OCTOBER 19, 1963 * VOL. 89, NO. 16

A Classification of Virus Particles Based on MorphologyJUNE D. ALMEIDA,* Toronto

"Virus, Virus, shining bright,In the phosphotungstic night,What immortal hand or eye,Dare frame thy fivefold symmetry."(With apologies to William Blake [1757-1827])

OWING to new and improved methods forelectron microscopy and improvements in the

instrument itself, it is now possible to use morpho-logical differences as one basis on which virusparticles can be classified. This communicationpresents such a classification, but the importantsteps which made this possible will first be re-viewed briefly.The way in which virus particles were initially

prepared for examination with the electron micro-scope was simply by allowing a drop of an aqueousvirus suspension to dry on to a grid. Contrast,which permitted the outline of particles to be de-termined, was achieved because the virus particlesproduced greater scattering of the electrons thanthe areas between them. By this technique Kauschiet at. showed in 1939 that tobacco mosaic virus hada rod-like form. Next, the basic outline of severalplant viruses was established by Stanley and Ander-son in 1941; in every instance these were eitherrod-like or spherical. In 1942 Green et at. by thissame technique showed that vaccinia virus had abrick-like form. Influenza virus was seen by Tayloret at. in 1943 and the Shope papilloma virus bySharp et at. in 1942.1A distinct advance occurred in 1945 when Wil-

harms and Wychoff devised a method whereby ametallic mist could be directed from a point sourceand at a known angle on to particles prepared asabove. The shadows thus cast were devoid ofmetallic deposit. By knowing the angle from whichthe metal was sprayed and from studying the sizeand form of the shadows that were cast by differ-

This work was aided financially by grants from the NationalCancer Institute of Canada and the U.S. Public HealthGrant No. C4964.*Division of Biological Research, Ontario Cancer Institute,Toronto 5.

ABSTRACT

Recent improvements in electron micro-scope techniques which allow the study ofvirus fine structure have permitted thegrouping of many viruses on a purelymorphological basis. Briefly the techniquesused in electron microscopy for the studyof viruses are reviewed. and the symmetryproperties of virus particles as revealed bynegative staining are discussed somewhatmore fully.

Finally, virus particles are grouped ontwo bases, firstly the site of formation ofthe virus within the cell as seen by thinsectioning techniques, and secondly thesymmetry property of the virus as seen bynegative staining. Consideration of thegroupings obtained in this way reveals thatthe biochemical and physical properties ofa virus can be deduced from the readilyestablished morphological characteristics.

Canad. Med. Ass. 3.Oct. 19, 1963, vol. 89

Canad. Med. Ass. 3.Oct. 19, 1963., vol. 89 ALMEIDA:CLASSIFICATION OF VIRUS PARTICLES 789

of cells (Figs. 1 c to e) .1 The greatest use of thistechnique from the present point of view is toestablish the location of the virus within the cell.The most recent important advance in methods

for studying viruses with the electron microscopeoccurred in 1959 when Brenner and Home2 ap-plied the method of negative staining to virusparticles. This method entails the treatment ofvirus suspensions with a solution of some electron-dense substance, usually phosphotungstic acid. Thenegative stain not only penetrates between all theparticles in a preparation, but also into the mostminute irregularities on their surfaces; and hence,since the electron beam penetrates sites where thephosphotungstic acid is not present, the surfaceconfiguration of particles is thrown into sharp relief.Adenovirus was one of the first to be visualized bythis method. This was done by Home et al.3 and itshowed the surface of the particles to be studdedwith projecting subunits that were geometricallyarranged. This finding, as will be shown, openedthe door to establishing morphological differencesbetween different viruses, differences that werehitherto undisclosed, and, as we shall see, it wasto provide a new means of classifying viruses ona morphological basis.4' 13At first the method of negative staining was ap-

plied only to purified suspensions of virus. Sincepurification procedures may injure or even destroyvirus particles, the use of purified suspensionslimited to some extent the usefulness of the nega-tive staining technique. However, methods de-veloped by Almeida and Howatson5 and byParsons6 at the Ontario Cancer Institute have over-come many of the difficulties involved with puri-fied suspensions because their methods permitnegative staining to be utilized directly on prepara-tions of infected material without any pri3r puri-fication procedures being required. These methodshave the advantage of not only subjecting virusparticles to the least possible stress,7 but also per-mitting virus particles to be visualized in situ. Forexample, Fig. 7b and Fig. 7c were obtained by thefirst of these methods.5

