380. Dr. F. F. Blackman and Mr. A. M. Smith.  [Deec. 19, 
chamber would ever pass away into F and G, but the whole of it could be 
analysed for its CO:-content. 
As the current rate is 300 cc. per hour and each pipette holds 200 c.c., 
this would mean an analysis about every 40 minutes. In practice an 
analysis about every hour is thought sufficient, and for the other 20 minutes 
the current solution drops over into the measuring cylinders. 
(h) Analysis of Gas-bubbles liberated by the Plant—When the CO,-content 
of the current through the assimilation chamber is low there may be an 
almost complete absence of “ bubble-formation”; all the oxygen formed 
being carried off in solution in the liquid. With more active assimilation 
bubbles are liberated in the chamber; these are never pure oxygen, but 
contain always CQg2 in proportion to the CO2-pressure in the liquid. 
It follows that just as in strong COs, bubble-counting alone gives an 
exaggerated value of assimilation ; so also analysis of the diminution of CO, 
in the liquid alone gives too great a value for assimilation, because some of 
the CO, gone from the solution has really diffused into the gas-bubbles that 
are forming. For the correct measure of assimilation this physical loss must 
be estimated and subtracted from the total apparent loss of COs. 
To provide for this the gas-bubbles given off from the plant are separated 
from the Jiquid, continuously collected and analysed at appropriate times by 
means of the valve Y and the gas collector Z. The water-current leaving 
the top of the assimilation chamber by d passes immediately through a 
hollow metal drum Y on its way to the outflow tube . This drum (see the 
large scale drawing of Y at the top of fig. 1) contains an outlet tube (e) in 
its roof, through which the gas that rises above the water, and the gas alone, 
is drawn off to the gas collecting vessel Z. Z is connected with a 
reservoir of mercury, which is lowered so that there is a strong suction 
exercised through f, ¢ to the valved outlet in the top of the drum. 
In the drum is the very light, loosely fitting, hollow metal float p. If the 
drum is full of water this float rises as high as possible and presses the 
little metal disc 7 (mobilely suspended by a strip of oiled silk) against the 
nozzle of the gas-outlet tube. This prevents any escape of water in the 
direction ¢, /, in spite of the suction from the mercury reservoir. When a 
certain amount of gas from the chamber has collected round the float in the 
top of the drum, the float sinks by its own weight, and the disc 7 drops and 
allows the gas to be sucked out through ¢ till the rising water carries p and r. 
up again to close the nozzle. The nozzle-tube e can be screwed down through 
the roof by the flange w till the valve mechanism is working well, and then 
locked in position by the screw collar ¢.* 
* The roof of the drum is not soldered in but fits tightly into a rabbet (et shown), 
