Sir

Your Editorial on resistance to non-proliferation of nuclear weapons is, in my view, a good example of selective presentation of data (Nature 461, 11; 2009). As first director of the US Defense Threat Reduction Agency, from 1998 to 2001, I consider a few corrections are in order.

Although the predicted lifetimes of the plutonium pits of US weapons are 40–80 years, the thousands of non-nuclear components in the pit's environment are less stable. Radiation damage, loss of plasticizers and binders, and other changes mean that they could eventually compromise the performance of highly optimized cold-war weapons, designed to minimize weight and use of plutonium and uranium.

Replacing these parts with exact equivalents is often not possible because the materials and production processes no longer exist or because their production is precluded by changes in environmental and safety regulations. Confidence in performance and the willingness of politicians to accept expert judgement about it are likely to suffer.

The proposed Reliable Replacement Warhead (RRW) was designed using nuclear systems that were more robust and had higher margins against failure, thus relaxing the stress on new non-nuclear systems intended for replacement and future production. As a bonus, the safety and surety features of the weapon were improved. It is not a stalking horse for nuclear testing but would increase military and congressional confidence in weapons performance; this assurance is vital as falling numbers render each weapon more important for deterrence.

Unlike the unlamented Robust Nuclear Earth Penetrator, which failed tests of both utility and credibility, the RRW does not represent a lowering of thresholds against nuclear use, nor an opening to worrying policies of pre-emption.

You should not chide the United States for desiring to do what the British, French, Russian and Chinese governments are also doing or planning to do — namely, modernizing their stockpiles with new weapons systems.