Neuroanatomical　connectivity　and　functional　connectivity　are　to　be　studied　for　understand　the　mechanism　of　neural　systems　accepting　the　sensory　inputs　and　combining　different　information.　We　examined　the　structural　features　of　three　mammalian　cerebral　cortex　networks　and　a　number　of　randomized　control　networks　expressed　as　graphs　and　patterns　of　functional　connectivity　to　which　they　give　rise　when　implemented　as　dynamic　systems.　We　found　that　the　cerebral　cortex　of　macaque,　cat　and　rat　have　smaller　characteristic　path　length　and　diameter　of　graph　but　higher　reciprocal　fraction　and　cluster　index　in　structure　connectivity,　which　shows　features　characteristic　of　small-world　networks.　At　the　same　time,　the　cerebral　cortex　networks　have　higher　entropy　and　complexity　in　function　connectivity　compared　with　the　same　size　and　density　random　networks.　Multidimensional　Scaling　analysis　showed　that　the　cerebral　cortex　areas　with　similar　functions　connect　with　each　other　more　closely.　©　2008　IEEE.