a degenerate helkx, ,such as a straight line or a circle.

Notice that a hel.ix accounts rather naturally for the long
repeat distance of the structure, as this would correspond

to one turn of the helix.  Thlis restriction - that the

phosphate-sugar groups are spatially uniform - is a great
help in model building  as it reduces very considerably the
number of possibilities that have to be explored.  Indeed at
first we were unable to build any satisfactory model consistent
with our assumptbons, but eventually  we arrived at a structure
which we now bel&ve to be correct in its broad outlines,

  This particular,model does not contain just one DNA        "

chain, but a pair of'them, wound round a common axis,  These

two chains are Hnked together by their bases,  A base on one
chain is joined by weak ph,vsical bonds  to a base at the
same level on the &&her chain, and all the. bases'are, paired
of bn this way  rrtight along the structure.  This is'shown
diagrammatically in Fig. . The general appearance of the
structure is shown Pn a symbolic mariner.... Fig.   , in which

the two ribbons represent the phosphate sugar chains, and the
pairs of bases holdine them together are symbolfzed as
horPeonta7 rods.  It wltll be found that t~lls~figure looks
exactly the same upside down ,  and to preserve this feature
                                                                           *

we have bui2.t our model so that the actual sequence of atoms
in one phosphate-sugar' chain is In the opposite direction
to the corresponding sequence in the other, This\rls shown


symbolically  by the two small arrows.

   Now it is found that one cannot build this model with

any bases one pleases; only  aertain pairs of the four bases

will fit into the structure,   In any pair there must always
be one big one (purinej and one little one(pyrimidine). If

one tries to put in two purines -  two big ones, that is, -

there is not sufficient room for them.  ConverselPy a pair made

of two pyrimidines is too small to bridge the gap between the

two chains.  Moreover when one examines in detail hQvv the
hydroaen bonds are formed between the bases it is found that

(making ce,rtain plausible assumptions)  the pairing is even

more restricted.  The only possible pairs that ~111 fit in


are

and

Adenine   with   Thymine.

Guanine   with   Cytosine.

  The way these pairs are formed 9s shown in Figa        9
The dotted lines show the weak physical bonds, knom as
hydrogen bo;ds,i which hold the two bases of a pair together.
                                                                           

(hydroP;en bondsare, for example, the main forces holding
                                                                           1

different water'molecures together, and it is because of t'nem

that wqter 5.s.a IPquid at room temperatures a?d not a gas)
                     
               /      ; !                                                                .
  lk !Thkse s'pec3.fJ.c pairs can be built into the structure
          i

either wa,y round. ve can have adsnine on the first chain
                                                                           ,,,

paired wi&h thymine oh the second,   or vita versa,  Hut if'
               ,.
we do .bave adenine at some point on one of the chains, it is
          ,               


essential to have thymine paired with it on the other. It,
                                     I
iS impossible to fit in guanine or cytosine or a second

adenine.  In the same way guanine must always be paired with

cgtosine.

On t*he other hand the model places no restriction on the  

sequence  of pairs of bases as one proceeds along the structure4
                                                                           ; ; 
Any JU speciiic pair can follow any other specific pair. .This
is because a pair of bases is flat, and as in this mode'1 they

are stacked one above another like a pSle of mini it'does
not matter which pair goes above whPch,                    I

This spec'lfic pairing of the bases 1s the direct rasult .

of the assumption that both phosphate-sugar chains are'helical;

                                                                                    
'This !lmplPes that the distance apart of two sugar groups at
                                                                           ',
the same level (bne belonging to each chain) is always the         .
                                                                           1
same, no matter 'where one is along the chain.  It follows that.

the bases, which are of course linked to the sugars, have
always the same amount of space in which to fit, as can be

seen from studying Fig.          . If it were not for thSs      

                                                                           

restriction the bases c&nd hy,dro?yen-bond together in many'
                                                                           \
different ways.  It is the regularity of, the phosphate-sugar       '

zhains , ,therefore,  which is at
          * g

Evidence for the model

The experimental evidence in support of a mod@.

the root of the

  
,speci

ftL.c

.

of

general type is now considerable.  lXeasurenents of the density,
          I                                                                           ,!'.>
and water content of the DNA fibres,  taken with. the evidence

showing how the PibreS can be extended in leng%h, strongly


suggest that there are two DNA chains in the structural unit.

