In general, many Tracks go to Buses, while Buses go to the Master, which then goes to actual the physical output. You send the Tracks to Buses to treat similar instruments, well, similarly. So backing vocals go to the same bus where you can add the same EQ, compression and even reverb so they jell (a pseudo-technical term) together. Same for the guitars. Same for the drum. Other tracks can go directly to the output bus. This methodology developed from recording consoles, which was the way to wire actual analog hardware. The engineer would, for example, send all the drums through a buss w/ a pair of 1176's on it to punch up the entire kit, or guitars through the SSL bus comp on an SSL board. And most studios didn't have an unlimited number of hardware effect units, so it was a bang for buck way to get the most out of hardware.
The same goes for sends. Sends developed as a way to "tap off" the signal and send it somewhere else - like monitor sends for live music or headphone sends in the studio. For mixing, it is a way to put another level of treatment on a sound and share it between several sources. Say a nice reverb. You could slap it on the guitar bus, but your lead guitar is going to be louder than your rhythm and the more reverb it generates make it seem more distant, while the rhythm guitar sounds almost dry. Make a send buss (which feeds back into the guitar buss), slap the reverb on it, and you can send different amounts of the dry signal to the reverb buss, so the rhythm guitar sounds all psychedelic while the lead is dryer and comes to the front of the sound stage. Again, this development was prodded by the fact you didn't usually have a couple of lexicon reverbs just waiting to be used, so if you could make one do double-duty that was good.
It just adds another layer of flexibility. When you are just starting out mixing, you probably won't use it esp. on basic singer-songwriter stuff or even rocknroll. Some of it you will incorporate into your workflow, and most importantly use as a problem solver at times once you understand it. In my guitar example above, you can spend a lot of time deciding whether to bring the reverb sound back into the guitar buss, back to the main, or put it on another buss with different compression/EQ. Not that it will change the sound much, but it is interesting to figure out. You probably won't do that if you are getting paid by the project (unless you want it to eat into your hourly wage), but if you are getting paid for absolute sound quality or doing it for your own project on your own damn time why not?
@