Kinetic investigation was made of the ammonia decomposition using15N as a tracer on doubly-promoted iron catalyst kept at 470°C. The present work was mainly concerned with the two elementary steps of the reaction; that is the decomposition of ammonia to form adsorbed nitrogen and hydrogen atoms (step I) and the desorption of the adsorbed nitrogen (step III) . Observed values for (stoichiometric number of the rate-determining step) were closely distributed near to unity, indicating the step (III) is rate-determining. The ratio of the rate of step (I) to that of step (III) was varied from 4 to 11 depending on the ammonia pressure. These values are much lower than those in our previous work made at 430°C. The comparison of these results suggests that the step (I) is rate-determining at higher temperatures.
Under chronic exposure of ionizing radiation on the growing barley plants for long term, size of mutated sectors induced in a spike would vary from a whole spike to only one spikelet as a unit of screening. But it may be reasonably assumed that mutations in a spike are induced much more at the later stage of the development because of their more initial cells of the spike. Therefore, a peak of the variation of the mutated sector size is expected to be found at the smallest unit. Experimental results showed that almost all the mutants were found as one plant in an M1 spike progeny or as one segregating line in M3 family (cf. Table 1, 2) The following three patterns were classified according to the ascertainment of mutations M2M3 (++) : mutations are ascertained as mutated plants in a spike progeny in M2 and also as segregating lines in M3 family derived from normal M2 plants of the progeny. M2M3 (+-) : mutations are ascertained as mutated plants in a spike progeny only, and normal plant in the progeny give not any segregating lines in M3. M2M3 (-+) : mutations are ascertained only as segregating lines in M3 family. The probability of the occurrence of each case were calculated for various size of the mutated sectors (cf. Table 4) . The expected proportion of the mutational events that are ascertained in M2 to all the mutation ascertained is given in the equation (6) . The proportion of the mutations which are not ascertained throughout M2 and M3 due to the occurrence of normal genotype in a mutated spikelet can be estimated from the equation (7) . For an numerical example, experimental data for five treatments including two exposure stages and three exposure rate were presented and the mutation rate were estimated with correction of bias due to sector size by applying the present method (cf. Table 5) .