This is a continuation of the discussion started on my main sandbox page. In particular, this page is intended to discuss the specifics of constructing a circle center, and restoring the conditions to the Poncelet–Steiner theorem, when provided two non-intersecting circles devoid of their centers, but having a common centerline between them provided, or when provided three non-intersecting circles devoid of their centers.
Given is a line m (in black) and four arbitrary points A, B, C, and D on the line. Given also is an arbitrary circle in the plane; observe the center is not indicated as it is not utilized in this construction.
We may construct points E and F, known as the fixed points of involution. These points are defined by the four points on the line m, and are independent of the circle. The order of these points is important as the relationship is particular. The six points hold a proportionality relationship: .
Taking points A, B, C, and D out of order, the fixed points E and F would be constructed in different locations. Though, with the appropriate relabeling of points, the proportionality would still hold.
The fixed points may be constructed thus:
In the above construction, it is assumed that segments AB and CD are exterior one another; i.e. they do not overlap. This has the consequence of placing points E and F interior their respective line segments. If segments AB and CD overlap or one is interior the other, then points E and F may have a less predictable and exterior placement.
It is the case that points E and F are both projective harmonic conjugates of one another with respect to both segments AB and CD, simultaneously. Alternatively phrased, line segment EF is the segment to which both A and B are projective conjugates of one another, and both C and D are conjugates of one another.
The purpose of this construction is to find the intersection points of a circle, and a line defined by a point in the plane and a point on the circle. In other words, from the perspective of a point in the plane, a line to any point on the circle intersects at a new point to be constructed. The point on the circle is projected through the circle, along the line, to the other side of the circle. In order to complete this construction the arc of the circle is not required, nor a circle of any kind in the plane, but five points on the arc of the circle.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly.
Provided are five points in space - points A, B, C, D, E arranged counter clockwise about the circle - such that all five exist on the circumference of the same circle. There exists also a point M in space.
We wish to project any point on the circle (point B is chosen in the construction below) to the opposite side of the circle as viewed from the perspective of point M. That is to say, we wish to find the intersection point between the circle defined by the five points, and the line MB, where point B must necessarily be one of the five points on the circle circumference.
From the perspective of point M, both points B and B' are colinear with point M, and existing on the circle. Thus point B' is the intersection point between the circle and line MB. This was achieved without the need of the arc; only five points on the circle were required.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly. The circle is strictly identified by five unique points on its circumference, as per the previous construction, but only two of which - B and C - are depicted.
The polar is the line connecting points of tangency. In the image below, the tangent lines that pass through point M to the circle defined by five points (points not depicted) will intersect the circle at the points of tangency. A single line passing through the points of tangency is known as the polar.
Choose two (of the five) points on the circle arbitrarily. In this example points B and C have been selected. The remaining three points A, D, and E have been hidden in this construction, but are still necessary for the referenced sub-constructions.
Observe that this construction is not significantly different than the one provided on in my first sandbox, specifically the construction of tangent lines to a circle. Polars, tangent lines and points of tangency are all intimately interrelated. Any line passing through point M and passing through the circle suffices for that construction, but necessitates the use of both intersection points to the circle. Here, one of the intersections must be constructed as a projection of another. Additionally here we are using only one pair of lines through point M rather than two - a variant discussed in the referenced construction.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly.
This construction will be used for circles defined by five points on the arc, or any circle in which line-circle intersection points cannot be directly found. The five points are not depicted, but will be required for the referenced sub-constructions. The referenced sub-constructions also require the use of a provided circle somewhere in the plane (its center point is not necessary).
Given a line MN which passes through the circle defined by five points, we may find the points of intersection between the circle and the line.
The goal of this section is to construct points on the circle which is only identified by the presence of other points also already on the circle. The circle is not compass-drawn. A minimum of three points on the arc are required to identify a unique circle, but in order to construct additional points on the circle using straightedge only, more information than three points is required. Various scenarios are discussed below.
