]\gt ; IV .
A Memoir on Abstract Geometry .
By Professor CAYLEY , F.R.S. Received October 14 , \mdash ; Read December 16 , 1869 .
I to the Society the present exposition of some of the elementary principles of an Abstract -dimensional Geometry .
The science presents itself in two ways , \mdash ; as a legitimate extension of the ordinary two-and three-dimensional geometries ; and as a need in these geometries and in analysis generally .
In fact whenever we are concerned with quantities connected together in any manner , and which are , or are considered as variable or determinable , then the mature of the relation between the quantities is frequently rendered more intelligible by regarding them ( if only two or three in number ) as the coordinates of a point in a plane or in space : for mole than three quantities there is , from the greater complexity of the case , the greater need of such a representation ; but this can only be obtained by means of the notion of a space of the proper dimensionality ; and to use such representation , we the geometry of such space .
An important instance in plane geometry has actually presented itself in the question of the determination of the number of the curves which satisfy given conditions : the conditions imply relations between the coefficients in the equation of the curve ; and for the bettel understandin of these relations it was expedient to consider the coefficients as the coordinates of a point in a space of the proper dimensionality .
A fundamental notion in the general theory presents itself , slightly in eometry , but already very prominently in solid geometry ; .
we have here the difficulty as to the form of the equations of a culve in space , or ( to speak more accurately ) as to the expression by means of equations of the twofold relation between the coordinates of a point of such curve .
The notion in question is that of a -fold relation , \mdash ; as shed from any system of equations ( or onefold lelations ) serving for the expression of it , and as giving rise to the problem how to expless such relation by means of a system of equations ( or onefold relations ) .
Applying to the case of solid geometry my conclusion in the general theoly , it may be mentioned that I regard the twofold relation of a in space as being completely and precisely explessed by means of a system of equations , when no one of the functions , . . .
is a linear with constant or variable integral coefficients , of the others of them , when surface whatever which passes through the culve has its equation expressible in the form with constant or variable integral coefficients , , . . .
K. It is hardly necessary to remark that all the functions and coefficients are taken to bc rational functions of the coordinates , and that the word integral has reference to the coordinates .
General Explanations ; Retation , Locus , \mdash ; Article Nos. 1 to 36 .
1 .
Any quantities may be represented by means of quantities as the ratios of of these to the remaining quantity , and thus in place of the absolute values of the quantities we may consider the ratios of quantities .
2 .
It is to be noticed that we are throughout concerned with the ratios of the quantities , not with the absolute values ; this being understood , any mention of the ratios is in general unnecessary ; thus I shall speak of a relation between the quantities , meaning thereby a relation as regards the ratios of the quantities ; and so in other cases .
It may also be noticed that in many instances a limiting or extreme case is sometimes in cluded , sometimes not included , under a general expression ; the general expression is intended to include whatever , having regard to the subject-matter and context , be included under it .
3 .
Postulate .
We may conceive between the quantities a relation* .
4 .
A relation is eithel regular , that is , it has a definite manifoldness , or , say , it is a -fold relation ; or else it is irregular , that is , composed of relations not all of the same mani- foldness .
As to the word ' ' composed see post , No. 14 .
5 .
The ratios determined ( not in general uniquely ) by means of a -fold relation ; and a lelation cannot really be mole than -fold .
But the notion of a more than -fold relation has neveltheless to be considered .
A relation may be , either in mere appearance or else according to a provisional conception thereof , more than -fold , and be really -fold or less than -fold .
Thus a relation expressed by or more equations is in general and primd facie more than -fold ; but if the equations are not independent , and equivalent to or fewel equations , then the relation will be -fold or less than -fold .
Or the given relation may depend on parameters , and so long as these are arbitrary be really more than -fold ; but the parameters may have to be , and be accordingly , so determined that the relation shall be -fold or less than -fold .
A more than -fold relation is said to be superdeterminate .
6 .
A system of any number of onefold relations , whether independent or dependent , and if more than of them , whether compatible or incompatible , is telmed a ' Plexus , ' viz. if the number of onefold relations be , then the plexus is -fold .
A -fold plexus constitutes a relation which is at most -fold , but which may be less than -fold .
7 .
