Argthtjtr

I have the following matrix

usepackage{amsmath}

usepackage{amssymb}



$$

M=begin{bmatrix}

frac{partial f_1}{partial x_1}(p) & frac{partial f_1}{partial x_2}(p) & cdots & frac{partial f_1}{partial x_n}(p) \

 frac{partial f_2}{partial x_1}(p) & frac{partial f_2}{partial x_2}(p) & cdots & frac{partial f_2}{partial x_n}(p) \

$vdots$ & $vdots$ & $vdots$ & $vdots$ \

 frac{partial f_m}{partial x_1}(p) & frac{partial f_m}{partial x_2}(p) & cdots & frac{partial f_m}{partial x_n}(p) \



end{bmatrix}$$

which complies nicely. However, in the third line of the matrix, I would like to change the third "vdots" to a "ddots". If I do so, I get the following error:

LaTex Warning: No author given.

(The directory on my computer)

!Missing $ inserted.

< inserted text >

$

I.72 $vdots$ & $vdots$ & $ddots

$ & $vdots$

?

Could anybody help please?

Curiously enough, vdots is also legal in text mode, because it is defined as

DeclareRobustCommand{vdots}{%

  vbox{%

    baselineskip4p@ lineskiplimitz@

    kern6p@hbox{.}hbox{.}hbox{.}%

  }%

}

This fact should not be exploited, better use vdots only in math mode.

On the contrary, ddots is surrounded by mathinner, so it has to appear in math mode:

DeclareRobustCommand{ddots}{%

  mathinner{%

    mkern1mu

    raise7p@vbox{kern7p@hbox{.}}%

    mkern2mu

    raise4p@hbox{.}%

    mkern2mu

    raisep@hbox{.}

    mkern1mu

  }%

}

What happens with your input can be understood once you realize that bmatrix internally uses array, where each cell implicitly has $ at the beginning and $ at the end. Thus in

 $vdots$

the first $ exits from math mode, producing an empty formula; then vdots is typeset in text mode and finally another empty formula is produced from the trailing $ and the implicit $ at the end.

With $ddots$ it is the same, but now TeX finds mathinner in text mode and raises an error.

How to cure this? Just don't use $ inside bmatrix or anything that's already in math mode.

Caveat 1. I've seen several pieces of MathJax code where $ seems to be used as a mean to produce “weird” symbols: something like $alpha$ = $beta$ or worse. The code should be $alpha=beta$, of course.

Caveat 2. Never use $$ in a LaTeX document. See Why is [ ... ] preferable to $$ ... $$?

You may also want to increase the vertical spacing between the first two rows.

documentclass{article}



usepackage{amsmath}

usepackage{amssymb}



begin{document}



begin{equation*}

M=begin{bmatrix}

  frac{partial f_1}{partial x_1}(p) &

  frac{partial f_1}{partial x_2}(p) &

  cdots &

  frac{partial f_1}{partial x_n}(p) \[1.5ex]

  frac{partial f_2}{partial x_1}(p) &

  frac{partial f_2}{partial x_2}(p) &

  cdots &

  frac{partial f_2}{partial x_n}(p) \

  vdots & vdots & ddots & vdots \

  frac{partial f_m}{partial x_1}(p) &

  frac{partial f_m}{partial x_2}(p) &

  cdots &

  frac{partial f_m}{partial x_n}(p)

end{bmatrix}

end{equation*}



end{document}

Once you use [...], you're in display math mode and therefore don't need to supply in-line $...$ around math content:

documentclass{article}



usepackage{amsmath,amssymb}



begin{document}



[

  M = begin{bmatrix}

    frac{partial f_1}{partial x_1}(p) & frac{partial f_1}{partial x_2}(p) & cdots & frac{partial f_1}{partial x_n}(p) \

    frac{partial f_2}{partial x_1}(p) & frac{partial f_2}{partial x_2}(p) & cdots & frac{partial f_2}{partial x_n}(p) \

                 vdots                  &              vdots                  & ddots &              vdots                  \

    frac{partial f_m}{partial x_1}(p) & frac{partial f_m}{partial x_2}(p) & cdots & frac{partial f_m}{partial x_n}(p)

  end{bmatrix}

]



end{document}