On the Theory of the Polyedra .
" By the Rev. T. P. KIRKMAN , M.A. , P.R.S. , and Hon. Member of the Literary and Philosophical Society of Manchester .
Received May 10 , 1861 .
( Abstract .
) The complete resolution of the problem of the polyedra embraces the construction of all P-edra Q-acra , with an account of the symmetry of the solids when symmetrical .
Such construction being impracticable from the magnitude of the task , it is desirable that a method should be found of classifying and enumerating the P-edra Q-acra , so that from this knowledge of the inferior polyedra , the same can be obtained concerning the higher , without any constructions , and without any tentative process .
I have found that all attempts to enumerate a separate and welldefined family of the P-edra Q-acra , beyond that I have called The partitions of the R-pyramid ' ( Phil. Trans. 1858 ) , have been fruitless , and that the simplest method of solving the problem is to solve it entirely .
It is necessary first to have an accurate classification of polyedra as to symmetry .
This is1 .
Zoned symmetry ; 2 .
Zoneless axial symmetry ; 3 .
Mixed symmetry , both 1 and 2 ; 4 .
Neuter symmetry , neither 1 nor 2 ; 5 .
Asymmetrical .
1 .
Zoned symmetry.-A polyedron may have zones of one , of two , or of three configurations .
A zone divides the solid , when it is constructed with the greatest possible symmetry , into halves , of which one is the reflected image of the other .
Zoned polyedra are,1 .
Monozones ; 2 .
nm-zoned monaxines ; 3 .
Zoned triaxines ; 4 .
Zoned monarchaxines , having secondary axes ; 5 .
Zoned polyarchaxines , which have the axial systems of the regular polyedra .
The intersection of two zones is a zoned axis .
The zonal signature gives an account of the number of zonal summit faces and edges , whether polar or non-polar ; but not of the number of edges in the zoned features .
2 .
Zoneless symmetry.-This is m-ple repetition of configuration , in revolution about a zoneless axis .
An axis isjanal if two opposite eyes can read at the poles configurations C C ' , which are either identical , or one the reflected image of the other ; otherwise the axis is heteroid .
A zoned janal axis is objanal when C is C ' turned through two right angles .
A zoneless janal axis is contrajanal when C is the reflected image of C ' .
The polyedra of this symmetry are,1 .
Zoneless heteroid monaxines ; 2 .
Zoneless monaxine contrajanals ; 3 .
Zoneless triaxines ; 4 .
Zoneless monarchaxine janals , having secondary axes ; 5 .
Zoneless polyarchaxines , having the axial systems of the regular polyedra .
3 .
Mixed symmetry .
In this there is both a zone or zones and one or more zoneless axes .
The solids are,1 .
Monozone monaxines ; 2 .
m-zoned homozones , having m zoneless axes .
A zoned janal axis is hiomozone when the solid has zones of one configuration only ; otherwise the janal zoned axis is heterozone , i. e. the solid has zones of more than one configuration .
4 .
Neuter symmetry.-There is neither zone nor zoneless axis , but the edges of the solid fall into pairs ab , atb ' diametrically opposite , of which edges one is the reflected image of the other .
The polyedra of this symmetry are contrajanalanaxine polyedra .
5 .
The asymmetric polyedra are the most numerous of all .
The data required for the enumeration of the P-edra Q-acra are the following : A. A table of the hexarchaxine , tetrarchaxine , and triarchaxine P-edra Q-acra and Q-edra P-acra , both zoned and zoneless , with all the zonal signatures , and an account of the terminations of their principal axes , and with the zonoid signatures .
The zonoid signature in a zoneless or mixed symmetry gives the number of zoneless polar faces , summits , and edges .
B. A table of janal poles , containing an exact account of every polar feature terminating a janal axis , with the repetition if it be zoneless , and with the traces and zones if zoned .
C. A table of poles , containing an exact account of every polar feature terminating either a janal or a heteroid axis , with zonal signatures and traces when they exist , or with the zoneless repetition .
D. A table of faces and summits having a diagonal trace of a single zone , with the zonal signature .
E. A table of objanal monozone faces and slimmits , with their zonal and zonoid signatures .
F. A table of contrajanal anaxine pairs of edges , with the faces AB intersecting in them .
G. A table of all edges ( A B ) with their intersecting faces , the polar edges with their zones , the monozone zonal and epizonal edges each with its zone .
An epizonal edge is cut by a zone .