Given five distinct points on the arc of a circle, wherein the arc is not compass-drawn, additional points of the circle may be constructed. A variety of techniques are available.
This is a trivial implementation of an earlier construction, projecting a point through a circle defined by five points. We may choose arbitrarily or construct any point in the plane. Through this point we may project any of the five points already on the circle, the pair of which define a line, through to the other side of the circle colinear with the other two points. In this way, additional points may be added to the set of points known to be on the circle.
Using the previous construction to intersect an arbitrary line with a circle of five points, any arbitrary line through the circle may have both of its intersection points with the circle constructed. This is a trivial implementation of a previous construction.
It should be noted that this approach is unnecessarily complicated, and requires the implementation of other, simpler constructions such as the aforementioned projections through a circle in order to complete, which results in further point constructions on the circle, thus resulting in more than just two additional points of the circle. This construction is offered only as a possibility, not as a practical solution.
This construction utilizes the polar construction, both from three cutting lines and from two cutting lines. First, using any of four points on the circle, cutting lines may be constructed which intersect at a point which we may call a pole. From the pole and using the two cutting lines, the polar may be constructed. Utilizing the fifth point, the pole, and the known polar, we may establish a third cutting line and thus a sixth point of intersection.
This is an implementation of Pascal's theorem, in particular as it is used with circles. The theorem states that whenever a hexagon is inscribed in a circle (conic section, more generally), opposite sides may be extended to form three distinct intersection points, even if "at infinity", all of which are colinear. The line of these intersection points is known as a Pascal line. This construction makes use of this property.
Given five points on a circle, four sides of a hexagon are established; that is, vertices may be arbitrarily paired to define edges. Accordingly, one pair of known opposing sides are definable. A degree of freedom left for the geometer is in the order in which the five points are selected to determine hexagonal sides. A hexagon may have a side omitted between any pair of the established points, and furthermore, the hexagon need not be simple; Pascal's theorem allows for self-intersecting (non-simple) hexagons. The arbitrariness of these selections contribute to the placement of the new point we aim to construct.
The known opposite sides, which were arbitrarily chosen, may be used to find a single point of intersection on Pascal's line. Any arbitrary line may be chosen through this point to be the Pascal line, as no known second point on the line is determinable from the provided information; this is a second degree of freedom left to the geometer also contributing to the arbitrariness of the final new point. The remaining two unpaired sides of the hexagon may be extended to Pascal's line, intersecting it and establishing its opposite side. The two new constructed opposite sides intersect one another on the circle at the hexagons sixth vertex.
Given are points A, B,C, D, and E, all on the arc of the circle which is not here compass-drawn. The line segments AB, BC, CD, and DE (all in black), which are interior the circle, arbitrarily define five of the six sides of an inscribed hexagon. Further, sides AB and DE are arbitrarily chosen to be opposite one another.
This may be viewed as the reverse of a tangent line construction, referencing the section Construction of a Tangent/Polar to a Circle or Arc by way of subsection from a point on the arc of a circle. In a previous construction, five points on a circle were used in order to create a tangent line to the circle through one of the points.
In that construction, these five points were chosen arbitrarily on an arc of a compass-drawn circle in the plane. The arc does not need to exist for the tangent through a point to be constructed, as the arc was never used, but as a curve upon which to arbitrarily place five points. If the points are provided and known to be on the arc already, the same construction may be achieved without the circle arc being drawn.
The construction in this section, however, attempts to construct one of the five points on the arc of the circle, using four other points which includes a point of tangency and a tangent line. This construction utilizes the geometric properties of the pentagram, and constructs the same set of lines in a different order, from a different initial set of conditions to arrive at a different end goal. In this case, we start with a tangent line (and the point of tangency) and three additional points on the arc of the circle, to arrive at a fifth point.
