DNA of Marsh and Long Ancestors

This article discusses the DNA shared by Cousin Chris G, who reached out to me as a descendant of Eliakim Marsh (1816-1881) and Susan Long (1819-1882) of Westfield, (now in Union County), New Jersey.

Chris G's father is a third cousin, once removed to my father and his siblings.  As we approach the third cousin level, DNA may or may not be shared.  We checked for shared DNA at GedMatch.com so that we could see all the shared segments.

One of my uncles shares only a tiny segment (3 cM) of DNA with Chris G's father.  This would not have been reported as a match at the three major testing companies (23andMe, FamilyTreeDNA, and AncestryDNA), but GedMatch allows you to see tiny segments.

When we look at segments above 5 cM, the other siblings share larger segments with this third cousin:
My father shares two segments totaling 42.3 cM.
My aunt shares two segments totaling 46.9 cM.
My uncle shares one segment of 17.7 cM.

Next we looked at other people who also share these same segments.  Anyone who matches my father and Chris G's father on the same segment will descend from Eliakim Marsh and Susan Long, or one of the ancestors of Eliakim or Susan.

My father and Chris G's father share a segment of 25 cM on chromosome 5.  Someone who tested at FamilyTreeDNA matched both men on this same spot with a slightly smaller segment of 15 cM.

We had to travel back in this distant DNA cousin's tree many generations until we were in New Jersey.  The common ancestors are the 3rd great grandparents of Eliakim Marsh:  John Marsh (1661-1744) and Elizabeth Clarke (1664-1739).  They lived in Rahway and Elizabeth.

But- we may also have Denman ancestors in common.  I have not confidently traced back beyond Eliakim Marsh's great grandfather, Philip Denman.  This distant cousin also has Denman ancestors in Westfield, New Jersey.

Plus, Eliakim's mother, Abigail Willis, is another tail.  She could share some ancestral lines with her husband.

We may end up with a situation seen with the Morris County DNA cousins, where we share multiple lines of ancestry and cannot isolate the DNA to a particular ancestor.

My question is:  would this DNA cousin, who is probably a seventh or eighth cousin through these Marsh or Denman lines, share a segment of DNA 15 cM long?  Shouldn't the segment have broken up into smaller, and perhaps not distinguishable, fragments?  Is it possible that someone who was born in the 1660s still has a large segment of their DNA detectable in their descendants?

Thanks to everyone who participated in this effort through DNA testing and/or researching.