These data being first obtained for the P-edra Q-acra and for the Q-edra P-acra , the numbers of the following solids and faces and summits are exactly known for both P-edra Q-acra and Q-edra P , acra ; and we suppose that all the inferior polyedra are registered in like manner:a .
The m-zoned monarchaxines with their zonal signatures , and their principal poles ; b. The zoned triaxines with their zonal signatures ; c. The m-zoned homozones with their zonal and zonoid signatures , and their exact principal poles ; d. The zoneless monarchaxine janals with their repetition and zonoid signatures , and their principal poles e. The zoneless triaxines ; f. The monozone monaxines with their zonal signatures and their poles ; g. The zoneless monaxine contrajanals with their repetition and their poles ; h. The m-zoned monaxines with their zonal signatures and poles ; i. The monozones , with their zonal signatures ; j. The zoneless monaxine heteroids with their axes and repetition ; k. The contrajanal anaxine polyedra ; 1 .
The asymmetric polyedra ; m. The monozone A-gons and A-aces with their traces and zonal signatures ; n. The objanal monozone A-gons and A-aces with their zonal and zonoid signatures ; o. The janal anaxine A-gons and A-aces ; p. The asymmetric A-gons and A-aces .
And these numbers being registered along with the data preceding for all signatures , we have a complete classification and enumeration of the P-edra Q-acra and Q-edra P-acra , which can be continued to any values of P and Q. The difficulty lies in the obtaining of the data ABCDEF G. We begin with ( C , D , and G ) .
Analy/ sis of a polar or monozone sutmmit of a P-edron Q acron V. Let pt be a polar or monozone p-ace , which , if polar , is either a zoned or zoneless termination of an axis .
If 2-zoned or of r-ple zoneless repetition , there is a sequence of configuration r times read about the pole .
Any summit m through which lies the triangular section pnmn of the solid is a deltotomous summit about p ' ; and if mn be an edge , it is a deltotomous edge , about p ' , or of p. If the deltotomous edge mn be in two facesff ' , of which f has no deltotomous summits about pt except mn , while f ' either has more than two deltotomous summits about pt , or is a triangle mnr having r collateral with p ' , mn is a primary effaceable of pt .
If the deltotomous edge mnz be in two faces ff , of which neither has any deltotomous summits about pT besides mn , nor is a triangle mnr having re collateral with pt , mn is a secondary e , aceable of p. The sumimit p ' has e ( &gt; " 0 ) effaced effacebles .
: These can be restored in one way only , since each ( nn ) must complete a face f which has no deltotomous summit besides inn .
Let them be restored .
We have the perfect suzmmit p ( i. e. a summit which has no effaced effaceables ) of a ( P + e)-edron Q-acron VI , which has E~e effaceables .
The process whereby the data C , D are obtained is the construction in groups of all perfect polar p-aces and monozone p-aces having a diagonal zonal trace , of all the ( P+e)-edra Q-acra ( V ) , which can be reduced by effacement of e effaceables to a polar or monozone p-ace of a P-edron Q-acron V. These polyedra V , can be all constructed about their perfect poles and registered ; for no result of effacement is small enough to be employed in the construction of a polyedron ( V , ) .
Each perfect p-ace p ' constructed has E(~e ) effaceables , which are registered with the signatures of symmetry ; so that all results of effacing e(O~ ) of the E effaceables about any registered perfect p-ace are exactly known without repetition , and without enumeration of two summits , of which one is the reflected image of the other .
The process of this construction of perfect p-aces pf is the converse of the 7reduction ofp ' In the reduction of p ' , we remove all the rays of p ' , whereby we lay bare either a polyedron or a reticulation , which has among its linear sections the E effaceables ofp ' .
If the effaceables of the reticulation are all secondary , it is a full reticulation , which is an agglutination of polyedra cohering by those secondary effaceables , which are linear sections , and the only linear sections , of the reticulation .
If the reticulation has no effaceables , it is a plane reticulation , being simply a polygon partitioned by certain diagonals .
If there be one or more primary effaceables , we have a mixed 1etieclation .
The mixed reticulation always reduces:in one way only by sections in its external primary effaceables , to a subject reticulation , which is either mixed or full or plane , or else to a polyedron .
And , by continuing this reduction , we always obtain finally a full reticulation , a plane reticulation , or a polyedron , " 2-zoned , or of ri-ple zoneless repetition .
, The primary effaceables of a mixed reticulation are the joints or seams by which it can be either talken to pieces or constructed , in one way only .
And the portions into which it breaks up by these joints , are either polyedra , or full reticulations , or plane reticulations , of which we have an account in our register .
