214 DISCOVERY REPORTS 



space, that is, the fiher chamber, must enlarge. Water will be sucked in and will pass 

 from the mid-ventral space through the tips of the setal combs. It cannot pass in 

 laterally because of the solid mat of setules. Some of the particles in the stream will be 

 entangled on the tips of the setal combs. The remainder will pass into the chamber and 

 these, as I have explained, cannot get out. They will be caught on the setules of the 

 filter setae. 



When the limbs move up again towards the body, this, in itself, will diminish the 

 volume of the filter chamber and so tend to force out its contained water. However, a 

 real suction pump exists which actively sucks water out of the chamber. This is formed 

 by the first and second trunk limbs, and their inter-limb space I call the suction chamber 

 (Fig. 5 A). The valve of the pump is provided by the epipodite of the first trunk limb 

 which, as I have explained, completely covers laterally this inter-limb space. When the 

 second trunk limb moves backwards it will tend to increase the volume of the suction 

 chamber and hence water will be sucked in. Now this cannot pass in laterally as in the 

 case of the more posterior trvmk limbs for the epipodite will flap against the second 

 trunk limb and act as a closing valve. Some will pass in from the mid-ventral space and 

 will be filtered on to the setae of the second trunk limb, but the majority will be sucked 

 out of the filter chamber. 



From Fig. 2 it can be seen that the suction chamber is roofed by the filter setae of the 

 first trunk limb. This is most clear on the right-hand side of the figure. When the 

 second trunk limb is dragged backwards its tip will probably touch the limb in front. 

 The suction chamber is now closed ventrally and also laterally by the epipodite. Water 

 sucked in must hence pass either from the median space or through the roof of the 

 chamber, that is, from the filter chamber. From Fig. 5 A it can be seen that this roof 

 forms a much greater area than the median wall and so it is probable that the majority 

 of the water passes in through the roof. 



Water drawn from the filter chamber into the suction chamber will pass out back- 

 wards when the second trunk limb moves forwards relatively to the first. When this 

 happens the space between the two limbs will diminish and so water will be forced out. 

 The epipodite will now flap outwards and allow a free passage posteriorly exactly as in 

 the case of the combined exopodites and epipodites of Nebalia. 



The peculiar re-curvature of the first row setae of the second trunk limb is difiicult 

 to explain. I can only say that if they projected directly forwards in the same direction 

 as the homologous setae of the more posterior trunk limbs, they would project right 

 through the filter apparatus and interfere with its functioning. The direction of the 

 setae of the second trunk limb is, on the whole, at right angles to those of the first trunk 

 limb. The curved setae thus lie across the filter setae, and I suggest that they act as a sort 

 of bufi^er to the latter preventing them being forced backwards too far when the filter 

 chamber is diminishing. They could not act in this way, even with their re-curvature, if 

 they pointed in the same direction as the filter setae as they would slip in between the latter. 



The particles of food filtered off on to the walls of the filter chamber have now to be 

 transferred to the mouth. One of the most striking things about Nebaliopsis is the 



