A CLOSED ADJUSTABLE HEAD INFILTROMETER

G.C. Topp and W.D. Zebchuk

Land Resource Research Institute, Agriculture Canada, Ottawa, Ontario K1A 0C6

Contribution no. 84-36, received 18 July 1984, accepted 3 March 1985

Tbpp, G.C. and W.D. Zebchuk. 1985. A closed adjustable head infiltrometer. Can. Eng. 27: 99-104.

A closed-top infiltrometer was designed to measure infiltration rate, at negative head, intoa 0.25-m-diameter singlecylinder. After initial wetting of a soil at positive head it was possible to reduce the head to - 50 mm and continueinfiltration without causing air to flow into the water. The reduction of the applied head resulted in a steady rate ofinfiltration to be achieved veryrapidly. The useof a negative applied headduringinfiltration allowed the measurement ofinfiltration in structured soil without free water flowing in macrovoids such as cracks and biopores. The infiltration ratemeasured at an appliedheadof - 50 mmfor foursoiltypescompared favorably with the saturated hydraulic conductivitymeasured by an air-entry permeameter. The constant-head infiltrometer requiredabout three times as much waterbut wasmuch easier to use than the air-entry permeameter.

INTRODUCTION

Infiltration is a very important processas it is the factor which determines the

ratio of water entering soil to that which isrunoff from any particular rainfall event.As a result infiltration is a process that hasbeen extensively measured (Amerman1983; Constantz 1983). Hillel (1980) hasportrayed a correspondence between infil-trability after a considerable period ofinfiltration and saturated hydraulic conductivity (see his Fig. 2.1). Amerman(1983) pointed out that this infiltrabilityrelates to the hydraulic conductivity of arewetted or satiated soil which may be lessthan the saturated hydraulic conductivitybecause of air entrapment. A further complication to the measurement of infiltrability is the variation of the soilmacropores being measured. Dixon (1975)has shown that small changes in thehydraulic head applied during pondedinfiltration caused large changes in theinfiltrability of the soil. He attributed thesechanges in infiltrability to the exclusion ofwater flow from successively smallermacropores as the applied hydraulic headis reduced. In addition, Bouma andDekker (1978) and Bouma et al. (1978)have shown the importance of deep infiltration through large pores or "short circuiting" under both ponded and non-ponded boundary conditions.

Most of the methods for field measure

ment of saturated hydraulic conductivity inthe absence of a water table involve an

infiltration process (Bouwer and Jackson1974; Reynolds et al. 1983; Topp and Binns1976).Thus there is an implicit assumptionof correspondence between infiltrabilityand saturated hydraulic conductivity(Ksat). The measurement of infiltrabilityand the determination of Ksat from infiltration in well-structured soils depends verymuch on the surface or boundary through

- 250 mm -

Figure 1. The schematic cross section of the closed-top constant head infiltrometer. FR, floodingreservoir; CR, controlled-head reservoir (each having 4-L capacity); FV, floodingvalve; CV, valve on CR; SP, air-escape standpipe; ST, calibrated sight tube; C,cylinder; MR, Marriotte regulator; MA, Mariotte head adjustment tube; OR, O-ringseal; L, cast acrylic lid; WF, wetting front in the soil.

which water enters the soil, as well as themagnitude and degree of control of theapplied head of water.

Dixon (1975) has presented an effectivebut elaborate scheme for control of head

and measurement of infiltration with

closed-top infiltrometers. More recently,Constantz (1983) has shown a simplerclosed-top infiltrometer with head controlbut the range of applied heads was limited

to positive values. The infiltrometer usedin the present study is a simplified versionof that used by Dixon (1975). It is usefulfor measuring infiltration on smaller areaslike the equipment used by Constantz(1983).