Every relation whatever is expressible , and that precisely , by means of a plexus ; but for the expression of a -fold relation we may require a more than -fold plexus .
whole difficulty of the subject is ( to speak ) in the analytical representation of a relation ; without it , the theories of the text cannot be exhibited with equivalent generality ; and I have for this reason presented them in an abstract formwithout analyticaJ expression or commentary .
But it is perfectly easy to obtain analytical iuustrations ; a omefold relation is expressed by an equation ; and ( although -fold relation is not in general expressible by equations ) any independent equations \amp ; c. constitute a -fold relation .
Thus ( No. 4 ) , an instance of irregular relation is , viz. this is satisfied by satisfaction either of the onefold relation , or of the twofold relation .
And post , Nos. 14 and 21 , the relation composed of the two onefold relations and is the onefold relation the relation regated of the same two relations is the twofold relation PROFESSOR CAYLEY .
ON ABSTRACT GEOMETRY .
8 .
Postulate .
We may conceive a m-dimensional space , the indetermination of the ratios of coordinates , and locus in of the point , the unique determination of these ratios .
More generally we may conceive any number of spaces , each of its dimensionality , and existing apart by itself .
9 .
Conversely , any quantities may be taken as the cooldinates of point in a -dimensional space .
10 .
The coordinates may have a elation ; it appears ( ante , No. 5 ) that the case , or where the lelation is more than -fold , is not altogethel excluded ; but this is not now under consideration .
The two limiting cases and will be pre- sently mentioned ; the remaining case is ; the system of points the cooldinates of which satisfy such a relation constitutes a -fold or dimensional locus .
And is the manifoldness , the dimensionality , of the locus .
11 .
If , that is , if the ratios are determined , we have the point-system , which , if the determination be unique , is a single point .
The explession alocus\ldquo ; may extend to include the point.system , and therefore also the poin .
If , that is , if the cooldinates are not connected by any relation , we have the original -dimensional space .
12 .
We may say that the -dimensional space is the locus in not oldy of the points in such space , but of the locus determined by any relation whatever between the coordinates ; and in like mannel that any dimensional locus in such space is dimensional space , a locus in of the points thereof , and of evely locus determined by a relation between the coordinates , implying the -fold relation which corresponds to the dimensional locus .
13 .
There is not any locus corresponding to a relation which is really mole than -fold ; hence in speaking of the locus corresponding to a given relation , we either assume that the lelation is not more than -fold , or we mean the locus , if any , corresponding to such relation .
14 .
Any two or more relations may be composed together , and they are then factols of a single composite relation ; viz. the composite relation is a relation satisfied if , and not satisfied unless , some one of the component relations be satisfied .
15 .
The foregoing notion of composition is , it will be noticed , altogether different from that which would at first suggest itself The definition is defective as not explaining the composition of a relation any number of times with itself , or elevation thereof to power ; which however must be admitted as part of the notion of composition .
16 .
A -fold relation which is not satisfied by any other -fold relation , and which is not a power , is a prime relation .
A relation which is not prime is composite , viz. it is a relation composed of prime factors each taken once or any other number of times ; in particular , it may be the power of a single prime factor .
Any prime factor is or multiple according as it occurs once or a greater number of times .
17 .
A relation which is either prime or else composed ofprime factors each of the same manifoldness is al.egular relation ; a -fold relation is ex termini regtllal .
An relation is a composite relation , the prime factors whereof are not all of the same manifoldness .
18 .
A prime -fold relation cannot be implied in any prime -fold relation different from itself .
But a prime -fold relation may be implied in a prime more-than-k-fold relation , \mdash ; or in a composite relation , regular or irregular , each factor whereof is more than -fold ; and so also a composite relation , regular or irregular , each factor whereof is at most -fold , may be implied in a composite relation , regular or irregular , each factor whereof is more than -fold .
In a somewhat different sense , each factor of a composite relation implies the composite relation .
19 .
A composite relation is satisfied if any particular one of the component relations is satisfied ; but in older to exclude this case we may speak of a composite relation as being satisfied distribuhvely ; viz. this will be the case if , in order to the satisfaction of the composite lelation , it is necessary to consider all the factors thereof , or , what is the same thing , when the reduced relation obtained by the omission of any one factor whatever is not satisfied .
