MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Hello everyone,

I'm currently working over a certain class of ODE of the form

$D_N \; \phi(x) = \lambda^N \phi(x) \quad, \quad N>1$

where $D_N = \delta_N \delta_{N-1} \cdots \delta_1$ and $\delta_k = (\frac{d}{dx}-A_k(x)), \quad 1 \leq k \leq N$.

For various reasons my interest at the moment is focused on a class of objects, that I call $m$-order Wronskians defined as $ \quad w_m = \det[W^{(m)}] \quad $ with the entries of the $m \times m$ matrix $W^{(m)}$ being

$ W^{(m)}_{ij} = \frac{d^{i-1}}{dx^{i-1}}\phi_j(x) \qquad 1 \leq i \leq m\ , \quad 1\leq j \leq m $

where $\left\lbrace \phi_j(x) \right\rbrace_{j=1}^N$ is a set of $N$ independent solutions of the ODE.


1 - The $N$-order Wronskian $w_N$ is a nonzero constant, given I set $\sum_{i=1}^NA_i=0$ (which means that the coefficient of the $(N-1)$-order derivative in the ODE vanishes). This obviously reflect the independence of the solutions $\phi_j$ .

2- The $(N-1)$-order Wronskian $z(x) = w_{N-1}$ is a solution of the adjoint equation

$ D_N^{\dagger} \; z(x) = \lambda^N z(x) \quad $ where $\quad D_N^{\dagger} = \delta^{\dagger}_1 \delta^{\dagger}_2 \cdots \delta^{\dagger}_N \quad$ and $\quad \delta^{\dagger}_k = (\frac{d}{dx}+A_k(x))$


some argument led me to argue that $m$-order Wronskians ($\forall m\ ,\ 1\leq m \leq N-1$) should satisfy some kind of ODE or pseudo-ODE (with which I mean equations that might contain both differential and integral operators). Sadly I could neither demonstrate this nor find any reference concerning this subject

Does someone know some reference to this subject?

Thank you very much for your time.


share|cite|improve this question
One small technical comment: you may want to assume $N > 1$ for your first "known fact". If $N = 1$ and $\lambda \neq 0$, the $N$-Wronskian is not necessarily constant. – Willie Wong Nov 3 '11 at 13:05
Yeah, thanks, I forgot to mention this detail, I'll add it. – Stefano Nov 3 '11 at 14:05
up vote 1 down vote accepted

I don't have a reference, a while back I learned in a paper by Percy Deift something that my help; it is easy to fill in the details with a bit of linear algebra. Write the equation as a system $\frac{d}{dx}F = M F \quad $ where $F$ is a function of $x$ taking values in $R^{n}$, and $M$ is a square matrix valued function. If $F_1, F_2$ are solutions, then $G= F_1 \wedge F_2$ the wedge product of the two solutions satisfies $\frac{d}{dx}G = M_2 G$ where $M_2 \colon R^{n} \wedge R^{n} \to R^{n} \wedge R^{n}$ is the linear map such that $M_2(a \wedge b) = Ma \wedge b + a \wedge Mb$ for all $a, b \in R^{n}$. The components of $G$ in the standard basis of $R^{n} \wedge R^{n}$ are the 2-minors of $\{\frac{d^{i-1}}{dx^{i-1}}\phi_j(x)\}_{1 \leq i \leq n\ , 1\leq j \leq n}$ ; i.e. the 2 order Wronskian derived from $F_1, F_2$ is one of the coordiantes of $G$. The other Wronskians appear in higher wedge products of solutions.

share|cite|improve this answer
Thank you very much! Yesterday I begun following this way, but I still hadn't found a result. – Stefano Nov 4 '11 at 10:40

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.