I know that this post may sound trivial for some of the readers, but, there are people (and not only students or unexperienced researchers, but even PhDs, who claim to have some research experience) who still do not know (or know, but don't use) the stuff that follows (and I can tell you this from my own experience).

Every paper should have a References section. In this section you usually place the papers where the results you used in your work appeared first. However, it is important not only to put a paper in such a section, but to cite it accordingly: whenever a result appears in the paper it should be accompanied by a link to the paper where it was published; otherwise someone may not be able to identify correctly the source of that result, or, even think that the result belongs to you. Also, for each result you should always cite its original source: it is rather frustrating for an author to see that someone uses his/her results but doesn't cite the papers in which these results were introduced, citing instead some other papers (maybe of some other authors) in which the results were only used. Clearly the latter case may be a consequence of the first one: once a result or a definition is used but not linked to its original source, someone may not be able to find out in which paper that result/concept was introduced, thus may not be able to cite it correctly. Finally, all these rules apply for self-citing: even if, at some moment, you use a result you introduced in a previous paper, cite that paper, don't write that result as if you introduce it at that moment.

As a reviewer I apply the following policy: using, in a paper, a result without citing it leads to reject decision on the paper; using a result from an article placed in the References list, but without specifying explicitly in the text the link between that result and the article from the References, or citing the wrong source of a result, leads to a major revision decision or a reject and resubmit decision (I am influenced in my decision by the general merits of the paper). Of course, in every case I try to find and indicate to the author the original paper where the used result was published. Finally, a special case arises when the author cites a handbook or a survey for some of the results used (usually results seen as common knowledge); in this case, I usually try to find the correct reference (which in this case may be more difficult) and suggest the author to use it, while the decision is minor revision.