THERE should be a chemical weapons ban, but not the Chemical Convention now under negotiation in Geneva. Instead, the 1925 Geneva Protocol, which prohibits use of chemical weapons, should be amended to ban possession as well.The draft convention should be scrapped for three reasons: It is not verifiable. The verification regime under negotiation addresses only those types of cheating for which there is hope of detecting violations, such as production in commercial facilities and failure to destroy declared stockpiles. Yet there are several ways to cheat for which there are no adequate or effective means of detection, including: 1. Diversion of chemicals from commercial facilities. Many common chemicals usable in or as weapons are produced for legitimate industrial purposes. Phosgene and hydrogen cyanide, for example, are used in industry, but can also be used as weapons. In many countries they are produced in large quantities at multiple sites. Highly accurate accounting of feedstocks going in and product emerging from such facilities is impossible; transport losses, spillage, and other discrepancies can result in up to 10 perc ent error. If a country were to cheat by diverting common chemicals, it is very unlikely that it would be detected. 2. Production of agents not listed in the convention. Production of novel chemical agents is technologically feasible and essentially unverifiable. Inspectors will be looking for chemical compounds specified in the schedules - lists of chemicals - contained in the convention; they will not have the tools to look for chemical compounds of which they may not even be aware. Toxins, for example, pose a particularly difficult problem. Genetic engineering and other advanced techniques can be used to separate a nd mass-produce deadly chemicals from biological origins. There is no way that such activities can be precluded, nor can they be detected via current technical means. 3. Production in a clandestine facility. A plant capable of producing 100 tons of chemical agent per year can fit into a space 40 ft. by 40 ft. A cheater can hide the facility underground, in a secret section of a pesticide plant, or any suitable location. If a suspicious country demanded a challenge inspection of the site, and access were allowed, the illicit activity might be discovered. But how can a covert facility be pinpointed for challenge inspection? Photography cannot be relied upon, as chemical weapons facilities may have no unique features that would reveal their purpose. To find a covert plant, one would have to depend on chance intelligence from human sources. 4. Retention of undeclared chemical stockpiles. Chemical weapons are easy to hide; bulk chemical agent is even easier. A nation might cheat on the convention by placing weapons or agent in moving railcars, underground sites, or non-suspicious buildings. If inspection were requested, the weapons or agent could be rapidly moved before inspection. Challenge inspections pose significant risks. A country might select for inspection a sensitive United States facility - one dealing with stealth technology or nuclear weapons, for example. Even if it is clear that the intent is espionage, the US would have no right of refusal. Likewise, intrusive inspections of the US chemical industry would yield opportunities for industrial espionage. Inspections will also cause erosion of constitutional rights under the Fourth Amendment and loss of profit. Implementation is too costly. Financial costs of the convention's verification scheme are exceedingly high. Thousands of facilities in some 70 countries will be subject to routine inspection, in addition to an undetermined number of challenge inspections. These inspections will be costly not only because of their large number, but because of the proposed international bureaucracy. Akin to the International Atomic Energy Agency (IAEA), it will oversee verification. The IAEA, which inspects only about 500 facilities in 50 countries, has a yearly budget of about $160 million. The budget for the Chemical Convention will be significantly higher because of initial capital expenditures on equipment, a possible fleet of aircraft for quick challenge inspections, possible regional offices, greater numbers of facilities and inspectors, etc. Per-country costs for inspected countries will be increased by the requirement for national authorities to supply compliance data, and by costs incurred by inspected companies. The complexity and cost of the draft Chemical Convention will not assure effective verification. Let's take a short cut by following the model of the Biological and Toxin Weapons Convention, a ban that admits it cannot be verified. The easiest, quickest way to achieve a chemical ban is by amending the Geneva Protocol of 1925 to prohibit possession of chemical weapons. To put teeth into a chemical ban, nations should concentrate on two issues: sanctions against nations who threaten to use or are found to have used chemical weapons, and improving techniques to verify if and when chemical weapons have been used. There should also be intensive efforts to improve protection against known and possible new chemical weapons, and strong nonchemical defenses should be maintained. Deterrence and protection will be essential, given that verifying a chemical ban is not possible.