Oxygen hy Sea- IFceds. 247 



partly because of the varialiility in the intensity of the 

 heat of the sun, which induced a more or less rapid increase 

 or diminution of the temperature in the upper part of the 

 apjtaratus, which the convection currents could not perfectly 

 antagonise, and partly because the experiments had to be 

 continued during the night, when the apparatus could not 

 be attended to. 



The significance of this factor I endeavoured to determine 

 by erecting a second apparatus prepared like the first, but 

 without any plants, and exposed in every way to similar 

 external conditions. A similar current of sea-water was 

 passed through the apparatus from the tank, and in it the 

 gaseous molecules given off by the water itself in virtue of 

 changes of temperature accumulated. In this manner the 

 part plaj'ed by the water as such was determined, and in 

 any case, where even a very small accumulation of gases 

 took place the experiment was repeated. When a hardly 

 appreciable quantity of , gas had accumulated in the second 

 apparatus, the influence of the changes of temperature on 

 the water was regarded as insufiicient to modify the read- 

 ing in the apparatus in which the experimental plant or 

 plants were exposed, and this reading was accordingly 

 accepted. 



It need hardly be noted that, in admitting a current of 

 sea-water into the vessels through the india-rubber tubing, 

 great care was taken so to regulate the conditions that 

 bubbles of air were not passed in along with the water, nor 

 were swift surface eddies set up, as these might have caused 

 an abnormal amount of air to pass into the circulating 

 water. 



After a sufficient amount of gas had accumulated in the 

 collecting tube, the entire apparatus was removed and 

 immersed in a tubful of sea-water which had been prepared 

 for some time so as to settle and acquire a temperature 

 equal to that of the water in the apparatus. The depth of 

 water in the tub was such that the lower end of the collect- 

 ing tube of the experimental apparatus could be removed 

 from the jar without coming above its surface. The 

 accumulated gas was now decanted into a clean eudiometer 

 which was carefully transferred, its open end being closed 

 by the thumb, to a chemical stand, and on this it was clamped 



