NATURAL PHILOSOPHY. 209 



top with such tenacity. It explains Bertholet's experiment on the 

 forced dilation of fluids, in which a deaerated fluid, sealed when hot, 

 does not shrink in cooling for a long time, but at last breaks and col- 

 lapses, indicating that it has borne a great tension before yielding. 

 Prof. Henry's elegant experiments with soap-bubbles, in which, by 

 measuring the tension of the enclosed air, he is able to deduce first, 

 the compressing force, and then the cohesion of the fluid film Avill 

 furnish an independent confirmation of the same general views. We 

 may remark that the heterogeneous structure of the outer layers 

 would destroy the mobility of their parts, and give a film-like charac- 

 ter to the fluid surfaces, while all within the film would have free mo- 

 bility. This, with the additional fact of a drawing inward of the out- 

 ward layers by the unbalanced cohesive action of the layers near the 

 surface, explains the great variety of formal phenomena exhibited by 

 drops, bubbles, and fluid surfaces. 



About four years since, I conceived the idea of directly measuring 

 fluid cohesion, by rupturing a pure fluid column in a cylinder with a 

 moving piston. By filling the cylinder with the fluid to be tested, and 

 immersing the piston, by the aid of a valve closing at will, the force 

 requisite for starting the piston will be the cohesion of the column, on 

 allowing for atmospheric pressure. Of course the fluid must adhere to 

 the cylinder more strongly than it coheres in itself, else the adhesion 

 only would be measured. Nor must it contain any air-bubbles, as the 

 presence of one such, however small, will give a start to the break, by 

 presenting a weak surface. 



I anticipate that exceeding small air-bubbles will have the effect of 

 making the indications irregular, as the smallest bubbles will only start 

 a break on the application of very considerable force. 



I will now apply this discussion to steam, lioiler explosions. The condition 

 requisite for ebullition in boiling Avater is simply that air-bubbles in the 

 heated portions shall present on their borders the weakly coherent sur- 

 faces requisite for evaporation to be established. Perfectly deaerated 

 water, with a limited surface, would not boil at all, but would steadily heat 

 up until it reached that point at which it would flash explosively into steam. 

 Now, one chief chief cause of local explosions is clearly of this descrip- 

 tion. The boat stops at a wharf ; the "doctor," or pump supplying 

 water to the engine, being worked by the engine itself, stops its 

 water supply when the engine stops. The water in the boiler goes on 

 boiling until all the air bubbles are boiled off from the water, and their 

 air is mixed with the steam above. Then there ceases to be any evapora- 

 ting surface, except that on the top layer, which is farthest from the heat- 

 ing surface, and quite inadequate to the consumption of the heat supplied. 

 Then the mass of water begins to heat up, and it goes on storing up the 

 unconsumed caloric, until the water is far hotter than the head of 

 steam would indicate. The engineer then starts the engine ; this starts 

 the pump, which throws a stream of air charged with water directly 

 ir>to the glowing fluid. The heat instantly finds its outlet by an over- 

 whelming evaporation on the newly supplied bubble surfaces, and a 

 tumultuous ebullition follows. The gathered store of heat flashes off a 



