]\gt ; [ 259 ] XIL On the Contact of Surfaces .
WILLIAM SPOTTISWOODE , .A .
, Treas .
Received January lS , \mdash ; Read February 22 , 1872 .
IN a paper published in the Philosophical Transactions ( 1870 , p. 289 ) I have considered the contact .
at a point , of two curves which are coplanar sections of surfaces , and have examined somewhat in detail the case where one of the curves , viz. the section of , is a conic .
In the method there employed , the condition that the point should be sextactic , involved the azimuth of the plane of section measured about an axis through ; and consequently , arded as an equation in the azimuth , it showed that the point would be sextactic for certain definite sections .
It does not , however , follow , if conics having six-pointic contact with the surface be dralvl in the planes so determined , that a quadric surface can be made to pass through them all .
In fact it will be shown in the sequel that when this is possible , the quadric in general degenerates into a pair of planes .
The investigation therefore of the memoir above quoted was not directly concerned with the contact of surfaces , although it may be regarded as dealing with a problem intermediate to the contact of plane curves and that of surfaces .
In the present investigation I have considered a point common to the two surfaces and , an axis drawn arbitrarily ough P , and a plane of section passing through the axis and capable of ution about it .
Proceeding as in the former , and forming the equations for contact of various degrees , and finally rendering them independent of the azimuth , we obtain the conditions for contact for all positions of the cutting plane about the axis .
Such contact is called circumaxal ; and in particular it is called uniaxal , biaxal , \amp ; c. according as it subsists for one , two , \amp ; c. axes .
If it holds good for all axes through the point , it is called superficial contact .
It would at first sight seem that there should be a similar theory as to the number of axes about which contact must subsist in order that it may subsist about all axes , or be superficial .
It is , indeed , found that if two-pointic contact be biaxal , or if three-pointic be triaxal , \amp ; c. , the contact will be superficial .
But this would prove too much , as it would give four conditions instead of two for two-pointic , six conditions instead of for three-pointic , \amp ; c. superficial contact ; and , in fact , it turns out that there arc always in two-pointic contact one , in three-pointic two , \amp ; c. axes ( viz. the ents to the branches of the curve of contact through the point ) about which the contact is circum- axal per se , so that the theory in one sense disappears .
But as it at first had a semblance of existence , it may still be worth while to have laid its ghost .
At the conclusion of S 3 it is shown that the method of plane sections may , in the cases pnterest amportance , aced bSPOTTISWOODE OONTACT OSURFACES eral method of curved sections .
In the concluding section a few general considerations are given relating to the determination ofsurfaces having superficial contact of various degrees with given surfaces ; and at the same time it is indicated how very much the general theory is affected by the particular circumstances of each case .
The question of a quadric having fourpointic superficial contact with a given surface is considered more in detail ; and it is shown how in general such a quadric degenerates into the tangent plane taken twice .
To this there is apparently an exceptional case , the condition for which is given and reduced to a comparatively simple form ; but I must admit to having so left it , in the hope of iving a fuller discussion of it on a future occasion .
The subject of three-pointic superficial contact was considered by DUPIN , 'Developpe.ments de Geometry , ' p. 12 ; and , as I have learnt since the memoir was written , a general theorem connecting superficial contact and contact along vari branches of the curve of intersection of two surfaces ( substantially the same as that given in the text ) was enunciated by M. MOUTARD* . .
Preliminary Formuloe and Transformations .
Let be the equations of two surfaces whose contact is the subject of investigation .
Let their degrees be and respectively ; and let , as usual , ( 1 ) Also let , . .
, . . . . . .
represent the corresponding differential coefficients of Further , being arbitrary quantities , let ( 2 ) Also , forming the detelminants in the usual order , let ( 3 ) * I'oncelet , ' Applications a la Geometry , ' 1864 , tom .
u. p. 363 . .
W. SPOTTISWOODE ON THE OF SURFACES .
are the six coordinates of the line of intersection of the planes or the axis as it may be termed ; and represent five planes passing through the axis , the azimuth of the last plane about the axis being determined by the quantity : This being so , the quantities , . .
, . .
be found to satisfy the following relations:\mdash ; ( 4 ) while for every point of the plane we shall have also .
( 5 ) Again , let Then if we write column 1column 2column 3column 4 for each of the columns in succession of the expression for , it will be found that the following transformation may be effected , ' ' .
