370 Pub. Puget Sound Biol. Sta. Vol. 2, No. 54 



lor its liydrogen-ion concentration with Hynson, Westcott and Dunning's 

 set for the colorimetric determination of the hydrogen-ion concentration 

 of the blood.* During the summer of 1919 both the Hynson, Westcott 

 and Dunning's set, and the indicators and buffers used in 1918 were em- 

 ployed. In comparing samples with the standards, they were held in 

 diffused day-light in the box with a white background supplied for the 

 purpose, or against a white paper. Ten cc. of sea-water was placed in 8 

 small test tube, and after adding the indicator, was compared with ten cc. 

 of the standard buffer solution with the same amount of indicator added 

 and contained in a similar tube, or to the standard buffer set. When 

 possible both methods were used, the more reliance being placed in the 

 standard buffer set. The standard buffer solutions were always kept 

 stoppered to prevent contamination from the air and the carbon dioxide of 

 the observer's breath. All readings were corrected for the "salt error" of 

 Sorensen (1909) and Sorensen and Palitzsch (1910). ]McClendon (1917, 

 1918) was followed in making this correction. The chlorine of the sea- 

 water was not taken daily to determine the concentration but the same 

 correction was made (Shelford and Powers 1915) for all determinations 

 for both 1918 and 1919. The readings are given in the second decimal 

 but no great accuracy can be claimed for the second decimal place, since 

 the same factor was used in correcting for the "salt error" for all deter- 

 minations ; and the second decimal place was always an estimation, as the 

 standard buffers read only to the first decimal place. No corrections were 

 made for the variation in temperature (Sorensen and Palitzsch 1910, 

 Henderson and Colin 1916, and McClendon 1917, 1918) as the temperature 

 variations were slight in most cases and corrections for this factor were 

 considered M'ithin experimental error. In most cases where the tempera- 

 ture was high there was also a low hydrogen-ion concentration which was 

 generally outside the range of the standard buffer set, thus making the 

 second decimal less dependable; and without exception the correction for 

 the temperature factor would have been in the second decimal place. The 

 temperature at different depths was taken with a Nagretti-Zambra revers- 

 ing thermometer, and at the surface by a chemical thermometer reading to 

 one tenth of a degree. Samples from the surface and near the surface 

 were collected for oxygen determinations by connecting a small glass bottle 

 hj means of a rubber hose to a large bottle from which the air was ex- 

 hausted, thus eliminating any possible leakage of air into the sample. 

 Samples from greater depths were taken by means of a water bottle open 

 at both ends and automatically closed by sending down a messenger. Great 

 care was taken to avoid contact with air. Both the hydrogen-ion concen- 

 tration and the oxygen content were determined from the same sample. 



* The writer wishes to thank Professor C. M. Child for tlie use of his 

 hydrogen-ion colorimetric set in .standardizing- buffers used in the summer of 

 1918. ^ ">. ' ^W\ 



