Vspace in a table in a two column page

I made a table spanning the top of a two column page, but there is a problem in the last column where $j$ is very close to the hline above, while other elements in the same row are of the same vertical height. The same problem happens to the "transition" header in the last column. And is there a way to make the first row a little wider(the row with $\lambda$ and transition)? How to fix these issues?

\newcolumntype{M}{>{\centering\arraybackslash}m{\dimexpr.096\linewidth-2\tabcolsep}}

\begin{table*}[t!]
\footnotesize
\centering
\begin{tabular}{|| M | M || M | M || M | M || M | M || M | M ||}
\multicolumn{2}{c}{\bf Lyman} & \multicolumn{2}{c}{\bf Balmer} & \multicolumn{2}{c}{\bf Paschen} & \multicolumn{2}{c}{\bf Brackett} & \multicolumn{2}{c}{\bf Pfund} \\
\hline
$\lambda$(nm) & transition & $\lambda$(nm) & transition & $\lambda$(nm) & transition & $\lambda$(nm) & transition & $\lambda$(nm) & transition \\[1mm]
\hline
93.78 & 6$\rightarrow$1 & 410.2 & 6$\rightarrow$2 & 1094 & 6$\rightarrow$3 & 2630 & 6$\rightarrow$4 & 7400 & 6$\rightarrow$5 \\[1mm]
94.98 & 5$\rightarrow$1 & 434.0 & 5$\rightarrow$2 & 1282 & 5$\rightarrow$3 & 4050 & 5$\rightarrow$4 &  & \\[1mm]
97.25 & 4$\rightarrow$1 & 486.1 & 4$\rightarrow$2 & 1875 & 4$\rightarrow$3 &  &  &  & \\[1mm]
102.6 & 3$\rightarrow$1 & 656.3 & 3$\rightarrow$2 &  &  &  &  &  & \\[1mm]
121.6 & 2$\rightarrow$1 &  &  &  &  &  &  &  & \\[1mm]
\hline
$\frac{\lambda_{1,j}}{\lambda_{1,j+1}}$ & $j$ & $\frac{\lambda_{2,j}}{\lambda_{2,j+1}}$ & $j$ & $\frac{\lambda_{3,j}}{\lambda_{3,j+1}}$ & $j$ & $\frac{\lambda_{4,j}}{\lambda_{4,j+1}}$ & $j$ & $\frac{\lambda_{5,j}}{\lambda_{5,j+1}}$ & $j$ \\[4mm]
\hline
1.01 & 5 & 1.05 & 5 & 1.17 & 5 & 1.54 & 5 &  & \\[1mm]
1.02 & 4 & 1.12 & 4 & 1.46 & 4 &  &  &  & \\[1mm]
1.05 & 3 & 1.35 & 3 &  &  &  &  &  & \\[1mm]
1.18 & 2 &  &  &  &  &  &  &  & \\[1mm]
\hline
\end{tabular}
\caption{Expected hydrogen wavelengths, transitions and wavelength ratios.}
\end{table*}


Here is the preamble.
\documentclass[aps,prl,reprint,twocolumn,secnumarabic,nobalancelastpage,amsmath,amssymb,nofootinbib]{revtex4}

        % nofootinbib is another document class option that allows you to put
% footnotes on the page where they occur rather than at the end of the
% paper.  This makes for easier reading!

% secnumarabic is a particularly nice way of identifying sections by
% number to aid electronic review and commentary.

% amsmath and amssymb are necessary for the subequations environment
% among others

\usepackage{graphics}      % standard graphics specifications
\usepackage{graphicx}      % alternative graphics specifications
\usepackage{longtable}     % helps with long table options
\usepackage{url}           % for on-line citations
\usepackage{bm}            % special 'bold-math' package
\usepackage{comment}
\usepackage{fancyhdr}% http://ctan.org/pkg/{fancyhdr,graphicx,lastpage}
\usepackage{color}
\usepackage{atbegshi}
\usepackage{textcomp}
\usepackage{siunitx}
\usepackage{lipsum}
\usepackage{float}
\usepackage{array}
\usepackage{tabularx}   % for 'tabularx' environment
\newcolumntype{C}{>{\centering\arraybackslash}X}
\usepackage{booktabs}   % for '\toprule', '\midrule', etc
\usepackage{amsmath}    % for '\dfrac' macro

