I noticed strange behaviour with the newtx package. I'm using MiKTeX 2.9 with TeXnicCenter 2.0 Beta 1 64 Bit with pdflatex.

Here's what I'm talking about: enter image description here

Here's the code:

\usepackage{newtxtext,newtxmath} %\usepackage{txfonts}


d_{Rx}(t) = \begin{cases}
  1, & \text{if } \frac{I_0\left(\sqrt{l_1^2\gamma_1}\right)}{I_0\left(\sqrt{l_0^2\gamma_0}\right)}\exp\left(-\gamma_1+\gamma_0\right)\geq 1,\\
  0, & \text{otherwise}.


Is this normal? I don't think it should be like that. Is anyone else experiencing this? Am I doing something wrong?

  • 1
    If you have access to (or can download) the MathTime Professional 2 (mtpro2) fonts, you won't be experiencing this problem. Note that while the full mtpro2 package isn't free, its "lite" subset, which is all that most people ever seem to use, is in fact free of charge.
    – Mico
    Nov 22, 2013 at 22:34

1 Answer 1


It's hard to believe, but the depth of the digit 0 is 0.12pt (at normal size) in the font used by newtxmath, while it's 0pt for the digit 1. This creates a bigger depth in the contents of the square root in the denominator, which is sufficient for the choice of a bigger square root symbol.

The difference seems tiny, but in your case, due to the presence of the superscripts, the difference in depth turns out to be 0.08758pt.

You can solve the issue, which is also due to being in cases by using \dfrac or a more complicated workaround:





enter image description here

  • I think giving a positive depth to 0 counts as a (minor) bug in whichever font that 0 is from. A more serious bug is the difference in stroke thickness between some sizes of the radicals.
    – Dan
    Nov 23, 2013 at 6:22
  • @Dan I agree. I guess the non zero depth comes from automatic calculations based on the AFM file data.
    – egreg
    Nov 23, 2013 at 9:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .