Crimes reported to police throughout Maryland fell by 9 percent during the first six months of 1999, continuing a downward crime trend for the fourth year in a row, according to recently released figures compiled by Maryland State Police.

Police and criminologists attribute the decrease, which mirrors a nationwide trend, to a crackdown on repeat offenders, programs that put police officers in closer touch with the communities they patrol and the continuing thriving economy.

The number of crimes reported from January through June fell the most--12 percent--in Montgomery and Prince George's counties compared with the first six months of last year. Crimes in the Baltimore area and Western Maryland fell 8 percent for the same period,

In the Southern Maryland counties of Calvert, Charles and St. Mary's, reported crimes dropped by 7.6 percent overall, but those numbers broken down by crime show that rapes rose by 8 percent and arson increased by 22 percent.

In Howard County, crimes increased by 4 percent. The overall crime figures refer to the number of reports for homicide, rape, robbery, aggravated assault, burglary, theft, car theft and arson.

Statewide, homicides fell by 11 percent, rapes dropped by 16 percent and car theft and arson fell 12 percent, according to the state police figures.

Adam Gelb, policy director for Lt. Gov. Kathleen Kennedy Townsend (D), who oversees crime issues in the state, said he attributes at least some of the decline to state initiatives that have focused on the relatively small group of repeat offenders who commit a disproportionately large number of crimes.

Such programs include more frequent drug testing of people on probation and parole and teaming probation officers with local police to better enforce probation restrictions.

But Gelb noted that crime statistics also reflect changes in how much trust residents have in their local police and, in turn, how comfortable they feel reporting crimes. For example, if a police department gains a reputation for treating victims of domestic violence with sensitivity, Gelb said, the number of assaults reported in that jurisdiction may increase simply because more are reported.

"In a lot of ways, you're recording police behavior more than actual crimes occurring," Gelb said.

Specialists in crime statistics also caution against comparing one county's figures with another's because, although local police departments are supposed to classify crimes by standard definitions, it's not known whether all do. A purse-snatching that might be recorded as a robbery in one jurisdiction might be classified as an assault in another.

"Every county could be doing it slightly differently," said Henry Brownstein, director of the graduate program in criminal justice at the University of Baltimore and a specialist in crime statistics. "They're supposed to use the same definitions, but the definitions aren't always precise."

Brownstein said the statistics are most helpful as a measure of general trends within one county. Although the number of crimes reported can depend on everything from activity in local drug markets to the overall health of the economy--low crime generally follows low unemployment--Brownstein said police also should be given credit.

More police departments are using computers and other technology to trace where crime is happening and are focusing on those areas. In Maryland, police work with community leaders in crime-ridden neighborhoods designated as "hot spots."

"Up until recently, criminologists used to think police didn't have much to do with crime trends at all," Brownstein said. "But now . . . they can track better what's going on, and they're responding to patterns."

Sometimes police aren't exactly sure what the numbers mean. In Howard County, the 4 percent increase in overall crimes reported is fueled most by a 62 percent jump in robberies during the first six months of this year compared with the same time last year, said Sgt. John Superson, a Howard police spokesman.

But the 124 robberies reported in the first half of this year were not as unusual as the remarkably low 76 robberies during the same time last year, Superson said. This year's robbery figure is actually on par with four of the five previous years but appears to represent a significant increase because last year's figure was unusually low, he said.

Police don't know why so few robberies were reported last year, Superson said.

"The telltale sign" of robbery trends, he said, "will be in the last six months of this year."