In the image, a point A exists on the circle, and is a point or tangency for the tangent line (drawn in black). The circle is not compass-drawn, but is illustrated (in dashed gray) for expositional purposes; we may not directly intersect with the arc of the circle. Also on the circle are points B, C, and E. We wish to construct a point D between the latter two points. The required segments of the pentagram are drawn (in black): segments AC, CE, and BE.
Using a modified approach to the construction intersecting a line with the circle of an arc, it is possible using the same set of lines constructed in reverse order, to find two additional points (a fourth and a fifth) on a circle when only three are provided, two of which having tangent lines through them.
This section allows us to construct points on the arc of a circle that is only identified by a single point in the plane, in a coaxial system. A coaxial system is a set of circles, even if they are yet to be constructed, any pair of which share the same radical axis. They share the same centerline as well, though this is not the defining attribute. Two compass-drawn circles are provided by hypothesis; these establish a radical axis and therefore define a coaxial system. For any point in space - excepting those on the radical axis - there exists a single circle passing through it which belongs to the same coaxial system. This section will demonstrate methods for constructing additional points on this circle, given the one initial point and the two provided circles.
Provided below are two compass-drawn circles, arbitrarily labeled l and r, for their left and right relative positions.
Additional new points on the circle may be systematically constructed in sequence. Each newly constructed point D may be relabeled the new point C in the next iteration of the application of this construction, thereby producing a sequence of new points on the circle. Each constructed point becomes a new launch point, leading to one additional new point. This process may continue until and unless points C=D, at which point the reiterative process may terminate. If insufficient points have been constructed, alternative solutions must be considered.
In each iteration of the construction, one of the tangent lines to circle l from the point C had already been previously constructed, and leads back to a previously constructed point. Only one tangent line is useful and has the potential to lead to new points. This is true for all constructed points on the circle, except the first. The original point C, which defined the circle in the coaxial system, offered two tangents to circle l, neither of which had previously been constructed, thus offering two unique new points. This may be where the original point T becomes useful, and may be relabeled point A. Switching the roles of points A and T of the first iteration will produce a new set of iterations, which may continue until again points C=D.
Alternatively, we may also resolve the problem by swapping the roles of circles l and r, and construct tangent lines to the alternate circle. Placing points A on circle r, and finding its polar to circle l instead, may also resolve some limitations.
It should also be noted that, in addition to the original provided point C, only four new points need to be constructed, in the general case, or three new points and a known tangent line, or two new points and two tangent lines, before the constructions in the previous section Finding additional points on a circle may be utilized instead, which do not require a second circle or a coaxial system to work.
Furthermore, the next subsection offers a completely unique method based on projective points, of constructing new points on this circle in the coaxial system, that does not suffer the same limitations.
Given is a point point C in the plane.
Additional points may be constructed as necessary. This construction may be repeated, using many arbitrary but unique lines m through point C. It is also valid to substitute point C in the plane with any other point also on the same circle, such as any of the points previously constructed in the sequence of such points.
Only three points on the circle and two tangent lines, four points on the circle and one tangent line, or five points on the circle, are required before the constructions of the previous section Finding additional points on the circle may be utilized, which do not require a second circle or a coaxial system.
Given are two non-intersecting circles in the plane - l and r - through which a common centerline, c, is drawn. The centers of these circles may be constructed. In the construction below, it is presumed that the provided circles are external one another, though this is not required.
In such a scenario where the line m is tangent to one or both circles, the construction simplifies greatly. If, for example, the line m is tangent to circle l, the following may be done. The point A=B may be labeled P, circle l may be labeled circle p. The intersection points of circle p with the centerline may be labeled and . Steps 2 and 3 may be skipped, for the associated circle l. Line PUp is trivially constructed. The alternate circle r for which line m is not tangent must still undergo the full construction to derive line QUq, after which steps 4 and 5 may be followed. A similar argument can be made for each circle in kind; the simplest scenario is if line m is tangent to both circles.
