1) Seasonal variation of the results in the micro determination of carbon and hydrogen was discussed statistically from the results of the test analyses.
2) The micro-balances used had an accuracy of σw_??_2λ, sucrose of 2_??_4mg was used chiefly as the test material, and its contents of carbon and hydrogen were determined after the method of PREGL with some modifications using an automatic controlled combustion apparatus(3)
3) In 406 analyses (during 2.5 years), C% found showed a good consisteney with that calculated, namely mean error=+0.025% and σc=0.15%. No special care for the air conditioning was taken, yet no seasonal variation of the data was observed (Table 1 ). As somewhat higher results (ca. 0.2%) were expected in summer by the diffusion of the humidity into the absorption tube, it was supposed that the sample might absorb humidity from the air while it was being weighed.
4) In 108 analyses (during 1 year), H% found showed a good consistency with that calcu-lated on the average (mean error=-0.03%), but the fluctuation of the data (OH=0.19%) was somewhat larger than that of C%. Besides, H% found was high in summer and low in winter. These facts might be explained as follows.
a) The amount of water introduced into the reaction tube while the platinum boat was taken out and in it might be canceled, on the average, by the amount of water temporarily taken hold by PbO2.
b) But the amount of the former was high in summer and low in winter, and that of the latter might not be constant in every determination.
