Soil and Environmental Analysis: Physical Methods - Chapter 12 pps

As a consequence of these effects, the length of individualroots, their rate of extension, and increases in root diameter can be changed bysoil physical conditions, thus affecting the ov

Root Growth: Methods of Measurement

David Atkinson

Scottish Agricultural College, Edinburgh, Scotland

Lorna Anne Dawson

Macaulay Land Use Research Institute, Aberdeen, Scotland

The morphology of a plant root system is a function of its genetics and the ronment in which it grows (Smucker, 1993; Aiken and Smucker, 1996) Mor-phology is also affected by interaction with soil microorganisms, e.g., arbuscularmycorrhizal fungi (Hooker et al., 1992) Both individual plant roots and wholeroot systems can and do show substantial variation within the potential range oftheir characteristics Soil physical factors, particularly temperature, aeration, wa-ter potential, and mechanical impedance, are frequently the cause of limits to theexpression of genetic potential The morphology of the root system can thus beregarded as representing the integrated effects of three factors This chapter firstreviews those root properties that are most likely to be influenced by soil physicalconditions and then describes methods that can be used in the field or laboratory

envi-to describe particular attributes of root systems It illustrates some of the uses envi-towhich particular methods have been put and some of the limitations of their use.Bohm (1979) has given a more complete description of methods for measuringroots, and Atkinson (1981) has reviewed those methods relevant to tree crops Theimpact of soil biological and chemical factors, and of the growth of the aerial parts

of the plant, on root growth have been reviewed in general terms by Russell(1977)

In both field and laboratory, many of the methods give information on arange of parameters For example, when a root system is observed directly (e.g.,through an observation panel), measurements can be made of length, diameter,

longevity, and branching It therefore seems more logical to divide studies on rootsystems by type of method rather than by root system property Consequently, inthis chapter, the major groups of available methods and the significance of themeasurements they facilitate are discussed together, but in the context of the need

to determine how plant function can be influenced by soil physical conditions

A Root System Properties

Root systems are branched structures composed of a number of individual rootswith normally up to four orders of branching Individual roots are themselvesmade up of large numbers of cells The processes relating to root developmenthave been characterized at both cellular (e.g., Scheres et al., 1996) and molecularlevels (e.g., Chriqui et al., 1996) The size, shape, and form of these cells, thenumbers in a particular tissue (e.g., xylem or cortex), and their function (Smucker1993) may be altered by the growing environment Major soil physical factors,such as soil water potential and soil mechanical resistance, can affect root prop-erties such as cell wall extensibility and wall pressure in a number of ways Cellwall pressure is closely related to the rate of root growth, while osmoregulation isclosely related to changes in soil water potential but less completely related tomechanical resistance As a consequence of these effects, the length of individualroots, their rate of extension, and increases in root diameter can be changed bysoil physical conditions, thus affecting the overall volume of soil exploited byroots, via effects on horizontal spread and the depth of penetration, which in turninfluence the resources of water and nutrients available to the plant Other pa-rameters that can vary include the angle at which roots grow through soil (e.g.,their susceptibility to geotropism)

The longevity of roots varies between species (Atkinson, 1985) and betweenroot types in a species (Hooker et al., 1995) The rate of production, and the lon-gevity, determine the total root length and average root length density, i.e., thelength of root (LA) under an area of soil surface or the length (LV) in a volume ofsoil These factors are important to the ability of the root system to obtain nutri-ents for plant growth In addition to possible effects of soil factors on morphology,root function (e.g., nutrient uptake per unit root length, surface area, or volume)may be altered as a consequence of effects on the types and ages of root present.However, the exact effects of physical conditions on the above parameters areincompletely described or understood, and considerable plasticity clearly exists inrespect of most properties (e.g., Reynolds and D’Antonio, 1996) Roots are nor-mally considered in relation to their ability to supply water and nutrients to theplant, but they are also required to anchor the plant (Coutts, 1983) and to producehormones, which may regulate the growth and performance of both root and shoot(Aiken and Smucker, 1996) The root system of most plants exists in nature in

a symbiotic association with fungi (mycorrhizas), and so assessments of effects of

soil conditions on roots should also consider effects on mycorrhizas Effects ofphysical factors on root characteristics are summarized in Table 1.

B Potential Effects of Physical Properties

Many root system parameters can be influenced by a change in soil physical ditions (Table 1) Root effects depend upon many factors, including the nature ofthe changed soil variable, the species under investigation, and conditions in otherparts of the soil General principles were reviewed by Greacen and Oh (1972)

In a study of the effect of zones of contrasting bulk density on root system opment in oats, the effect of a given value of bulk density varied according to itsrelation to the density of other areas in the soil column (Schuurman, 1965) Com-paction did not reduce branching, although it did influence root survival Thelength of branch roots, which was normally highest in the surface, was affected(Fig 1) Where the elongation of the main axis was reduced by a dense subsoil,its diameter increased and branching was stimulated

In addition to changes in overall root system length, mass, or volume, there can

be alterations in the partitioning of dry matter within the root system (e.g., byincreasing root branching or root number: Goss, 1977) In soil, root elongationwas reduced by 60% by a mechanical resistance of 1– 8 MPa in ryegrass and by

