236 DISCOVERY REPORTS 



PROTECTED THERMOMETERS 

 Change with Age in the Zero Point and Scale Corrections 

 Most of the information concerning the thermometers under review is best given in tabular form. In 

 Table i we have, therefore, listed the physical particulars for the protected deep-sea thermometers, 

 i.e. range, volume of mercury reversed at o° (usually referred to as Fq) and the periodic zero error, 

 together with data regarding the thermometers when actually in use. Unfortunately these data are 

 not as complete as we would wish ; for certain thermometers they are, however, considered sufficient 

 to give a reasonable indication of the reliability to be expected from that instrument. 



The table shows that the zero point generally rises with increasing age. This is a normal charac- 

 teristic of all conventional, or standard type thermometers and, as already mentioned, reversing 

 thermometers do not seem to provide any exception to this rule. Generally any major shift of zero 

 occurs in the early years of the life of a thermometer, a fact seemingly borne out by the figures avail- 

 able for our protected deep-sea instruments. If we consider the various age groups in order, the 

 following facts concerning the zero point appear. 



2g-groiip (1924-53). Four thermometers — nos. 854, 869, 871 and 876. Only one thermometer (869) shows a zero 

 drift of more than 0-03'' and, for all thermometers, the greater part of the drift took place between the years 1924 

 and 1949. Between checks in 1949 and 1953 two thermometers registered nil drift and the other two only o-oi°. 

 All of Richter and Wiese's make. First check at P.T.R., remainder by N.P.L. 



zS-groiip (1925-53). Three thermometers — nos. 971, 1023 and 1025. Drift of zero point for all three thermometers 

 very similar. Maximum drift 0-04°, minimum 0-03°, with the major change again occurring before 1949. Between 

 1949 and 1953 no change for one thermometer, o-oi° for the others. First check at P.T.R., remainder by N.P.L. 

 All of Richter and Wiese's make. 



20-gronp (1933-53). Three thermometers — nos. 1008, loii and 1257. Little change in period under review. 

 Maximum shift of zero was 0-03°, minimum 0-02°. Two thermometers showed major drift between 1933 and 1949; 

 the zero of the remaining one was constant between these dates but shifted 0-03° between 1949 and 1953. Of 

 Richter and Wiese's make and it is virtually certain that they were made, and originally calibrated by the P.T.R., 

 before 1925. This information cannot now be checked (see p. 233). One known check by P.T.R., two by N.P.L.; 

 thermometers nos. 10 11 and 1257 extensively used, especially the latter. 



ig-group (1934-53). Three thermometers — nos. CD 5556, 5559 and 5563. All from one batch of twelve made by 

 Negretti and Zambra. Maximum drift of zero, 0-05°, minimum 0-02°. One thermometer normal, with rise of zero 

 point over whole period; the remaining two, after a rise between 1934 and 1949, showed a fall of zero between 1949 

 and 1953. All checks by N.P.L. Although extensively used, the behaviour of the whole batch was generally erratic. 

 It is only fair to say, however, that after a lapse of many years Negretti and Zambra had not long begun again to 

 make reversing thermometers. 



iS-group (1935-53). Three thermometers — nos. 863, 883 and 1019. Zero points remarkably constant over period 

 but, as in the 20-year group, it is virtually certain that the calibration by the P.T.R. in 1935 was a re-calibration, 

 and that the thermometers were originally calibrated there in 1924 or 1925. All made by Richter and Wiese. One 

 check by P.T.R., two by N.P.L. 



id-group (1937-53). Two thermometers — nos. CG 11309 and CG 15243. Sole survivors of two separate batches 

 of six instruments each. Zero of no. 1 1309 shows a steady rise over the whole period; that of no. 15243 fluctuates. 

 Of Negretti and Zambra's manufacture. All calibrations by N.P.L. Both thermometers used extensively but 

 CG 1 1 309 was, at times, rather erratic. 



i^-group (1938-53). Fifteen thermometers — nos. CH 17798 etc., and CH 22676 etc. comprising eight and seven 

 thermometers respectively from two batches of twelve made in 1938 by Negretti and Zambra. First batch shows in 

 general a much more regular drift of zero than the second batch, of which all now under review had a nil correction 

 to zero at first calibration. No obvious reason for this difference is apparent ; both batches were made by the same, 

 and at that time the senior, operative on this work. All checks by N.P.L. Extensive use generally, with only a few 

 minor and unimportant failures. 