( 6 ) , let ' ' ' ( 7 ) then it will be found that where then it will be found that where Lastly , the operators , . .
are subject to the following identical relations , viz. :\mdash ; by means of which we may always eliminate one of the three operators entering into each of the expressions ( 8 ) .
In fact , the following values would express the result of such an elimination:\mdash ; ( 11 ) so that in the case where , we should obtain ' ( 12 ) There is one other mode of transformation which , on account of its utility , may properly find a place here .
If , be the same linear functions of , and l.espcctively , say . . . . .
( 13 ) then it will be found that ( 14 ) ' MR. W. ISWOODE ON THE CONTACT OF SURFACES .
263 Similarly , if be the same functions , the first quadratic in , 7 , ; the second lineo-linear in ; the third quadratic in , say , ( 15 ) then it will be found that , , .
, ( 1C ) ' And a similar process is also obviously applicable to functions of higher degrees .
S 2 .
Conditions of Contact .
In the memoir " " On the Contact of Conics with Surfaces above quoted , it was shown that the conditions for a 1 , 2 , . .
pointic contact at a point of the curves of section of the surfaces , made by a plane , may be explessed as follows : .
, ( 17 ) The cutting plane , say the plane , was supposed to pass through the point , say the point , to be capable of revolving about the axis whose six coordinates are , and to have an azimuth which , measured about the axis , is determined by the quantity The equations ( 17 ) may be supposed to be expressed in any of the forms to which they were reduced in , for instance , the form ( 8 ) , and dropping for the present the suffixes , so that shall be understood to represent any pair of the operators , . .
, . .
, the equations ( 17 ) may be written thus : . .
, ( 1S ) In the expansion of these expressions there will occur combinations such as where and represent any of the numbers 1 , 2 , 3 , 4 ; and when and are different , the compound operation is not in general the same as .
But in the case , it will be found on actual trial of the special forms that The same thing may be also proved by the following considerations , which arc applicable to all the forms of and .
In the first place is symmetrical , except regards algebraical sign , in respect first differential coefficients of and ; so that if be the expression obtained by replacing the differential coefficients of above lvlitten , we must also have ; ( 23 ) and the conditions for the existence of a -pointic contact at the point along some section through the axis will be expressed by eliminating : from the equations ( 22 ) , ( 23 ) , taken two and two together .
the condition for a three-pointic would be , ( 24 ) or Similally , for a four-pointic contact we should have , in addition to ( 25 ) , the following : and so on , for higher degrees , the azimuth of the plane section being always deter- mined by the value of at the point P. And it may be observed that , under the conditions supposed , there will in general be only one plane section alongwhich the contact will subsist .
MR. W. SPOTTISWOODE ON THE CONTACT OF SURFACES .
If the surfaces touch at the point , the equation is satisfied tically , and there will in general be two directions , determined by the equation .
which there will be three-pointic contact , as will be further noticed in the next section .
The condition for a four-pointic contact will then be obtained by elimi- nating from the equations ( 27 ) and ( 26 ) .
rrhese considerations may readily be extended to higher rees .
It will perhaps be worth while before further to evaluate the expression ( 24 ) ; viz. the condition to be fulfilled in order that when two surfaces meet at a point but do not touch , the curves of section made by some plane passing a given axis through shall have -pointic contact .
With a view to this , let us ake the following form for , viz. .
( 28 ) Then the expression to be developed ] be In this it will be found that the coefficients of , . .
vanish , and that the coefficients of are all the same , viz. Hence the whole expression may be reduced to the form , ( 29 ) it being understood that the operations do not affect the quantities so far as they appear explicitly in the above expression .
In order to calculate the coefficients of the powers and products of , we have , , ( 30 ) , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , each of which consists of two parts ; .
viz. the first involves the second differential coeffiand the second part of the same expression , suppose .
Hence the whole ession ( Hence the condition for a three-pointic contact in some plane about a given axis will be .
( 31 ) MR. W. SPOTTISWOODE ON THE CONTACT OF SURFACES .
267 This may be regarded as a condition to be fulfilled either by , the direction cosines of the axis , or by , the coordinates of the point .
Taking the first view , is the condition that the axis shall lie in the tangent plane of , and the condition that it shall lie in the tangent plane of ; hence ( 31 ) expresses the condition that the axis shall lie in the intersection of these planes .
On the other hand , arding a as given , the equation ( 31 ) will lepresent a surface whose intersections with U and will determine the points of three-pointic contact about a given axis .
The degree of this surface is ; and the number of points will therefore be .
Lastly , the equation ( 31 ) becomes independent of if .
