The universe of all public and institutional, nonmilitary, eating 

 places serving hot solid food (for consumption on the premises) 

 was subdivided, for each city, into four subuniverses , as follows: 



1. Eating places serving the public at large and principally 

 or importantly concerned with the service of food for 

 consumption on the premises: restaurants, cafeterias, 

 hotels, bars, etc. 



2. Eating places serving particular groups of the general 

 public at the place of their principal activity for con- 

 sumption there: in plant and in school feeding opera- 

 tions. 



3. Eating places serving particular "captive" groups in 

 quasi households: food serving facilities in hospitals, 

 nursing honnes, asylums, prisons, and "institutions" 

 generally. 



4. Eating places of all other types, located in semiprivate 

 organizations or in establishnnents open to the general 

 public but not principally nor importantly concerned 

 with the service of food for consumption on the premises: 

 food service in clubs, lunch counters, refreshment 

 stands, drugstores, variety stores, other retail estab- 

 lishments, transportation systems, etc. 



In all ten cities except Chicago, Illinois euid Cleveland, Ohio, the 

 sampling fraction varied from one subuniverse to the other so 

 that samples adequate for analysis purposes could be expected 

 to be produced for all four subuniverses, even though the number 

 of establishments in the four subuniverses varied greatly. (In 

 Chicago, Illinois and Cleveland, Ohio, the number of establish- 

 ments in the four subuniverses were such that, by taking the 