How MORPHOLOGY INVOLVES SYMMETRYTo proceed further with establishing a morpho-

logical basis for classifying viruses it is necessaryto delve more deeply into the nature and arrange-ments of the subunits which cover their surfaces.To do this involves a discussion of symmetry and

Fig. la.-A shadow-cast preparation of tobacco mosaicvirus showing the rod-shaped particles. X 30,000. (Courtesyof Mr. L. Pinteric.) lb.-A shadow-cast preparation ofcrystallized poliovirus. It is this ability to form crystalsthat made x-ray diffraction studies of viruses possible.x 36,000. (Courtesy of Ivir. L. Pinteric.) Ic. Thin sectionpreparation showing part of a cell infected with polyomavirus. Since this is a nuclear virus the particles are seenwithin the nucleus of the cell. X 35,000. ld.-Anothernuclear virus, Adenovirus Type 3. The micrograph showspart of a crystalline array of virus particles within thenucleus. The double nuclear membrane runs across the topof the micrograph. X 47,000. 1 e-Thin section preparationof a cytoplasmic virus, in this case molluscum contagiosum.The nucleus of the cell is compressed to one side and thecytoplasm contains many pox-type particles. X 25,000.

studies of the symmetry of virus particles that werepreviously made by means of x-ray diffraction oncrystals of virus particles (part of a crystal of polio-virus is shown in Fig. ib).

In 1956, Crick and Watson,8 chiefly from x-raydiffraction studies which they and others had madeon purified virus particles, predicted that simpleviruses would all be found to be composed of acentrally placed nucleic acid fraction and an ex-ternal protective protein coat or shell. They pre-dicted furthermore that the nucleic acid contentof a virus particle would be too limited to be ableto code information for the synthesis of a variety ofproteins and hence that the protein shell of a virusparticle would be found to be composed of smallidentical subunits of protein arranged identicallyon its surface. They predicted moreover that onlytwo types of geometric arrangement would occur.Firstly, there would be one type in which the iden-tical subunits would be arranged in the form of ahelix, with the particle displaying helical sym-metry. Secondly, in the instance of the so-calledspherical viruses, the cubic symmetry found by x-ray diffraction in crystals of these viruses wouldextend to the individual virus particles them-selves; each virus would have subunits arrangedin cubic symmetry and this would mean that theform of the particles would be based on either thetetrahedron, octahedron, or icosahedron. The ad-vent of the negative staining of virus particles madeit possible to prove these predictions correct.

Since x-ray diffraction studies on purified viruscrystals were a very important factor in leading tothe predictions that virus particles, like crystals,would manifest symmetry, it was only to be ex-pected after the advent of negative staining thatsome of the terminology commonly used in connec-tion with crystals would be applied to virusparticles. In particular, since different crystalsdisplay different types of symmetry and are classi-fied to some extent by the kind of symmetry thatthey display, many viruses can now be classifiedas to whether they display cubic or helical sym-metry.9 What is meant by cubic symmetry inviruses will next be described.

CUBIC SYMMETRYIn 1957 Williams and Smith showed that when

tipula iridescent virus was shadow-cast from twodifferent points, the contours of the shadows seenwere those that would be thrown by an icosa-hedron;10 this was the first verification by electronmicroscopy of one of Crick and Watson's predic-tions.8The icosahedron is one of a class of solid geo-

metrical bodies, the regular polyhedra, each typeof which possesses the characteristics of havingmany faces all of which are identical with eachother. The icosahedron has 20 faces and 12 vertices.As is shown in the accompanying diagrams (Fig.2), it is possible to visualize such a body having

790 ALMEIDA: CLASSIFICATION OF Vnius PARTICLESCanad. Med. Ass. J.Oct. 19, 1963, vol. 89

Crick and Watson to the effect that the proteincoats of virus particles would be built of smallidentical subunits arranged in an identical fashionwith one another. In considering this matter weshall run into the difficulty that the term subunit,in relation to virus particles, has been used in twoways and this requires some clarification.With negative staining the protein coat of icosa-

