within the tub. The precision of this measurement is not great, a ton 

 inducing a variation of only 2-5- in. But that is sufficient, because if, 

 after a dive, an error of ^loo lb. were made in our estimate of the 

 ballast, it would be unimportant. The essential thing is to know, more 

 or less, whether the tubs are full of pellets or not. Let us suppose, 

 indeed, that through a small leak we have lost little by little a large 

 quantity of petrol and that the loss of static lift so produced has been 

 compensated, without our observing it, by our losing bit by bit all the 

 ballast. After a deep dive in these conditions we should not be able to 

 surface again, because we should not have at our disposal sufficient 

 ballast. It is true that the loss of petrol ought to have been perceptible 

 by verifying the level of water in the float, by means of the electric 

 sounder, but a double control is valuable just the same. 



May I be permitted here to say a few words on the subject of a 

 system of ballast control by a magnetic sounder such as I had pro- 

 posed for the FNRS 3. A small electro-magnet with a laminated iron 

 case and with an open magnetic circuit shows a certain self-induction. 

 If this magnet were to be plunged into the iron pellets, the magnetic 

 circuit would be closed and its self-induction would considerably 

 increase. If several magnets were disposed through the ballast tank 

 we could easily, by means of a suitable electric instrument mounted 

 in the cabin, determine in the course of a dive the level of the ballast 

 and thus calculate the reserve still at our disposal. This could be very 

 useful. 



At the time the Trieste was building, I had to do without this 

 measure for the following reason : I wanted it to be possible, in case of 

 the outlet being blocked, for the ballast tubs to be thrown overboard 

 altogether. If we had kept these tubs the shape of a high narrow tank, 

 lodged in the float, the risk would have been, in the case of the bathy- 

 scaphe listing heavily, that the tub would have become blocked in its 

 housing as it emerged, due to the leverage exerted by its length. As the 

 necessity for throwing overboard the whole tub occurs above all in 

 case of accident, we had to reckon with the possibility of the bathy- 

 scaphe listing heavily. To prevent all possibility of the tubs being stuck 

 in their housings, we had to construct large shallow tubs, having such 

 clearance that, even if inclined at 45° in relation to the float, they could 

 not be jammed. But with this shape the height of the level at any single 

 point would not give the necessary information. We should have had 

 to have a great number of magnetic sounders. 



[157] 



