I am wondering why the mathematic lines and the text in the first minipage are not aligned the same.
I would prefer the text part (below the first math part, equation (89)) to extend to the place where the first line of the equation 89 is located.

Could you help me please?
The code:
\documentclass[a4paper,12pt, reqno]{amsbook}
\usepackage[overlay]{textpos}
\usepackage{graphicx}
\graphicspath{F:/(some path here)/pdfs/}
\usepackage{setspace}
\usepackage{caption}
\usepackage{bm}
\usepackage{cancel}
\newcommand{\e}{\mathbf{e}}
\newcommand{\parn}[1]{\left({#1}\right)}
\newcommand{\p}{\partial}
\newcommand{\strain}{\bm{\varepsilon}}
\newcommand{\brak}[1]{\left[{#1}\right]}
\newcommand{\bff}[1]{\mathbf{#1}}
\begin{document}
\frontmatter
\mainmatter
\noindent
\begin{minipage}[t]{\textwidth}
\begin{minipage}[t]{.58\textwidth}% [t] means top
Then nabla is applied from the left
{\small\begin{align}
\nabla\times\strain\times\nabla &= \parn{\e_1\p_1 + \e_2\p_2 + \e_3\cancel{\p_3}}\times\brak{x_2\e_1\e_3+x_1\e_2\e_3}\notag\\
&=\p_1\parn{x_1}\e_3\e_3 - \p_2\parn{x_2}\e_3\e_3 = \bff{0}\quad \surd
\label{eq:}
\end{align}}
Michell-Cesaro formula; which is a path integral; is path independent for compatible strain fields therefore we take the simplest integration path\;(Fig. 6)
\end{minipage}%<- This is obligate
\hspace{0pt}
\begin{minipage}[t]{.4\textwidth}%
\vspace{0pt}%<- Fake line
\captionsetup{font=scriptsize}
\centering
\includegraphics[width=0.7\textwidth]{sol10c.pdf}%[width=0.48\textwidth]
\captionof{figure}{Inetgration path.}
\label{fig:digraph}
\end{minipage}
\end{minipage}
\end{document}
textpospackage here. And theX_1in the image should be neglected. – Vahid Nov 24 '12 at 16:25\strainand\parnmean. On the other hand it simply seems that the first line in equation 89 is overfull and the unique remedy I see is to make more room for it by enlarging the minipage. – egreg Nov 24 '12 at 16:36\strain,\parn,\p,\e,\brak? How are they defined or what package provides them? – Gonzalo Medina Nov 24 '12 at 16:36