\newcommand \widebox [1]{\setlength\fboxsep{6pt} \boxed { #1}}
\newcommand \ev [1]{\left\langle #1 \right\rangle}
\newcommand \evd {\operatorname{E}\evdel}
\newcommand \bbrac [1] {\left ( #1 \right)}
\newcommand \bl {\biggr | }
\newcommand \espa {\\[2.5 mm]}
\newcommand \vbrac [1] {\left [ \,#1 \,\right]}
\newcommand \dpart [2] {\frac{\partial{#1}}{\partial{#2}}}
\newcommand \kai [2] {\chi_{#1}^{\, (#2)}}
\newcommand \ud {\uparrow \,\downarrow}
\newcommand \du {\downarrow\,\uparrow }

\fancyhf{}
\pagestyle{fancy}
\cfoot{\thepage}
\AtBeginDocument{%
\AtBeginShipoutNext{%
\rfoot{}%
}%
}


• m is a tricky beast but you don't need it here as you only have one line in each cell, try with p instead. – David Carlisle Apr 23 '15 at 17:47
• Sorry, I'm not sure which $m$ you are referring to. Can you specify it? – Lawerance Apr 23 '15 at 18:03

I wouldn't use the m column type for this table. In its place, I would use a centered form of the X column provide (provided by the tabularx package), in part to let LaTeX handle the chores of determining column widths. Separately, I'd use the line-drawing macros of the booktabs package to get well-spaced horizontal lines; plus, I'd omit all vertical lines.

The only "tricky" aspect of the table is that whereas the contents of most cells should be centered, the contents of the first 5 data rows of the first column would benefit from being aligned on the decimal separator. This may be achieved by (i) using the S column type (provided by the siunitx package) for the first column and (ii) encasing those items in the first column that should be centered in pairs of curly braces.

Finally, rather than add 1 mm of extra row height manually to each row, I believe it's easier to set the macro \extrarowheight to that value.

\documentclass[twocolumn]{article}
\usepackage{tabularx}   % for 'tabularx' environment
\newcolumntype{C}{>{\centering\arraybackslash}X}
\usepackage{booktabs}   % for '\toprule', '\midrule', etc
\usepackage{amsmath}    % for '\dfrac' macro
\usepackage{siunitx}    % for "S" column type

\begin{document}
\begin{table*}
\setlength\tabcolsep{2pt}  % default value: 6pt
\setlength\extrarowheight{1mm}
\small
\begin{tabularx}{\textwidth}{@{} S[table-format=3.2] *{9}{C} @{}}
\toprule
\multicolumn{2}{@{}c}{\bfseries Lyman} &
\multicolumn{2}{c}{\bfseries Balmer} &
\multicolumn{2}{c}{\bfseries Paschen} &
\multicolumn{2}{c}{\bfseries Brackett} &
\multicolumn{2}{c@{}}{\bfseries Pfund} \\
\cmidrule(r){1-2} \cmidrule(lr){3-4} \cmidrule(lr){5-6}
\cmidrule(lr){7-8} \cmidrule(l){9-10}
\multicolumn{1}{C}{$\lambda$ (nm)} & transition &
$\lambda$ (nm) & transition & $\lambda$ (nm) & transition &
$\lambda$ (nm) & transition & $\lambda$ (nm) & transition \\
\midrule
93.78 & 6$\to$1 & 410.2 & 6$\to$2 & 1094 & 6$\to$3 & 2630 & 6$\to$4 & 7400 & 6$\to$5 \\
94.98 & 5$\to$1 & 434.0 & 5$\to$2 & 1282 & 5$\to$3 & 4050 & 5$\to$4 \\
97.25 & 4$\to$1 & 486.1 & 4$\to$2 & 1875 & 4$\to$3 \\
102.6 & 3$\to$1 & 656.3 & 3$\to$2  \\
121.6 & 2$\to$1 \\
\midrule
{$\dfrac{\lambda_{1,j}}{\lambda_{1,j+1}}$} & $j$ &
$\dfrac{\lambda_{2,j}}{\lambda_{2,j+1}}$  & $j$ &
$\dfrac{\lambda_{3,j}}{\lambda_{3,j+1}}$  & $j$ &
$\dfrac{\lambda_{4,j}}{\lambda_{4,j+1}}$  & $j$ &
$\dfrac{\lambda_{5,j}}{\lambda_{5,j+1}}$  & $j$ \\
\cmidrule(r){1-2} \cmidrule(lr){3-4} \cmidrule(lr){5-6}
\cmidrule(lr){7-8} \cmidrule(l){9-10}
{1.01} & 5 & 1.05 & 5 & 1.17 & 5 & 1.54 & 5 \\
{1.02} & 4 & 1.12 & 4 & 1.46 & 4 \\
{1.05} & 3 & 1.35 & 3  \\
{1.18} & 2 \\
\bottomrule
\end{tabularx}
\caption{Expected hydrogen wavelengths, transitions and wavelength ratios.}
\end{table*}
\end{document}


