1892 - 93 .] Mr J. Y. Buchanan on Beep-Sea Water-bottles. 241 
circumstances should he rejected, and are therefore much better left 
unrecorded. 
A very important operation in oceanographic research is taking 
the temperature of the water at a series of depths, and collecting 
samples of it for examination. The more easily and expeditiously 
this can be done, the more likely is it to he done, and the greater 
will be the number of samples and temperatures appearing in each 
series. With the thermometer and water-bottle, each attached inde- 
pendently to the line, a certain interval must necessarily be between 
them. This necessitates stopping the line when the thermometer 
comes up, reading it, hoisting the water-bottle up so as to be in 
position for tapping off the water, resetting the water-bottle, lower- 
ing again a little, resetting the thermometer, and then letting go. 
Recent experience has shown me that it would be a very great con- 
venience if an arrangement could be made so that, the water-bottle 
and thermometer came up at once and at the same level, suitable 
for reading, tapping, resetting, and then letting go. The differential 
motion of the connecting-rod, with regard to the body of the water- 
bottle, seemed to offer the means. The brass tube, meant to hold 
the reversing thermometer, is pivoted at its lower end to a socket 
which slides over the connecting-rod, and can be pinched to it by a 
screw. The cover of the brass tube carries a groove or collar, into 
which a brass hook engages wRen the instrument is upright and the 
stop-cocks open. The brass hook gears into the projecting extremity 
of the upper stop-cock lever, and consequently has a differential and 
downward motion with regard to the connecting-rod when it is 
falling and the stop-cocks are closing. This downward motion dis- 
engages it from the collar on the upper part of the thermometer 
tube, which springs outwards and turns through 180°, and engages 
with the lower part of the connecting-rod, being now in the reversed 
position. The method of fixing the thermometer in its reversed 
position is very simple. A hook is screwed into the heel of the 
tube, and in the place of its rotation. When upright, this book is 
on tbe inside, and a large brass ring wbicb encircles tbe connect- 
ing-rod engages in it. When tbe thermometer falls outward, the 
book pulls tbe ring outwards, and it presently drops down over 
tbe thermometer, encircling both it and the connecting-rod, and 
necessarily keeping the thermometer in its place. The instru- 
VOL. XIX. 28 / 1 / 93 . Q 
