# How to properly typeset multiletter quantities, e.g. Re or Nu (Reynolds, Nusselt numbers)

Is there a canonical way to typeset such quantities in equations and inline math?

\documentclass{article}
\begin{document}
$Re$, $Nu$ are clearly not an option.
$\mathit{Re}$, $\mathit{Nu}$ appear ok to me.
\end{document}

• I'm never happy with this issue. Italics make it too trivial and hard to read as it might mean R*e, upright makes it operator. Lately I decided to switch to Real and Imag for the complex numbers and stick to the upright version. Not that it is the optimal but best of the worst in my opinion. – percusse Aug 21 '13 at 14:12
• although the question is about typesetting, it's also highly "area" dependent. a mathematics or physics forum would have more participants knowledgeable about the area. – barbara beeton Aug 21 '13 at 14:34
• My copy of of the classic text by Schlichting, "Boundary-Layer Theory" (McGraw Hill 1979, 1st ed. 1955), has them in sans serif: $\mathsf{R} = \rho VD / \mu$, and avoiding the issue altogether with a single letter notation. – Steven B. Segletes Sep 5 '13 at 10:18
• The 10th (last) edition (in German) uses $\mathrm{Re}$ or $\text{Re}$. I think we wont be able to avoid the use of two-letter-symbols for that nowadays. – LaRiFaRi Sep 5 '13 at 10:28
• @LaRiFaRi Please, never ever use \text for this sort of thing! It does crazy things. As well, I would prefer \mathup since that works expectedly in beamer for instance. – yo' Sep 5 '13 at 11:43

When in doubt, I normally consult Knuth's publications (if I expect to find a paper that is likely to have an example of the issue at hand) and secondly the relevant Journal author guidelines (in that order).

Sometimes the latter will have some badly made templates, but most probably would have strict and stroppy editors. One caveat is that if you aiming at eventually publishing the document the editor is likely to have the final say and I would suggest it is unwise to get into an argument about typography with the editor.

I guess that the OP is dealing with Fluid Mechanics. The Journal of Fluid Mechanics has instructions as to how to typeset these dimensionless numbers, including a template and instructions.

Since these are essentially numbers I would recommend the approach that they are typeset in math italic font. It also looks better if you are describing any of these numbers inline.

The above cited journal has them defined as:

\newcommand\Rey{\mbox{\textit{Re}}}  % Reynolds number
\newcommand\Pran{\mbox{\textit{Pr}}} % Prandtl number, cf TeX's \Pr product
\newcommand\Pen{\mbox{\textit{Pe}}}  % Peclet number


which, I agree with egreg is not very wise. However, since they provide the .cls is not the end of the world and the template works.

Personally I recommend you use the \DeclareMathOperator from amsmath (if you worry about spacing-personally I wouldn't). Also my own preference is to use a notation such as, N_{\mathrm{Re} rather than a double symbol such as Re.

As a final word, I think consistency is the key and that you should use the notation that is most familiar with your readers.

• JFM is definitely a good reference. – AlexG Mar 13 '14 at 14:12

Update

I collected the ideas in the comments and answers and wrote myself some macros for the numbers I need. Maybe anybody else wants to use them. Works fine with sub- and superscripts and in multiplications a.s.o.

\documentclass{article}
\usepackage{mathtools}
\usepackage[%
,math-style=ISO
,bold-style=ISO
,sans-style=italic
]{unicode-math}

