from conventional construction. I suspended the valve between four 

 spiral springs (see Fig. 25). Thus it can move vertically without any 

 friction. The rubber ring itself is large enough for the small lateral 

 displacements which these springs allow to take place without incon- 

 venience. 



The weight of the flap is such that it will lift itself as soon as the 

 difference of pressure provided for is reached. 



We could have contemplated the construction of metallic flaps. 

 At first sight, such a valve appeared to be perfect. But just then there 

 came into our memory the phantom of the very learned Mayor Otto 

 von Guericke of Magdeburg, with his hemispheres, or, an example less 

 distant in time, the terrible suckers of the giant squid which dragged 

 to his death one of the brave sailors of the Nautilus. 



If two metallic surfaces which are very flat, touch, and if one tries 

 to separate them, it may occur that a vacuum is formed between them 

 and opposes their separation. In the circumstances called 'ordinary' 

 this adhesive force can naturally not exceed the pressure of the atmos- 

 phere, that is, 14-2 lb. per square inch. (This force must not be confused 

 with the adhesion which occurs between two pieces machined with an 

 extreme precision.) But, once again, a bathyscaphe is not always in 

 'ordinary' circumstances. At a great depth the 'sucking effect' can 

 assume tremendous proportions — at 3^ miles, for example, 600 

 atmospheres. This is not pure theory. The poor sperm-whales could 

 say something about the wounds they receive at great depths from the 

 suckers of the squids and which would be impossible if the effect of 

 suction were limited to one atmosphere, which Otto von Guericke 

 noted. 



It may be imagined what would be the consequences if suction 

 effect occurred between our flaps and their seatings, even if it were over 

 a small area of their surface of contact. It must be rendered completely 

 impossible. Fig. 25 shows the construction chosen. It is a supple 

 rubber membrane which rests on the seating. If the flap tends to rise, 

 the membrane is lifted first from its periphery. It can be seen at once 

 that in these conditions no suction effect is to be feared. 



It is true that in the beginning I was told that a valve with a flexible 

 membrane was far too unorthodox and that it was dangerous thus to 

 plunge into a new type of construction. I replied that I had inherited 

 from my parents a pump of which the flaps were composed of supple 

 membranes, that this pump had been working for nearly seventy 



[178] 



