I thought the different sized gaps may be related to paragraphs. But now I am wondering if those paragraphs identified in the OCR process are interfering in the indexing when a paragraph has erroneously been placed between two lines.
The other thing that puzzles me about this article is that the full article title is duplicated in the title provided (by the "vendors") for the article on the page display, so why is the search for "leaving examination" not finding it there, if the paragraphs are indeed the cause of the search problem. Or is it possible that "RESULTS OF THE LEAVING" has been identified by the OCR process as the main title and "EXAMINATION" identified as the subtitle ??