Three arc (or spiral?) jets were observed at Teide Observatory (Tenerife, Spain) in 1995. The events were separated approximately one month interval and occurred on August, September and October when Comet Hale-Bopp 1994 O1 were at ~7 AU from the Sun. It seem that the process developed as follows. The surface layer of the comet had a region (or regions) enriched by CO. By the thermodynamic reason any solid mixture of CO and H₂O must form an eutectic or, roughly speaking, grains of water ice embedded into solid CO. Since the abundance of CO was ~=0.1 of that of H₂O then the nuclear material was a conglomerate of icy dust glued together by solid CO and CO grains glued by water ice.

At large heliocentric distances only CO can sublimate. This stabilizes the temperature of the surface initially at the level of sublimation equilibrium of CO, i.e. 30-40 K. When CO is sublimating the water ice grains are accumulated at the nuclear surface. The surface layer becomes a porous medium of H₂O ice grain. The plane which separates the CO-H₂O mixture and H₂O "dust" is moving inward. This process changes the thermal regime of the surface layer. The temperature of the outer surface is no more controlled by sublimation of CO. It starts to grow up to the radiation equilibrium temperature T_rad ~= 153 K at r ~7 AU. This temperature is just those of crystallization of amorphous ice. Therefore the wave of crystallization will be propagating inward following the front of sublimation of CO.

After the heat wave, generated by crystallization, reaches the CO enriched layer under water ice grains the pressure of CO vapour increases and the mantle of water ice grains is blowing up. The outer parts of the icy dust layer are moving first and get the maximal velocity. Then the velocity decreases together with gas pressure because the mantle becomes thinner and more permeable. After all the water ice particles leave the active region the naked CO+H₂O surface returns to the sublimation equilibrium at 40 K.

This process is impossible at the heliocentric distances less than 5 AU because sublimation of water ice changes the thermal regime of the nuclear surface. Icy grains are no more able to form the heat protection mantle. This function takes usual mineral dust. So any dust jet or envelop at short heliocentric distance should have another nature: it should be of mineral dust which is blown up by water vapour.

A method is proposed to solve the inverse problem: to determine the kinematic picture of jet formation, the velocity of ejecta as a function of time, parameters of rotation of the nucleus etc directly from the observed movement of a jet. This method is applied to the jets observed at 7 AU and the results are presented and discussed.