The study of isolated phloem in Heracleum has been extended to intact functioning sieve tubes. Techniques of phloem dissection combined with Nomarski interference optics have been developed to permit useful visual observations, photomicrography, and ciné photographs of sieve tubes which are apparently normal. In these preparations, plastids and organelles called by us ‘marker particles’ are visible in rapid bouncing motion, and the state of dispersal of these particles is related to the amount of damage done to the preparations. The movement of the marker particles and their subsequent fixation shows that they are apparently attached to or restrained by an invisible network in situ. The network is very sensitive to disturbance and readily collapses around the sieve plates to form slime plugs upon damage to the sieve element. The marker particles do not move through the cell nor across sieve plates in mature Heracleum. In young cells cyclosis is observable along the periphery and this suggests that a vacuole may then be present. In undamaged mature cells there was evidence neither of a vacuole nor of trans-cellular tubules of any size optically detectable. The motion of the particles was greater than Brownian movement and appeared to be under some physiological control. Their movement probably indicated the presence of an operating transport phenomenon, either because solution was moving past them in the sieve tube or because they were themselves attached to a contractile network actively in pulsatory motion. Nearby companion and parenchyma cells showed normal cytoplasmic streaming. Proposed mechanisms of translocation involving cytoplasmic streaming do not seem to be applicable to the phloem of Heracleum. The movement of the marker particles seemed to agree best with a mechanism of ‘activated’ mass flow.