I am working with a U.S. Census Tract layer. I am just wondering why there are tracts entirely within another tract when, in general, the territory is divided by partitioning the space into a number of simple rectangular regions, divided up by axis parallel (more or less) splits.
Census tracts must align with county boundaries - that is part of their designation. From the US Census Bureau:
"Generally, census tracts have between 2,500 and 8,000 residents and boundaries that follow visible features. When first established, census tracts are to be as homogeneous as possible with respect to population characteristics, economic status, and living conditions."
So, since all of Kansas's counties were established in the mid-1800s, before census tracts came about in 1906, the Census Bureau had to work around the established county boundaries. Many of Kansas's counties have very low populations but since the tracts can't cross county lines, those counties have to be the extent of a single tract. To divide a mid-size county according to the Census guidelines (keeping tracts as homogeneous as possible) one solution might be to put the area around a town into one tract and the surrounding rural areas in another.
This seems to be pretty common in surrounding states with grid-like counties.