Laboratory Investigations into the State of Nutrition

We use laboratory methods of investigating the state of nutrition with the hope that they will supply data which are impersonal, divested of local and temperamental bias, and expressed in terms which convey the same meaning all the world over.We hope also that they will be specific, not only detecting malnutrition, but also giving an indication of the faults in the diet which caused the malputrition.When such data have been collected we must ask two questions, first, whether the data are trustworthy, and secondly, what they mean. Level of haemoglobin in the blood Reliability of the technique CONFERENCE PROCEEDINGS '948observers (Table I , nos. 2 1 and 3 2 ) differed consistently from one another in their readings by nearly 8 %, while the readings made by less happy observers might be out by as much as 14%.The means of the readings made by most of the observers were below the true value.It might be hoped that individual observers would be consistent in their own readings, and Mackay, Wills & Bingham (1946) have produced some evidence that the relation between their'readings actually did not change.At the London Hospital we compared the visual readings of observer 22, who read I yo low in no.