Addendum The OP has informed me that he/she is using the revtex4 document class, which appears to be incompatible with the booktabs package and its rule-drawing macros. :-(

Here's an alternative solution, which works with the revtex4 document class as well as with many other document classes. It uses a tabular* environment instead of a tabularx environment, and it inserts (typographic) struts at selected locations to obtain a better vertical separation between the lines drawn by \hline and \cline and the material above and below the lines.

I still wouldn't use any vertical lines with this solution, by the way.

\documentclass{revtex4}
%% define a few struts
%% (from code by Claudio Beccari in TeX and TUG News, Vol. 2, 1993)
\newcommand\Tstrut{\rule{0pt}{2.9ex}}       % "top" strut
\newcommand\Bstrut{\rule[-1.2ex]{0pt}{0pt}} % "bottom" strut
\newcommand\TBstrut{\Tstrut\Bstrut}         % "top and bottom" strut

\usepackage{array}      % for '\extrarowheight' macro
\usepackage{amsmath}    % for '\dfrac' macro
\usepackage{siunitx}    % for "S" column type

\begin{document}
\begin{table*}
\setlength\tabcolsep{0pt}  % default value: 6pt
\setlength\extrarowheight{1mm}
\begin{tabular*}{\textwidth}{ S[table-format=3.2] @{\extracolsep{\fill}} *{9}{c} }
\hline
\multicolumn{2}{c}{\bfseries Lyman\TBstrut} &
\multicolumn{2}{c}{\bfseries Balmer} &
\multicolumn{2}{c}{\bfseries Paschen} &
\multicolumn{2}{c}{\bfseries Brackett} &
\multicolumn{2}{c}{\bfseries Pfund} \\
\cline{1-2} \cline{3-4} \cline{5-6} \cline{7-8} \cline{9-10}
{$\lambda$ (nm)\TBstrut} & transition &
$\lambda$ (nm) & transition & $\lambda$ (nm) & transition &
$\lambda$ (nm) & transition & $\lambda$ (nm) & transition \\
\hline
93.78 \Tstrut& 6$\to$1 & 410.2 & 6$\to$2 & 1094 & 6$\to$3 & 2630 & 6$\to$4 & 7400 & 6$\to$5 \\
94.98 & 5$\to$1 & 434.0 & 5$\to$2 & 1282 & 5$\to$3 & 4050 & 5$\to$4 \\
97.25 & 4$\to$1 & 486.1 & 4$\to$2 & 1875 & 4$\to$3 \\
102.6 & 3$\to$1 & 656.3 & 3$\to$2  \\
121.6 \Bstrut & 2$\to$1 \\
\hline
{$\dfrac{\lambda_{1,j}\Tstrut}{\lambda_{1,j+1}\Bstrut}$} & $j$ &
$\dfrac{\lambda_{2,j}}{\lambda_{2,j+1}}$ & $j$ &
$\dfrac{\lambda_{3,j}}{\lambda_{3,j+1}}$ & $j$ &
$\dfrac{\lambda_{4,j}}{\lambda_{4,j+1}}$ & $j$ &
$\dfrac{\lambda_{5,j}}{\lambda_{5,j+1}}$ & $j$ \\
\cline{1-2} \cline{3-4} \cline{5-6} \cline{7-8} \cline{9-10}
{1.01\Tstrut} & 5 & 1.05 & 5 & 1.17 & 5 & 1.54 & 5 \\
{1.02} & 4 & 1.12 & 4 & 1.46 & 4 \\
{1.05} & 3 & 1.35 & 3  \\
{1.18\Bstrut} & 2 \\
\hline
\end{tabular*}
\caption{Expected hydrogen wavelengths, transitions and wavelength ratios.}
\end{table*}
\end{document}

