It looks like the ones showing up in the RSWho browse are the ones that also have good I/O connections; I can't be sure but it looks like the 22nd one in the list has a good connection, and the first drive in the browse is 192.168.0.122.
So it could be as ASF suggested that those are the only ones with EtherNet/IP properly enabled.
I use two free network tools for stuff like this:
TCPING and
Angry IP Scanner.
Both let you query a device using a TCP connection instead of the blunt instrument of an ordinary ICMP PING.
EtherNet/IP devices will always allow a TCP connection request on Port 44818. If a device doesn't answer on TCP Port 44818, then it doesn't have an EtherNet/IP protocol stack running.
I usually also check for TCP Port 502 (Modbus TCP) and TCP Port 34964 for ProfiNet RPC (Profinet doesn't really use TCP for IO, so I'm less familiar with how effective that is).
I have seen some devices that can only establish one messaging connection. This was common in low-powered DeviceNet sensors, but I've seen it in EtherNet/IP as well.
So if you have a PLC trying to connect to it but it hasn't established a cyclic I/O connection yet (like it's got a mismatch or configuration issue), the device is blind to PC-based configuration tools.
See if the RSLinx browse changes when you disconnect the ControlLogix from the network.