers makes possible the ence.
destruction of mankind. Science used co-operatively (A) A survey of the entire ocean in three dimensions.
by all nations for the increase of human knowledge (1) The ocean floor. An international programme was
and the improvement of man's productive capacity proposed to develop a detailed map of the floor of the
can give all men on Earth a satisfactory and worth- world ocean, including sub-bottom reflecting layers.
·while life. Scientists bear a responsibility both to (2) Waters of the ocean. An international pro-
foster the constructive use of science and to help in gramme should be devised to survey a.nd map
preventing its destructive use. the three-dimensional distribution of temperatures,
The deliberations of the Conference were carried salinity, density, dissolved oxygen, and nutrient salts,
out in plenary sessions and in meetings of working under average conditions, of the ocean and synoptic
groups. Similar suggestions for co-operative research surveys to develop the broad picture of seasonal and
• Sir John Crawlord (Australia); Prof. Rans Thirring (Austria); shorter-period changes in more limited areas, as well
l'L"Of. C. Pavan (Brazil); Prof. G. Nadjakov (Bulgaria); Prof. G. as the study of the interactions among the major
Burkhardt (Federal Republic of Gennany); Sir Edward Bullard,
Prof. A. Haddow, Sir Ben Lockspeiser, Prof. J . Rotblat (Great Britain) ; bodies of water in the ocean.
Prof. F. B. Straub (Hungary) ; Dr. G. Bernardini (Italy); Prof. T. (3) Ocean life. An international survey and map-
Toyoda (,Japan); Prof. B. V. A. Rllling (Netherlands); Academician ping showing the major biological provinces of the
A. A. Blagonravov, Academician N. N. Rogolubov, Academician M. l\1.
Dublnin, Prof. V, M. Khvostov, Academician N. M. Sissakian, Prof. ocean and determination of the fertility of the waters
N. A. Talensky, Academician I . E . Tamm, Academician A. V. Topchiev
(U.S.S.R.); Prof. Harrison Brown, Dr. William Consolazlo, Prof. at all levels in the food chain and the standing crop
Paul Doty, Prof. Bentley Glass, Prof. C. O'D. Iselin, Dr. Martin of food materials available for human use should be
Kaplan, Prof. Chauncey Leake, Prof. Linus Pauling, Prof. Jay Orear,
:Prof. W. Pickering, Mr. Gerard Piel, Prof. I. Rabl, Prof. Eugene undertaken.
R.ablnowlteh, Dr. Roger Revelle, Prof. Alexander Rich, Prof. Walter (B) Earth's c1'U8t and mantle. Deep drilling pro-
Rosenbllth, Dr. Eugene Staley, Dr. Alvin Weinherg, Prof. Eugene gramme. The objective of drilling through the
Wigner, Prof. J. R. Zacharias (United States).

© 1961 Nature Publishing Group