would break out. At Fort Holabird, the home of the Army Intelligence Command, wiser heads prevailed and the idea of a computerized barometer of racial tensions was quietly shelved.

The collection of incident reports on minor altercations also was justified on the theory that they might give early warning of where and when riots would break out. In point of fact, the reports did not give the Army analysts any better advance notice than top Justice Department officials, who relied more on personal contacts with community leaders.

When the decision to call out federal troops was contemplated, the President, the Attorney General, and the Under Secretary of the Army did not turn to Army intelligence for information and advice. They relied instead on the observations and judgment of their own personal representatives at the scene. Since federal troops could not be deployed until state authorities had demonstrated their lack of resources, there was time for a personal reconnaissance by high government officials and their aides. While the civilians were making their reconnaissance, stand-by riot units were alerted and sometimes prepositioned at nearby bases.

The military's own after-action reports of the riots of the late 1960's clearly demonstrate that the Army's greatest advance requirement was not for personality or organizational data pertaining to the politics of the ghettos, but for reconnaissance and liaison data. What it needed most were city packets of maps and other records describing potential approach routes, the height of overpasses and the strength of bridges, information on potential riot areas and bivouac sites, the addresses and telephone numbers of local law enforcement officials, and the frequencies used by police and National Guard radio networks.

Unfortunately, reconnaissance and liaison work often took a back seat to the investigation of individuals and organizations, because investigating was what military intelligence was most accustomed to doing. This sometimes left troop commanders gravely unprepared for riot duty. For example, when General Throckmorton entered Detroit at the head of a joint force of 10,000 Army and National Guard troops, he had only a gasoline station map to guide him. (He and his troops did not even share the same company's maps). Back at the Pentagon, the officers in the war room struggled to follow the action in Detroit on a map which recorded contours and elevations in great detail, but omitted city streets.

The second urgent need of riot units was for tactical intelligence: information on mobs, fires, and potential targets of looting or damage. Here again, Army intelligence's obsession with subversion limited the production of useful intelligence. The first items which the intelligence analysts plotted on the situation maps at the Pentagon were power plants, radio stations, and armories-the traditional targets of Communist insurgents bent on seizing governments. Meanwhile, apolitical rioters in Detroit happily looted stores and supermarkets.

During the riots, there were reports of sniping which seemed to confirm the view that conspiracies lurked behind the riots. Most of these reports turned out to be false or exaggerated. When the smoke cleared, it was the civilians who had been shot with government bullets. Insofar as there were snipers, they were attended to without the help of dossiers or mug books.

Unfortunately, the riot manuals used by the Army and FBI today still depict the typical mob as an instrument of revolution, manipulated by wellcoordinated agitators. Military and civilian planners, transfixed by the rhetoric and aspirations of "revolutionary" groups, continue to ignore their own experience and the indisputable fact that America has yet to experience an organized ghetto riot.

Mass demonstrations. The second major justification for the Army's civil disturbance intelligence program was that it was needed to give early warning of the intentions, capabilities, and probable courses of action of the leaders of mass demonstrations.

During the late 1960's, there were five mass demonstrations of particular concern to the Army. These were the anti-war March on the Pentagon in October 1967, the Poor Peoples' Campaign (including Resurrection City and Solidarity Day) in the Spring of 1968, the demonstrations at the Democratic

National Convention that summer, the Counter-Inaugural demonstrations in January 1969, and the anti-war Moratorium March on Washington in November 1969. In each instance, the key problem for those government officials who were determined to avoid violence and permit lawful protest was to strengthen the leadership of non-violent organizers and to structure the mode of protest so as to defuse tensions and limit confrontations between demonstrators and troops.

There was no need for extensive files on or the infiltration of the anti-war and civil rights groups which organized the major marches on Washington. Most of these groups were open about their plans, as indeed they had to be to lead the thousands of unorganized persons they hoped to bring together. Newspaper reports of press conferences, handbills, and other easily obtained publications outlined most of their objectives and tactics. The civil rights groups worked closely with the Community Relations Service of the Justice Department; and the anti-war groups disclosed their plans while negotiating for parade and demonstration permits. The most serious danger of organized violence came from fringe groups like the Yippies and SDS, and they were watched closely by civilian law enforcement agencies. The most serious danger of unorganized violence came from government officials who deliberately predicted violence in order to deter moderate protesters from participating.