\newcommand{\Arch}{\operatorname{\mathit{A\kern-.06em r}}} % http://de.wikipedia.org/wiki/Archimedes-Zahl
\newcommand{\Biot}{\operatorname{\mathit{B\kern-.06em i}}} % http://de.wikipedia.org/wiki/Biot-Zahl
\newcommand{\Cauc}{\operatorname{\mathit{C\kern-.07em a}}} % http://de.wikipedia.org/wiki/Cauchy-Zahl
\newcommand{\Damk}{\operatorname{\mathit{D\kern-.06em a}}} % http://de.wikipedia.org/wiki/Damk%C3%B6hler-Zahl
\newcommand{\Eule}{\operatorname{\mathit{E\kern-.03em u}}} % http://de.wikipedia.org/wiki/Euler-Zahl
\newcommand{\Four}{\operatorname{\mathit{F\kern-.10em o}}} % http://de.wikipedia.org/wiki/Fourier-Zahl
\newcommand{\Frou}{\operatorname{\mathit{F\kern-.07em r}}} % http://de.wikipedia.org/wiki/Froude-Zahl
\newcommand{\Gras}{\operatorname{\mathit{G\kern-.05em r}}} % http://de.wikipedia.org/wiki/Grashof-Zahl
\newcommand{\Karl}{\operatorname{\mathit{K\kern-.11em a}}} % http://de.wikipedia.org/wiki/Karlovitz-Zahl
\newcommand{\Knud}{\operatorname{\mathit{K\kern-.11em n}}} % http://de.wikipedia.org/wiki/Knudsen-Zahl
\newcommand{\Lewi}{\operatorname{\mathit{L\kern-.05em e}}} % http://de.wikipedia.org/wiki/Lewis-Zahl
\newcommand{\Mach}{\operatorname{\mathit{M\kern-.10em a}}} % http://de.wikipedia.org/wiki/Mach-Zahl
\newcommand{\Nuss}{\operatorname{\mathit{N\kern-.09em u}}} % http://de.wikipedia.org/wiki/Nusselt-Zahl
\newcommand{\Pecl}{\operatorname{\mathit{P\kern-.08em e}}} % http://de.wikipedia.org/wiki/P%C3%A9clet-Zahl
\newcommand{\Pran}{\operatorname{\mathit{P\kern-.03em r}}} % http://de.wikipedia.org/wiki/Prandtl-Zahl
\newcommand{\Rayl}{\operatorname{\mathit{R\kern-.04em a}}} % http://de.wikipedia.org/wiki/Rayleigh-Zahl
\newcommand{\Reyn}{\operatorname{\mathit{R\kern-.04em e}}} % http://de.wikipedia.org/wiki/Reynolds-Zahl
\newcommand{\Schm}{\operatorname{\mathit{S\kern-.07em c}}} % http://de.wikipedia.org/wiki/Schmidt-Zahl
\newcommand{\Sher}{\operatorname{\mathit{S\kern-.07em h}}} % http://de.wikipedia.org/wiki/Sherwood-Zahl
\newcommand{\Stro}{\operatorname{\mathit{S\kern-.07em r}}} % http://de.wikipedia.org/wiki/Strouhal-Zahl
\newcommand{\Webe}{\operatorname{\mathit{W\kern-.14em e}}} % http://de.wikipedia.org/wiki/Weber-Zahl

\begin{document}
\begin{align*}
\Arch &=\frac{\increment\rho g L^3}{\rho\nu^2} \\
\Biot &= \frac{\alpha \cdot L}{\lambda_\mathup{s}} \\
\Cauc &= \frac{\rho \cdot \omega^2l^2}{E} \\
\Reyn &\approx \Damk^2\Karl^2 \\
\Lewi &= \frac{\Schm}{\Pran} \\
\end{align*}
\end{document}


This yields e.g.:

Please feel free to edit my kerning or to add more numbers!

The book Detailtypografie from Forssman and de Jong explains:

Symbols for physical (and technical) values are set italic. Likewise characteristics (dimensionless numbers), whose symbols consist of several letters, are set italic (here the text italic must be taken).

Loosely translated by me... sorry for bad English. Examples given by them are Eu, Re, Fr, Sr, Ma, and We.

Translated to LaTeX (the Math-chapter has been set in LaTeX by Johannes Küster), this would mean \textit{...}. Or am I wrong?

The problem however is, that the text italic changes with the surrounding font. I don't like this behavior, as it is hard to read sans serif symbols in serif formulas. For my taste, a symbol should stay the same in the whole document.

Then maybe I should use \mathit{...}? The problem is, there is no way to distinguish this from two multiplied variables.