The X-ray patterns have a larpe number of places where the
diffraction intensity i.s zero and these occur exactly where

one expects them from helical structures of this type, Moreover

the X-ray diffractton data approximates quite closely to

cylindrical symmetry, as Pt should.  Recently Wilkins and his
co--workers have given a brilliant analysis of the detafls of
the X-ray pattern of the crystalline form, and have shown that
they are consistent with a structure of this type, though in
thI.s form the bases are not perpendicular to the fibre axis,

but tilted awe.y from it*

,As f the structure is a relatively stiff one it easily

explains the extended shape of the DNA in solution.  It is

also consistent with the tftration   curve.  ThPs has irreversible

features which suggest that the bases are hydro,gn-bonded

together.  However the most striking support for the specific

paifing :7f the bases comes f,rom the recent analytlcal data.
These show that for every species so far examined   - and

there are over fo\rty of them ~nxdsx -  the number of adenines

in some given DNA is closely  equal to the number of thymines,

and the nwnl?er of guanines equal. to the number of cytosbnes,
althoupl? the cross--ratio (between say adenine and guanine)

can vary considerably from species to species.  This remakable
                                                                           .                  

fact, which is exactly what one would expect from a model
containing only the spacific pairs, was first pointed out by

Dr. Chargaff,   Indeed, since the sequence of bases along a


single chain is believed to be irrema2aaL this result is very

df.f"icuIt to explain except b,y spe;cific pairinK.

It ml,qh+ be thought that while this model might be correct
for the DNA extracted from the cell and made into fibres the
DNA in the cell was in a radically different 981~ form. This

see!ns unlikely since ft is difficult to see how the very charact-
aristic features of the model could be produced merely by
the extraction procedure. However Dr, Wlkins has shown
that it is possible to get very similar X-ray pictures from
intact biological material, such as sparm heads and bacteriophage,
so that there seems little doubt that the structure is

biologPcally significant.


  The present pos;ftion is tl-lerefore that while the details

of the structure remain to be worked out -  and :.MAl this is

done the model cannot be considered as proved - it seems
very nrobable that the following statements will stand the

test of time:

1. The structure consists of two chains

2. The chains are helfcal and wound round a common axis.

3. They are held together by h,ydrogen bonds between

specific pairs of bases.

4. The structure occurs in biologically intact material,


13
/   A nosslble replication mechanism

NOW the exc4Xl.?g thing about a pnodal of thfs type fs

that ft Imrx!ediately su.ggests how the DNA mig:blt produce an

exact copy 07 itself.  This is beg/&se tha model consists of

two parts, each oif which is the complement  8 the other. The

basic idea is thak the two chains in the structure unwind

and separate,   ,Each chai.n then acts as a sort of mould onto

which a new compl,ementar.y chamn aan be synthesised,  When this

process is complete there will be two pairs of chains where
we only had one betore.  Moreover, because of the specific

pairl.ng of the bases the sequence of the pairs of bases will
%ave been duplicated ',exactly.

`//

As an analogy cons$der two photographic films, one a

                               t

posit&e ard the other h negative of the same scene.   Now if


one @ves the positive to one person , and asks him to print

a negative from i.t, and also .qSvcs the orfqkal negative to

another persorl, nqd ask him to print a positive from it they

W-illE  end up with two pairs o-P photographs, each like the

orfghnl pair.  We shall, in effect have made an exact Copy

of our orIgina pair I.n one step.

To see bon: this works out l.9 the case of DNA let us consider
t-he process in rather more detail,  i3iwe we have to synthesise

two new chains w xq  reqllirt3 Some new material.  The exact

precursors of D.?JA are not. known, but 2et us assume for simplicity

t'nat it is buf.lt up from nucleotides, which is the Tame given

to the small SBolecules which contain one phosphate, one sugar

and one base.

Imagine,then,  that we have a sZn,gle helical chain of

DNA, and t?aat floatiqq EnroundA.t, inside the cell, there is d

supply of the four sorts of nucleotides.  Every now and then

a lose nucleotfde wi.12 attach itself by its base to one of

the bases of the DNA chain.  NOW Pf th%s happens to two
                                                                           ';

adjoining bases; and i.f the loose nucleotides are the type

which can form specific pairs with those already there, they

will be bn just the right positLo to be joir?ed together, and,

eventually ,  to form part of the new chain.  If one or both

of them is not the correct type to go in at that point it

will be impossible to join them together and before long they

will diffuse,elsewhere.  Thus only  the nucleotides with the


At the moment this IAaa aust bs rwnrh9d simply a!5 a

suspect &hat the replication starts almast as soon as the

most alosely m3sociated~.wfth regU.cation ep strCzlctum 0i


exactly `t'ne -type one would need to carry out a specific

replica-t~an S~taaga process, namel"y one showing both varfety

and compleme~tarity,  \Ve can 0nI.y hope that the simplicitsy

of the i&?a will stimulate SoTi3eone to devise some new experiments

which will either prove or disprove it.