This is a continuation of the discussion started on my main sandbox page. In particular, this page is intended to discuss the specifics of constructing a circle center, and restoring the conditions to the Poncelet–Steiner theorem, when provided two non-intersecting circles devoid of their centers, but having a common centerline between them provided, or when provided three non-intersecting circles devoid of their centers.
Given is a line m (in black) and four arbitrary points A, B, C, and D on the line. Given also is an arbitrary circle in the plane; observe the center is not indicated as it is not utilized in this construction.
We may construct points E and F, known as the fixed points of involution. These points are defined by the four points on the line m, and are independent of the circle. The order of these points is important as the relationship is particular. The six points hold a proportionality relationship: .
Taking points A, B, C, and D out of order, the fixed points E and F would be constructed in different locations. Though, with the appropriate relabeling of points, the proportionality would still hold.
The fixed points may be constructed thus:
In the above construction, it is assumed that segments AB and CD are exterior one another; i.e. they do not overlap. This has the consequence of placing points E and F interior their respective line segments. If segments AB and CD overlap or one is interior the other, then points E and F may have a less predictable and exterior placement.
It is the case that points E and F are both projective harmonic conjugates of one another with respect to both segments AB and CD, simultaneously. Alternatively phrased, line segment EF is the segment to which both A and B are projective conjugates of one another, and both C and D are conjugates of one another.
The purpose of this construction is to find the intersection points of a circle, and a line defined by a point in the plane and a point on the circle. In other words, from the perspective of a point in the plane, a line to any point on the circle intersects at a new point to be constructed. The point on the circle is projected through the circle, along the line, to the other side of the circle. In order to complete this construction the arc of the circle is not required, nor a circle of any kind in the plane, but five points on the arc of the circle.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly.
Provided are five points in space - points A, B, C, D, E arranged counter clockwise about the circle - such that all five exist on the circumference of the same circle. There exists also a point M in space.
We wish to project any point on the circle (point B is chosen in the construction below) to the opposite side of the circle as viewed from the perspective of point M. That is to say, we wish to find the intersection point between the circle defined by the five points, and the line MB, where point B must necessarily be one of the five points on the circle circumference.
From the perspective of point M, both points B and B' are colinear with point M, and existing on the circle. Thus point B' is the intersection point between the circle and line MB. This was achieved without the need of the arc; only five points on the circle were required.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly. The circle is strictly identified by five unique points on its circumference, as per the previous construction, but only two of which - B and C - are depicted.
The polar is the line connecting points of tangency. In the image below, the tangent lines that pass through point M to the circle defined by five points (points not depicted) will intersect the circle at the points of tangency. A single line passing through the points of tangency is known as the polar.
Choose two (of the five) points on the circle arbitrarily. In this example points B and C have been selected. The remaining three points A, D, and E have been hidden in this construction, but are still necessary for the referenced sub-constructions.
Observe that this construction is not significantly different than the one provided on in my first sandbox, specifically the construction of tangent lines to a circle. Polars, tangent lines and points of tangency are all intimately interrelated. Any line passing through point M and passing through the circle suffices for that construction, but necessitates the use of both intersection points to the circle. Here, one of the intersections must be constructed as a projection of another. Additionally here we are using only one pair of lines through point M rather than two - a variant discussed in the referenced construction.
In the image below, a light blue dashed circle is apparent. This circle is not constructed by compass; the arc does not exist in the plane. It is only shown for convenience of exposition. We may not intersect lines with it directly.
This construction will be used for circles defined by five points on the arc, or any circle in which line-circle intersection points cannot be directly found. The five points are not depicted, but will be required for the referenced sub-constructions. The referenced sub-constructions also require the use of a provided circle somewhere in the plane (its center point is not necessary).
Given a line MN which passes through the circle defined by five points, we may find the points of intersection between the circle and the line.
The goal of this section is to construct points on the circle which is only identified by the presence of other points also already on the circle. The circle is not compass-drawn. A minimum of three points on the arc are required to identify a unique circle, but in order to construct additional points on the circle using straightedge only, more information than three points is required. Various scenarios are discussed below.