The purpose of this paper is to describethe design and operation of a closed-topinfiltrometer on which there is easy controlof the applied head of water. The infil-

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985 99

TABLE I. PHYSICAL CHARACTERIZATION OF THE AP HORIZONS OF THE SOILS wetted soil at H = +50 mm and at

H = —50 mm. The manometers on thetensiometer were read at the times whensteady infiltration rates had been achieved.

Characteristics of the Soils and SitePreparation

The soils chosen for evaluation of the

CHI were five of the soil associations stud

ied earlier by Topp et al. (1980) —Uplands, Castor, North Gower, Dalhousieand Rideau which ranged in texture from amedium sand to a heavy clay. Table I givesa selection of some of the data obtained to

characterize the Ap horizon of these soils.Twenty sites were chosen for measurementon each of the soils. On 10 of the sites the

infiltration rates were measured by theCHI. On the 10remaining sites Ksat valueswere obtained by air-entry permeameter(AEP) as described by Topp and Binns(1976).

At the time of measurement the sites

were in grass or hay except for the Uplandssoil which was cropped to small grains. Ateach site the soil surface was cut free of

vegetation and the upper 20 - 30 mm ofsoil was carefully removed. The cylinder(used for either AEP or CHI measurements) was driven into the soil with ahand-operated sliding hammer and driverassembly. The top of the cylinder was leftabout 10 mm above the soil surface. The

soil just inside the perimeter of the cylinder was tamped with a 6-mm rod to closeany holes that may have been opened bydriving in the cylinder. Either the CHI orthe AEP assembly was clamped onto theupper flange of the cylinder for measurement. To minimize disturbance to the soil

surface by the incoming water, metal disks100 mm in diameter were placed directlyunder the openings from the CHI reservoirs. The AEP was fitted with an attached

water deflector.

Org. Sand ClaySoil Texturet Structured C (%) (%) (%)

Uplands ms Single grain 1.2 86 6Castor vfsl Weak, medium granular 4.2 23 18

North Gower 1 Moderate, fine subangular blocky 3.6 29 28

Dalhousie cl Moderate, fine subangular blocky to fine granular 9.5 16 33Rideau c Weak to moderate, fine subangular blocky 3.4 5.5 53

tms, medium sand; vfsl, very finesandy loam;1, loam;cl, clay loam;c, clay.tMcKeague and Wang (1982).

trability obtained by this constant-headinfiltrometer (CHI) operating at a negativehead is compared to Ksat measured by anair-entry permeameter (AEP).

MATERIALS AND METHODSFigure 1 shows the arrangement used

for infiltration intoa singlecylinderhavinga diameter of 250 mm and driven into thesoil to a depth of 150mm. The cylinderused was one already described by Toppand Binns (1976) as part of the air-entrypermeameter. The infiltrometer was constructed with two reservoirs FR and CR

each having a volume of 4 L and withvalves FV and CV opening through thecover plate. This assembly was clampedand sealed to the cylinder after it had beendriven into the soil. Also connectedthrough the cover plate is an air-escapestandpipe SP. A Marriotte-type air pressure regulator as shown on the right side inFig. 1 was used to control the pressurehead on the water.

Infiltration was initiated from the flood

ing reservoir FR by opening valve FV.Water entered under a positive head ofabout 200 mm. This caused rapid and uniform wetting of the soil surface and expulsion of the air via the air-escape standpipeSR When water was observed to rise in the

standpipe to above the level of the bottomof the control reservoir CR the flow was

switched to allow input from CR at controlled head (i.e. close valve FV and openCV). The Marriotte regulator was set sothat hw-ha = 4-50 mm by changing theheight of the Marriotte adjustment tubeMA. For best operation infiltration mustproceed under positive head until theupper 100 mm of the soil has been wetted.After about 5 min of infiltration at a

positive head the applied head wasadjusted to zero and the rate of fall of thewater level in the calibrated sight tube STon reservoir CR was recorded until a

steady rate had been achieved, usually inless than 15 min. Then the head was

adjusted to —50 mm; again the rate of fallof the level in ST was recorded until

steady, usually within 15 min. After initialexperimentation and troubleshooting ofthe procedure, we chose two hydraulic

100

head settings at which to record the infiltration rate. Thesewereat hw - ha = 0 andat hw—ha = -50 mm. The infiltrationrates were computed from the rates of fallof the water level in ST by multiplyingthese rates by the ratio of the area of reservoir CR to the area of soil surface.

