I have a server(fedora 12) with two nics on it, directly connected to say 192.168.1.0 and 192.168.2.0 the route table looks like this

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.1.0     192.168.1.1    255.255.255.0   U     0      0        0  eth0
192.168.2.0     *              255.255.255.0   U     0      0        0  eth1

eth0 = 192.168.1.15
eth1 = 192.168.2.1

and a directly connected user (Mythdora) on the 192.168.2.0 network with ip 192.168.2.2 and route table like so

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.2.0     *               255.255.255.0   U     0      0        0  eth0

the cable is a crossover and it works all three nics work -- connected my laptop to either end and assign it a valid 192.168.2.0 ip the pings will work. In fact if I disconnect the server side and plug the eth cable into the laptop and have thte box ping the laptop continually remove the eth cable and plug it back into the server both sides ping... unfortunately the box realizing it's connected to a different pc wipes its route table after say ten minutes or so.

if I do a trace route from a box on the 1.0 network to the servers 192.168.2.1 interface never get a reply from it.

as a note at one point I could ping the server from the 192.168.2.2 box but the server couldnt ping the 192.168.2.2 box.