( 32 ) which will consequently express the conditions that a three-pointic contact may subsist in some plane about any axis .
The degrees of these equations are , and respectively .
Points for which such contact will subsist for any axis do not in general exist when and do not touch ; but the condition for their existence will be found by eliminating from the equations S 3 .
Jfodes of Contact .
Hitherto we have considered only the contact of the curves of section of the surfaces made by definite planes passing through an axis .
If , however , in the equations ( 18 ) , which express the conditions for the contact of these curves , we equate to zero the coefficients of the various powers of the quantity , which determines the azimuth , we shall obtain a new series of conditions .
And the fulfilment of these conditions will ensure the subsistence of contact , of the degree under consideration , independently of the azimuth of the cutting plane ; or , in other words , for all plane sections round the point whose planes of section pass through the axis , such contact may be called circumaxal ; and , in palticular , contact which holds good in this manner for a single axis might be termed uniaxal contact ; that which holds good similarly for two axes might be telmed biaxal contact ; and so on for a greater number of axes .
But before entering into this question , it will be as to establish a theorem .
to the number of sections necessary to ensure uniaxal contact .
Returning to equations ( 18 ) ; the second , , expresses the condition for two-pointic contact .
Suppose that this holds good mole than one value of : , say , and .
Then , , down the equation for each of these values , we may eliminate the coefficients obtain the lesultant , . .
( 33 ) But as and are by hypothesis diffel.ent , the above equation cannot be satisfied , and consequently the coefficients of and in the equation under consideration must separately vanish .
But the evanescence of these coefficients expresses the conNDCCCLXXII .
2 ditions fniversal tpointic contact .
Hence ipointic contact ubsisk f axis .
It will be shown in the sequel , as is well known from other considerations , under these circumstances the contact will hold good for all axes through the point A similar result follows in the case of three-pointic contact .
If the third equation of ( 18 ) holds good for three values of , say , then writing down the equation for the three values successively , we shall be able to eliminate the three coefficients of the powers of and obtain the resultant , , ( 34 ) which cannot be satisfied , since by hypothesis the three values of are all different .
Hence the coefficients of the equation in question must separately vanish .
In other words , if a three-pointic contact subsist for three positions of the cutting plane about an axis , it will subsist for all positions about that axis .
The same law may obviously be extended to contacts of higher degrees .
The axis may be drawn , as before stated , in any direction through the point ; it may therefore be made to coincide with a tangent to the curve of intersection of and at the point .
But in that case it is obvious that two-pointic contact would subsist for two positions ( in fact for all positions ) of the cutting plane without involving the conditions for the ordinary contact of the two surfaces as a consequence .
It is perhaps desirable to show that the formulae here employed take cognisance of this circumstance , as well as of the corresponding circumstances in the cases of contact of higher degrees .
Suppose , then , that two-pointic contact subsists for two positions of the cutting plane about the axis , say for the two planes , ; then , adopting the last form of the group ( 8 ) , we have the two equations ( 35 ) Adding to these the two identical equations , and eliminating , we obtain the resultants ' ' .
( 36 ) ma .
W. SPOTTISWOODE ON TIIE CONEACT OF SURFACES .
269 And if we regard these equations as determining a particular direction for the axis , they express the condition that it must coincide with the tangent line to the curve of interof and at the point ; so that in this particular case the equations ( 35 ) involve , . .
as a consequence .
Again , in the case of three-pointic contact , we may take the following form , viz. . . . .
( 37 ) Then , since the operation vanishes identically , we obtain , by operating with it upon respectively , and then , the following lesultant : .
( 38 ) But this is the condition that ( 37 ) may be resolved into linear factors .
Supposing it so resolved into the product .
then one of these factors must vanish in virtue of ( 37 ) .
If , then , the contact subsists for three positions of the cutting plane , we may write . . . . . .
( 39 ) to which we may add , in virtue of the identical equations , the following , ; . . . . .
( 40 ) whence , eliminating , . .
, we obtain ' ' ' ' ( 41 ) showing that if the planes all intersect in a tangent to the curve of intersection , the conditions , . .
are not of necessity fulfilled .
It is perhaps unnecessary to pursue this part of the subject further .
Returning from this digression to the equations ( 18 ) , it may be observed that if there be two-pointic circumaxal contact about the point .
when , the equation will be satisfied by two values of ; in other words , the curve of intersection of and V will have a double point at , and along each of the branches the contact will be three-pointic .