And when the composite relation is satisfied distributively , the several factors thereof are satisfied alternatively ; viz. there is no one which is oughout unsatisfied .
20 .
A composite onefold relation is never distributively implied in a prime -fold relation\mdash ; that is , a prime -fold relation implies only a prime onefold relation , or at least only implies a composite onefold relation improperly , in the sense that it implies a certain prime factor of such composite onefold relation .
Conversely , every -fold relation which implies distributivel .
a composite omefold relation is composite .
21 .
Any two more relations may be aggregated together into , and they are then constituents of , a single aggregate relation ; viz..the aggregate lelation is only satisfied when all the constituent relations are satisfied .
The aggregate relation implies each of the constituent relations .
22 .
There is no meaning in aggregating a relationwith itself ; such aggregation only occurs accidentally when two relations aggregated together become one and the same relation ; and the aggregate of a relation with itself is nothing else than the original relation . .
A onefold lelation is not an aggregate , but is its sole constituent ; a more than onefold relation may always be considered as an aggregate of two or more constituent relations .
The constituent relations determine , they in fact constitute , the aggregate relation ; but the aggregate relation does not in any wise determine the constituent relations .
Any relation implied in a given relation may be considered as a constituent of such iven relation .
24 .
The regate of a -fold and a -fold relation is in general and at most fold relation ; when it is fold relation , the constituent relations are independent , but otherwise , viz. if the aggregate relation is , or has for factor , a less than fold equation , the constituent relations are dependent or interconnected .
PROFESSOR CAYLEY ON ABSIRACT GEOMETRY .
25 .
Passing from relations to loci , we may say that composition of relations corresponds to the congregation of loci , and the aggregation of relations to the intersection of loci .
26 .
For , first , the locus ( if any ) corresponding to a given composite relation is the conrregate of the loci corresponding to the several prime factors of the given relation , the locus corresponding to a single factor being taken once , and the locus corresponding to a multiple factor being taken a number of times equal to the multiplicity of the factor .
27 .
And , secondly , the locus ( if any ) corresponding to aiven ) relation is the locus common to and contained in each of the loci corresponding to the sevelal constituent relations respectively ; or , what is the same thing , it is the intersection of these several loci .
28 .
It may be remarked that a -fold locus and a -fold locus where ( or where the aggregate relation is more than -fold ) have not in any common locus .
29 .
Any onefold relation implied in a given -fold relation is said to be in involution with the -fold relation , and so in a system of onefold relations , if any relation be im- plied in the other relations , , what is the same thing , in the relation aggregated of the other relations , then the system is said to be in involution ; a system not in involution is said to be asyzygetic .
30 .
Consider a given -fold relation , and , in conjunction therewith , a system of an ) number of onefold relations each implied in the given -fold relation .
We may omit from the system any relation implied in the remaining relations , and so until we arrive at an asyzygetic system .
Consider now any other onefold relation implied in the given -fold relation ; this is either implied in the system of onefold relations , and it is then to be rejected , or if it is not implied in the system , it is to be added on to and made part of the system .
It may happen that , in the system thus obtained , some one relation of the original system is implied in the remaining lelations of the new system ; but if this is so the implied lelation is to be rejected ; the new system will in this case contain only as many relations as the original system , and in any case the new system will be asyzygetic .
Treating in the same manner every other onefold relation implied in the given -fold relation , we ultimately arrive at an asyzygetic system of onefold relations , such that every onefold relation implied in the given -fold relation is implied in the asyzygetic system .
The number of onefold relations will be at least equal to ( for if this were not so we should have the given -fold relation as an aggregate of less than onefold relations ) ; but it may be greater than , and it does not appear that there is any superior limit to the number of onefold relations of the asyzygetic system .
31 .
The system of onefold relations is a precise equivalent of the given -fold relation .
Every set of values of the coordinates which satisfies the given -fold relation satisfies the system of onefold relations ; and reciprocally every set of values which satisfies the system of onefold relations satisfies the given -fold relation .
But if we omit any one or more of the onefold relations , then the reduced system so obtained is not a precise equi- valent of the given -fold relation ; viz. there exist sets of values satisfying the reduced system , but not satisfying the given -fold relation .
