**3. Detectors**

NaI(Tl) scintillation detectors are more suitable for this measurement than semiconductor detectors because of their greater stopping power. The most precise measurement of TBN reported is 1.6% for a neutron dose of 0.45 mSv (Ryde *et al* 1989) using a 4 GBq 252Cf fission source. A commonly employed technique in the measurement of body nitrogen is to measure the ratio of the emissions from nitrogen and hydrogen. This ratio is much less sensitive to variations in body size, neutron fluence and detector characteristics, which affect the signal from each element alone. It also permits the determination of TBN from partialbody irradiation (of the torso and thighs, thereby minimizing the radiation dose to radiosensitive tissues such as the eyes) assuming that hydrogen comprises one-tenth of body weight (Vartsky *et al* 1979). This requires correction since the proportion of body weight due to hydrogen has been estimated to vary from 9.5 to 10.8% in a large population of patients.

Chlorine may be determined from its emission at 8.57 MeV after deduction of the underlying background noise due to random summing and scattered gamma rays from nitrogen (Mitra et al. 1993). It may also be determined from its prominent emission at 6.11 MeV, but a high-resolution semiconductor detector (Ge(Li) or hyper pure Ge) must be employed to distinguish this emission from the emission from oxygen at 6.134 MeV.