Similarly , if there be three-pointic circumaxal contact about the point .
when in addition to the former we have , then the equation will be satisfied by three values of ; that is , the curve of intel.section will have a subsist for a definite number of these series of curves , in order that it may subsist for all such series .
In the latter case we shall call the contact superficial ; with two-pointic contact , and taking the form for , we obtain , on equating to zero the coefficients of respectively , ( 42 ) and applying to these the transformation ( 14 ) , we deduce the following forms , or . . . . . .
( 43 ) If this contact holds good for a second axis , we shall have also .
( 44 ) But since the two axes by hypothesis do not coincide , ( 43 ) and ( 44 ) cannot both be satisfied except on the conditions . . . . .
( 45 ) These conditions are in reality only two in number , in consequence of the identical l.elation .
This shows that if two-pointic contact be biaxal it will be supel.ficial .
But inasmuch as the directions of these axes are arbitrary , we may take for one of them the tangent to the curve of intersection of and through ; hence , setting this axis aside , and reckoning only arbitrary axes , we may state that if two-pointic contact be uniaxal it will be .
This , of course , is merely the ordinary property of common contact .
At the risk of being tedious on so simple a question , I venture to point out that result substantially the same may be deduced by a geometrical consideration from the equation , without the intervention of the suppositions ( 45 ) .
Take *Since the above I have founct that a similar theorem was enunciated by M. NOUTAmD .
PoNCBLIiT , YApplication d'Analyso a la Geometry , ' 1864 , MR. W. SPOTTISWOODE ON TACT OF SURFACES .
two axes passing through , and a pair of planes passing through each ; then , if two-pointic contact subsist for each pair of.planes , the contact will be biaxal , as was shown at the commencement of the present section We shall now have three planes in all , ( say the planes , forming a solid angle ; and in virtue of the equation with which we started , we shall have ( 4C ) But as these planes by hypothesis do not pass through one and the same straight line , the determinant of these equations cannot vanish .
Hence the system ( 46 ) can hold only on the conditions .
But we may take , as before , the tangent to the curve of intersection at as one of the axes .
Hence we come to the same conclusion as before .
Passing to the case of three-pointic contact ( and that two-pointic superficial contact subsists at the point P ) , and equating to zero the coefficients of the powers of in the equation , and adopting the same form as before , we shall obtain . . . . .
( 47 ) . . . .
which , by means of the transformation ( 16 ) , may be reduced to the forms:\mdash ; ( 48 ) whereof three only are independent .
And if the contact be triaxal , we should have ( taking the first of these forms ) , ( 49 ) Eliminating the coefficients , we obtain , by the usual method , . . . . .
But as by hypothesis the three axes are all distinct , this equation cannot be satisfied ; and therefore ( 49 ) can coexist only on the conditions .
Hence if the contact be triaxal it will be superficial .
But we may take for two the axes of triaxal contact the tangents to the two branches of the curve of intersection through and for every position of the cutting plane about each of these axes the contact wffl be three-pointic , viz. two consecutive points of the branch to which the axis is a tangent and one point of the other branch will lie in the plane ; whence it follows that , reckoning only arbitrary axes as before , if three-pointic contact be uniaxal it will be supefficial .
And this method has application to all degrees of contact .
The equations ( 48 ) would apparently determine two axes about which three-pointic contact would be circumaxai ; but that this is not the case will appear from the actual solution of one of them .
In fact the solution of the third equation depends quantity , in order to develop which we have the following values:\mdash ; ' ' ' ' ' Hence , by the method of compound determinants , in the expression the term independent of the coefficient of ' ; while the coefficient of will be found to be W. ISWOODE ON TIIIE CONTACT OF SURFACES .
that the whole expression sought suppose .
This being the case , the solutions of the equations may be written in the following forms:\mdash ; , ( 51 ) which involve is therefore a surface which cuts in a curve , at each point of which there is an axis , about which there is three-pointic contact .
It may be shown also , by the following geometrical constluction , that if -pointic contact be triaxal it will be superficial .
If we take three axes , , and draw through each three planes ; then if three-pointic contact subsist for each triplet of planes , the contact will be circumaxal for each axis , and therefore triaxal .
If we take a fourth axis , , the planes will pass three and three through each of the axes , will serve for the planes required , viz. the planes say the planes , . . .
Taking the forms for , the conditions for three-pointic contact along each of these planes will be .
( 52 ) Eliminating , . .
, . .
, we obtain the lesultant , , . .
, . .
, .
( 53 ) , . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
, . .
, . .
, .
, . .
, . .
: , . .
