THIS BRITISH STUDY WAS CONDUCTED IN TWO PARTS. THE FIRST PART COMPARED THE CHARACTERISTICS OF THE LIFE TERMERS WITH A SAMPLE OF 96 ARSONISTS SENTENCED TO 18 MONTHS OR MORE WHO CAME UP FOR PAROLE REVIEW BETWEEN APRIL 1973 AND APRIL 1974. THE SECOND PART STUDIED 139 ARSONISTS GIVEN SIMILAR DETERMINANT SENTENCES AND RELEASED FROM PRISON DURING THE YEARS 1970-1972, PLUS 8 ARSONISTS WHO WERE SENTENCED TO 5 YEARS OR MORE AND WERE DISCHARGED IN 1973. FIVE FACTORS WERE FOUND TO ENTER INTO A MULTIPLE REGRESSION EQUATION PREDICTING TYPE OF SENTENCE (FIXED-TERM OR LIFE): PSYCHOPATHY, SEXUAL ABNORMALITY, VALUE OF DAMAGE CAUSED BY THE ARSON OFFENSE, HISTORY OF ARSON, AND PERHAPS HISTORY OF CRIME IN GENERAL. ONLY PSYCHOPATHY AND SEXUAL ABNORMALITY WERE VALIDATED IN THE ANALYSIS OF THE LARGER SECOND GROUP. IT WAS FOUND THAT THE MEN SERVING LIFE SENTENCES OR LONG TERM SENTENCES (5 YEARS OR MORE) HAD MANY CHARACTERISTICS IN COMMON BUT THAT NEITHER GROUP RESEMBLED THE SHORT TERM MEN. NONE OF THE FIVE IDENTIFIED FACTORS PROVED PREDICTIVE OF EITHER SENTENCE LENGTH OR RECIDIVISM. THE BEST SINGLE PREDICTOR OF FUTURE CONVICTION WAS THE NUMBER OF PREVIOUS CONVICTIONS, A COMMON FINDING IN CRIME STUDIES. THE FINDING THAT DIAGNOSES OF SEXUAL ABNORMALITY AND PSYCHOPATHY ARE MORE COMMON AMONG LIFERS THAN AMONG MEN SERVING LONG DETERMINATE SENTENCES MAY ARGUE THAT LIFERS ARE MORE DANGEROUS. HOWEVER, FIXED-TERM RELEASEES WHO EXHIBIT THESE CHARACTERISTICS ARE NOT MORE LIKELY TO BE REARRESTED THAN OTHER RELEASEES. TABLES PRESENT STUDY DATA. REFERENCES ARE PROVIDED. (GLR)