Observations of Comet 9P/Tempel 1 indicate that the boundary of the thick, fluid-like deposit had retreated by up to 50 m during one orbital period (Veverka et al., 2013; Thomas et al., 2013). According to Kossaki (2015), sublimation of buried ice may be responsible for the observed effect, but only if the dust has very high thermal conductivity. This problem is investigated using new laboratory experiments about the sliding of dust on an inclined ice surface. New numerical simulations were also performed using a model that uses values of material parameters derived from in-situ investigations of comet 67P/Churyumov-Gerasimenko. We have found, that sub-dust sublimation of ice can be responsible for the observed retreating of the boundary only if the concentration of ice in the deposit is lower than the average for the nucleus. Thus, the material of deposit was probably ejected from some place of the nucleus and out-gassed in the space. It could be also desiccated before sliding down from slope of a surrounding rise. This is likely only for the dust mantle of limited thickness. It means that several slides down are necessary for forming of the observed thick deposit.