Problem 195
Medium Equations Not Solved for y'

Find the singular solution (envelope) of the one-parameter…

Find the singular solution (envelope) of the one-parameter family of curves $y^2=(x-c)^3$, and verify it against the differential equation of the family.

Explanation of Method

The envelope is found by the C-discriminant method: eliminate \(c\) between \(F=y^2-(x-c)^3=0\) and \(\partial F/\partial c=0\). It is then checked that the candidate envelope satisfies the family's differential equation, and that it is a genuine (tangential) envelope and not merely an algebraic coincidence.

Step 1: C-discriminant elimination

With \(F=y^2-(x-c)^3\), \(\dfrac{\partial F}{\partial c}=3(x-c)^2\). Setting this to \(0\) gives \(c=x\) (\texttt{sympy.solve}); substituting back into \(F=0\) gives \(y^2=0\), i.e. \[ y=0. \]

Step 2: Differential equation of the family

Differentiating \(y^2=(x-c)^3\) implicitly gives \(2yy'=3(x-c)^2\). To eliminate \(c\), use \(y^2=(x-c)^3\Rightarrow (x-c)=y^{2/3}\), so \((x-c)^2=y^{4/3}\) — verified directly in \texttt{sympy} by substituting \(c=x-y^{2/3}\) into \((x-c)^2\) and simplifying to \(y^{4/3}\) exactly (no stray branch factors). Hence \[ 2yy'=3y^{4/3} \quad\Longrightarrow\quad y'=\frac{3}{2}y^{1/3}\quad (y\ne0), \] confirmed both symbolically and by a numerical spot check (finite-difference slope of \((x-c)^{3/2}\) at several random \((c,x)\) pairs matches \(\tfrac32y^{1/3}\) to within \(10^{-6}\)).

Step 3: \(y=0\) satisfies the differential equation

Substituting \(y=0\) into \(y'=\tfrac32y^{1/3}\) gives \(0=\tfrac32\cdot 0=0\): the constant function \(y=0\) trivially satisfies the ODE, so it is at least an algebraic (particular, non-generic) solution.

Step 4: Genuine tangency check (not just algebraic coincidence)

To confirm \(y=0\) is a true envelope — tangent to every member of the family, not merely a curve that happens to satisfy the ODE — consider one branch \(y=(x-c)^{3/2}\) (valid for \(x\ge c\)) and its slope \(y'=\tfrac32\sqrt{x-c}\), computed directly with \texttt{sympy.diff}. Evaluating \[ \lim_{x\to c^+} y' = \lim_{x\to c^+}\tfrac32\sqrt{x-c} \] with \texttt{sympy.limit} gives exactly \(0\). Since the member curve passes through \((c,0)\) — the same point where it meets \(y=0\) — with matching slope \(0\), the line \(y=0\) is tangent to every member of the family at its own point of contact \((c,0)\). This confirms \(y=0\) is a genuine singular solution/envelope in the rigorous tangency sense, not merely a curve that happens to satisfy the ODE algebraically.


Result

Final answer:

Singular solution (envelope): \[ y = 0. \] Differential equation of the family (re-derived independently, carefully eliminating \(c\) via \(c=x-y^{2/3}\) to avoid fractional-power branch errors): \[ y' = \frac{3}{2}y^{1/3}. \] The envelope \(y=0\) satisfies this ODE, and the limit computation \(\lim_{x\to c^+}\tfrac32\sqrt{x-c}=0\) confirms every member of the family is tangent to \(y=0\) at its contact point \((c,0)\), so \(y=0\) is a genuine singular solution and not just an algebraic coincidence. The source's derivation is confirmed correct in full.