We first suppose given all plane and full reticulations , as well as the inferior polyedra , with all their signatures .
By means of these , imposed as marginal charges on a subject reticulation , we can construct and register exactly all polar and monozone mixed reticulations , with their repetition and zonal signatures , and with their effaceables , which are always edges that have been loaded with marginal charges .
Every reticulation , plane , full , or mixed , is registered with its maryinal signature .
The marginal signature exhibits to the eye all that is requisite to be known either for the coronation of the reticulation by a p-ace p , whereby it becomes a polyedron , or for the construction on it of higher reticulations .
The marginal signature to be constructed can always be exactly written ; and there is a given number of ways of constructing it on the subject signature , which can be registered , and the number is obtained by inspection of our signatures .
We always see in a mixed marginal signature what is rernoveable by sections in the external primary effaceables , and in a full reticulation we see what is removeable by external secondary effaceables .
The rules for charging a subject marginal signature , so as to construct another signature upon it , are:1 .
No primary effaceable which is external in the subject is to be external in the constructed ; therefore the subject compartment standing on every external primary effaceable must receive at least one charge , which charge will be a compartment of the new signature .
2 .
Solid charges ( polyedra or full reticulations ) are imposed on plane marginal triangles of the subject .
Plane reticulations , which give plane compartments , are imposed on solid marginal edges of the subject by a marginal triangle of the charge , which is lost in the operation , or supposed to be cut away .
Thus a marginal triangle is lost by every charge imposed , whether plane or solid .
We handle all plane reticulations only by their marginal triangles , given by their signatures , and thus we escape the necessity of keeping any account of their summits .
3 .
A plane submargin of the subject is never charged .
4 .
Every charge adds a primary effaceable in a mixed reticulation , and may add any number of secondary effaceables , if a solid charge .
5 .
If the penesolid 402 , which consists only of two marginal triangles , is employed to form a compartment of the new signature ( which is the compartment ( 2 ) , or ( 2 ) r times/ repeated in the r-ple marginal signature ) , the resulting reticulation is closed to further marginal increments , and can only be crowned .
The zonal and zonoid signatures of reticulations give an account of all zoned and polar effaceables , so that after coronation of a reticulation we can always enumerate the results of effacement , with their symmetry and signatures .
Any effaceable can be effaced independently of others .
Asymmetric mixed reticulations are never constructed , unless such as can be crowned by a line ; we thus escape the construction of the most numerous class of mixed reticulations .
And asymmetric coronations , or asymmetric results of effacement , are never registered , unless the coronation be by a line , or the effacements be about a perfect edge .
An edge is a 2-ace summit which has always e( 0 ) effaced effaceables that can be restored in one way only .
WThen they are restored we have a perfect 2-ace , which is to be constructed and registered , whether symmetrical or not .
The removal of an edge lays bare in all cases either a polyedron or a penesolid , which may be either plane , full , or mixed .
A penesolid is a reticulation which can be crowned with a line , thus becoming a polyedron .
The order of the marginal features is not regarded in the marginal signature of a reticulation , as we keep no exact account of the faces about the crowning p-ace ; but the marginal signature of a penesolid is always so written that the exact faces of the crowning line shall be known by inspection of the signature .
A penesolid may have any number of effaceables which are registered with the marginal and zonal or zonoid signatures .
The results of effacement about the crowning edges are the data C G , and these are exactly given by our methods .
The marginal signatures of penesolids are of the form [ 2a2b ] , [ 2a2b ] , [ 3a4c ] , [ 3aaAc ] , l[aA6b .
[ AaB .
l , &amp;c. The first has two marginal plane triangles and a+ -b submargins ; the second has one marginal plane triangle ; the third has none .
The fourth has a+ c plane szutbargins and a solid submargins .
We see that polyedra are removeable by the external primary effaceables in the third , which have a ( 3+ 1)-gonal and a ( 4+ 1)-gonal face ; and that the fourth has been constructed by charging the plane marginal triangles of a subject penesolid , one with a polyedron having a 3+1-gonal face , and the other either with a polyedron having an ( A+ l)-gonal face , or with a full penesolid having the solid margin A. The edge that crowns the first is the intersection of an ( a+3)-gon and a ( b+3)-gon .
The edge that crowns the third may be the intersection of an ( a +3)-gon and a ( c+ 6)-gon , or of an ( a+4)-gon and a ( c-.5)-gon , &amp;c. The polar summits being known for P.edra Q-acra and Q-edra P-acra , their reciprocal polar faces are of course known also .
