TRANSACTIONS OF SECTION B, 747 
either 5 or 10 mil‘imetres in height, as may be necessary, shows a pale but well- 
defined ‘cap’ in air containing any proportion of acetylene less than the lowest 
explosive proportion. When the hydrogen flame is exposed to the air to be tested 
for acetylene in a darkened space, it is at once tinged yellowish-green. The 
bluish pale cap has the following heights with varying proportions of acetylene, 
-when the hydrogen flame is 10 millimetres in height :— 
0°25 per cent. gives 17 mm. cap. 
05 ” ” 19 ” ” 
10 ” ” 28 ” ” 
2-0 ” ” 48 ” ” 
When the hydrogen flame is reduced to 5 millimetres :— 
2:5 per cent. gives 56 mm. cap. 
2°75 ” ” 7 ” ” 
A convenient portable form of apparatus was shown by the writer, which 
enabled air to be passed readily over the standard hydrogen flame in a darkened 
vessel, and which quickly furnished the reading of the height of the cap. 
In Determining the Limits of Explosibility, when acetylene is mixed in 
gradually increasing proportion with air and kindled, the writer adopted a simple 
method referred to at the last meeting of the Association. It was found that air 
must contain at least 3 per cent. of acetylene before it can be kindled by a flame 
and the mixture caused to burn throughout. As the proportion of acetylene is 
increased, the explosive character is augmented. When 22 per cent. of acetylene 
is present, carbon begins to separate during the burning. The amount of carbon 
which separates increases until the explosive character of the mixture disappears ; 
this point is reached when 82 per cent. of acetylene is present in the air. 
The limiting percentages in air which are explosible are accordingly as follows, 
and may be compared with those already determined by the writer for other 
combustible gases :-— 
Acetylene . . ° ‘ » 8to $2 
Hydrogen . 5 c . Aa gyal es 
Carbon monoxide . : 3 ae elihaege 
Ethylene - F - é ape. 29 
Methane : ‘ 4 ; sa Oli Les 
It will be seen that acetylene gives a wider range of explosive proportions 
than any other of these gases does. Probably this is due to its endothermic 
nature, which leads to the gas being able to generate heat by its own decom- 
position: heat thus generated would undoubtedly aid in causing explosion, and 
would thus extend the limits of explosive mixtures, 
4. The Accurate Determination of Oxygen by Absorption with Alkaline 
Pyrogallol Solution. By Professor Frank Ciowss, D.Se. (Lond.) 
It was found repeatedly in my laboratory that during the absorption of 
oxygen from the Brin gas a considerable volume of carbon monoxide was evolved, 
although this did not occur in absorbing oxygen from air. If the evolution of the 
gas was known to take place, and the carbon monoxide was subsequently absorbed 
by cuprous chloride solution before reading off the residual nitrogen, the estima- 
tion of the volume of oxygen was correct; if this precaution was not taken the 
estimation was open to serious error. Repeated trials with varying proportions 
of eae and potassium hydrate showed that the evolution of carbon monoxide 
might be entirely prevented by using a sufliciently large excess of potassium 
hydrate. With the following proportions no fear of this source of error need be 
felt, even when pure oxygen is being absorbed :—160 grams of potassium hydrate 
and 10 grams of pyrogallol in 200 cubic centimetres of solution. 
3c2 
