July 25, 1878] 



NATURE 



349 



ciently reliable for the purpose ; a new, more accurate, and more 

 delicate instrument was therefore a great desideratum, and it 

 seems that Messrs. Negretti and Zambra have now solved this 

 problem in a satisfactory manner. 



The construction of the new thermometer will be readily 

 understood by reference to Fig. 3. The bulb is cylindrical, 

 and mercury is the thermometrical fluid. The neck of the bulb 

 is contracted in a peculiar manner at A, and upon the shape and 

 fineness of this contraction the success of the instrument mainly 

 depends. Beyond A the tube is bent and a small catch reservoir 

 is formed at B, for a purpose to be presently explained. At the 

 end of the tube a small receptacle c is provided. When the 

 instrument is held bulb downwards it is seen to contain sufficient 

 mercury to fill the bulb, tube, and a part of the reservoir c, 

 leaving sufficient space in C for the expansion of the mercury 

 when the temperature rises. In this position no scale would be 

 possible, as the apparent movement of the mercury would be 

 confined to the space c. When the thermometer is held bulb 

 upwards the mercury breaks off at A, but by its own weight 

 flows down the tube, filling c and a portion of the tube above c, 

 this portion being in proportion to the existing temperature. 

 The scale accordingly is divided from c upwards. To prepare 

 the instrument for observation it is only necessary to place it 

 bulb downwards, the mercury of course assuming the tempera- 

 ture surrounding the instrument in the same way as any ordinary 

 thermometer. When at any time or at any place it is required 

 to read off the temperature, all that has to be done is to turn 

 the thermometer bulb upwards, and to keep it in that position 

 until the reading has been taken. This may be done at any 

 time afterwards : for the quantity of mercury in the lower part 

 of the stem which records the reading is too small to be appre- 

 ciably influenced by a change of temperature, unless it be very 

 considerable, while that in the bulb will continue to contract 

 with greater cold and to expand with greater heat ; and in the 

 latter case some mercury will'pass the 'contraction A and may 

 fall down and lodge at b, but 

 it cannot go further as long as 

 the bulb remains upwards ; and 

 thus the quantity recording the 

 temperature in the stem will 

 remain unaltered. It is seen, 

 therefore, that this new instru- 

 ment is simply a recording 

 thermometer, which acciurately 

 and delicately records the tem- 

 perature at the time and place f 

 when and where it is turned 

 over. It is not intended and 

 cannot be used as a self -regis- 

 tering maximum or minimum 

 thermometer. 



Of course some contrivance 

 must be provided for turning " 

 the thermometer bulb upwards , 

 at any depth in the water. For 

 this purpose the instrument is r 

 fitted into a wooden frame - 

 loaded with shot, free to move f 

 from end to end of it, and with " 

 its weight so regulated as to 

 render the whole just buoyant t^ 

 in sea-water. t 



In using the thermometer a \ 

 cord is passed through the hole : 

 in the frame nearest the bulb, 

 and the instrument is fastened 

 by this cord to the sounding- 

 line. In descending, the ther- 

 mometer will be pulled down 

 with the bulb downwards ; but 

 upon being pulled up, the in- 

 strument, owing to the resist- 

 ance offered by the water, will 

 turn over and come up bulb uppermost ; the temperature of the 

 spot where it turned over will then be indicated. The illus- 

 trations we subjoin will fiirther elucidate this matter. 



As regards the thermometer itself, it was necessary, in order 

 to make it perfectly accurate, to protect it against pressure, even 

 if intended for shallow seas, as well as for the deepest. For 

 whether used in deep or shallow water, unless so protected, its 



Fig. I. — Descending. 



indications would always be more or less faulty. Like an ordi- 

 nary thermometer, it is devoid of air, and thus differs materially 

 from Six's, which, containing compressed air, has a certain 

 internal power of resistance. Hence the new instrument would 

 be more affected by pressure than Six's, however thick the glass 

 of the bulb. By the simple expedient of placing the entire ther- 

 mometer into a glass shield or sheath, and hermetically sealing 

 the latter, the effect of external pressure is entirely eliminated. 

 The shield must of course be a strong one. It need not be ex- 

 hausted of air. Its effect, however, will undoubtedly be to 

 render the thermometer it protects against pressure less sensitive 

 towards changes of temperature ; in other words, it v;ill make 

 it sluggish. To counteract this sluggishness, some mercury is 

 introduced into that portion of the shield -which surrounds the 

 bulb, and is confined there by means of a partition cemented 

 in the shield round the neck of the thermometer bulb. The 

 action of this mercury is that of a heat-conductor from the exte- 

 rior of the shield to the interior of the tliermometer, and the 

 efficacy of this arrangement ias been proved by experiment, 

 the instrument thus protected being, in fact, far superior in 

 sensitiveness to Six's thermometer. 



As long as the shield withstands the pressure, the thermometer 

 will be unaffected, and there is abundant evidence already to 

 show that a shield of the above description will resist the pres- 

 sure even at the bottom of the deepest ocean ; doubtless it will 

 be compressed a little at great depths, but the pressure will fail to 

 have any appreciable effect upon the thermometer itself. This 

 method of shielding is quite efficient, and thermometers thus 

 protected need not be tested for pressure in the hydraulic press ; 

 all that is necessary is to test them very accurately for sensitive- 

 ness and for errors of graduation. The new instrument is in- 

 tended to be a standard instrument and permits the reading off 

 of at least two-tenths or even one-tenth of a degree. The test for 

 sensitiveness determines how many seconds the instrument requires 



Fig. 3. Fig. 2.— Ascending. 



to note a change of 5° rise or fall, and the time has been found 

 to be from five to ten seconds. 



A considerable number of these instruments have already been 

 tested at Kew Observatory with perfectly satisfactory results, 

 which place beyond doubt their value as standard deep-sea 

 thermometers. 



Thus, provided the turning-over gear is found to answer, this 



