No one is going to accuse you of plagiarism because you didn't cite EVERY paper on a given topic. That isn't plagiarism. I just did my first official review, and I could probably have listed 20 things off the top of my head they didn't cite that they could have, let alone what I would have turned up if I actually searched - its not a big deal, and I didn't say anthing about it because citing 20 more papers with similar findings in slightly different paradigms wouldn't add anything. At worst, a reviewer will point out a competing viewpoint, or a more thorough investigation that might be a "stronger" citation. Or to a crappy paper on the topic that they wrote and they want you to cite just so they can inflate their own citation count (not joking...it happens).
Everything is inter-connected in this field, and if people really cited ALL relevant research in a remotely developed area, then a typical paper would have 500 pages of references alone, and nothing would ever get done. It is one thing to make assertions without backing them up, ignore opposing viewpoints in the literature, or copy ideas or even entire paragraphs from other papers. The former two will get you rejected, the latter will get you in serious trouble. However, no one is going to bat an eyelash at you citing 3 examples instead of 4. If you want proof, pick a random sampling of articles in a top tier journal. Read through their intros and pick out some parts of it. I guarantee you that if you do a quick lit search on those topics you will find a TON of papers they could have cited but didn't. Obviously, you should be familiar with the general body of literature, but to try and cite everything would be insane. You would spend the rest of your life working on this one paper.
As for your second question, I'm not sure I understand. If you go and read those 3 papers and cite them yourself for a similar point that isn't plagiarism by any stretch of the imagination. References exist primarily so you CAN do things like that! If a paper leads you to a number of other interesting references, there's no need to cite that paper for leading you there. You're not taking their ideas, you're just using references for their intended purpose. You're always better off going back to original sources in those situations, you generally don't want to "cite someone citing something else" unless its a review, synthesis, etc.
Now if you just use the citations and don't read the papers, that is bad. You occasionally see "X as cited in Y, 1972", but generally only for very old or obscure articles that its unreasonable to expect someone to track down the original for. Looking up the original yourself is perfectly acceptable, and there's no need to cite the paper that led you to it unless it is also relevant to the discussion