A simple laboratory experiment wascarried out to demonstrate the nature of

pressure head profiles during infiltration atnegative applied head. The results fromthis experiment will be used to explain theoperation of a constant head infiltrometeroperating at a negative head. An acryliccylinder, 76 mm in diameter and 250 mmlong, was filled with 240 mm of fine sandysoil. Six very fast response tensiometersconnected to water-filled capillary manometers were inserted through the cylinderwall at positions 10, 20, 30, 40, 50, and60 mm below the soil surface. The upperend of the cylinder was fitted with a capthrough which water was introduced froma constant head burette.

Infiltration into this soil column was

initiated in a method directly analogous tothat described for the field procedure.Infiltration rates were recorded into the

dry soil when H = +50 mm andH = - 50 mm and before the wettingfront reached the bottom of the cylinder.Infiltration was stopped and the water wasallowed to redistribute overnight (18 h). Asecond infiltration was carried out on the

TABLE II. CONSTANT HEAD INFILTROMETER RESULTS (CHI) AT TWO APPLIED HEADS,H = 0 AND H = - 50 mm AND K^ MEASURED BY AIR-ENTRY PERMEAMETER (AEP)t

md"1

CHI CHI KsatSoil atH = 0 SDF atH = -50 SDF AEP SDF

Uplands 1.78a 1.69 — — 1.15ft 1.53

Castor 0.34\ab 1.61 0.266ft 1.80 0.562a 1.32

(revised) 0.405a 1.41 0.367a 1.44

North Gower 5.64b 1.97 2.64c 2.99 11.4a 1.41

(revised) 8.03a 1.53 6.94a 1.48

Dalhousie 13.8a 1.45 10.4a 1.79 15.3a 1.42

Rideau 31.9a 1.29 20.2b 1.32 15.4ft 1.39

tEach entry in columns 2,4 and 6 are geometric means of 10measurements (except "revised" entries) and theassociated standard deviation factors (SDF) are given in columns 3, 5 and 7. The "revised" entries resultedafter removing extraneous data as outlined by Dixon (1965).a-c Similar letters associated with each soil indicate means that are not significantly different at the 5% level.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985

UPLANDS rCHIat-0 xAEP

1 23456789 10

MEASUREMENT NO.

(a)

CASTOR • CHI at-50 mm x AEP1.0

0.8

•o 0.4

-t r -r"

X

—I—" 1 1 1 •

•

F" 1 1

X -

X

"X

•

•

X

X X X

X

X•

T

T !

I 11 A

(.

•

••

1 1 -

•1

- 1

-

L_ _._!_ i i J_

±

0.2

0.11 23456789 10

MEASUREMENT NO.

(b)

• CHI at-50 mm XAEP

123456789 10

MEASUREMENT NO.

(c)

show CHI at H = - 50 mm. On the right of each diagram geometric means andidentified by arrows are those found to belongto a different population (Dixon 1965).remaining data in Figs. 2 (b) and 2(c).

RESULTS

Table II presents geometric meanvalues and standard deviation factors

(SDF) of infiltration rates measured by theCHI at two applied hydraulic heads of H =0 and H = - 50 mm and the corresponding Ksat mean values and SDF measuredby the air-entry permeameter. The presentation of geometric means and standarddeviation factors was based on the assumption that infiltration rates and hydraulicconductivity are log-normally distributedfor a given soil. Only CHI measurementsat H = 0 mm were acquired initially forthe Uplands soil. The site was plowedbefore we returned to complete another setof CHI measurements.

