The initial stages of the photolysis of cellulose under vacuum can be followed by the gas production. A photolysis cell is described which allows a rigorous temperature control and gives a reproducible photolysis surface of the solid carbohydrate sample. A mass spectrometer was used to analyze the gases. To measure accurately the small amounts produced in a small volume requires careful calibration of this instrument, together with a fixed routine for the experimental procedure. Using this apparatus, we have photolyzed hydrocellulose under vacuum with 253.7 nm light. Only H2, CO, and CO2 were evolved, the rate of H2 production increasing to a constant value, that of CO increasing more slowly without becoming constant, and that of CO2 decreasing rapidly to a constant value. No induction periods were observed. The gas evolution was not caused by residual O2 or CO2 adsorbed on the polymer, but the rates were altered by leaving the photolyzed sample under vacuum overnight or heating the unphotolyzed sample under vacuum at 100°C. The rates of gas production increased markedly when the light filter was changed to allow a combined photolysis by 253.7 and 184.9 nm light. The results illustrate that stringent precautions must be taken to filter off the latter line in polymer photolyses with Hg arc light if the results are to be meaningfully interpreted in terms of 253.7 nm light photolysis. On photolyzing hydrocellulose previously boiled in NaOH solution, a larger H2 production rate and a smaller CO2 rate (compared with untreated hydrocellulose) were observed. It is suggested that this standard alkali treatment is the reason for the differences between our results and the results of the photolysis of cellulose reported in the literature.