• Great answer! I think I will probably adopt your answer, although I have turned in my lab report. I had a problem with your codes. When I run it separately in a new file, the plot works just as what you showed here. But when I put in the packages and your codes of table in my original codes, all the horizontal lines and vertical lines are missing. I don't understand why it happens. I am putting my preamble in my question so that you can take a look. Also, I am going to attach a current picture that has your program but with some problems. Thanks! – Lawerance Apr 24 '15 at 4:04
• I checked it was something incompatible with my documentclass, which is revtex4. – Lawerance Apr 24 '15 at 4:16
• @Lawerance - AFAICT, the revtex4 class is incompatible with the booktabs package. :-( I've posted an addendum to provide a solution that should work fine with the revtex4 documenet class. – Mico Apr 24 '15 at 5:05
• You are amazing. Thanks for the detailed explanation and improvement. Have a nice day! – Lawerance Apr 24 '15 at 6:43
• @Mico - revtex is incompatible with booktabs, but caters for alternatives: \toprule, \colrule and \botrule. I suggest the OP tries them first. – Arash Esbati Apr 24 '15 at 7:24

David is referring to the m-specifier in

\newcolumntype{M}{>{\centering\arraybackslash}m{\dimexpr.096\linewidth-2\tabcolsep}}


How about a layout like this:

\documentclass[a4paper,10pt]{article}
\usepackage{tabularx,booktabs,geometry}
\geometry{textwidth=15cm}

\begin{document}
\noindent
\footnotesize
\begin{tabularx}{\linewidth}{@{}*{5}{>{\centering}p{.9cm}>{\hfil}X<{\hfil}}@{}}
\toprule
\multicolumn{2}{@{}l}{Lyman} & \multicolumn{2}{l}{Balmer} &
\multicolumn{2}{l}{Paschen}  & \multicolumn{2}{l}{Brackett} &
\multicolumn{2}{l}{Pfund} \\\cmidrule(r){1-2}%
\cmidrule(lr){3-4}\cmidrule(lr){5-6}%
\cmidrule(lr){7-8}\cmidrule(l){9-10}
$\lambda$(nm) & transition & $\lambda$(nm) & transition &
$\lambda$(nm) & transition & $\lambda$(nm) & transition &
$\lambda$(nm) & transition \\\cmidrule(r){1-1}%
\cmidrule(lr){2-2}\cmidrule(lr){3-3}\cmidrule(lr){4-4}%
\cmidrule(lr){5-5}\cmidrule(lr){6-6}\cmidrule(lr){7-7}%
\cmidrule(lr){8-8}\cmidrule(lr){9-9}\cmidrule(l){10-10}
%%
93.78 & 6$\rightarrow$1 & 410.2 & 6$\rightarrow$2 &
1094  & 6$\rightarrow$3 & 2630  & 6$\rightarrow$4 &
7400  & 6$\rightarrow$5                               \\
94.98 & 5$\rightarrow$1 & 434.0 & 5$\rightarrow$2 &
1282  & 5$\rightarrow$3 & 4050  & 5$\rightarrow$4 &   \\
97.25 & 4$\rightarrow$1 & 486.1 & 4$\rightarrow$2 &
1875  & 4$\rightarrow$3 &                             \\
102.6 & 3$\rightarrow$1 & 656.3 & 3$\rightarrow$2 &   \\
121.6 & 2$\rightarrow$1 \\ \cmidrule{1-10}
%%
$\frac{\lambda_{1,j}}{\lambda_{1,j+1}}$ & $j$ &
$\frac{\lambda_{2,j}}{\lambda_{2,j+1}}$ & $j$ &
$\frac{\lambda_{3,j}}{\lambda_{3,j+1}}$ & $j$ &
$\frac{\lambda_{4,j}}{\lambda_{4,j+1}}$ & $j$ &
$\frac{\lambda_{5,j}}{\lambda_{5,j+1}}$ & $j$ \\ \cmidrule{1-10}
1.01 & 5 & 1.05 & 5 & 1.17 & 5 & 1.54 & 5  \\
1.02 & 4 & 1.12 & 4 & 1.46 & 4             \\
1.05 & 3 & 1.35 & 3                        \\
1.18 & 2                                   \\
\bottomrule
\end{tabularx}
\end{document}