, . .
which is the condition that the six planes should all touch a cone of the second But the planes in question pass three and three through four lines ; and as it is impossible through any one line to draw more than two planes touching a cone of the second will no longer express the conditions for two- , three- , . .
pointic contact of the curves of section made by the plane or the plane , but the contact of the curves of section made by the surface W. And as the surface is perfectly arbitrary , the formulae will apply to any curve drawn at pleasure from the point on the surface U. It is to be borne in mind that in expanding the expression for three-pointic contact we shall obtain ; but in the only case which possesses much interest , viz. when the two-pointic contact at the point is superficial , we have ; and consequently , ( 56 ) which is of the same form as the expression derived in the case of plane sections .
And as the operators , . .
are unchanged , and are subject to the same identical relations as before , the conditions of contact now considered will be susceptible of the same transformations ( the transformations ( 13-16 ) excepted ) as those considered before .
these , therefore , we may draw the following conclusion:\mdash ; Consider two surfaces , , having superficial two- , three- , . .
pointic contact at a point ; from any number of curves arbitrarily on ; two , three , . .
consecutive points of these curves will , in consequence of the superficial contact , lie also on V. This being so , if for any three , four , . .
of the curves an additional consecutive point lies on then the same will be the case for all the curves , and there will be superficial three- , four- , . .
pointic contact between and at the point P. This may be also stated in the following form : If two surfaces , , have two- , three- , . .
pointic superficial contact a point , and if ough P we dxaw any number of surfaces arbitrarily , the curves of section on and which correspond to one another will , in consequence of the superficial contact , have two- , three- , . .
pointic contact .
This being the case , if any three- , four- , . .
corresponding curves have three- , four- , . .
pointic contact , then all will have three , four- , . .
pointic contact ; and there will be three- , four- , . .
pointic superficial contact between and at the point P. MR. W. SPOTTISWOODE ON THE CONTACT OF SURFACES .
275 This theorem for the case of three-pointic contact was given by ' Developpe ments de Geometry , ' p. 12 .
S4 .
On Surfaces having Superficial Contact with given It is well known that at any point of a surface we may in general determine a plane touching , or , in terms of this memoir , having two-pointic superficial contact with U. This suggests the question whether surfaces of other degrees may not determined superficial contact of higher degrees with ai a point P. The number of conditions for a 1 , 2 , . .
-pointic superficial contact has been shown above to be 1 , 3 , 6 , . .
the number of independent constants in the equation of a surface of the decree 1 , 2 , . .
, is 3 , 9 , 19 , . .
; so that , the equations which express the conditions of contact for determining the constants of , we shall meet with the following cases .
First,.if the numbel of conditions be equal to the number of constants , there will be a determinate surface having a superficial contact of the degree under consideration ( say ) with at the point P. Secondly , if the number of conditions exceed that of the constants by unity , the constants may be eliminated , and the result will be an equation between the coordinates ; in other words , an equation to a surface which will cut in a curve at every point of which a surface may be drawn having -pointic superficial contact with U. Thirdly , if the number of conditions exceed that of the constants by 2 , we may eliminate the constants in two ways , and obtain two lesulting equations , which will represent two surfaces mutually cutting in a finite number of points , at each of which a sulface may be drawn having -pointic superficial contact with U. Lastly , if the number of conditions exceed that of the constants by more than two , we shall obtain a number of resulting equations equal to that excess .
From these , together with the equations , the variables may be eliminated ; so that the number of resultants less 2 will represent the number of conditions which must subsist among the constants of , in order that it may be possible to draw such a surface .
This being the case , there will be a determinate surface of the degree having -pointic superficial contact with ( 1 ) at any point on , or ( 2 ) along a certain curve on , or ( 3 ) only at a finite number of points on , according as the expression ; or , clearing denominators , accolding as .
( 57 ) fDCCCLXXII .
276 MR. w. SPOTTISWOODE ON THE CONTACT OF SIIBFACES .
Now it is obvious from the signs of the terms on the left-hand side of this equation that the result cannot be positive if be less than ; beginning therefore with , we obtain ; or , resolving into factors , we have the three cases Next , let ; then , substituting this value in the equation ( 57 ) , we obtain ; or resolving into factors so far as possible , we have the three cases the first and last of which give no solutions in positive whole numbers .
This appears to exhaust all the solutions of ( 57 ) in positive whole numbers .
Reca pitulating the foregoing results , we may form the following Table:\mdash ; Such is the general theory .