On semi-logarithmic plots, Fig. 2 displays the CHI and AEP measured data. Tothe right of the measured data are the geometric mean values along with the rangerepresented by ± one standard deviationfactor. For four of the soils the data from

both CHI and AEP were within a total

range of less than one order of magnitude.For the North Gower soil the values from

both methods ranged from 0.55 to 26.6 m-day1 or a factor of 49 (Fig. 2c). Com

parisons of the ranges represented by ±one standard deviation factor about the

geometric mean show that in all casesexcept for the North Gower soil the rangesgiven by both methods are overlapping(Fig. 2). The geometric means show thatfor two soils, Uplands and Rideau the infiltration rates were higher than the measuredKsat. For the remaining three soils thereverse was true (Table II). Analysis ofvariance has shown that the infiltration

rates at H = 0 and Ksat were the same forthe Castor and Dalhousie soils. At H =

— 50 similar values were found for the

Dalhousie and Rideau soils. The visual

assessment of ranges and statistical analysis show that the CHI and the AEP givevalues for infiltration rate and Ksat whichare comparable.

A look at the variance as given by themagnitude of the SDF in Table II showedthat infiltration rates were in general morevariable than Ksat. If there is a correspondence between Ksat and infiltration rate weexpected approximately equal variancesfor the two methods. A closer examination

shows that for the North Gower loam and

the Castor very fine sandy loam the SDF

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985

DALHOUSIE • CHI at -50 mm xAEP

40

32

1 16

<C/>

X

x T

y|x if

l|-x ;1

x • I -

J- "

• » i i i i 1 1 1 1——

1 23456789 10

MEASUREMENT NO.

(d)

RIDEAU •CHI at-50 mm xAEP

1 23456789 10

MEASUREMENT NO.

(e)Figure 2. Ksat onlogarthmic scale as measured byCHI and AEP at 10 locations onthe five soils. Figure 2(a) shows CHI at H = 0 and the others

one standard deviation are presented. The dataThe filled triangles are geometric means of the

were considerably greater for the CHI thanfor the AEP. These two soils were struc

tured but the structure was not stronglywater resistant. The initial flooding of thesoil to initiate measurement by both theCHI and the AEP takes place at positiveheads. In the case of the CHI, discs wereplaced on the soil surface to deflect theenergy of the incoming water. In the AEPthe deflector was mounted on the lid. The

splash of water off the deflector discs in theCHI may have disrupted the soil structurecausing some sealing of the soil surface.Any sealing of the soil surface during measurement would cause both lowered infil

tration rate and greater variance. Recognizing this problem we have now fitted theCHI with a water deflector similar to that of

the AEP.

By conducting both methods at eachsite it was possible to compare their ease ofoperation. Requiring only one person, theCHI is an easier method than the AEP

which requires two persons. The CHImethod takes more total time for each

measurement (>20 min) because of theneed to achieve a steady flow. The waterrequirements of the CHI are about three

101

times that of the AEP.

DISCUSSION

Visual examination of the data showsthat some of the data from the CHI measurements (Fig. 2b, c, d) are not onlywidely spaced as discussed above but alsoappear to be in two sets, especially for theNorth Gower soil. Dixon (1965) has outlined a procedure for designating datavalues which are extraneous or lie outsidea statistically allowed range. In applyingDixon's (1965) procedure we calculatedthe ratio of the range of magnitudes of thedata set to the standard deviation. Thiscalculated ratio was compared to the limiting values given by Dixon (1965). In ourcase an estimate of standard deviation was

obtained from the AEP measurements. Intwo of the soils extraneous values were

found and those data points are identifiedin Fig. 2b, c by small arrows. In all casesthe excluded values were low relative to

the remainder of the data. The sealing ofthe soil surface during measurement, asdiscussed, above, might have been the factor which caused some of the CHI data to

be extraneous.