Detailed personality and organizational information also was not necessary to estimate the size of the crowds. Advance estimates were obtained from bus companies and railroads (sometimes in an unnecessarily intimidating manner) by FBI and Army agents. Government agents also made vehicle counts on the highways coming into Washington while military planes took aerial photographs of the assembling crowds.

Infiltration of demonstrating groups was standard practice for many government agencies, including the Army, but there is little evidence that the reports obtained were of much use. (Had the groups been intent on doing violence, the reports might have been different, but even then, there is no reason to believe that the information had to be collected by the Army).

The approach of civilian officials responsible for supervising Washington's response to mass demonstrations was to go out into the streets and see for themselves. For example, during the November 1969 Moratorium demonstrations Army General Counsel Robert E. Jordan III waded into crowds at DuPont Circle with his own team of radio-equipped lawyers. Jordan was not about to trust reports from agents he did not know so long as it was possible for him to see the situation for himself. During the March on the Pentagon in 1967, top military and civilian officials observed the events from their windows and over closed-circuit television filmed from the building's roof.

Chicago was a different problem. There the Yippies and city officials conspired to create a situation in which violence was inevitable. Early warning intelligence on the intentions of the various demonstrating groups was not very helpful because the authorities deliberately destroyed the leadership capacity of moderate groups by refusing to permit them to march near the convention center or camp in the parks. At the same time, the Yippies skillfully inflamed official expectations of violence with outrageous threats that were totally beyond their ability to carry out. On this occasion, analysts at the Army's Counterintelligence Analysis Branch took note of the dwindling numbers of protesters determined to go to Chicago and the size of the Illinois National Guard to rightly predict that federal troops would not be required. Their prediction, however, was not conveyed to President Johnson who, in any case, would probably have ignored it just as he ignored the counsel of the Attorney General. Political risks, rather than intelligence estimates, were the decisive factors in the President's decision.

The key point which should be borne in mind about the mass demonstrations is that while they required reconnaissance and liaison data for planning and tactical intelligence about the route or place of protest, the mode of demonstration, and the leadership capacities of the organizers, they did not require political surveillance by the military. What background information the Army needed to brief its generals about the politics of the situation was easily obtained from civilian officials and the press.

Disruptions of the Military. The third justification for the Army's surveillance of civilian protest groups was the need to protect military posts and activities from disruption and obstruction by anti-war and anti-military groups. The obstruction of a troop train in Oakland, California, the disruption of a celebrated court-martial in Texas, a large antiwar protest at Fort Dix, New Jersey, and the threatened confrontations at military installations throughout the country at the time of the October 15, 1969 anti-war Moratorium, all have been cited by intelligence officers as grounds for keeping tabs on militant anti-war groups.

The issue here clearly is not the military's right to know if some group plans to storm the gates of one of its forts, but the extent to which military intelligence agents should be sent out into the civilian community to investigate the plans of protest groups. Few people would question the right of Army intelligence officers to read about such plans in the newspapers, or to receive advance warning of them from civilian law enforcement agencies. Similarly, there is no reason to question the Army's right to have observers stationed on post, or on the streets near the post, to give early warning of the demonstrators' arrival. The problem comes when military intelligence wants to infiltrate a group because it lacks confidence in the information received through liaison from civilian agencies. This is particularly likely to occur overseas where the local police services are in the hands of foreign governments which may not be concerned about demonstrations by Americans on or about U.S. military installations.

Interference with court orders. During the late 1950's and early 1960's, the U.S. Army was called out twice and alerted on several other occasions to assist with the enforcement of federal court orders directing the desegregation of schools in the South. At Little Rock Central High School in 1957 and at the University of Mississippi in 1962 federal troops faced hostile crowds of citizens angry over the federal intervention. In Little Rock, Army intelligence agents and FBI agents worked together to investigate rumors of possible attacks on federal troops and the nine black children they were charged with protecting.

