Mathematica Stack Exchange is a question and answer site for users of Mathematica. 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'm trying to compute the eigenvectors for:

$$ M = \left( \begin{array}{ccc} 1 & 4 \\ 4 & 100 \end{array} \right) $$

Both myself and Mathematica report the eigenvalues as:

$$ \lambda_1 = \frac{1}{2} \left(101+\sqrt{9865}\right) \approx 100.161 \\ \lambda_2 = \frac{1}{2} \left(101-\sqrt{9865}\right) \approx 0.838647 $$

But when I ask for eigenvectors, the answer changes depending on whether I input the numbers as integers or floating-point numbers.

When I ask for Eigenvectors[{{1, 4}, {4, 100}}] I get:

$$ v_1 = \left(\frac{1}{8} \left(-99+\sqrt{9865}\right), 1\right) \approx (0.0403383, 1) \\ v_2 = \left(\frac{1}{8} \left(-99-\sqrt{9865}\right), 1\right) \approx (-24.7903, 1) $$

When I ask for Eigenvectors[{{1.0, 4.0}, {4.0, 100.0}}] I get:

$$v_1 \approx (0.0403055, 0.999187) \\ v_2 \approx (-0.999187, 0.0403055)$$

When I calculate by hand, I get a solution which matches the first query.

So, am I going crazy and overlooking some important maths (not unlikely, very tired...) or is this a bug?

Possible related issues?:

share|improve this question
1) We do not discuss Wolfram|Alpha questions/issues here. They're explicitly off-topic. 2) The bugs should be used only if the behaviour has been confirmed by the community as anomalous and a bug. – R. M. Mar 21 '14 at 22:31
Normalise/@... on exact case, will get you same results as inexact case. – ciao Mar 21 '14 at 22:35
This is what Mathematica does. I don't see Alpha referenced int the question. – Mark Adler Mar 21 '14 at 22:35
@MarkAdler Did you follow the "query" links? :) – R. M. Mar 21 '14 at 22:40
This question appears to be off-topic because it appears to be concerned with the behavior of Wolfram|Alpha. – m_goldberg Mar 21 '14 at 23:21
up vote 4 down vote accepted

Eigenvectors for inexact arguments are normalized:

Eigenvectors[{{1, 4}, {4, 100}}]
% // N
Normalize /@ %% // N
Eigenvectors[{{1.0, 4.0}, {4.0, 100.0}}]

{{1/8 (-99+Sqrt[9865]),1},{1/8 (-99-Sqrt[9865]),1}}



share|improve this answer

Eigenvectors are just a direction, so the ones you have are equivalent up to a multiplicative factor. Both answers are correct. The second one is normalized to length 1.

share|improve this answer

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.