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


I am currently working on a paper that involves an upper bound of the largest root of a polynomial. With the help of the Mean Value Theorem, I believe a colleague and I have proved the following: if f^(n)(p) > 0 (the n-th derivative evaluated at "p", for n=0,1,2..d, where d is the degree of the polynomial), then p>r, where r is the largest root of the polynomial.

The only question I have is the following: is this well-known or has been cited/proven elsewhere? It just seems basic enough that someone in the past has referenced it in a paper/Lemma. Though we could be mistaken and have a flaw in our own proof...



share|cite|improve this question
Doesn't this basically just say that if all the coefficients are positive, then the roots are all negative? (by shifting $p$ to $0.$)? If so, this predates the Descartes' rule of signs... – Igor Rivin Aug 10 '11 at 20:02
It sounds like a consequence of Rolle's Theorem to me. You might reword it as a request for a reference. If it is true and published, you may get a reference. Otherwise (though I doubt it) the result is new or wrong and you may find out which anyway without getting a reference. I suggest you add the reference-request tag yourself, and don't worry in print about whether there is a mistake: if there is, it will be found. Gerhard "Ask Me About System Design" Paseman, 2011.08.10 – Gerhard Paseman Aug 10 '11 at 20:04
up vote 3 down vote accepted

For me, it is a folklore result for people interested in Descartes' rule of signs, even though I am not aware of any paper stating it explicitly. It follows from Descartes' rule of signs.

Theorem (Descartes' rule of signs). The number of positive roots of a real polynomial is not larger than the number of sign changes in the sequence of its coefficients.

You can prove the following generalization:

Lemma. Let $P$ be a degree-$d$ polynomial and $r\in\mathbb R$. The number of sign changes in the sequence of the $P^{(n)}(r)$ for $n=0$ to $d$ is an upper bound on the number of roots of P larger than $r$.

Proof. Let $Q(x)=P(x+r)$. Then $Q^{(n)}(0)/n!$ is the coefficient of $x^n$ in $Q$. Therefore, the number of signs changes in the sequence $(P^{(n)}(r))_n$ equals the number of sign changes in the sequence of coefficients of $Q$. By Descartes' rule of signs, this number of sign changes upper bounds the number of positive roots of $Q$. $\square$

Remark. It is a consequence of Rolle's Theorem since Rolle's Theorem is used to prove Descartes' rule of signs.

share|cite|improve this answer
Didn't Descartes predate Rolle? – Igor Rivin Aug 10 '11 at 21:28
Historically, perhaps. Foundationally, not in my calculus text (the one where analytic geometry is in a separate text). Gerhard "Priority Matters In Dessert Lines" Paseman, 2011.08.10 – Gerhard Paseman Aug 11 '11 at 3:20
You are right, Descartes' rule of signs appears in La Géométrie in 1637, and Rolle proved his theorem in 1691. I should have written "Rolle's Theorem can be used to prove Descartes' rule of signs." What I meant is that those results are very close to each other, thus your two initial comments are right. – Bruno Aug 11 '11 at 8:24
Should the coefficient be represented as $\frac{Q^{(n)}(0)}{n!}$? Regardless, the proof still works. It's been so long since I used it that I forgot about Descartes' Rule of Signs. Thank you very much for all the input. – Mike Aug 11 '11 at 21:36
You're of course right, I forgot to divide by $n!$. – Bruno Aug 12 '11 at 7:49

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.