In more detail
A PDF starts with %PDF- and a version number and ends with %%EOF. In between are numbered objects (pages, fonts, images, content streams), a cross-reference table giving each object's byte position so a reader can jump to any page, and a trailer pointing to the document root. The current version is PDF 2.0 (ISO 32000-2); PDF/A is a restricted variant for long-term archiving. Merging joins the page lists of several files, splitting extracts a range, and numbering stamps a number on each page.
Why it matters
Edits are often saved as incremental updates appended to the end of the file, so earlier versions of the content can remain inside it. Drawing a black box over text does not remove the text underneath: it can still be selected or extracted, so use a proper redaction tool. PDFs also carry Metadata such as author and creating software. The Document Forensics page examines these traces.
Example
The skeleton of a PDF, as seen in a text editor:
%PDF-1.7
1 0 obj
<< /Type /Catalog /Pages 2 0 R >>
endobj
...
startxref
...
%%EOF
2 0 R is a reference to object 2, the page tree.