220 Proceedings of the Royal Society of Edinburgh. [Sess. 
On the Occasional Abnormal Behaviour of the Densimeter. 
After the apparatus had been in use for some time, the density values, 
when compared with those derived from pyknometer measurements, became 
incorrect. The discrepancies, though small at first, would frequently in- 
crease quite suddenly and cause a maximum difference of about db'002 in 
the density value. Numerous experiments proved that this error could 
not be due to any defect in the design of the instrument, but was the 
result of some change in the condition of the interior surface of one or both 
of the glass tubes. A certain oily appearance could sometimes be detected. 
At first it was thought that the cause might be some growth, similar to 
that which has been observed by Dr Yeley in burettes which have been 
in use for some time. This theory received some support from the fact 
that the densimeter gave perfectly normal values after it had been re- 
cleaned. The improvement was not, however, maintained by allowing any 
air that passed into the apparatus to filter through cotton-wool ; we may 
therefore safely conclude that the abnormal condition of the tubes was 
not due to a growth of bacteria introduced from this source. Further, the 
tubes occasionally assumed a decided oiliness after they had been cleaned, 
so rapidly as to suggest that the trouble was due not to any organic 
growth, but to some volatile constituent of the vaseline used for lubri- 
cating the taps.* 
With the object of verifying this, all vaseline was removed, the appar- 
atus cleaned as already described, and finally steamed out. In some 
experiments honey, and in others glycerine, was employed as a lubricant ; 
these, as a rule, produced a decided improvement, but sooner or later the 
glass tubes again became affected. The last possibility was that the oiliness 
might be due to some volatile emanation from the black vulcanised rubber 
connections at the top. This possibility was put to the following test. 
A flask was made chemically clean by treating it with a mixture of ethyl - 
alcohol and fuming nitric acid, and repeatedly rinsing with distilled water ; 
it was then partly filled with distilled water ; whenever the flask was shaken, 
the water wetted the upper parts and then drained in a perfectly normal 
manner. A short length of the rubber tubing was now introduced ; a very 
* Mr Gunther suggested that bacteria were already present in the waters. This, how- 
ever, appears improbable ; for it was found that, after the india-rubber tubes had been 
purified and the use of vaseline discarded, another portion of the water with which the 
tubes were charged when much trouble had been experienced, could be kept in the appar- 
atus for three or four weeks without producing any ill effects ; micrometer readings taken 
at various times during that period were always normal. 
It may also be added that, although the author has examined many hundreds of samples 
of sea-water, he has never observed the development of any similar oiliness within the 
containing bottles. 