Where federal troops are lawfully assigned to enforce court orders, it is difficult to question the necessity or propriety relevant intelligence services. However, it is important that the role be viewed as a security function, and not as a license to collect personality and organizational data for use beyond the immediate mission.

Para-military and military resistance. The ultimate justification for focusing the civil disturbance intelligence effort on civilian political groups was the counter-insurgency theory. As the Assistant Chief of Staff for Intelligence said to his staff in the domestic war room at the height of the Detroit riot: "Men, get out your counter-insurgency manuals. We have an insurgency on our hands."

General Yarborough was wrong in that assessment, but his error does not relieve this Subcommittee of its obligation to determine the legitimate counterintelligence and combat intelligence needs of the military in times of rebellion or civil war, and to express them precisely in statutory terms.

One way to do this might be to define the protected rights in terms of the First Amendment distinction between the single expression of beliefs and the incitement or solicitation of others to take criminal actions. Another way would be to conceive of the military's informational needs in a martial law situation in terms of the laws of treason, espionage, sabotage, murder, and assault. A third would be to focus on the need of tactical combat units to know the intentions, capabilities, and probable courses of action of hostile forces, including guerrillas within their own lines. So long as the war-time needs of the military are viewed in these lights, S. 2318 should pose no threat to legitimate military operations.

Counterintelligence

The dominant mission of Army intelligence off the battlefield has traditionally been that of counterintelligence. Strictly speaking, the term implies countering the operations of foreign intelligence agencies bent on espionage, sabotage, and disruption. In practice, Army intelligence has recognized no such restrictions. In the absence of sufficient espionage and counter-espionage

investigations to keep its agents busy, the Army has charged them with investigating the more amorphous subjects of "subversion" and "disaffection." Nowhere in the lexicon of the intelligence corps is there a precise definition of either term. "Subversion" suggests active efforts by foreign agents or domestic revolutionaries to undermine the loyalty of soldiers to the Army and the country. "Disaffection" suggests the state of mind that makes a soldier an easy mark for a subversive solicitation. Together, the terms have been used to authorize outrageous abuses of investigative discretion. Military intelligence agents, like their civilian counterparts, have proven incapable of distinguishing between shades of dissent, or even between moderate, liberal, radical, and revolutionary groups. They have also demonstrated their ignorance of the basic First Amendment distinction between advocacy of the desirability of violence and solicitation of others to commit crimes. They have not viewed "subversion" within the military as the active solicitation of military personnel to violate lawful orders in the services of some foreign power or domestic revolution, but have equated it with any attack on the legitimacy of the military or those elements of the social and political order with which it has traditionally identified.

Similarly, the use of the investigative category of "disaffection" has led to the uninhibited surveillance of servicemen and their political affiliations within the civilian community. In theory, the term implies susceptibility to solicitation by enemy agents. In practice, it has supplied an investigative rubric under which many commanders have justified the use of intelligence agents to enforce their particular brand of discipline and morale. (See the section on "Command Intelligence" below).

It would greatly assist analysis of the subject matter of this bill if the term "counterintelligence" were confined to its narrowest and most precise definition, and such rubbery terms as "subversion" and "disaffection" stricken from the vocabulary entirely.

Criminal Intelligence

In defining the military's legitimate intelligence needs it also should not be forgotten that there are laws which military investigators must enforce. For example, Army intelligence agents are currently charged with investigating possible instances of espionage and sabotage; Army criminal investigators must look into potential violations of the Uniform Code of Military Justice. Thus, like all agencies charged with enforcing laws, the military services may sometimes collect "criminal intelligence."

Criminal intelligence, like counterintelligence, is a rubbery word that can be stretched to justify a wide range of investigative abuses. Properly conceived, it should be used only to refer to the opening stages of a criminal investigation by an authorized law-enforcement unit. Thus, when the Army's Criminal Investigations Division suspects that anti-war civilians are soliciting servicemen to disobey lawful orders, there should be no doubt about the legitimacy of its investigative interest.

