1. Experiments with suspended particles on living Tornaria show that particles move everywhere from the raised fields into the nearest part of the system of gutters (sulci). 2. A continuous current runs in the gutters towards the mouth. 3. Both phenomena are traced to the transverse beating of the cilia along the margins of the gutters. 4. The festooned or tentaculate parts of the ciliated bands serve the collection of particles; the plain-edged gutters are simply propulsive and transmissive; the lateral bays are traps for stray particles. 5. The mouth is small, on the anterior side of the floor of the oral gutter, beneath a canopy formed by the overhanging anterior wall which bears the pre-oral band. There is no adoral band. 6. The oesophagus is ovate-cylindrical, uniformly ciliated, not dilated to form an oral cavity, and without a ventral loop. 7. Above and below the mouth, in the sulcar depression, are two thickened tongue-shaped patches bearing long cilia. The ‘sub-oral’ is regarded as inhalant, driving food-particles inwards over the ventral lip; the ‘supra-oral’ as exhalant, sweeping outwards the surplus water. 8. An additional discharge of surplus water probably takes place before reaching the mouth between the overlapping edges of the ventral sulcar convergence, and is probably swept below by Spengel's ‘ciliated streak of the ventral saddle’. 9. The feeding mechanism is compared and contrasted with that of Bipinnaria, as described by Gemmill (Asterias) and by Tattersall and Sheppard (Luidia). 10. The relations of the ciliated bands to locomotion and feeding are examined in general. It is urged that in all cases the cilia beat transversely to the course of the band.