hedral virus particles is seen to be studded withextremely minute projections; these have beentermed capsomeres. The capsomeres are themorphological subunits, and they fit together tocover the whole of the particle. The most efficientand economical shape for subunits to have, if theyare to be fitted together to cover a flat surface, isthat of a hexagon, as in a honeycomb.'2 Thesurface of an icosahedron presents a problem inthis connection, for its covering must extend oververtices. However, the problem would be solved ifeach capsomere that was directly over a vertex wasa pentagon, for hexagons could then be fittedagainst each of their five sides and so the proteinshell would maintain an icosahedral form. Anillustration of this concept is depicted in Fig. 3e,in which the central white subunit is a pentagonforming one of the vertices of an icosahedron. Thisarrangement of hexagonal and pentagonal capso-meres seems to hold for the whole series of icosa-hedral viruses that have as yet been identified(Fig. 4); the simplest virus of this series, bacterio-phage .x174 (Fig. 3a), has 12 capsomeres andthe largest number of subunits is found on tipulairidescent virus which has 812 of them.Two variations of the arrangement of hexagons

and pentagons described above should now beconsidered. The first relates to how the form of avirus of this type would be affected if for somereason pentagonal capsomeres were not present intheir proper numbers and as a consequence hexa-gons were fitted together more extensively than in

Fig. 3.-This plate shows examples of negatively stainedcubic viruses all at a magnification of x 300,000. 3a.-Bacteriophage /.yl74 has the simplest arrangement ofsubunits of viruses showing cubic symmetry. It has 12subunits, one placed at each vertex of an icosahedron.3b.-Turnip yellow mosaic is the only virus known. atpresent that belongs to the series having a subunit placedcentrally on each of the 20 triangular faces of an icosa-hedron; these together with the 12 subunits on the verticesof the icosahedron give a total of 32. 3c.-While Fig. 3bshows the normal negative-stained appearance of the turnipyellow mosaic, this Figure (3c) has been photographicallyreversed to give greater prominence to the subunits.(Courtesy of Dr. H. E. Huxley.) 3d-A group of negativelystained wart-virus particles. Wart virus belongs to thepapova group of viruses and there is at present controversyas to whether this group has 42 or 92 subunits. 3e.-A modelconstructed of hexagons and pentagons according to theplan shown in Fig. 2. There are 30 white hexagons and 12white pentagons representing the arrangement of capso-meres that would be present on a virus with 42 subunits.This may well be the arrangement for wart virus, shown inFig. 3d of this plate. 3f.-A tubular form of polyoma virusshowing what happens when an icosahedral virus is in-capable of forming "pentagons". Also of interest are thesubunits that have detached from the top of the tubule asthey remain in a hexagonal form. 3g.-The arrow indicatesa particle of Adenovirus Type 12 that clearly exhibits theicosahedral shape of the virus. By counting the number ofsubunits lying between two vertices it is possible to cal-culate the number of subunits composing the protein shell.3h.-Varicella virus, an example of the compound cubicgroup of viruses. The geometrically arranged capsid is seenless clearly as it is covered by an outer fringed membranederived from the cell.

Canad. Med. Ass. J. ALMEIDA: CLASSIFICATICOct. 19, 1963, vol. 89

)N OF VIRUS PARTICLES 791

Fig. 3

792 ALMEIDA: CLASSIFICATION OF VIRUS PARTICLES Canad. Med. Ass. J.Oct. 19, 1963, vol. 89

the usual particle. Without pentagons to coververtices the particles could be visualized as con-tinuing to increase in some direction by addingmore hexagons in the region where the pentagonswere missing. It is possible that this is the explana-tion for the fact that certain viruses13 that areordinarily icosahedral sometimes manifest tubularforms, as shown in Fig. 3f.The other variation is in the arrangements of

hexagons alone. As pointed out by Caspar andKing,9 only two arrangements are geometricallypossible. Basically they are distinguished by thepresence or absence of a hexagon placed centrallyon each triangular facet of the icosahedron (Fig.4). So far, only one virus, turnip yellow mosaic(Fig. 3b), has been shown to have hexagons placedcentrally on the triangular facets.14 Each of thetwo arrangements described gives rise to a series(Fig. 4) depending on the number of subunits be-tween any two vertices, and the total number ofsubunits covering the particle can be calculatedby the use of two simple formulae:

(a) 10 ( n-i ) 2 + 2, for the more common serieswhich have no central subunits on theirtriangular facets,

and(b) 30( n-i ) 2 + 2, for the series with centrally

placed subunits,where n = the number of subunits on one side ofa triangle. For practical purposes this means thatif it is possible to count the number of subunitsbetween any two vertices, or, as it is more usuallydescribed, five-fold axes, then it is possible tocalculate the number of subunits or capsomeres onthe protein shell or capsid. It may be of interest tostudy the particle of adenovirus shown in Fig. 3g.The icosahedral form of the virus is very clearlyshown, and it should be possible for the reader tocount the number of subunits lying betweenvertices and hence estimate the number of capso-meres on the virus.

