It may be several weeks or even months before I shall ask you to drench Germany in poison gas, and if we do it, let us do it one hundred per cent. In the meanwhile I want the matter studied in cold blood by sensible people and not by . . . psalm-singing uniformed defeatists . . .

SO SPOKE Winston Churchill to his military chiefs of staff in a "secret minute" written during the worst of the 1944 German "buzz bomb" cruise missile raids on London. Earlier, after the fall of France, he had directed that poison gas be used on German troops if they crossed the Channel to invade Britain.

Both orders were made despite Britain's solemn 1925 treaty promise not to be the first to use poison gas in war against any country which had signed the treaty (Germany had signed). In neither case was gas used--in the first, because German troops did not invade Britain; and in the second, because Churchill's military planners feared the consequences of German retaliation in kind. In both, the treaty promise appears to have been considered but pushed aside by Churchill.

This book is filled with such episodes. It is the best, up-to-date account of gas and germ warfare available for the lay reader. It follows the worthy tradition set by Frederick J. Brown and Seymour M. Hersh in their separate accounts published in 1968. It adds to Brown and Hersh by describing the use of gas, toxins and herbicides since 1968 in Vietnam, Yemen, Afghanistan, Laos and Kampuchea. It also summarizes recent international negotiations to end gas and germ warfare. And it contains much newly declassified information about World War II, including an amazing account of British-U.S. collaboration, urged on by Churchill, to produce anthrax bombs to be used if necessary on German cities.

Current controversies make the book particularly relevant. It reveals, for example, the complete failure of British and U.S. intelligence to detect Germany's World War II development of nerve gas, a new, extraordinarily dangerous gas we did not then have. If a poison gas disarmament treaty were negotiated, detection of such a development in violation of the treaty would be extremely important. The World War II experience suggests that national intelligence would not suffice for detection. Indeed, U.S. and Soviet negotiations toward a new gas disarmament tready had foundered by 1980 on our demand for on- site inspections. President Reagan now proposes to renew these negotiations-- while at the same time producing a new generation of nerve gas weapons.

Recent accounts of the use of poison gas and toxins-- "yellow rain"--in Afghanistan, Kampuchea and Laos are also summarized in the book. In the 1925 Geneva Protocol, the Soviets, like the British, promised not to be the first to use gas and bacteriological weapons against other parties. Because Soviet troops are in Afghanistan in very large numbers, gas use against tribesman there can likely be pinned on the Soviets. The book suggests, however, that the gas may be CS, a form of tear gas which we used against the enemy in Vietnam claiming it did not violate the Protocol. Moreover, Afghanistan is not a party to the Protocol. Soviet use would arguably not violate the Protocol's terms for that reason, as well as because of the kind of gas used.

In Laos and Kampuchea, Soviet presence is less visible and responsibility less clear. Nor are Laos and Kampuchea parties to the Protocol. However, the toxins apparently used are the subject of a second treaty, the 1972 prohibition on the development, production and transfer of biological and toxin weapons. The Soviet Union, Laos and Kampuchea are all parties to this treaty, as are over 100 countries. A likely violation of this treaty has taken place in Southeast Asia if the accounts published in the book and by the State Department are true. But which countries are responsible is not yet clear.

For me, the greatest significance of the book is in testing the effect of the Geneva Protocol to inhibit gas and germ warfare. This treaty is the oldest and most significant ban on first use of a weapon of mass destruction. Consideration of a "no-first- use" agreement for nuclear weapons has been urged recently by McGeorge Bundy, George F. Kennan, Robert McNamara, and Gerard Smith. Would such an agreement for nuclear weapons have a significant restraining effort on its signatories?

The book asks what restraining effect the Geneva Protocol had in World War II. Putting aside a cloak-and-dagger account of the assassination of a German general by British agents, poison gas and germ weapons were not in fact used in warfare by any party to the Protocol in World War II--though the British came close. Harris and Paxman conclude that they were not "because of the precise military circumstances prevailing at the time," not because of the treaty. "At no point was the fact that chemical weapons were banned under international law a major condideration not to go ahead and use them (except possibly in the personal antipathy of Roosevelt)."

In my view, Harris and Paxman expected too much of the Protocol. It did in fact help establish an international norm against poison gas warfare just as the First Commandment helped establish a norm against murder. By Harris and Paxman's account, the international condemnation of poison gas affected American and British leaders; Churchill in the end was probably restrained in part by the abhorrence for gas on the part of Eisenhower and Roosevelt as well as British military men.

The Protocol came to symbolize the abhorrence which many felt after World War 1. The Protocol's lack of any internationally ordered sanction was attributable to the absence of world order in general, not to a failure in draftsmanship. If retaliation was the primary sanction acting to deter to poison gas and germ warfare in Europe during World War II, the Protocol provided the international standard of conduct.

Until a new world order can provide effective sanctions, fear of retaliation is likely to remain the principal deterrent for any ban on first use of weapons of mass destruction. But the norm of conduct--the prohibition on first use-- needs nevertheless to be established, just as a standard for murder is needed before the question of punishment for murder becomes relevant.