MAJOR ADVANCES WERE MADE IN THE STATISTICAL ANALYSIS OF OFFENSE SPECIALIZATION BY WOLFGANG ET AL. THROUGH THEIR USE OF STOCHASTIC MODELING. HOWEVER, THEIR RESULTS DID NOT DISTINGUISH BETWEEN OFFENSE DYNAMICS THAT REFLECTED THE MARGINAL DISTRIBUTIONS OF THE SAMPLE AS A WHOLE (AND THUS MAY BE CONSIDERED RANDOM) AND DYNAMICS THAT SIGNIFICANTLY DEVIATED FROM THIS DISTRIBUTION. THIS STUDY USED THE FINITE MARKOV CHAIN MODEL (A STOCHASTIC MODEL) TO ARBITRARILY RESTRICT A SAMPLE OF 750 YOUTH FROM THE COOK COUNTY JUVENILE COURT (ILLINOIS) TO THOSE WHO HAD COMMITTED FIVE OR MORE OFFENSES. THE STUDY THEN ANALYZED THE FIRST FIVE OFFENSES (THEREBY LOSING A GREAT DEAL OF DATA). AS THE SEQUENCE NUMBER OF THE TRANSITION INCREASED, THE ACTIVE PORTION OF THE SAMPLE BECAME SMALLER AND HIGH-ORDER TRANSITIONS OF THE MOST ACTIVE DELINQUENT REMAINED UNANALYZABLE. ALTHOUGH THE TRANSITION PROBABILITIES INDICATED THAT SPECIALIZATION IN MOST OFFENSE TYPES WAS NOT TYPICALLY A DOMINANT FEATURE OF A YOUTH'S OFFENSE HISTORY, THE RESIDUAL STRUCTURE OF THE MATRICES SHOWED DEFINITE EVIDENCE OF SOME SPECIALIZATION TENDENCIES FOR THE WHITE AND NONWHITE YOUTH SAMPLES. HOWEVER, IT STILL LEFT THE DYNAMIC STUDY OF DELINQUENCY IN A VERY RUDIMENTARY POSITION. THE MARKOV MODEL REPRESENTED SPECIALIZATION IN THE AGGREGATE, WITH THE ASSUMPTION THAT EVERYONE IN THE SAMPLE IS CHARACTERIZED BY IDENTICAL TRANSITION PROBABILITIES. THIS ASSUMPTION OF HOMOGENEOUS TRANSITION RATE IS UNREALISTIC. FINALLY, SPECIALIZATION IN TERMS OF A MARKOV MODEL HAS A VERY LIMITED SENSE: THE CONSECUTIVE COMMISSION OF LIKE OFFENSES. EVEN THOUGH A YOUTH RARELY COMMITS THE SAME OFFENSE TWO TIMES IN A ROW, MOST OF THE YOUTH'S OFFENSES MIGHT BE OF THE SAME TYPE. IF SPECIALIZATION WERE REDEFINED AS MUCH, A LARGE POPULATION OF BOTH RACIAL GROUPS OF THE SAMPLE MAY BE SAID TO SPECIALIZE. PERHAPS THE USE OF RESIDUAL ANALYSIS IN THE DYNAMIC MODELS COULD INCREASE THE SENSITIVITY OF THE INTERPRETATION OF THE DYNAMICS OF DELINQUENCY AND HELP TO DISCERN PROCESSES NOT READILY APPARENT BY A VISUAL INSPECTION OF THE TRANSITION PROBABILITIES. NOTES AND REFERENCES ARE GIVEN. (MJW)