This may suffice as an account of the data ( C , D , G ) , except that we shall return to the construction of the plane and of the full reticulations .
We consider next the data B. Analysis of the polara janal sunmits p'p , of a P-edron Q-acron W. Let p ' and p , be 7r-aces terminating a janal axis .
The ordinary effaceables of either summit , which are what we have already considered about a polar p-ace , can be restored in one way only .
The janal symmetry remains , though modified by such restoration .
There are in faces about p ' a certain number of triaces collateral withp , .
These are ( w^ 0 ) rhomlbotomouts triaces .
Let each be made collateral with p ' , and let the w similar triaces in faces about p , be made collateral with p ' .
They are now all rhiombotonmous tessaraces , sueyb hat quadrilaterals p1 p p , p , are dlrawn through pairs p p , of the tessaraces .
We have now a pair of janal r'-aces perfect both in their ordinary and their peripolar efftceables , which is registered in our table of perfect janal poles .
A peripolar effaceable is any ray drawn from a principal janal pole to a rhombotomous tessarace , which is always collateral with either pole , and may lose either of its effaceables .
If the two polar summits be entirely removed , there is laid bare either a janal polyedron or a janal reticulation at either pole .
The two opposite reticulations have the same marginal signature , which differs in nothing that needs here be noticed from that of a polar reticulation .
This is reducible at either pole by the external primary effaceables , or if it be a janal full reticulation , by its external secondary effaceables , either to a janal polyedron , or to a janal full reticulation , or to aprimitive , or to afundamentaljanal reticulation ; and this reduction is possible in one way only .
A fundamentaljanal reticulation is made thus:-Let a primnary polar plane reticulation , i. e. one which has no diagonals but the bases of its marginal triangles , and no summits besides those of those triangles , be laid on a mirror , with the triangles a little raised from the mirror .
The reticulation and its image form afundamental janal reticulation .
If the primary be turned in any way about an axis perpendicular to the mirror , while the image remains stationary , we have a primitive janal reticulation .
The edges in the mirror , common to both polar faces , are the submargins of the polar faces , on which primitive submargins rhombotomous tessaraces can be planted at pleasure in janal coronation .
The marginal triangles of the fundamental are all doubled , and the bases of the pairs are zonal effaceables .
Some only of the marginal triangles of the primitive are doubled , and the bases of the pairs are zonoid effaceables .
There is only one way in which a given janal reticulation can be ' constructed on a fundamental or on a primitive .
The rules for construction differ little from the rules for polar reticulations , and we conceive that the marginal charges are imposed alike in the opposite polar faces , thus preserving janal or contrajanal symmetry .
Inspection of our signatures gives us the result in all cases .
The janal reticulation is registered with its marginal signature , which of course is the same at either pole , with its zonal signatures , principal and secondary , and with its zonoid signature if it have secondary zoneless poles .
An account is kept , in all constructions on a fundamental or primitive , of primitive plane suibaryins , because it is on these only that rhombotomous tessaraces can be deposited injanal coronation .
These are the submarginal edges ( not in triangles ) of the fundamental ol primitive .
The symmetry of a janal subject is modified in various ways by marginal charges , by coronation , or by effacements ordinary and peripolar .
The modifications are the following : On the r-zoned hetcrozone janal subject we construct results,1 .
-zoned heterozone , 2..-zoned homozone , 3 .
r-ple zoncless monarchaxine janal , 4 .
v2-ple monaxine moioozone , 5 .
--ple monaxine contrajanal .
On the " -zoned homozone subject we construct results,1 .
--zoned homozone , 2 .
v2-ple zoneless monarchaxine janal , 3 .
2-ple monaxine contrajanal .
On the ri-ple zoneless monarchaxine janal subject we construct only '-ple monarchaxine janal results .
On the r'i-ple monaxine monozone subject we construct results,1 .
r-plc monaxine monozone , 2 .
r-ple monaxinc contrajanal .
On the ri-ple monaxinc contrajanal subject we construct only r-ple monaxine contrajanal results .
In all these cases the constructions are given exactly by inspection of our signatures for every divisor r of ri : and we can both complete the table of perfect janal poles of ( P+e)-cdra Q-acra which we require , and register the janal results of e ordinary effacements in every possible way upon each of them , for every kind of symmetry .
IIaving thus registered the results of ordinary effacement about janal axes in a table of janal poles perfect only in peripolar effaceables , we have next to register the results of peripolar , ffacement .
