Three techniques for the serological diagnosis of Rocky Mountain spotted fever were compared by testing 417 sera from 178 patients who very probably did not have rickettsial infections and 88 sera from 41 patients who very probably had Rocky Mountain spotted fever (SF). The techniques were complement fixation (CF), indirect fluorescent antibody (IFA), and microagglutination (MA). To avoid possible degradation during unnecessary purification, the antigens were prepared by methods that were as simple as possible. In the CF tests of 417 sera from patients with nonrickettsial diseases there was only one titer of 8 and none at higher dilutions, whereas with the IFA and MA tests 4-8% of the sera reacted with SF antigens and 4-20% reacted with murine typhus (MT) antigens; the evidence indicated that these reactions were not caused by specific rickettsial antibody. With the SF sera, it could be seen that the IFA test was the most sensitive and the MA test was the least sensitive at each interval after infection. Moreover, the IFA results showed the least number of confusing cross-reactions with MT antigens and the MA test showed the most. The relative advantages of the three tests in serodiagnosis of rickettsial diseases are discussed.