Given five distinct points on the arc of a circle, wherein the arc is not compass-drawn, additional points of the circle may be constructed. A variety of techniques are available.
This is a trivial implementation of an earlier construction, projecting a point through a circle defined by five points. We may choose arbitrarily or construct any point in the plane. Through this point we may project any of the five points already on the circle, the pair of which define a line, through to the other side of the circle colinear with the other two points. In this way, additional points may be added to the set of points known to be on the circle.
Using the previous construction to intersect an arbitrary line with a circle of five points, any arbitrary line through the circle may have both of its intersection points with the circle constructed. This is a trivial implementation of a previous construction.
It should be noted that this approach is unnecessarily complicated, and requires the implementation of other, simpler constructions such as the aforementioned projections through a circle in order to complete, which results in further point constructions on the circle, thus resulting in more than just two additional points of the circle. This construction is offered only as a possibility, not as a practical solution.
This construction utilizes the polar construction, both from three cutting lines and from two cutting lines. First, using any of four points on the circle, cutting lines may be constructed which intersect at a point which we may call a pole. From the pole and using the two cutting lines, the polar may be constructed. Utilizing the fifth point, the pole, and the known polar, we may establish a third cutting line and thus a sixth point of intersection.
This is an implementation of Pascal's theorem, in particular as it is used with circles. The theorem states that whenever a hexagon is inscribed in a circle (conic section, more generally), opposite sides may be extended to form three distinct intersection points, even if "at infinity", all of which are colinear. The line of these intersection points is known as a Pascal line. This construction makes use of this property.
Given five points on a circle, four sides of a hexagon are established; that is, vertices may be arbitrarily paired to define edges. Accordingly, one pair of known opposing sides are definable. A degree of freedom left for the geometer is in the order in which the five points are selected to determine hexagonal sides. A hexagon may have a side omitted between any pair of the established points, and furthermore, the hexagon need not be simple; Pascal's theorem allows for self-intersecting (non-simple) hexagons. The arbitrariness of these selections contribute to the placement of the new point we aim to construct.
The known opposite sides, which were arbitrarily chosen, may be used to find a single point of intersection on Pascal's line. Any arbitrary line may be chosen through this point to be the Pascal line, as no known second point on the line is determinable from the provided information; this is a second degree of freedom left to the geometer also contributing to the arbitrariness of the final new point. The remaining two unpaired sides of the hexagon may be extended to Pascal's line, intersecting it and establishing its opposite side. The two new constructed opposite sides intersect one another on the circle at the hexagons sixth vertex.
Given are points A, B,C, D, and E, all on the arc of the circle which is not here compass-drawn. The line segments AB, BC, CD, and DE (all in black), which are interior the circle, arbitrarily define five of the six sides of an inscribed hexagon. Further, sides AB and DE are arbitrarily chosen to be opposite one another.
This may be viewed as the reverse of a tangent line construction, referencing the section Construction of a Tangent/Polar to a Circle or Arc by way of subsection from a point on the arc of a circle. In a previous construction, five points on a circle were used in order to create a tangent line to the circle through one of the points.
In that construction, these five points were chosen arbitrarily on an arc of a compass-drawn circle in the plane. The arc does not need to exist for the tangent through a point to be constructed, as the arc was never used, but as a curve upon which to arbitrarily place five points. If the points are provided and known to be on the arc already, the same construction may be achieved without the circle arc being drawn.
The construction in this section, however, attempts to construct one of the five points on the arc of the circle, using four other points which includes a point of tangency and a tangent line. This construction utilizes the geometric properties of the pentagram, and constructs the same set of lines in a different order, from a different initial set of conditions to arrive at a different end goal. In this case, we start with a tangent line (and the point of tangency) and three additional points on the arc of the circle, to arrive at a fifth point.
