\documentstyle[paasms4,psfig,pptwocol]{preprint} \input pub.sty \begin{document} \title{THE TEMPERATURE OF EXTENDED GAS IN ACTIVE\\ ~\\ GALAXIES---EVIDENCE FOR MATTER-BOUNDED CLOUDS} \author{A.~S.\ Wilson\/\thanks{Also Astronomy Department, University of Maryland, College Park, MD 20742}\morethanks{Visiting Astronomer, Cerro Tololo Inter-American Observatory, operated by the Association of Universities for Research in Astronomy, Inc., under contract with the National Science Foundation}\\ \\ Space Telescope Science Institute\\ 3700 San Martin Drive, Baltimore, MD 21218\\ \and L.\ Binette\/\thanks{Also European Southern Observatory, Casilla 19001, Santiago 19, Chile}\\ \\ Instituto de Fisica, UFRGS\\ Campus do Vale, CxP 15051, 91501-970 Porto Alegre - RS, Brazil\\ \and T.\ Storchi-Bergmann\/\samethanks{2}\\ \\ Instituto de Fisica, UFRGS\\ Campus do Vale, CxP 15051, 91501-970 Porto Alegre - RS, Brazil} \tobe{20 June 1997}{The Astrophysical Journal (Letters)} \recacc{27 January 1997}{31 March 1997} \maketitle \abstract{ We report measurements of the electron temperature at about a dozen locations in the extended emission-line regions of five active (Seyfert and radio) galaxies. Temperatures (T$_{[OIII]}$ and T$_{[NII]}$) have been determined from both the I(\mbox{[OIII]}$\lambda$4363)/I(\mbox{[OIII]}$\lambda$5007) and I(\mbox{[NII]}$\lambda$5755)/I\mbox{[NII]}$\lambda$6583) ratios. T$_{[OIII]}$ lies in the range (1.0--$1.7)\times10^{4}$~K. We find a strong trend for T$_{[OIII]}$ to be higher than T$_{[NII]}$, with the difference typically being $\approx5,000$~K. Because the critical density for collisional de-excitation of the $^{1}$D${_2}$ level in NII is lower than that of the same level in OIII, the deviations of the measured intensity ratios from those expected for T$_{[OIII]}$ = T$_{[NII]}$ in the low density limit are unlikely to result from collisional de-excitation. The measured values of T$_{[OIII]}$ and the differences between T$_{[OIII]}$ and T$_{[NII]}$ are very similar to those found in Galactic planetary nebulae. It is argued that the dominant form of energy input to the clouds is photoionization, but detailed modelling indicates that the temperature difference is too large to be accounted for in terms of photoionization of ionization-bounded clouds. We propose instead that both matter- and ionization-bounded clouds are present in the extended emission-line regions, with most of the \mbox{[OIII]} emission originating from a hot zone in the matter-bounded clouds and essentially all of the \mbox{[NII]} from the ionization-bounded clouds.}