754 Scientific Proceedings, Royal Dublin Society. 
liquids are at any other temperature. This is a great draw-back ; — 
for, in practice, it is very troublesome to bring liquids to a given 
temperature, especially if they are hotter than the standard and 
have to be cooled. 
Perhaps the most serious, because altogether unavoidable, 
source of error is that due to capillarity or surface tension. All 
estimations of density are based on the supposition that the 
distance to which the hydrometer sinks is due to its weight alone; 
but this is not quite correct. ‘The reading of the instrument is 
also affected, and to a very appreciable extent, by the capillary 
attraction of the liquid. As this capillary attraction is generally 
different for different liquids, it follows that a hydrometer will 
necessarily indicate different densities when placed in liquids 
having the same density, but differing in capillarity. 
From all this it is apparent that specific gravities determined 
by the common hydrometer must, in the 
majority of cases, be considered as little 
better than fair approximations. 
Nicholson’s hydrometer, though it has 
no scale, is still adversely affected by capil- 
larity. Indeed, this source of error was 
pointed out by Nicholson himself: more- 
over, the instrument loses nearly all its 
value from the fact that its weight must be 
known, and as this weight is lable to 
change through use, a delicate balance must 
be at hand to determine it when necessary. 
The hydrometer (see figure) which the 
writer proposes is free from the above- 
mentioned defects. It has no arbitrary 
scale ; its weight need not be known; the 
effect of capillarity is totally eliminated, 
and it may be used for any temperature. 
The illustration shows clearly the shape of 
the instrument. The method of using it 
is very simple. It is immersed in any 
liquid, and weights put on the dish at top 
until it sinks to a marked point between the second and third 
bulbs ; additional weights are now put on until it sinks to 
