We will study equations of the form x = t\dot x + f(\dot x) (1). And the geometry hidden behind them.

In the (t, x, p)-space, equation (1) defines a surface M. Note that intersection of M with the planes p = C (where C is a constant). are straight lines: x = Ct + f(C).

Similar to what we have done in: https://menagerieofmathematics.wordpress.com/2025/01/13/clairauts-equation-space-of-1-jets/

we will find the integral curves on M (and then, find the integral curves on the (t, x)-plane).

First, we find the direction field (i.e. the field of traces of the contact planes) on M through the relations:

x = pt + f(p)

dx = pdt + tdp + f'dp (condition of belonging to the tangent plane of the surface of M)

dx = pdt (condition of belonging to the contact plane)

We obtain the 1-form (t + f')dp = 0. On the (t, p)-plane, this form describes the structure of a field of lines. Note that a unique line (line parallel to the t-axis) is assigned by the 1-form to each point EXCEPT for the points t = -f’ (these points are assigned a space of lines, i.e. dp can be of any value).

Why does this occur? Looking back at the relations, we see that when t = -f’, the tangent plane coincides exactly with the contact plane. These points are exceptional and the set of these points is the discriminant curve t = -f'(p).

On the surface M, the discriminant curve is described by the relation t = – f'(p) and x = pt + f(p). On the (t, x)-plane, this curve can be described by the parameterization \gamma(C) = (t = -f'(C), x = Ct + f(C)).

On the (x, p)-plane, the integral curves are given by p = C (where C is constant). On the surface M, the integral curves are given by (t, Ct + f(C), C). Projecting the curves onto the (t, x)-plane gives us the integral curves x = Ct + f(C).

So, what is the relation between the discriminant curve and the integral curves on the (t, x)-plane?

Claim: the integral curves x = Ct+ f(C) are the tangent lines of \gamma(C).

The velocity vector at each point on \gamma(C) is given by:

\dot {\gamma}(C) = (-f''(C), \frac{d(-Cf'(C) + f(C))}{dC} = (-f”(C), -Cf”(C))$.

Note that the 1-form dx = Cdt vanishes on these vectors. Hence, the family of tangent lines along the curve \gamma(C) can be described by this 1-form. We find at each point (-f'(C), Cf'(C) + f(C)), the tangent line is x = Ct + f(C).

A Curve and its Family of Tangent Lines

Identifying the plane as the space of all lines (except for the vertical lines) and the dual plane of points:

And also

A shift in perspective – how they all connect together.

The duality between a point and the space of lines centered at the point:

Taking this to the 3-dim. space. Duality between a hyperplane (passing through the origin) and its normal vector (up to multiplication by a non-zero constant):

2 responses

  1. May 19th: the Euclidean Bundle and its Projective Dual – Menagerie of Mathematics Avatar

    […] Let be an arbitrary point on the red ellipse. It follows that and the associated tangent line at that point is , where . This is true for any point on the red ellipse. By circling around the red ellipse once, the corresponding tangent line also induces a circling around the blue ellipse. The other way applies as well. I discussed more about this in my previous post on the Legendre transformation: https://menagerieofmathematics.wordpress.com/2025/01/25/the-geometry-of-the-legendre-transformation-… […]

    Like

Leave a Reply

Is this your new site? Log in to activate admin features and dismiss this message
Log In