r 341 1 



No. 44. 



NOTES ON THE FILTRATION AND OTHER ERRORS IN THE DETER- 

 MINATION OP THE HYDR0C4EN ION CONCENTRATION OF SOILS. 



By W. R. G. ATKINS, Sc.D. 



(Read May 27. Printed July 5, 1924.) 



Introduction. 



In many investigations in which a soil extract is examined it is necessary to 

 enquire into the most suitable proportions of soU. and water to be taken for 

 the determinations, and such an enquiry is necessary both for the eleetrometrie 

 and colorimetrie methods of measuring hydrogen ion concentration. Since 

 exact work by the latter method necessitates clear solutions, it is important to 

 study -the means available for obtaining these. Attention must also be paid 

 to the choice of the most suitable indicators for the range required. 



This paper embodies the experience obtained by the writer on these matters 

 during the last four years, and is brought forward now because it seems possible 

 that some of the errors consequent upon variations in technique have not been 

 fully appreciated. 



Proportion of soil to water. 



There is a general impression that for purposes of hydrogen ion determination 

 the precise proportions of soil and water taken is a matter of little import,ance. 

 Observations to this effect have been made by various workers, notably by 

 Sharp and Hoagland (1916) and by Arrhenins (1922, 2). This arises in t)art 

 from the low solubility of the constituents of the soil which liberate hydrogen 

 or hydroxyl ions, so that a relatively minute amount of soil may suffice to 

 produce a saturated solution. There is, furtheiinore, the apparent anomaly 

 that dilution may have but little effect upon the hydrogen ion concentration. 

 This is fully considered by Clark (1922), but it may be said that from the 

 writer's experience markedly acid soils of a sandy nature are more likely to 

 give dilution errors than are clay or humic soils which are more highly 

 buffered. Joseph (1923) has shown that for the alkaline soils of the Sudan 

 the proportions of the mixture cannot be neglected. He adopts the one to 

 five ratio advocated for general soil work by Hoagland, Martin, and Stewart 

 (1921). 



With these reservations the precise weighing out of the proportions is not 

 as a general rule necessary for work to an accuracy of p-ff 01, as may be 

 gathered from the figures given in the following tables. 



Table I shows a comparison of the pH values obtained by extracting soils 

 (a) for three to four hours, with two parts of water to one of soil by weight, 

 and (&) by pouring off the extract and adding distilled water to the residue 

 so as to make the mud-water mixture up to the original volume ; this was shaken 

 up at intervals during a further fifteen hours, making eighteen hours m all. All 

 the samples were centrifuged till clear and examined colorimetrically. 



SOIENT. PBGC, R.D.S., VOL. XVII, NO. 44. 3 U 



