TRANSACTIONS OF SECTION D. 729 



protrudes into the septal tissue until it meets a cavity which communicates with 

 the septal canal and with the anterior segment. 



Up to the present the author has not heen ahle to detect any aperture at the 

 extremity of this epidermic tube ; but he is almost certain that such an aperture 

 really exists, and may open under the action of certain muscular fibres. 



The septal canals, as stated above, are the way through which the eggs 

 reach the fifth segment ; but the fact that they exist in the second septum, whilst 

 the segment in front of this has no gonad, shows obviously enough that they must 

 have another function. 



The author ventures to offer the following e.xplanation as to the function of all 

 these well-guarded septal pores, and of the epidermic tubes of the fifth and sixth 

 septa. 



It is a fact easily observed that the worm swells or dilates its body when it 

 wants to adhere to the sheath in which it lives. And in this way it opposes such 

 a powerful resistance, that it is quite impossible to pull it out of its tube without 

 breaking its body in pieces ; and when a part of the body has heen cut off, the 

 remaining segments do not relax at all, hut remain as turgid and resisting as 

 before. This shows that the various segments of the body may swell or relax 

 quite separately. 



The septal organs are the valves which allow the coelomic fluid to flow in or 

 out when they are open, but impose an insuperable resistance to its exit when the 

 worm wants to dilate one or two segments separately. 



The paired tubes of the fifth and sixth septa are very likely intended to take in 

 a small quantity of water from the outside, to be mixed with the coelomic fluid, 

 when a larger quantity of it happens to be required. 



To sum up, although these researches and experiments are not finished, we have 

 sufficient reason to consider the curious septal organs of Ozvenia as valves intended 

 to regulate the pressure in the separate chambers of the perivisceral cavity, and to 

 eventually divide entirely from one another those which at a given moment the 

 animal desires to dilate under the contraction of the muscular coat of the skin. 



If this view is correct, we must regard the body of Oicenia as a very elaborate 

 hydraulic mechanism. 



12. On a simple and efficient Collecting Reservoir for the Surface Tow-net. 



By W. Garstang. 



13. On the Statistics of Wasps. By Professor F. Y. Edgeworth. 



The number of wasps in a nest may be inferred from the number issuing per 

 minute : if (I) we know the average time occupied by a wasp in the cycle of 

 operations between two successive exits, (2) we assume that the whole population 

 is occupied in keeping up the traffic. 



(1) The author has collected some statistics bearing on the first datum. The 

 average time occupied by the wasps which he has observed in loading is six 

 minutes — varying from two minutes, when the load consists of liquid sweets (of 

 Sir J. Lubbock's observations), to ten minutes, when dried marmalade has to be 

 hewed. The average interval between the departure of a laden wasp and the 

 return of the same wasp for another load is — with much less variation — nine 

 minutes. Accordingly tlie mean periodic time for a wasp employed in collecting 

 sweets may be assumed to be about fifteen minutes. This result is verified and 

 corrected by other methods applicable to all kinds of employment — e.(/., stopping the 

 entrance of a small nest and noting the times of arrivals. The corrected figure is 20. 



If the number of wasps issuing per minute is X, the total number would be 

 20 X X, if .assumption (2) held. But how inadequate it is is shown from the fact 

 that the same nest within a short period shows very different rates of traffic. 

 Thus a nest which at the beginning of a week had a traffic (entrances + exits) of 

 twenty per minute, had, after three days, a traffic of sixty per minute ; and, after 

 two more days, a trartic of only twelve per minute, we can at best infer that the 



