ACOUSTICS AND GRAVITATION. 



91 



(open below) was allowed to rest directly on the porous cup (top above c, 

 figure no); but no resulting difference appeared. Other conditions clearly 

 prevail, for in this case the mouth of the swimmer is all but closed by the top 

 of the porous cup and yet no marked change of diffusion follows. 



To account for this anomalous result, the temperature graph may be con- 

 sulted. It then appears that the kinks in the B graph closely follow the 

 changes of temperature, if the latter are plotted positively downward. The 

 phenomenon thus seems to be associated with the solubility of air in water. 

 In this way a reason for a degree of ineffectiveness of a wide neck may be also 

 surmised; for, as a result of changes of the volume of the charge of air, the 

 water, more or less laden with air by solution at different temperatures, is 

 alternately expelled from or enters the diver. The mechanism is thus depend- 

 ent on solution and convection, and only secondarily on diffusion. 



The results for the vessel F, figure 112, in which the diver was much lighter 

 and more mobile, bears out the same surmise even more fully. The relative 

 scale is here somewhat larger and agreement between the temperature fluc- 

 tuations and the kinks in the F graph more complete. This graph also 

 shows a steady ascent, the mean value of which may be estimated as over 

 0.80 per cent per day. Here again there is no adequate effect attributable 

 to diffusion of air through the porous cup, which may, therefore, be discarded 

 as useless. 



Thus it follows that in the air-air diffusion the solution effect is of predomi- 

 nating importance. Indeed, if the parts of the two graphs between February 25 

 and February 27 are compared, the g/m increase amounts to 5,600 per day; the 

 corresponding temperature decrease is 4.6 per day. Hence the g/m rise should 