32 .
In fact consider a -fold ation the aggregate of less than all of the onefold relations of the asyzygetic system , and in connexion therewith an omitted onefold relation ; this omitted relation is not implied in the aggregate , and it constitutes with the aggregate not fold , but only a -fold relation .
This happens as follows , viz. the omitted lelation is a factor of a composite onefold relation distributively implied in the aggregate ; hence the aggregate is composite , and it implies distributively a composite onefold lelation composed of the omitted relation and of an associated onefold relation ; that is , the aggregate will be satisfied by values which satisfy the omitted relation , and also by values which ( not satisfying the omitted relation ) satisfy the associated relation just leferred to .
33 .
Selecting at pleasure any of the onefold relations of the asyzygetic system , being such that the aggregate of the relations is a -fold relation , we have a composite -fold relation wherein each of the remaining onefold relations is alternatively implied ; viz. each remaining onefold relation is a factor of a composite onefold relation implied ibutively in the composite -fold relation .
Hence considering the onefold relations , viz. any elations of the asyzygetic system , each one of these is implied alternatively in the aggregate of the remaining relations ; and we may say that the onefold relations are in convolution .
34 .
More generally any or more , or all the relations of the asyzygetic system in , that is , any relation of the system is alternatively implied in the aggregate of the remaining relations , or indeed in the aggregate of any relations ( not being themselves in conYolution ) of the remaining relations of the asyzygetic system .
It may be added that , besides the relations of the system , there is not any onefold relation alternatively implied in the asyzygetic system .
35 .
The foregoing theory has been stated without any limitation as to the value of and it has I think a meaning even when is ; but the ordinary case is Considering the theory as applying to this case , I remark that the last proposition , viz. that no reduced system is a equivalent of the given fold relation , is generally true only on the assumption of the existence or quasi-existence of sets of values satisfying a more than -fold relation .
For let be , and , on the contrary , assume , as we usually do , that it is not in general possible to satisfy a more than wfold relation between the coordinates ; the number of relations in the system may be ; and if this is so , then selecting any relations of the system , it may very well happen that the given -fold relation is not satisfied by any sets of values other than those which satisfy the relations , \mdash ; that is , that the relations are a precise equiva- lent of the given -fold relation .
But even in this case the consideration of the entire system of the onefold relations is not the less advant ; and I say in general that the given -fold relation has for its precise and complete equivalent the asyzygetic system of onefold relations .
PROFESSOR CAYLEY ON ABSTRACT GEOMETRY .
36 .
[ In illustration of the foregoing Nos. 29 to 35 , I remark that , for the functions or equations \amp ; c. , if we have identically here the factors , are integral functions of the coordinates , and where some one of these factors , say , , is a constant ( or if we please ) , then the system of functions or uations is in involution ; or , to speak more accurately , the function or equation is in involution with the remaining functions or equations , . . .
But when the , . . .
are no one of them constant , then we have a convolution .
If is in involution with the remaining equations , . . .
, then is implied in these equations , and the relations and equivalent to each other .
But in the case of a convolution where then the relation the equations , . . .
imply , that is , or else ; or , what is the same thing , the relation is a relation composed of the two relations and . .
In the -folcl relation expressed by the more than equations . .
any of these equations which are not in conyolution , and uniting thereto any one of the remaining equations , we have a conyolution of equations ; and when a fold relation is precisely expressed by means of a system of or mole equations , . .
then every equation implied in the given relation , or , what is the same thing , the equation of any onefold locus passing through the locus given by the -fold relation is in involution with the equations , . . .
, that is , we have identically , . . .
being integral functions of the coordinates .
] Omal Relation ; Order .
Article Nos. 37 to 42 .
37 .
A -fold lelation may be linear or omal .
If , the corresponding locus is a point ; if the locus is a -fold , or dimensional omaloid ; the expression omaloid used absolutely denotes the onefold or dimensional omaloid ; the point may be considered as a fold omaloid .
38 .
A -fold relation which is not lineal or omal is of necessity composite , composed of a certain number of ?