When analysis of variance was performed on the data after elimination of the

extraneous data there was no significantdifference between infiltration rates at

H = - 50 and Ksat for the four soils onwhich the CHI was used at H = - 50 (seethe "revised" entries in Table II and the

filled triangle means in Fig. 2b, c). Thusthere appears to be equivalence betweenKsat by the AEP and the infiltration measured by the CHI at H = -50 mm. Itremains, however, to be determined if theCHI data which were judged extraneouswere a result of sealing of the soil or fromsome other unidentified artifacts. Further

research will be necessary to answer thisquestion.

We had anticipated greater difference instructured soil between Ksat by the AEPand the infiltration rate by the CHI. TheAEP measurements were made with an

applied positive head of >800 mm. Thepresence of cracks and biopores wouldhave offered paths for the downward flowof free water leading to an apparentincreased Ksat value as a result of "shortcircuiting" as defined and described byBouma et al. (1978). The CHI measurements at H = —50 mm, however,excluded the possibility of the flow of freewater in cracks and biopores. In addition,the reduction of the applied head to - 50mm would tend to reduce the hydraulicgradient to less than unity. Therefore theinfiltration rates at H = - 50 mm would

tend to be less than the actual hydraulic

102

-150 -100

H,h (mm H20)-50 0 50

(a)

---30

Z (mm)

Z(mm)

Figure 3. Selected hydraulic heads, H (solid lines) and pressure head, h (dashed lines) versusdepth profiles. Profiles are drawn for applied hydraulic heads of -I-50 mm and- 50 mm for wetting front depths of A - 20 mm, B -110 mm and C -180 mm. Theequation relating H, h and z is H = h + z.

conductivity of the saturated soil. Thismeans that the hydraulic conductivity ofthe saturated soil lies between the Ksat byAEP and the infiltration rate at H =

—50 mm. Above we have shown that the

data from the two methods were comparable indicating that both methods gave agood measure of Ksat.

Others have found that single-ring infil-trometers have given steady infiltration

rates which are greater than Ksatas a resultof lateral flow beyond the bottom of theinfiltrometer ring. Using tensiometers, a300-mm-diameter inner ring and varyingdimension of outer buffer rings, Ahuja etal. (1976) have analyzed the percentage ofinfiltration which was attributed to lateral

flow. For conditions similar to those used

in our experiment they found that the lateral flow was "as much as 20% of the

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985

vertical flow in some soil horizons in earlystages of infiltration". It was possible thatour CHI values, especially at H = 0, weregreater than Ksat by 20% or by a factor of1.2 of Ksat. It was not possible to separatethe effect of lateral flow on our experiments from the other factors discussed

above.

Dixon (1975) has discussed the measurement of infiltration with a single-ringclosed-top infiltrometer at applied negative hydraulic heads. In his case, air wasbubbling up through the water on the soilsurface because the negative head was setbefore a plastic seal was removed to initiate flow. He noted that for the one soil

when infiltration was measured at H =

- 50 mm it took 15 min before bubblingceased. During Dixon's infiltration the soilwas always in a condition of increasingpressure head. We were able to take advantage of the hysteretic nature of soil waterbehavior by first beginning the infiltrationusing a positive applied head and thenadjusting the head to a negative value afterseveral minutes. This gave a number ofadvantages as follows: (1) Infiltration wastaking place in a soil where the pressurehead of the wetted soil near the surface was

changing in the direction of decreasingpressure head; this prevented air entrythrough the wetted soil and prevented bubbling up or backflow of the air; in our casethe bubbling of air would have appeared asan error in infiltration. (2) It was mucheasier to start the infiltration measurement

where positive heads were applied than ifusing negative heads throughout. (3) Ongoing from the initial infiltration atpositive head to infiltration at negativehead we usually found that steady rateinfiltration occurred within a few minutes

of lowering the applied head; this resultedin much shorter time intervals between

measurements than was reported by Dixon(1975).