But it is probable that it undergoes modifications in each particular case ; it certainly does so in the only case fully examined here , viz. that of a quadric having four-pointic contact with U. In fact , inasmuch as the equations , whereby the constants are ultimately determined , are linear the solution is in every case unique .
But four-pointic contact will subsist if we consider the quadric to consist of the tangent plane taken twice ; and as the solution is unique , no other quadric can in general be drawn having four-pointic contact .
Further , since a tangent plane can in general be drawn at every point on , the quadric of four- pointic contact ( viz. the tangent plane taken twice ) exists genelally ; and the condition difference , conditions \mdash ; constants restricting it to a curve on must be satisfied .
W. SPOTTISWOODE ON THE CONTACT OF SURFACES .
identically .
Similar remarks apply to contact of the degrees 6 , 8 , 10 ; but not apparently to that of the degree 15 .
It will be worth while to examine the case of four-pointic contact more in detail ; and for this purpose the special transformation of the operation , employed in the memoir before quoted , appears to be best adapted .
following method of effecting that transformation is perhaps more expeditious and direct than the process used in the memoir itself .
Taking for the form , we have ' But by the equations ( 11 ) of the memoir , ; ' ' whence since by hypothesis .
Again , ' whence , remembering that ( on the supposition ) we derive the following : ' ' ' DCBA So that finalIy the expression , combined with , is reduced to .
W. SPOTTISWOODE ON THE CONTACT OF RFACES .
which agree with the results obtained in the memoir .
In order to determine a quadric which shall have superficial four-pointic contact with the surface , let be the value of when , 7 , are written for respectively in the last line and the last column of ; let be the value of when are written for respectively in the last line 01 the lasl column , nnrl in the last column or the last line of ; and let be the value of when are written for respectively in the last line and the last column of H. Further , let , . . .
( 08 ) differentiations being effected shown memoir permissible ) without reference to , D. Lastly , if , . .
, , . .
, , . .
be the values of , . .
when becomes , lespectively , let ' ' ' ' DABC ' ' ( 59 ) That is to say , X , are the determinants formed from the matrix opposite to them by each of the columns in order ; and are the negatives of the determinants formed from the matrix opposite to them by omitting each of the columns 4 , 5 , 6 , 7 in order , and always retaining the columns 1 , 2 , 3 .
This being premised , the conditions which the coefficients of the . . .
( 60 ) must satisfy in order that four-pointic contact may subsist between the two curves of section of the surfaces made by the plane will be , as proved in the memoir above quoted , ( 61 ) and the contact will be circumaxal if the foregoing equations are made independent of .
If , therefore , we represent by , . .
, . .
; , . .
, . .
; , . .
, . .
; , . .
, . .
the coefficients of the powers of in , . .
, . .
respectively , we shall have four equations in the place of each one of the above \mdash ; apparently twelve and it is not difficult to see that omitting the factor , and employing the relations \amp ; c. , this expression may be reduced to the following form : ' , ( 63 ) which vanishes identically , since \amp ; c. The four equations derived from each of ( 61 ) are consequently reduced to three , and the whole number of equations connecting the ten coefficients of to nine , the proper number for the determination of their nine ratios .
This being the case , we may take as the equations for the ratios a : the following , viz. ; and the quantity to which 1 is proportional will then be which , the factor , is equal to Calling the coefficient in the last expression , and forming similar MR. W. SPOTTISWOODE ON THE CONTACT OF SURFACES .
expressions in \amp ; c. , we shall obtain the following values for the ratios viz.\mdash ; ( C4 ) But , , .
Hence , omitting the common factor , we have ; .
( 65 ) and proceeding in a similar manner with the equations in , it will be found that , when the quantity does not vanish , ; ( 66 ) that is to say , that the only quadric which in general has a superficial four-pointic contact with at any given point is the tangent plane taken twice .
From the relations given above , it appears that if any one of the equations ( 67 ) is satisfied , then all are satisfied ; so that it will be sufficient to study any one of them .
Although I have not succeeded .
in reducing the expression to any very simple form , it may still be worth while to show how it may be extricated from the condition of a compound determinant .
With a view to this , we may write down the values of , . .
in full , viz. :\mdash ; . .
, . . . .
This being so , let ' ' ' ' Also let represent the expressions obtained by writing for respectively in the first lines of .
Then it will be found , by the method of compound determinants , that whence ; similarly , both of which , when equated to zero , are comprised in the system .
( 69 ) Similarly , taking the determinants ' we should alrive at the system .
( 70 )