The effect of a peripolar effacement is , that a rhombotomous tessarace becomes a rhombotomous triace , and that the secondary zone , if there is one , is destroyed , a different janal symmetry being introo duced .
Our signatures give us an exact account of all rhombotomous tessaraces introduced in coronation , and we can readily enumerate the results of peripolar effacement by mere inspection of those signatures .
All fundamental and primitive reticulations are given by general formula , as are also the primary plane reticulations on which they are founded , in terms of their marginal signatures , and of their zonal and zonoid signatures .
Every kind of janal symmetry may be seen in these fundamentals and primitives .
Thus , on the supposition , as before , that we have tables of inferior polyedra , and of the plane and full reticulations , symmetrical or not , we can obtain completely the data B for P-edra Q-acra , and for Q-edra P-acra .
The data E are what the 2-zoned homozone polyedra become when r=2 .
They are given by the theory of construction of the homozone poles for the general value of r ; and each summit is registered with its zonal and zonoid signature and with its effaceables ; and the results of effacement , ordinary or peripolar , are accurately known .
The data F are what the r-ple monaxine contrajanals become for r= 1 .
We never descend so low as -= 1 , except in the construction of penesolids , viz. contrajanal anaxine penesolids .
The janal anaxine pairs which crown these and other penesolids are data F. They are registered as perfect edges with their effaceables , and the janal results of effacement are also data F. The data A are easily obtained .
Every polyarchaxine reduces in one way only to a regular polyedron , on the principal faces of which it is constructed , always by its principal poles .
The effaced effaceables of a polyarchipolar summit ( which exists either on a polyarchaxine or on its reciprocal ) are all ordinary , and can be restored in one way only .
We conceive them restored about all like archipoles .
These poles being removed , a polyedron or a reticulation is laid bare , which has a marginal signature differing in nothing that needs here be noticed from that of a polar reticulation .
By inspection of this signature , we can construct on it a given number of polyarchaxine reticulations of given signatures , and the process differs in nothing from the construction of a polar reticulation , except that we conceive it effected in all the principal faces of the subject .
We can crown , and register all summits with zonal or zonoid and marginal signatures and with effaceables , so that the results of ( e ~ 0 ) symmetrical effacements about all the principal axes are readily enumerated .
These results are the data A. It remains only that we return to the construction of the full and plane reticulations , from which all our marginal charges , if they be not polyedra , are selected .
The full reticulations are reducible always either to a nucleus line or to a nucleus polyedron , by sections in their external effaceables .
The marginal signature shows always the edges removeable by such sections , and construction proceeds by the rule that no external effaceable of the subject shall be external in the result .
All effaceables are here secondary .
The modifications of symmetry of polar monozone and asymmetric subjects and constructions are expressed in general formulaw , and the results are always registered with all signatures without ambiguity or repetition .
We never construct janal full reticulations except what have , as the word janal implies , a symmetry : and such constructions are always polar , except the objanal monozones and the contrajanal anaxine full penesolids .
The only difference between these full constructions and that of mixed reticulations is , that no marginal triangles are handled or lost in the process ; and that in the building of an r-ple repetition on a subject of ri-ple repetition , we descend to the value r= 1 , which gives the asymmetric fiull reticulations by the formulefor the general value of r. The plane reticulations have lastly to be considered .
All plane r-gonal penesolids ( rOf ) havingf--(~ 2 ) diagonals , symmetrical or not , are given by general formulae in terms of their marginal signature , [ 2a2b ] , ( r=4-4 a+ 6 ) , that is , in terms of ( rfab ) .
The line which crowns this signature is the intersection of a ( 3 + a)-gon and a ( 3 + b)-gon .
All polar plane reticulations and all monozones which have less than three epizonal edges in the zone , reduce in one way only to a polar or monozone primary , which has no diagonals except the bases of its marginal triangles .
On this primary as subject the reticulations in question are constructed by one operation , by charging its marginal triangles with marginal triangles of inferior reticulations , whereby two marginal triangles disappear for every charge imposed .
The number and the symmetry of the constructions are always given by inspection of our signatures .
The monozones which have more than two epizonal edges in the zone , reduce by section in the central epizonal or epizonals to two or to three inferior zoned reticulations .
We construct , conversely , a given marginal and zonal signature thus in every possible way , the number of constructions being always given by inspection of our tables .
All polar reticulations having this zone and marginal signature will be formed by the process .
The polar being already registered , the monozones are obtained by subtraction .