Since the morphological subunits, the capso-meres, of virus particles with cubic symmetry areeither hexagons or pentagons, they are not, ofcourse, identical. Accordingly if Crick and Watson'sprediction8 of identical protein subunits identicallyarranged is to hold, there must be smaller proteinsubunits than capsomeres and those "hexagons" thathave broken off from the tube form in Fig. 3f doappear to be made up of even smaller units. These,it is suggested, could be the corner posts of thehexagons and pentagons. In Fig. 4 the hexagonsand pentagons are shown with knobs at each cornerand it is apparent that if these are the basic uniteach subunit is now arranged identically with thosearound it. Each basic or structural unit is shownas having two strong bonds and one weak bond(Fig. 4); this concept of strong and weak bondsis based on the observation that, if disrupted, thesubunits remain either in the hexagonal or penta-

v§-Qcz

.2K.

I>4Thr.0Fig. 4

Fig. 4.-A diagrammatic representation of the close-packedarray of hexagons and pentagons on the surface of virusparticles having cubic symmetry. The series 12, 42, 92,- - - - - 812 is the one that has most frequently been found.The diagram shows the arrangement of subunits on onlyone triangular facet of the icosahedron, but the relationshipof the triangular facet to the whole icosahedron can be seenby comparing this to Fig. 2. It should be noted that "penta-gons" are present only at vertices. The basic structural unitof these viruses is represented by the corner posts of thehexagons and pentagons and each one of these is identicallyarranged with regard to every other one.

gonal form as is shown in Fig. 3f. In vie.v of theseobservations and this reasoning it is feasible toconsider that the capsid of viruses showing cubicsymmetry need contain only one basic type of pro-tein subunit, that represented in effect by thecorner posts of the pentagons or hexagons.

Viruses such as those we have been discussingare designated as simple cubic viruses and consistsimply of nucleic acid and protein. Some otherviruses showing cubic symmetry, however, add an

Fig. 5.-This plate illustrates negatively stained helicalviruses all at a magnification of X 300,000. Sa-Negativelystained rods of tobacco mosaic virus. The central hole of thehelix is clearly seen but the spacing between the turns ofthe helix 23 A' is such that the actual helix is not resolvedin untreated virus. 5b.-A particle of Parainfiuenza III.This virus belongs to the compound helical group andclosely resembles measles and mumps virus. The outermembrane is covered with fine filamentous projections andat one point, indicated by an arrow, a short length of theinternal helical component of diameter 170 A' can be seen.Although it is not completely established, the helix ofviruses of this type would seem to be a single one. 5c. Asmall portion of the helical component from a completelydisrupted measles virus particle. Sd-A second type ofcytoplasmic compound helical virus, influenza. Three of thepleomorphic particles can be seen in the upper part of themicrograph. In contrast to the type of virus shown in Fig.6b influenza has coarser projections on the outer membraneand does not disrupt spontaneously, having to be treatedwith a solvent such as ether before the 90 A' diam. internalhelix can be seen. Se. Internal component of rabies virus.The outer membrane of the complete virtis is similar tothat shown for p"r.i.nfluenza i. Fig Ib: however in thiscase the internal helix has a 90 A' diameter and is doublein nature. The arrows indicate two points where the coni-plete helix can be seen to divide into two separate filaments.

ALMEIDA: CLASSIFICATION OF Vuius P.urncu.s 793

Fig. B

794 ALMEIDA: CLASSIFICATION OF VIRUS PARTICLES

outer membrane at some Stage of their develop-ment and these we shall call compound cubicviruses (Fig. 3h).

