SECT. I] CHEMICAL INSTRUMENTATION 119 



appreciable extent by bottom sediment and suspended materials. Unlike 

 fluorescein, it is quite stable with respect to photodecomposition. 



The wavelengths of maximum excitation and fluorescence permit the design 

 of a measuring procedure that, for most practical purposes, is free from inter- 

 ference by naturally fluorescing and light-scattering substances. 



The Turner Fluorimeter (G. K. Turner Associates, East Palo Alto), when 

 fitted with the appropriate cut-off optical filters between exciting light source 

 and test solution and between test solution and photomultiplier detector, shows 

 a detection limit (1% full scale) of 0.002 parts of dye per billion parts of water. 

 In turbid inshore waters this is reduced by background fluorescence to 0.008 

 parts per billion. The fluorimeter can be obtained with a flow cuvette and re- 

 corder, making continuous in situ measurement possible by pumping the 

 sample through the instrument. 



Carpenter (1960, unpublished MS) made a comparison of rhodamine B and 

 radioactive tracers. Using the figures of 20,000 curies of gamma emitter, at a 

 cost of $2.50 per curie, and a detection limit of 2 x 10~ 9 curie per liter as sug- 

 gested by Folsom and Vine (1959) for a deep-water tagging experiment, 

 Carpenter estimated that 0.1 kg of dye is equivalent to one curie of gamma 

 activity or 2 metric tons of dye for an equivalent experiment, with a direct 

 material cost reduction of 250%. In addition the cost and hazards of handling 

 20,000 curies of gamma activity would be considerably greater than handling 

 two tons of pigment. 



It now appears that the use of rhodamine B with fluorimetric detection is 

 very well suited to a variety of tagging experiments in most natural waters. 



4. Measurement of Dissolved Gases 



The Winkler Method has been the universally used procedure for measuring 

 dissolved oxygen in sea-water. Its simplicity is both an asset and a liability. 

 The procedure can be carried out at sea by relatively unskilled technicians. 

 Nevertheless, the procedure involves several chemical reactions all of which 

 are liable to interference by side reactions, if operating procedures are not 

 strictly followed, and by reactions with a variety of dissolved substances which 

 in analyses of sea-water are tacitly assumed to be absent. Also, the usual 

 standardization procedure is not in terms of dissolved oxygen, but rather 

 utilizes a standard dichromate, iodate or biniodate solution, for which an 

 oxygen equivalent must be assumed. Furthermore, the common practice in 

 oceanography of computing "per cent saturation" or "oxygen deficit" implies, 

 first of all, that saturation values for given temperatures and chlorinities are 

 well known and, secondly, that the water samples found by analysis to be under- 

 saturated were at some previous time just saturated. 



It now appears that our knowledge of the accuracy of the results of Winkler 

 titrations, the accuracy of published saturation tables, and the validity of the 

 assumption of previous saturation under the conditions specified in saturation 

 tables may limit the usefulness of oxygen data as now measured. 



Carritt (1954) pointed out that, whereas oxygen saturation tables provide 