At the same time, there should be no doubt that civilian law enforcement agencies within the United States should be given full responsibility for the civilian side of such investigations. Except in rare instances, military investigators should concentrate on persons subject to the UCMJ and limit their surveillance of civilians to the territorial boundaries of military posts.

Respect for the primacy of civilian agencies off-post would also mean that activities like the anti-war coffeehouses that sprang up around military installations in the United States during the late 1960's would be off-limits to military intelligence, but that criminal investigators assigned to nearby posts might legitimately receive reports from civilian law enforcement agencies on cases in which the coffeehouse proprietors are suspected of unlawfully soliciting GI's to disobey laws or commit crimes. It would also mean agents could not be loaned to civilian agencies for undercover duty except in joint investigations and could not be used for intelligence duty at national party conventions.

Overseas, the absence of a well established force of civilian investigators enforcing American law increases the pressure for off-post investigative activities in cases involving the joint political activities of American soldiers and civilians. In Germany, for example, Army intelligence apparently infiltrated

the Goessner Industrial Mission because American clergymen there were counselling servicemen. The clergymen insist that they were not counselling the soldiers to desert or otherwise deny the authority of their officers, but what if they had? Would it be improper for military intelligence to undertake such an investigation?

The use of intelligence agents rather than criminal investigators suggests that the Army did not expect evidence of criminal conduct to be uncovered, so perhaps the offense lies in using the wrong team of Army agents. But suppose Army criminal investigators had been used instead? Should S. 2318 contain any restrictions on the nature and scope of their inquiries? Security Intelligence

The fourth category of domestic intelligence activities of the military I have chosen to call "security intelligence." By this term I mean to refer primarily to background information on politically active groups amassed to facilitate loyalty determinations under the personnel security program. However, I also recognize that the term can be used to encompass information about potential disruptions directed at military installations or activities by anti-war and anti-military groups. Since the bulk of the latter category has been discussed under civil disturbance intelligence, I shall confine my discussion here to the so-called "subversives files" maintained by various military intelligence agencies for general reference and to characterize the "loyalty" of political groups with which persons being considered for security clearances may be affiliated.

The "subversives files" of the Army Intelligence Command are described at pages 21-23 of the subcommittee's 1972 staff report entitled "Army Surveillance of Civilians: A Documentary Analysis." There are two facts about these files which should be borne in mind. First, they are crammed full of unverified newspaper clippings, handbills, and membership lists, which, in the main, are useless. Security clearance adjudicators who want to know if a particular organization is considered "subversive" find it easier to refer to short one to three-page "characterizations" prepared by the FBI or Army analysts. Second, characterizations which attempt to define the poltical legitimacy of various groups support the denial of security clearances on the basis of "guilt by association." The only fair way to proceed is to grant or deny a clearance solely on the basis of original investigations focused on the activities of the individual under consideration. Thus, the characterizations and the files which back them up are not needed if the Army intends to be fair to its

own.

Command Intelligence

Finally, a word should be said about "command intelligence." The term is my own invention, and by it I mean to refer to a variety of investigations ordered by commanders to help maintain order, discipline, and morale among their troops and to inform themselves about political protests within the civilian community which may affect the security of their commands and the efficiency of their operations. It is a catch-all category which sweeps across all other categories, but poses some special problems of its own.

For example, there is no reason to doubt the right of a military commander to receive information about attempts by civilian organizations to obstruct delivery of supplies to his troops, disrupt armed forces entrance and examining stations, or burn down ROTC buildings. Investigation of these matters for the purpose of criminal prosecution normally will lie in the hands of civilian authorities, but commanders are entitled to sufficient information about them to adjust military operations accordingly.

A more difficult set of circumstances exists when the commander is confronted with legal efforts by civilians to proselytize his troops. The communications are protected by the First Amendment, but they may affect the way in which the commander runs the unit or handles the grievances of his men. Communications from some civilian opponents of war to servicemen may evolve into criminal solicitations designed to promote disrespect for authority, evasion of duty, desertion, and other crimes.

Thus, the effort to define a commander's need for information about civilian communications to his troops quickly leads to the complicated questions about