VIRUSES HAVING hELICAL SYMMETRY

A second and very important group of virusescan now be distinguished because they possess amorphological character in common, that of havinghelical symmetry. One of the first studied and bestknown of these is tobacco mosaic virus, whichappears rod-shaped both by shadow-casting (Fig.la) and by negative staining (Fig. 5a). Althoughthe virus appears rod-shaped by these methods,it actually is in the form of a helix; this has beenshown in the electron microscope by the fact thatif the virus is completely degraded the proteincomponent can be reconstituted, and when this isdone the helical nature of the rod becomes apparentbecause the helix in this instance is not so tightlywound.'5 Caspar's diagrammatic representation ofthe substructure of tobacco mosaic virus is shownin Fig. 6. This shows that the "rod" is composedof identical asymmetric protein subunits arrangedin a helix.9 In this arrangement each subunit exceptfor one at each end is identically related to everyother subunit. The ribonucleic acid of the viruslies in an inner groove and the efficiency of theprotective protein shell is shown by the resistanceof the virus to ribonuclease and by the fact thatthe virus is more heat-stable than RNA isolatedfrom the virus.9 Many plant viruses are similar inform to tobacco mosaic virus but differ from it inhelix diameter and in some instances in flexibility.4All these are designated as simple helical viruses.Some of the more complex animal viruses also

exhibit helical symmetry. The difference betweenthese and the simple ones described above is thatalthough, like the simple helical viruses, theirnucleic acid protein complex is in the form of ahelix, they have in addition a surrounding mem-branous sac that is derived from the cell that thevirus infected.10 Viruses in this category we termcompound helical viruses.

There are three groups of compound helicalviruses known at present. These are distinguishedfrom one another by (1) whether the helix issingle or double, (2) the diameter of the helix,and (3) the nature of the projections of the outermembrane of the particle. Examples of all threeare shown in Figs. 5a-e.

COMPLEX VIRUSES

On morphological grounds there is a third groupof viruses that present either a more complicatedor less well understood structure than those thatare classed as having cubic or helical symmetry. Atpresent these are classified as complex viruses.However, even in this group elements of eithercubic or helical symmetry, or both, are present.

First to be considered here are the sophisticatedtailed phages. In the T2 phage the head is de-

Canad. Med. Ass. J.Oct. 19, 1963, vol. 89

Fig. 6

Fig. 6. A diagramn.atic representation of tobacco mosaicvirus. The asymmetric subunits are arranged identicallywith each other. The nucleic acid is shown as the smallerhelix contain.d in a gronvo of th' structural protein units.(Courtesy of Dr. D. L. D. Caspar.)

scribed as a hexagonal prism with two hexagonalpyramids while the tail shows helical symmetry(Fig. 7a). The head of the B. megatheriurn phageis icosahedral.4' '.

Pox viruses with negative staining exhibit anappearance similar to an untidy ball of wool (Fig.7b), and as yet no definite symmetry propertie5have been resolved in them, although Home et al.3have shown a helical structure in orf virus whichbelongs to this group,'8 and more recently Fried-man-Kien, Rowe and Banfield'9 have shown asimilar helical structure in a pox virus isolatedfrom milker's nodules.A third complex virus is vesicular stomatitis virus,

the only known asymmetric animal virus.2.. Thisvirus appears to contain some kind of internalhelical arrangement but is sufficiently differcnt inother respects to warrant its being placed in thecomplex category (Fig. 7c).

Fig. 7.-This plate comprises examples of negatively stainedviruses that have been described as having complex morph-ology. All are shown at a magnification of X 300,000. 7a.-Micrograph of T4 bacteriophage. The lower particle showsthe normal appearance of the virus. The helical nature ofthe tail sheath is clearly visible. In the virus which ispartially shown in the upper part of the micrograph thesheath is contracted and the inner stem which is used to"inject" the nucleic acid content of the head into thebacterium revealed. 7b.-A negatively stsined vaccinia virusparticle still contained within a vesicle in the cell. Nodefinite symmetry is associated with vaccinia virus althoughother members of the pox group have been shown to havea helkal arrangement. 7c.-At the upper and lower rightof the micrograph, mature, bullet-shaped particles ofvesicular stomatitis virus are seen. The surface of thevirus is covered by fine filamentous projections. The incom-plete budding form of the virus shows an internal helicalarrangement.

ALMEIDA: CLASSIFICATION OF VIRUS PARTICLES 795

.L9Fig. 7

796 ALMEIDA: CLASSIFICATION OF Vnws PARTICLES Canad. Med. Ass. 3.Oct 19, 1963, vol. 89

aK

£1

'rn-a4(

IJ\

.0

'C4)420

0.

0)0)42'C

4)C..4)4)54

00.4)

4)4260

N42.-.'C42.0)

60.0c054$.,

0Co054.

