Date of Award


Degree Name

Master of Science



First Advisor

Dr. Duane Hampton

Second Advisor

Dr. Mohamed Sultan

Third Advisor

Dr. Daniel Cassidy

Access Setting

Masters Thesis-Open Access


The values reported in literature for the height of capillary rise in fine sands, silts, and clays are contradictory. Most of the values were based on mathematical models, which used estimated rather than measured input data. This work measured capillary rise values in laboratory experiments in fine-grained sands and silts.

Two uniform sands with grains 0.35--0.7 mm in diameter and 0.3-0.6 mm in diameter, silt with an average grain size below 40 microns, and 0.35--0.7 mm sand coated with a water-repellant spray were carefully packed into two-inch diameter glass columns. These columns were placed into clear plastic tanks with water levels held constant. The average height of capillary rise was 13.5 cm for the 0.35--0.7 mm sand, 14.85 cm for the 0.3-0.6 mm sand, at least 310 cm for the silt (the capillary rise was limited by the height of the column, 310 cm), and -5.75 cm for the 0.35--0.7 mm water-repellant sand.

All of the tests above were repeated using kerosene instead of water; in two silt columns, the capillary rise of kerosene was 210 cm. This value for kerosene gives a scaled value for water capillary rise of 360 cm. The silt tests took years to reach equilibrium; some columns failed.

Two equations were analyzed to determine if either one may be useful in predicting capillary rise. When analyzed, constants for the surface tension were replaced with actual values, and grain sizes were actually measured. The Polubarinova-Kochina (1952) equation, hc = 0.45 ((1 – n) / n)/ d10, where n is the porosity, hc is the capillary rise and d10 is the effective grain diameter (hc and d10 in cm). A better-known equation is hc = (2σ cosλ) / (ρwgR), where hc is the height of the capillary rise and σ is the surface tension of the fluid. Fetter (1994) uses R = 0.2 d10. Both equations predicted capillary rise heights that were smaller than values measured in the laboratory; Fetter’s values were larger. Fetter’s equation matched the measured values better when his R = 0.2 d10 was replaced with 0.1 d10.