Would anybody know what the deal was with the 2000 Census Traffic Analysis Zones?

I am working with CTPP 2000 Table 3 data. To do some spatial analysis, I turned to census boundary files for traffic analysis zones (http://www.census.gov/geo/www/cob/bdy_files.html). I quickly realized that 63 out 2639 zones for the Houston region are represented by 2+ non-adjacent polygons. Does this happen in other regions as well? Was this delineation done purposefully or perhaps these are simply errors stemming from TIGERLine 2000?

Any input will be much appreciated.

Thank you.

Dmitry Messen

H-GAC

dmessen@h-gac.com