Key points are not available for this paper at this time.
Mr Healy and Mr Westmacott describe a general technique for dealing with observations missing from block experiments analysed on automatic computers. When more than one observation is missing the technique is simpler than other methods hitherto described and has been proved in practice to be satisfactorily fast. It is applicable to any analysis in which least‐squares estimates are derived.
Healy et al. (Thu,) studied this question.