-fold linear or omal relations ; .
the -fold locus corre- .
to the -fold relation is a point-system of points , each of which may bc conidered as given by a sepalate -fold linear or omal lelation ; each which relation is a factor of the -fold relation .
The given / -fold relation , and the point-system colresponding thereto , are lespectively said to be of the ordel M. 39 .
The order of a point-system of points is thus , but it is of course to be borne in mind that the points may be single or multiple points ; and that if the system consists of a point taken times , anothel point taken / 3 times , \amp ; c. , then the number of points and therefore the order of the system is considered to be 40 .
If to a given -fold relation we unite an absolutely arbitrary fold linear relation , so as to obtain for the gregate a -fold relation , then the order of this -fold relation ( or , what is the same thing , the number of points in the corresponding IIDCCCLXX .
PROFESSOR CAYLEY ON ABSTRACT GEOIIETRY .
point-system ) is said to be the order of the given -fold relation .
The notion of ol.der does not apply to a more than -fold relation .
41 .
The egoing definition of order may be more compendiously expressed as follows : Giyen between the coordinates a relation which is at most -fold ; then if it is -fold , join to it an arbitraly linear lelation so as to render it ?
-fold ; we have a / -fold lclation a point-system ; and the order of the giyen tion is equal to the numbel of points of the point-system .
42 .
The relation ated of two or given lelations , when the notion of order applies to the aggregate relation , that is , when it is not mole than / -fold , is of an order equal to the of the olders of the constituent relations ; or , say , the olders of the given relations , . . .
, the order of the aggregate relation is .
Article Nos. 43 and 44 .
43 .
considered so far relations which inyolye only the coordinates ; the coefficients are pie , or , if literal , they elle absolute constants , which either do or do not satisfy certain conditions ; if they do not , the lelation assumed in the instance to be -fold is leally -fold , or , as we may expless it , the relation is really as well as formally -fold ; if they do satisfy certain lelations in virtue whereof the formally -fold elation is really less than -fold , say , it is fold , then the relation is in fact to be considered ab ?
nitio as fold relation : there is no question of a being in genelal -fold and becoming less than -fold , or suffering any other modification in its ; and the notion of more than -fold lelation is in the cedino theolby meaningless .
44 .
But a lelation between the coordinates may involve pal'ametelS , and so .
as these remain arbitrary it may be really as well as formally -fold ; but when the ters satisfy certain conditions , it may become fold , or may suffer some other modification in its form .
And we have to consider the theory of a relation between the coordinates . .
involving besides parametels which may satisfy certain conditions , or , say simply , a relation involving variable parameters .
If the number of the palameters be , then these parameters may be regarded as the ratios of quantities to a remaining quantity , and the relation may be considered as involving the parameters . .
And these may , if we please .
be regarded as coordinates of a point in their own -dimensional space , or we have to consider relations between the cordinates and the ( parameters or ) coordinates . .
It is to be added that a relation may involve distinct sets of parameters , say , we have besides the original set of parameters , a set of parameters involved homogeneously .
But this is a generalization the necessity for which has hardly arisen .
* The only exception is anfe , No. 5 , where , in illustration of the notion of a more than -fold relation , mention is made of'palameters.\ldquo ; PROFESSOR CAYLEY ON ABSIRACT GEOMETRY .
Quantics , Notation , aec .
ticle Nos. 45 to 55 .
45 .
A homogeneous function of the cooldinates is lepresented by a notation such as ( where indicates the coefficients and .
) the ) , nd it is said to be a quantic ; and in reference to the quantic the quantities or coordinates are also termed facients .
ally a quantic involving two more sets of cooldinates , facients , is lepresented by the similar notation . . . .
46 .
The quantic is unipartite , bipartite , tripartite , \amp ; c. , accol d as the number of sets is one , two , three , \amp ; c. ; and with respect to any set of coordinates , it is binary , ternary , quaternary , . . .
, according as the number of the coordinates is two , three , four , ; and it is linear , quadric , cubic , quartic , . . .
, according as the deglec in legard to the coordinates in question is 1 , 2 , 3 , 4 . . .
47 .
A quantic involving two or mole sets of COOl.dinates , and lineal in to of them , is said to be tantipartite ; , in particular , when thel.eale only sets , it is said to be lineo-linear ; we may even extend the epithet lineo-lineal to the case of number of sets .
