Few and conflicting reports have so far been published on the urinary excretion of luteinizing hormone (LH) in childhood (for reference, see Leone & Sciarra (1966 a) and Rifkind, Kulin & Ross (1967). Further investigation seemed therefore indicated.The daily excretion of LH was determined semiquantitatively in 24 normal, healthy and non-hospitalized prepubertal children (12 boys and 12 girls), aged 5\p=n-\10 yr. A complete 48 hr. urine specimen (1000\p=n-\2000 ml.) was collected from every subject without the use of a preservative and stored in a cold room at 4\ s=deg\ .Within 3 days of the urine collection, each sample was extracted by Johnsen's tannic acid method (1958) but using, as a rule, the highest concentrations of reagents suggested by the Danish author. Twenty-four hours before assay, the dry precipitates were ground to fine powder and redissolved in a 0\m=.\9% NaCl solution. As described by Sciarra, Pastorino & Leone (1968) the powders proved so soluble that they could be dissolved in very small amounts of saline 0\m=.\5\p=n-\ml.); consequently, urinary LH was concentrated up to 1000-4000 times with a 67 + 5 % recovery rate. After centrifugation, the supernatants were serially diluted with intervals of 1:1-5. LH was assayed by the immunological method of Wide & Gemzell (1962) but using the reagents of the Pregnosticon Test (Organon), the sensitivity of which was previously established at 1-6 i.u. LH (2nd IRP-HMG)/ml. (Leone & Sciarra, 1966&). The excre¬ tion results were corrected for the 33 % mean methodological loss by tannic acid extraction.In order to evaluate the results, reliability criteria for the method and regression of LH excretion on age for both males and females, with 95 % confidence limits, were calculated. Tests for deviation from parallelism were carried out. The accuracy of the method was in the region of 100 % ; its precision was high (mean = 0-12, range 0-08 to 0-15); its sensitivity was in the range of 0-6-1-2 i.u. LH/24 hr. urine vol. With regard to specificity, the test failed with several urinary extracts from two hypophysectomized young women; virtually no LH activity could be detected in the same extracts by the ventral prostate weight method as modified by McArthur (1952). However, the results were positive in pregnant women. Figure 1 shows that LH was detected in all the children investigated ; the mean urinary hormone level in boys was 2-9/i.u. LH/24 hr. (range 1-2-5-4) and 2-8 i.u./ 24 hr. in girls (range 0-6-5-4); it increased with age in both sexes and more rapidly in females than in males.These results, together with the observations of previous workers who demon-