In the image, a point A exists on the circle, and is a point or tangency for the tangent line (drawn in black). The circle is not compass-drawn, but is illustrated (in dashed gray) for expositional purposes; we may not directly intersect with the arc of the circle. Also on the circle are points B, C, and E. We wish to construct a point D between the latter two points. The required segments of the pentagram are drawn (in black): segments AC, CE, and BE.
Using a modified approach to the construction intersecting a line with the circle of an arc, it is possible using the same set of lines constructed in reverse order, to find two additional points (a fourth and a fifth) on a circle when only three are provided, two of which having tangent lines through them.
This section allows us to construct points on the arc of a circle that is only identified by a single point in the plane, in a coaxial system. A coaxial system is a set of circles, even if they are yet to be constructed, any pair of which share the same radical axis. They share the same centerline as well, though this is not the defining attribute. Two compass-drawn circles are provided by hypothesis; these establish a radical axis and therefore define a coaxial system. For any point in space - excepting those on the radical axis - there exists a single circle passing through it which belongs to the same coaxial system. This section will demonstrate methods for constructing additional points on this circle, given the one initial point and the two provided circles.
Provided below are two compass-drawn circles, arbitrarily labeled l and r, for their left and right relative positions.
Additional new points on the circle may be systematically constructed in sequence. Each newly constructed point D may be relabeled the new point C in the next iteration of the application of this construction, thereby producing a sequence of new points on the circle. Each constructed point becomes a new launch point, leading to one additional new point. This process may continue until and unless points C=D, at which point the reiterative process may terminate. If insufficient points have been constructed, alternative solutions must be considered.
In each iteration of the construction, one of the tangent lines to circle l from the point C had already been previously constructed, and leads back to a previously constructed point. Only one tangent line is useful and has the potential to lead to new points. This is true for all constructed points on the circle, except the first. The original point C, which defined the circle in the coaxial system, offered two tangents to circle l, neither of which had previously been constructed, thus offering two unique new points. This may be where the original point T becomes useful, and may be relabeled point A. Switching the roles of points A and T of the first iteration will produce a new set of iterations, which may continue until again points C=D.
Alternatively, we may also resolve the problem by swapping the roles of circles l and r, and construct tangent lines to the alternate circle. Placing points A on circle r, and finding its polar to circle l instead, may also resolve some limitations.
It should also be noted that, in addition to the original provided point C, only four new points need to be constructed, in the general case, or three new points and a known tangent line, or two new points and two tangent lines, before the constructions in the previous section Finding additional points on a circle may be utilized instead, which do not require a second circle or a coaxial system to work.
Furthermore, the next subsection offers a completely unique method based on projective points, of constructing new points on this circle in the coaxial system, that does not suffer the same limitations.
Given is a point point C in the plane.
Additional points may be constructed as necessary. This construction may be repeated, using many arbitrary but unique lines m through point C. It is also valid to substitute point C in the plane with any other point also on the same circle, such as any of the points previously constructed in the sequence of such points.
Only three points on the circle and two tangent lines, four points on the circle and one tangent line, or five points on the circle, are required before the constructions of the previous section Finding additional points on the circle may be utilized, which do not require a second circle or a coaxial system.
Given are two non-intersecting circles in the plane - l and r - through which a common centerline, c, is drawn. The centers of these circles may be constructed. In the construction below, it is presumed that the provided circles are external one another, though this is not required.
In such a scenario where the line m is tangent to one or both circles, the construction simplifies greatly. If, for example, the line m is tangent to circle l, the following may be done. The point A=B may be labeled P, circle l may be labeled circle p. The intersection points of circle p with the centerline may be labeled and . Steps 2 and 3 may be skipped, for the associated circle l. Line PUp is trivially constructed. The alternate circle r for which line m is not tangent must still undergo the full construction to derive line QUq, after which steps 4 and 5 may be followed. A similar argument can be made for each circle in kind; the simplest scenario is if line m is tangent to both circles.