48 .
Instead of the general notation . . . . .
we vrite , . . .
, where the coefficients now indicated by . .
and the degrees , . . .
49 .
In the cases where the particular values of the coefficients have to be attended to , write down the entire selies of coefficients , or at least refel .
thel.eto by the notation . .
and it is to be understood that the coefficients expressed or refelTed to are each to be multiplied by the appropliate numerical coefficient , viz. for the term this numerical coefficient is 50 .
It is sometimes convenient not to introduce these numerical iers , then use the notation or .
In ticular ( \amp ; c. denote lespcctiyely \amp ; c. ; but , \amp ; c. denote \amp ; c. , and so and denote respectively and 51 .
To show which are the coefficients that belong to the several terms respectively , it is obviously propel that the quantic should be once written out at full length ; thus .
, in of a ternary cubic function , we say let and the like in other cases .
52 .
A onefold relation between the cooldinates is expressible by means of an equation of the 53 .
The expression " " an equation\ldquo ; used without explanation may be taken to mean an equation of the form in question , viz. the equation obtained by putting a quantic equal to zero ; the quantic is said to be the nilfactum of the equation .
We may consequently say simply that a onefold relation between the coordinates is always expressible by an equation .
54 .
It is frequently convenient to denote the quantic or nilfactum by a single letter , to use a locutiom such as " " the equation \ldquo ; which really means that the single letter stands for the quantic , so that we are afterwards at liberty to write as an abbreviated expression for .
We may also speak of the equation or function , meaning thereby the equation , or the function U. 55 .
A -fold relation between the coordinates is ( as has been shown ) equivalent to a system of or more elations ; each of these is expressible by an equation and the -fold relation is thus expressible by a system of or more such equations .
Representing by the system of functions which are the nilfacta of these equations respectively , the -fold relations may be represented thus , ; or more completely , the ation being -fold , and the number of equations , by the notation fold ) PROFESSOR CAYLEY ON ABSTRACT GEOMETRy .
We may also speak of the system or relation , meaning thereby the system of functions , or the relation Resultant , Discriminant , .
Article Nos. 56 to 62 .
56 .
In the case , a iven -fold relation between the and the parameters leads to fold lelation between the parameters .
This is telmed the resultant relation of the given -fold lelation , when the additional specification is necessary , the resultant relation obtained by elimination of the coordi- nates 57 .
Consider a -fold relation between the coordinates and coordinates then , consideling as coordinates and the as parameters , we have corresponding to the given lelation a -fold locus in the -space ; and so if } , then , considering the as coordinates , but the as parameters , we have to the given lelation a -fold locus in the -space .
58 .
If , but if the esultant relation is satisfied , then the -fold relation becomes a ?
-fold linear relation between the coordinates . .
and is con- sequently satisfied by a single set of values of the coordinates .
Hence , considering the oiven -fold relation as .
the fold resultant relation , the .-fold relation will represent a single point in the -space , say , the common point .
59 .
A -fold relation , or the locus , or point-system thereby represented , may double or nodal point , viz. two of the points of the point-system may be coincident .
More generally a -fold relation , the locus theleby leprescnted , hayc double or nodal point ; for let the lelation if less than -fold be made -fold by adjoin- ing to it a linear fold relation satisfied by the coordinates of the point in question but otherwise arbitrary , then , if the point in question be a double or nodal point of the -fold relation , or of the point-system thereby represented , the point is said to be a double or nodal point of original -fold lelation , or of the locus thereby repre- sented .
60 .
A given -fold elation ( between the coordinates , or the locus thereby represented , has not in general a nodal point .
But if the relation involve parameters .
the a certain onefold relation be satisfied between the parameters , there will be a nodal point .
The onefold lelation between the lrameters is the discriminant relation of the given -fold relation .
61 .
In the case in question , } , the discriminant relation is the resultant relation of fold lelation which is the gregate of the givcn -fold relation a certain relation called the acobian relation , or when the distinction is cquired , the Ja'o- bian relation in regard to the . .
62 .
Consider a -fold relation veen t coordinates and the coordinates . .