S4)4260.

4)

420-

S04)00C

I.0

42

0

I*54K>'

0a

H42

544$550

Canad. Med. Ass. . ALMEIDA: CLASSIFICATION OF Vn.us PARTICLES 797Oct. 19, 1963, vol. 89

9,9,.(J)

COHOO*i4C0.CObCHH.c. *H Q)C) H H 4-)VC.HC . CU)

U) .4-) CO.U).C) ..4 C.' 00 o.m o

o c.U2

U) 0.-40-4-' .1-i 9,:). U).

H.. . H (p

O.0 H.CO

-.

.-.- .6) . O.. * *H..C.d. .'HH.

. . E-CH'-'+.. (VdU) 0 Z

9, U)9, U) * 6)H 6)0 U)

U). CH r 0 C CZ V ."-4 *H*H. .fl .

U) . C O.006) CC) HO 9, 0 .0.o CC.C *r-4'd d.6I +.lfr. .O0 U) Cc. C. CI.C .6)'H 'd _

QIH CO6)C"-\I 01 C.V.HCOU).p-( .U) CO'.1U) rl-PH6)6)H CO 06) HV CCCC. H.H.Q +. . 0 c.54.A.I.S *..V.3c.jI cii C.. 0 0. CO

.m

21 2CO6) 0H .

Cu 'VC) 6)cucd .H C

C) H .CO*H C 6) HC.

00 ___________________'V

-. C .

6) . Ce-PHO 0 OC)C HOF.0 . OCO'H C 0 CCCl) 9, C 0 C-H

.0C) .

H CO.Q

r-I9,, CO CC -P0 . C C)

. 6)

_________ C *.q. .0 CO S 9,*H cu

6) .H CO H 6)U) 6) C S. H.LH6)6) 'V C 6)CO6)6)'V

iHCfC .C) COr-I 'H 0 ..O5.6) a .o(pcu..r-

798 ALMEIDA: CLASSIFICATION OF VIRUS PARTICLES Canad. Med. Ass. J.Oct. 19, 1963, vol. 89

CLASSIFICATION

In the light of the foregoing we are now in aposition to consider how information derived froma study of viruses with the electron microscopecould serve to provide the basis for a useful classifi-cation of viruses. First, the study of thin sectionspermits viruses to be classified as to whether or notthey multiply in the nucleus or the cytoplasm ofthe cell. Secondly, negative staining permits theirfine structure to be determined and hence theirallocation into categories which relate to the typeof symmetry they manifest. By combining thesetwo criteria we then can arrange the viruses into aclassification as shown in Table I * and illustratedin Fig. 8.The organization of the proposed classification

shown in Table I and Fig. 8 is therefore as follows.The first distinction made is whether or notany particular virus multiplies in the nucleus orcytoplasm of the cell. This gives rise to two maindivisions. However, since the bacterial viruses can-not be classified in this way they have to be set ina special category, so that there are three maincategories in all.

Next, within each of these three categories themembers are further classified as to whether theymanifest cubic, helical, or complex symmetry. Thenext distinction that can be made relates to thosecategories where there are viruses that, althoughmanifesting the same type of symmetry, differbecause some are surrounded by an outer mem-brane and some are not; this gives rise to thedistinction that permits them to be classified ascompound or simple. Still further characterizationcan be provided by the number of morphologicalsubunits present in the protein shell of thoseviruses having cubic symmetry. The viruses havinghelical symmetry can also be classified still furtherby the nature of the helix and, in the instance ofthe compound helical viruses, by the morphologicalfeatures of the surrounding membrane.

It might now be asked whether groups ofviruses that have morphological features in com-mon have any other properties which they alsoshare in common. In other words, now that somuch is known about the structure of viruses, itmight be asked if there is any relation betweencertain structural features and biochemical proper-ties. In this connection it is of interest that a recentclassification of Hamparian, Hilleman and Ketler,21based on biochemical and physical properties ofviruses, is arranged into groupings very similar tothose shown here; for example, the group describedin this classification as simple cubic nuclear be-comes in its entirety the acid- and ether-stableDNA group of their classification. Compound cubicnuclear of this classification matches the acid- andether-labile DNA group. Simple cubic cytoplasmicmatches their heat-labile, ether- and acid-stable

*Table I is by no means exhaustive and lists only some ofthe commoner viruses.

