SIR,-We are grateful to Dr M Rosen and his colleagues (18 February, p 438) for a more complete description of the methods used in their trial of naloxone in neonates (23 July 1977, p 229). Following their censure we have earnestly endeavoured to give their paper a" proper" reading, but we feel that their letter fails to answer the points which we previously raised.

Our main concern was that a relatively new drug was used in this trial on babies who appeared to be perfectly healthy and that a large dose was administered in a situation in which pethidine antagonists would not normally have been considered. We wholeheartedly support their view that no drug should be given without a clinical indication and this is especially true in neonates. In their letter they mention that naloxone was not given routinely to babies included in the trial but only to those who had a clinical problem associated with pethidine used in labour. We are unable to find an explanation of exactly what this problem was. The criteria mentioned by the authors in their-paper seem only to emphasise what fine fettle the babies were in at birth and it is difficult to envisage what problem occurred in the first 60 s of life that necessitated the use of naloxone. Furthermore, the authors persistently refer to the subjects of the trial as" healthy mature neonates." It seems to us of fundamental importance that normal babies should not be exposed to the unnecessary risks associated with trials of new drugs. That drug effects may be more clearly demonstrated in healthy babies is hardly a justification for such conduct, for the effects of prematurity, asphyxia, and birth trauma should be excluded by subjecting the results of a large trial to conventional statistical analysis.