In old-growth forests, the complexity of forest structure affects the spatial difference in micro-environmental factors such as soil temperature and soil water content. If soil respiration in an old-growth forest is estimated with a low accuracy, it causes an error estimation of net ecosystem production in a large scale. Therefore, the heterogeneity in soil respiration was studied using soda lime in 1-ha study site in 2012 and AOCC method with IRGA in canopy-gap areas in 2013 in an old-growth beech forest, Mt Hakusan. The spatial pattern with the different vegetation and micro-environmental factors showed the high efflux in canopy and the low efflux in gap. All soil effluxes increased from spring (Jun.-Jul.) to summer (Aug.), and then decreased in autumn (Sep.-Nov.). The seasonal pattern showed the hysteresis loop that soil respiration in spring was greater than that at the same temperature in autumn. Diel soil efflux was greatly controlled by soil temperature but a diel lag between soil respiration and soil temperature led to diurnal hysteresis loop in some season.