Certain summation and transformation formulas for generalized hypergeometric series (Q2271981)

The main result may be stated as follows. Let \(n\in\mathbb{N}_0\), \(m_1,\dots, m_r\in\mathbb{N}\), and \(m= m_1+\cdots+ m_r\). Then \[ {_{r+2}F_{r+1}}\left[\begin{matrix} -n,a,f_1+ m_1,\dots, f_r+ m_r\\ b,f_1,\dots, f_r\end{matrix}\,\Biggl|\,1\right]= {(b- a- m)_n\over(b)_n} \prod^m_{j=1} {\xi_j+ n\over\xi_j}, \] provided that \((b- a- m)_m\neq 0\) and \(a\neq f_i\), \(i\in\{1,\dots, r\}\); the numbers \(\xi_1,\dots,\xi_m\) are zeros of a certain polynomial. Some particular cases are studied. Application of the theorem to Kampé de Fériet functions leads to a reduction formula for a special \(F^{q:r+1;0}_{q:r+1;0}[-y, y]\) in terms of \({_{p+r+1}F_{q+r +1}}[y]\), involving the numbers \(\xi_1,\dots,\xi_r\). A special case of the latter formula is \[ {_{r+1}F_{r+1}}\left[\begin{matrix} a,f_1+1,\dots, f_r+ 1 f_1,\dots, f_r\end{matrix}\,\Biggl|\,y\right]= e^q_{{r+1}F_{r+1}}\left[\begin{matrix} b-a-r,\xi_1+ 1,\dots, \xi_r+1\\ b,\xi_1,\dots, \xi_r\end{matrix}\,\Biggl|\,- y\right], \] which generalizes Kummer's first transformation formula for \({_1F_1}\).