Perhaps the greatest difficulty in the theory of the polyedra is the enumeration of the M asymmetric plane reticulations which have the marginal signature s= [ 2'w ] , where 2T means simply 2T ; and where T is the number of the margihal triangles , and w that of the submarginal edges , of which no two are contiguous .
The reticulations to be found will be registered ROF [ 2Tw]=M , where F-T=d is the number of diagonals which are not bases of marginal triangles , and R=2T+ s u , 2T+w is the number of the summits of the reticulation , of which s only are not summits in marginal triangles .
If we erase the d diagonals , and also the s diaces that they may leave ( summits of two edges ) , we have a primary reticulation , Rbf [ 2iTu ] , ( R'= 2T + ) , which may or may not be polar or monozone .
If now on the u submargins of RtOf we deposit 8 points ( diaces ) in every possible way , and then draw in the ( R-T=)(T+w=)(T+u+s )-gon enclosed by the T marginal triangles d diagonals in every possible way , none crossing another , so that one at least shall pass through each of the s points , we shall construct the asymmetric ROF1 in question among the results a certain number of times , namely , in every position of ROPF in which erasure of the d diagonals and of the s diaces will reproduce the primary I0( ) ; and if R/ Of be polar , 'there may be many such positions of 0OF for the same position of 1'Of .
And it is evident that we shall construct equally every plane reticulation R1O of every symmetry wliicll can reduce by the same process of cffacement to the same primary R'Of .
Nothing is easier than to determine the number of asymmetric constructions thus obtained of IROF [ 2Tt ] , if the number of all possible ways of drawing the d diagonals can be found .
We have to employ in turn every possible partition of the 8 points , of which one is S=a , +2+ a3 +* +a t ( , C u ) , ( au , + 1 ) , ( &gt; O 0 ) .
There is a given number of ways of depositing a , points on any one of the u submargins , a2 points on any other , &amp;c. All that is difficult is to determine , when a disposition of the a , + a , + ... +a points is made , in how many ways cl diagonals can be drawn in the ( T +w=)r-gon , so that one at least shall pass through each of the s points .
Let this number be rdq2 2 as 0,. . ?
It is given always by the equations following:-Let a1= 1 ; then , , , ... . .
= t , , ... -(- ) , , ... -(1 ) ( d-1 ) , , , ... .
Let an &gt; 1 ; then , ,d , laa , t3 ... =.d(l , ,-l),.a3 ... , -(.'-1 )(dl)( , ,_:-2 ) , ,,3. .
--(-l)d( , _-l ) T3 , ... , where drln(,.+d-1)1)( , -3 ) II(r-1 ) 11(d+ 1 ) I(,.-d-3 ) 1n(d ) is what I have called in my memoir " On the k-partitions of the It-gon and the l-ace " ( Phil. Trans. 1857 ) , the ( d+ 1)-divisions of the r-gon .
This number rda , za3 ... being thus given for every partition of s , the asymmetric plane reticulations ROF[2Tw ] are given , whatever be the primaries R'Of , symmetric or asymmetric , to which they are reducible by the above process of effacement .
And with these we have a complete solution , perfectly easy to calculate and register , of the problem of the classfieation and enumera .
tion of the P-edra Q-acra .
The memoir , which I have the honour to present to the Royal Society , contains all the general formule of this solution .
There is nothing to prevent our registering in all marginal signatures the exact order of the margins .
If this be done , we can crown every reticulation registered by a closed polygon A , made by connecting margins so that no linear section shall remain .
The faces collateral with A would be at the same time constructed , and could be registered ; that is , we could register all the faces collateral with any face in their order .
And in crowning by summits marginal signatures so registered , we could determine and register the faces of every p-ace of the P-edra Q-acra in their order .
The methods above given are applicable to this more laborious registration of results .
If this more tedious process be adopted , the construction of the P-edra Q-acra will be an easy matter .
I am not aware that anything has been printed on the subject of this theory beyond what I have attempted in the 'Philosophical Transactions , ' and in the ' Memoirs of the Literary and Philosophical Society of Manchester , ' except a short attempt made some three years ago by M. Poinsot to sketch a beginning of the investigation , which appeared in the 'Comptes Rendus .
' The attempt was well made , but the results given were not quite accurate .
I have it not at hand ; but I know that there is a defective enumeration of the simple solids there considered .
I have enumerated the polyedra of not more than 18 edges by this method , and I hope shortly to publish the classification and enumeration of the polyedra of 20 edges and under .

