Precisomundo, Ellen! No, really; that's a word. I've used it in Scrabble. Cough, cough.
"Endogamy" should be the word of the day for August 26. Strictly defined it's an anthropological or sociological term dealing with a custom of marrying within a cast, clan, or tribe. The definition is extended a bit (logically so) in genealogy, where it encompasses not only social mores but also geography: intermarriage driven by a physically limited population set.
As Ellen mentioned, early colonial New England is a classic example, as are various island populations. But the westward expansion in the U.S. saw similar environments. For example, Alabama did not open to settlers until the end of the War of 1812, and the first immigrants--often in conjunction with other families with whom they already had close association--began arriving in 1814, three years before the U.S Congress even created the Alabama Territory. Not exactly tons of 5,000-people towns with a broad spousal gene pool.
Most genealogists are used to, and comfortable with, family tree representations that look like massive organizational charts: neatly organized boxes marching two-by-two: two parents, they each having two parents, your having a total of eight grandparents and 16 great-grandparents and so on. Traditional trees and fan charts are dependent upon that tidy organization. In fact, it's difficult for many family tree software applications to represent things any other way and, I've found, difficult for some family researchers to mentally envision it organized differently.
In truth, though, many of our family trees--if inspected from the standpoint of biology and DNA--look more like hurricane spaghetti models (a subject near to me right now...literally). Relationships and ancestors overlap, and marriages within the same families are far from uncommon (as in Cindy's situation where two brothers may have married two sisters; reminds me of the old musical Seven Brides for Seven Brothers...go do those families' DNA charts).
The situation is exceedingly common, yet very difficult for us to deal with. Quick, without stopping to think about it: if your great-grandparents were 2nd cousins, what's the precise descriptive term for your grandfather's sister? And making a DNA consanguinity table for a situation where one set, whether 1st or 5th, of cousins marry isn't a difficult thing. But now assume a couple of generations before there was a another cousin intermarriage, and a generation later two brothers marry two sisters, and a generation later still a woman's husband dies and his brother takes her and her children in, and they marry, and have more children. A Jamie and Circe Lannister situation would at least be much more straightforward. ;-)
But the spaghetti model family tree is reality for many of us, whether we know it or not. With autosomal DNA coming into the picture in the past few years, we have to stay cognizant of it. The more triangulated matches we can document, the better able we are to analyze certain sharing "outliers" when they come along, and to use the information to add dimensions to the understanding of our family trees.