482 MR. J. B. LAWES, DR. GILBERT, AND DR. PUGH ON 
3. A simple glass shade is liable to introduce fewer sources of error than a com- 
plicated metallic framework with panes of glass cemented into it. The shade is easier 
to clean before commencing the experiment, and it is less likely to retain, at the termi- 
nation of it, any of the combined Nitrogen, either derived from the plant, or from that 
which has been supplied during growth. Lastly, the presence of oxidizable metallic sur- 
faces, affording a possible quantity of nascent hydrogen which might form ammonia with 
the Nitrogen of the air, is avoided. 
4. There is no organic matter present which can affect the result of the experiment. 
The only organic matter within the shade is that of a thin coating of the gutta-percha 
cement which has been described, by which the tube 7’ (fig. 1, Plate XIV.) is fixed into 
the hole n! (fig. 4, Plate XII.) at the bottom of the stone-ware lute. On analysis this 
cement was found to contain from 0-10 to 0:15 per cent. of Nitrogen. Hence, if the whole 
quantity of the cement in contact with the condensed water became decomposed, and 
yielded up its Nitrogen in such a manner as to become a product of the experiment, it 
would only so yield a few tenths of a milligramme of Nitrogen; but experiment proved 
that it did not suffer sensible decomposition when subjected, during a whole year, to 
exposure in the open air. 
5. In the passage of the air through the apparatus, the excess of pressure was upon 
the inside, instead of, as in the experiments of others, upon the outside of the enclosing 
vessel. In experiments of the kind in question in which the apparatus is exposed to 
the open air, and so subjected to climatic vicissitudes during a considerable period of 
time, the ordinary means of securing tightness in the laboratory cannot be depended 
upon; and an apparatus proved to be tight at one time may, as the result of a variety 
of causes beyond our control, be subject to leakage at another. But a leakage from the 
inside of the apparatus outwards cannot affect the result of our experiment; whilst a 
leakage in the opposite direction might introduce combined Nitrogen from the external 
atmosphere. In the arrangement which has been described, the excess of pressure is 
always on the inside during the passage of the air; and when the air is not passing 
there cannot be any important amount in the opposite direction due to changes of tem- 
perature and barometric condition, for it can never exceed that required to drive the air 
inwards through the bulb-apparatus M (Plate XITI., and fig. 1, Plate XIV.), which is 
altogether insignificant. 
6. That part of the apparatus which would be the most liable to leak, and which 
would be the most damaged by pressure, is subjected to the minimum amount of it. 
The entire pressure required to force the air through the apparatus, independently of 
that necessary to overcome friction, is 
5 x 1:85=9-25 inches of water 
to pass through the sulphuric acid in the bottles B and C (Plate XIIL., and fig. 1, Plate 
XIV.), and 
2:5 x 1:2=3-0 inches of water 