It has been seen that to a of values of the or , say , to a given point in the -space , there correspon a -fold locus in the -space , and that to a given set of values of the . .
or to a given point in the -space , there corresponds a -fold locus in the -space .
The -fold locus in the -space may have a nodal point ; this will be the case if there is satisfied between the a celtain onefold relation , the discriminant relation of the given -fold relation in regard to the . .
This onefold relation represents in the -space a onefold locus , the envelope of the -fold loci in the -space to the several points of -space .
The of the envelope is that to each point thereof there cor- responds in the -space a -fold locus having a nodal point .
Consecutive Points ; Omals .
Article Nos. 63-69 .
63 .
As the notions of proximity and remoteness have been thus far altogether nored , it seems necessary to make the following Postulate .
We may conceive a point consecutive ( or indefinitely near ) to a given point .
64 .
If the coordinates of the given point are . .
those of the consecutive point may be assumed to be . .
where , . . .
are indefinitely small in regald to . .
65 .
It may be remarked.that , taking the coordinates to be .
there is no obligation to have indefimtely small ; in fact whatever the magnitudes of these quantities are , if only X : , then the point will be the vely same with the original point , and it is therefore clear that a consecutive point may be represented in the same manner with magnitudes , however 1 , . . .
But we may assume them indefinitely small , that is , the ratios , . .
, where . .
are indefinitely small in legard to . .
will represent any set of ratios indefinitely near to the ratios . .
The foregoing quantities are termed the increments .
66 .
Consider a -fold lelation between the coordinates ( . .
} ; the increments are connected by a linear -fold relation .
The linear -fold relation is satisfied if we assume the increments proportional to the coordinates\mdash ; this is , in fact , assuming that the point remains unaltered .
We may write . .
since in such an equation only the ratios are attended to .
But it may be preferable to write .
In particular if , then the increments are connected by a linear -fold relation ; that is , the ratio of the inclements is uniquely determined ; and as the relation is satisfied by taking the propor- tional to the coordinates , it is clear that the values which the linear -fold lelation gives for the increments are in fact to the coordinates : viz. there is not in this case any consecutive point .
67 .
Considering the -fold relation as belonging to a -fold locus in the -space , so that al.e the coordinates of a point on this locus , then if in the linear -fold relation between the increments these increments are replaced by the coordinates of a point in the -space , then considering the original coordinates as para PROFESSOR CAYLEY ON ABSTRACT G3 meters , the locus of the.point is a k-fold omal locus : it is to be observed that , by what precedes , the linear k-fold relation is satisfied by writing therein the values , . .
, . . .
, that is , the -fold locus passes the original point ; the -fold omal locus is said to be the -omal of the original -fold locus at the ( point . .
which point is said to be the point of 68 .
If in the original -fold locus replace by . .
and combine therewith the -fold linear relation , we have between the coordinates -fold l.elation ( as parametels the cooldinates . .
; these parameters tisf.v the -fold relation , and in virtue heleof the 2 -fold relation ( whether is is not greater than ) is satisfied by the , . .
: . . .
; and not only so , but the point in question is a nodal or double point on the 2/ 1-fold locus .
It also follows that the -omal locus , considering in the linear relation as pala- meters satisfying the original -fold relation , has for its envelopc the locus .
69 .
We thus arlive at the notion of the double enel ation of a -fold locus , viz. such locus is the locus of the points , or , say , of the ineunt-points ; and it is also the envelope of the tangent-omals thereof .
have thus a theoly of duality ; I do not at present attempt to deyelope the theolv , but it is necessary to refer to it , in ordel to lemark that this theory is essential to the systematic develo meant of a ?
-dimensionnl geometry ; the oliginal classification of loci as onefold , tn.ofold , fold is incom- plcte , and must be supplemeltted with the loci connected with these loci lespectively .
And moreoyGl the theoly of the singularities of locus can only be tematically established by means of the same of duality ; the in regard to the ineunt-point must be of in connexion with the ities in to the tangent-omal .
These theories ( that is , the classification of loci , and th ishment and discussion of the singularities of each kind of locus ) , as extent is , should in the logical that which other lbons it may be expedient next to consider , the theol.y of mation , as depending on relations simultaneously the cooldinates and the . .