DIN EN ISO 80000 says:

Symbols for characteristics, like the Mach-number, symbol Ma, are set with two letters from the Latin alphabet, always with a large initial letter. It is recommended that such two-letter-symbols are separated by a space to other symbols, when appearing in a multiplication.

I have set some versions I have seen until know. I also saw a calligraphic Re or a Ma with reduced distance between M and a. But can't find those right now.

The distance between indices and symbol is differing between math italic and text italic. Don't know, what would be better.

This is not really an answer but an appeal to everybody, that there are some more thoughts needed. Seems like there is still no real solution around.

• As you observe, \textit is out of the question, because it inherits the font attributes of the font which is current outside math, so for example it could become bold italic. So I'd go for \mathit. The recommendation to detach these from adjacent variables translates into something like \newcommand{\Reyn}{\operatorname{\mathit{Re}}} – egreg Sep 5 '13 at 10:31
• yeah, me too, but how about the ISO rule of bigger distances to other symbols? Maybe together with setting M and a closer in Mach... Just ideas. – LaRiFaRi Sep 5 '13 at 10:33
• The "operator" nature of \Reyn as defined will produce thin spaces in $a\Reyn b$: try it. ;-) – egreg Sep 5 '13 at 10:35
• Oh cool. That works indeed. Would you do an answer from that? A cool option would be to reduce the distance a little bit. And a super cool extra, don't know how, to get rid of the little "R" tail disturbing for the Reynolds- or Rayleigh-number. – LaRiFaRi Sep 5 '13 at 10:50
• You can make Re a mathrel to get that spacing though it would have some minor disadvantages. – percusse Sep 5 '13 at 12:41

That depends on your tradition and what these are actually used for? If they are operators, I'd define them as such using \DeclareMathOperator (from amsmath). If it is just a name I'd use \mathrm or \mathit depending on tradition.

But I would also define macros for each of these such that I never write $\mathrm{Re}$ in the text.

BTW, if Re is just the real part operator, and you do not like the default look of \Re (most don't) just use

\renewcommand\Re{\operatorname{Re}}


as \Re should behave as an operator

• They aren't operators but act as variables (like x, y, z, t) taking values. Therefore I'd go for \mathit{...}. Thus, Reynolds number cannot be misinterpreted as the real part operator. – AlexG Aug 21 '13 at 13:49

It looked to me like the consensus in the literature is upright - to my surprise: Munson and Kundu use upright Re, as does Wikipedia. So \mathrm{Re} etc. in formulae.

But Torbjørn T. provided some italic references, which make more sense to me. My general point still holds - whatever you do, be clear.

The first time you introduce a variable (that isn't completely obvious in context/field) you should define it. Typesetting "Re" or "Ma" may mean that it's not as obvious as it might be - so define it in text: "where Ma is the Mach number of...". You may have to avoid ambiguity by using "Real()" rather than "Re()" if using complex variables as well. A reduced space between characters could be nice, but the reduction may have to be quite small (and therefore not very clear), and if it's not standard, wouldn't actually add clarity.

Apart from \Re, \Im you want to avoid confusion with chemical symbols.

• Thanks for your answer. Introducing variables is good practice, for sure, but I am very strict with the rule: "italic for every variable" (greeks, bolds, ...). – LaRiFaRi Sep 5 '13 at 10:55
• @LaRiFaRi, I'm less strict on the rule than on deferring to the literature. I tried to find an older work using italics to support their use (and my preference), but the look inside on Amazon skipped over those pages, and I'm too lazy to walk upstairs to the library! – Chris H Sep 5 '13 at 10:57
• The consensus in my bookshelf is italics. This includes Thorpe, Houghton, and (preprint of) Cushman-Roisin and Beckers. – Torbjørn T. Sep 5 '13 at 11:14
• @TorbjørnT. Good - I like Houghton, but it's at home. Maybe "pure" fluid mechanics uses upright and italics are preferred in a more applied sense? Probably I'm reading too much into it. – Chris H Sep 5 '13 at 11:16
• Could be, I don't know. I also have Stull which uses upright, but that has all math upright ... – Torbjørn T. Sep 5 '13 at 11:25