Mode of Operation of the InfiltrometerHow this infiltrometer (CHI) operated

at negative hydraulic head without airintrusion or backflow into the bulk water

above the soil requires explanation. For theexplanation we make use of the hypothetical wetting profile diagrams given inFig. 3. The similar profile diagram (Fig. 4)obtained during infiltration of the laboratory soil column validates the hypotheticalcurves in Fig. 3. The hydraulic head H(solid line) and the pressure head h(dashed line) versus depth profiles areshown for three possible positions of thewetting front during an infiltration process. In drawing these profiles we haveadopted the equation H = h + z and have

-150 -100 -50

h (mm H20)

0 50

I I I I I I I I I I I I I i \ Axi—i—r 1 1 1 1 Lx*

X^^

*

•sS*

^1?

'/ >x-x"

--50

•-^ X—"

Z(mm)

Figure 4. Pressure head, h, versus depth profiles obtained by tensiometers during infiltration intoa column of fine sandy soil where applied heads were - 50 mm (dots), - 42 mm (opentriangles) and 4-50 mm (crosses).

assumed that the soil within the cylinderhas uniform soil water properties.

In the case of shallow depth of wettedsoil as in Fig. 3a the hydraulic gradient(the inverse of the slope) is large over thewhole wetted region. The gravity component of the hydraulic gradient is smallcompared to the pressure head gradient atthis early stage of infiltration, i.e. pressurehead profiles and hydraulic head profilesare of similar shape. Thus a change ofboundary condition from H = + 50 mmto H = - 50 mm has little effect on the

overall gradient or head profile. If,however, the soil has large pores behindthe wetting front with an air-entry valuebetween 0 and - 50 mm H20 air will bepulled upward through the soil surface assoon as the applied head is reduced to H =— 50 mm.

For the condition portrayed in Fig. 3bthere is a 50-mm-deep region where thehydraulic gradient is unity (slope of 1) forthe profile starting at H = -1-50 mm.When the applied hydraulic head isdecreased to H = - 50 mm, however, thegradient decreased to some extent. For thiscondition the gradient decreased from 1 toan average of 0.6 over the upper 50 mm ofsoil. When H = - 50 mm at the surface, his increasing with depth with a maximum hat Z = -50 mm. In Fig. 3c, a 150-mm-deep region is portrayed as having flowoccurring under a hydraulic gradient ofunity at the right where the applied head isH = + 50 mm. When the applied head isreduced to H = -50 mm there is a

decrease in the hydraulic gradient. Thereduction in hydraulic gradient as depictedon the left in Fig. 3c is a reduction from 1to 0.6. The resulting profiles of h vs. Zshow that under the conditions on the rightin Fig. 3c h ^ 0 to a depth of Z =— 170 mm. Thus the soil is saturated to

this depth. After the applied head isreduced to H = - 50 mm the resulting hprofile shows an increasing h with depthuntil Z = -120 mm. When the applied

head is reduced to H = - 50 mm it is the

increasing h with depth as shown in Fig.3b, c which prevents air backflow into thebulk water above the soil even though it isat a negative pressure.

The laboratory experiment measuringinfiltration into a column of fine sand and

the measurement of pressure heads hasdemonstrated that h profiles do increasewith depth as hypothesized above (Fig. 4).The profile on the right was obtained during infiltration into air dry soil when theapplied head was H = 4-50 mm. The hprofile (dashed line) on the far left was thatobtained when the applied head was H =—50 mm during infiltration into air-drysoil. The middle h profile is that whichresulted when the applied head wasreduced to H = —42 mm during infiltration into the wetted soil. In both cases

where H was reduced to -50 mm or

—42 mm the resulting h profile shows anincrease with depth.

We suggest that the region showing anincrease in h with depth remains saturatedeven though a negative head is applied atthe soil surface. The direction of head

change prior to the condition shown on theleft of Fig. 3 is in a decreasing direction.The desorption curves for soil drainingfrom a saturated condition usually show nochange in water content for smalldecreases in h just below zero as a consequence of the hysteresis in the pressurehead vs. water content relationship. At Z= —110 mm in Fig. 3c where h = -12all pores having equivalent diameter<3 mm remain filled with water. Thus this

region, not at the surface but within thesoil, acts to prevent the passage of airupward. The soil above Z = —110 mmremains saturated even though a negativehead of —50 mm is applied at the surface.