RNA group. Compound helical cytoplasmic virusesin general match their acid- and ether-labile RNAgroup. In no case did a group placed together onmorphological grounds disagree with groups placedtogether on other bases and physical behaviour.

It is also of interest that when grouped in themanner proposed it becomes apparent that tumour-inducing activity is associated with two maingroups of viruses, the simple cubic nuclear andthe compound helical cytoplasmic. This will be dis-cussed in somewhat more detail in the accompany-ing article by McLeod and Ham.22

If it is true, then, that an arrangement based onmorphology can give us information about otheraspects of a virus or even cut down the range ofpossibilities to be tested, then a morphologicalbasis of classification has much to offer from apurely practical point of view. Since thin sectioningis a standa.'d technique wherever there are electronmicroscopes, and negative staining of the typementioned here is a straightforward techniquewhich makes virus purification unnecessary, theposition in a morphological classification of anyvirus that is under study can be readily determined.

I wish to thank the following for allowing me to repro-duce in the pictorial classification micrographs of viruseswith which they have worked:

Reovirus: Dr. S. Dales, Rockefeller Institute, N.Y.;Bittner: Dr. L. Dmochowski, Houston, Texas; Rous: Dr.R. R. Dourmashkin, London, England; T.Y.M.V.: Dr. H.E. Huxley, Camhridge, England; and Insect Nuclear Poly-hedrosis: Drs, Kenneth M. Smith and C. J. Hills, Cam-bridge, England.

Also, in alphabetical order, I would like to thank Mrs.NI. Betlem for technical assistance, Miss N. Boyd forsecretarial assistance, Mrs. P. Chibac for general assistance,Mr. R. S. Gilder for photographic and art work, Mr. L.Pinteric for helpful d:scussion and Dr. M. C. Williams forthe provision of viruses associated with skin lesions.

Most of all I would like to thank Dr. Arthur W. Hamfor encouragement in this undertaking and for allowingme the benefit of his extensive literary experience in pre-paring the manuscript.

REFERENCES

1. XVILLIAMS, R. C.: Advoaces Vii ts Res., 2: 183, 1954.2. BRENNER, S. AND HORNE, R. XV.: Biochim. Riophys. Acta,

34: 103, 1959.3. HORNE, R. XV. et al.: J. Molec. Biol., 1: 84, 1959.4. LXVOFF, A.. HORNE, H. AND ToURNIER, P.: cold Spy. Harb.

Symp. Quont. Bzol., 27: 51, 1962.3. ALMEIDA, J. D. AND HOWATSON, A. F.: J. cell Riol., 16:

616, 1963.6. PARSONS, D. F.: Ibid., 16: 620, 1963.7. ALMEIDA, J. D. et al.: Virology, 18: 147, 1962.8. Crnci., F. H. C. AND WATSON, J. D.: Nature (Londo ),

177: 473, 1956.9. CASPAR. D. L. D. AND KLUG, A.: cold Spr. Harb. Symp.

Qitant. Biol., 27: 1, 1962.10. XX.ILLIA.Is. H. c. AND SMITI-r, K. MI.: Riochim. Biophys.

Ada, 28: 464, 1958.11. KLUO, A. AND CASPER, D. L. D.: Adronces Virus Res.,

7: 225, 1960.12. COXETER, H. S. M.: Trans. Acad. Sci., 24: 320, 1962.13. HORNE, H. XV. AND WILoY, P.: Virology, 15: 348. 1961.14. HUXLEY, H. B. AND ZUBAY, G.: J. Molec. Biol., 2: 189,

19(0.15. NIxoN, H. L. AND XVooDs, H. D.: Virology, 10: 157, 1960.16. HORNE, R W. et al.: Ibid., 11: 79, 1960.17. HORNE H. IV. Sci, Amer., 208: 48, 1963.18. NAGINGTON, J. AND HORNE, H. XV.: Virology, 16: 248, 1962.19. FRI.MAN-KIEN, A. E., ROWE. XV. P. AND BANFII:I.D, XX..

(1.: Science, 140: 1335, 1963.20. HOWAYSON, A. F. AND XVHITMORE, G. F.: Virology, 16:

466, 1962.21. HAMPARIAN, V. V., HILLEMAN, i\1. H. AND KETLER, A.:

Proc. Soc. Exp. Biol. Med., 112: 1040, 1963.22. MCLEOD, D. L. AND HAM, A. XV.: can.ad. Med. Ass. J., 89:

799, 1963.