The temporal pattern of daily urinary nitrogen (UN) excretion was investigated in 21 young adult men and women who participated in two metabolic studies involving constant diets and lasting 8 to 11 weeks. For one group (16 subjects), nitrogen (N) intake was about 280 mg N/kg per day (mixed animal and plant protein sources) for 8 weeks; for the second group (five subjects), N intake was about 90 mg N/kg per day (egg protein) for 11 weeks. Two mathematical models were examined to determine how well they described the variations in UN. The first modeled only random variability while the second added a component of serial correlation (correlation between successive daily observations). After correcting the UN data for linear trend, to minimize effects of possible alterations in body composition, only two of the 21 subjects showed significant serial correlation of daily UN. Moreover, it is shown that any serial correlation undetectable in these data would not be of practical importance in estimating UN levels in short-term dietary experiments. It is concluded that the interpretation of metabolic N balance measurements involving constant N intakes over relatively short dietary periods in not complicated by cyclic, time-dependent variations in UN.