1892 - 93 .] Mr J. Y. Buchanan on Deep-Sea Water-bottles. 239 
exposes its edge to the passing water during descent, hut falls out 
at right angles to the axis of the instrument when the motion is 
reversed. The resistance which it thus offers to the passage of 
water produces a downward pressure sufficient to shut the stop- 
cocks and enclose the sample. This arrangement acts perfectly 
well, the only objection to it being that it requires a certain amount 
of practice and experience. For this reason the collection of inter- 
mediate waters on board the Challenger was never delegated, but 
was always superintended by myself. 
Where the “ messenger ’’ is used, no skill and no experience are 
required. The stop-cocks are supported by a line or wire from the 
upper end of the connecting-rod to the trigger. When the messen- 
ger falls on the trigger, the stop-cocks close by the weight of the 
connecting-rod, and, if necessary, attached weights. When used 
without the messenger, the point which requires experience and 
practice is to regulate the stiffness of the stop-cock, so that there 
shall he no danger of its closing by the weight of the connecting- 
rod and flap during descent, but shall certainly close when exposed 
to the entire pressure produced by the passage of the extended flap 
through the water. When the stop-cocks are fixed open until the 
messenger strikes the trigger, no such adjustment is necessary ; and 
the connecting-rod can be made as heavy and the stop-cocks as free 
as may be desired, so that there may be no question about the 
stop-cocks closing immediately they are released. This arrangement 
was used on board the yacht “Princess Alice,” owned and com- 
manded by H.S.H. Prince Albert of Monaco, during a cruise in the 
Mediterranean in September 1892. 
The “ messenger ” is by no means a recent invention ; it has been 
used by many early investigators, even in water of some depth. It 
was notably used by Aime in his very remarkable investigations in 
the Mediterranean in the years 1846-48. Its great merit is cer- 
tainty. It is often objected that it is too wasteful of time to be 
used in any considerable depth. But this is not so. The messen- 
ger, like any other body falling through water, attains a maximum 
and constant velocity before it has fallen more than a few fathoms. 
Its rate of descent can be perfectly accurately ascertained to a second 
by determining the time taken to reach the trigger at (say) 50 and 
150 fathoms. The difference between the two times gives the 