The use of a negative head applied tothe surface will not allow water to run

freely down into large voids such as wormholes and cracks which exist at or below

the wetting front, for example below Z =

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985 103

-170 mm. The negative head preventsflow ahead of the wetting front analogousto the "short circuiting" or bypass flowwhich was described and measured byBouma and Decker (1978) and Bouma etal. (1978). The infiltration measured underconditions of negative head at the soil surface as with the CHI is equivalent to Ksat.

CONCLUSIONS

This closed-top single-ring infiltrometer was easy for one person tooperate. Through comparison with Ksatvalues measured by AEP this CHI operating atH = - 50 mm gave good correspondence between CHI values and Ksat. Bystarting infiltration at positive heads it waspossible to take advantage of soil-waterhysteresis to prevent air backflow wheninfiltration was taking place at negativeheads and to achieve steady flow veryrapidly. Further investigation will berequired to explain the apparent large variation in CHI values on two of the soils.

REFERENCES

AHUJA, L.R., S.A. EL-SWAIFY, and

104

A. RAHMAN. 1976. Measuring hydrologicproperties of soil with a double-ring infiltrometer and multiple-depth tensiometers.Soil Sci. Soc. Am. J. 40: 494-499.

AMERMAN, CR. 1983. Infiltration measurement. Pages 201-214 in Advances in infiltration. Am. Soc. Agric. Eng., St. Joseph,Mich.ASAEPubl.no. 11-83.

BOUMA, J. and L.W. DEKKER. 1978. A casestudy on infiltration into dry clay soil. I.Morphological observations. Geoderma 20:27-40.

BOUMA, J., L.W. DEKKER, and J.H.M.WOSTEN. 1978. A case study on infiltration into dry clay soil. II. Physical measurements. Geoderma 20: 41-51.

BOUWER, H. and R.D. JACKSON. 1974.Determining soil properties. Pages611-672in J. van Schilfegaarde, ed. Drainage foragriculture. Agronomy 17. Am. Soc.Agron., Madison, Wis.

CONSTANTZ, J. 1983. Adequacy of a compactdouble-cap infiltrometer compared to theASTM double-ring infiltrometer. Pages226-230, in Advances in infiltration. Am.Soc. Agric. Eng., St. Joseph, Mich. ASAEPubl. no. 11-83.

DIXON, R.M. 1975. Design and use of closed-top infiltrometers. Soil Sci. Soc. Am. Proc.

39: 755-763.

DIXON, W.J. 1965. Extraneous values. Pages43-49 in C.A. Black, ed. Methods of soilanalysis. Agronomy 9. Am. Soc. Agron.,Madison, Wis.

HILLEL, D. 1980. Applications of soil physics. Academic Press, New York. 385 pp.

McKEAGUE, J.A. and C. WANG. 1982. Soilstructure: concepts, description and interpretation. Land Resource ResearchInstitute, Agriculture Canada, Ottawa,Ontario. Publ. 82-15, 42 pp.

REYNOLDS, W.D., D.E. ELRICK, and G.C.TOPP. 1983. A re-examination of the con

stant head well permeameter method formeasuring saturated hydraulic conductivityabove the water table. Soil Sci. 136:

250-268.

TOPP, G.C. and M.R. BINNS. 1976. Fieldmeasurement ofhydraulic conductivity witha modified air-entry permeameter. Can. J.Soil Sci. 56: 139-147.

TOPP, G.C, W.D. ZEBCHUK, and J. DUM-ANSKI. 1980. The variation of in situ mea

sured soil water properties within soil mapunits. Can. J. Soil Sci. 60: 497-509.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 27, NO. 2, FALL 1985