2 – 6 MPa in pea (Gooderham, 1977) Goss (1977), using a ballotini (glass sphere)system, showed that the effect of increasing pressure on root growth inhibitionvaried between species, with barley being the most sensitive of the species tested

Table 1 Root System Characteristics That Can Be Affected by Soil Physical Conditions

Anatomy Cell size, cortex width, balance of xylem cell types, epidermal wall

form, root shapeIndividual root

properties

Horizontal distribution, vertical distribution, length, mass, absoluteand relative distribution

Function Absorption of nutrients and water, anchorage, production of

bio-logically active molecules (e.g., enzymes, phenolics)

When roots are prevented from elongation, a resultant increase in the diameter isfound, mainly due to an increase in the cross-sectional area of the cortex (Wier-sum, 1957) Goss showed that even when root system mass is unchanged, lengthcan be reduced by 65% by mechanical impedance, while Logsdon et al (1987)demonstrated that an increase in root diameter can compensate, in part, for a re-duction in total length Appropriate measurements are clearly needed to establishsuch effects, although their physiological significance is still poorly understood.The use of penetrometers in such studies has been discussed by Bengough et al.(1994) An increasing concentration of roots at the ‘‘soil surface’’ in the laboratoryhad no effect on nutrient supply (Goss, 1977), but in the field, in the absence ofirrigation, it might be expected to have adverse consequences because the surfacewould quickly dry out This exemplifies why results cannot be directly transferredfrom laboratory to field There can also be changes in the internal root turgorpressure, which has been primarily studied in relation to water potential effects.

A novel approach, using a force transducer, allows the turgor in an impeded root

to be measured without the need to remove the root from the impeding ment (Clark et al., 1996)

environ-C Purpose of Measurements

The optimum method for the assessment of any root system will depend both onthe characteristics of the root system itself and on the reason for making the mea-surement The following are among the commonest reasons for measuring roots:

1 To assess the significance of changes in soil physical conditions on theplant

Fig 1 Effect of density of the topsoil (0 –25 cm) and subsoil (⬎25 cm depth) on opment of oat root systems Bulk densities above and below the boundary are given inmegagrams per cubic meter (From Schuurman, 1965.)

devel-2 To help to interpret a plant response to a particular soil treatment byunderstanding effects on water or nutrient supply

3 To improve the use of inputs (e.g., irrigation water or fertilizers) or tooptimize the effects of tillage and other soil management practices

4 To allow the development of better plant root systems by conventionalbreeding or genetic enhancement

D Significance of Root Features

Criteria for the selection of methods of assessing root performance are limited byour current understanding of the significance of particular root characteristics and

of the consequences of changes to them Information exists on the importance ofsome characteristics, and so we can identify circumstances in which particularmeasurements will be useful

1 Characteristics Influencing Water Supply

Soil water flux is the product of the hydraulic potential gradient between the soiland the root and the unsaturated hydraulic conductivity Typical maximum fluxrates seem to be around 2.5 mm d⫺1 (Russell, 1977) This should allow a rate

of root water uptake of around 160mL d⫺1cm⫺1 On this basis a root density

of 2 cm of root per cm2of soil surface area should be able to supply tional needs in Northern Europe The uniformity of root distribution and themean distance between roots both influence water supply Thus if an average rootlength density is clustered in one soil region, the flux to the root surface is likely

transpira-to be inadequate As soil water content and consequently unsaturated conductivitydecrease, a greater root density will be needed to supply the same total flux tothe shoot

The total volume of soil exploited by the root system directly affects thetotal amount of water and nutrients available to that plant This volume can berepresented by the horizontal spread times the maximum rooting depth (or thedepth containing 95% of roots), although the effective volume will be increased

by capillary rise Average root length density and root distribution within the soilvolume exploited are both important This, however, gives only a static picture Inthe field, the transpiration rate will change during the season, as a consequence ofweather and plant needs It is thus important to be able to assess the root system

as it grows, and to characterize factors such as root death; root longevity in somespecies is of the order of days, not weeks (Atkinson, 1985)

E Commonly Measured Characteristics

Despite the long list of root parameters that might be measured (Table 1), searchers have commonly determined a smaller set They are summarized in

re-Table 2

Root parameter Functional significance Method

Length Total system size

Potential for the absorption of nutrients

or waterSoil microbial activity(Temporal variation: frequent estima-tion needed)

Monolith methods (soil coring,etc.)

Rhizotron/minirhizotron ods (with assumptions)Profile wall methods (withassumptions and wherelimited precision is needed)Mass Total root system size Monolith methods (especially

meth-soil coring)Excavation (woody perennials)Number Growth regulator production Counts on soil cores

Profile wallsRhizotron/minirhizotronRoot : shoot ratio Relative allocation strategy Calculation from root and shoot

weightsLength density Limitations to soil nutrient and water

exploitation

Calculation from root lengthand soil volume

Specific length Within-root-system allocation strategy

Relative importance of soil exploitation

Calculation from root lengthand weight

Diameter Potential for mycorrhizal development

Regulation of water stressPotential for apoplast /symplastexchange

Growth potentialResponse to soil physical conditions

Direct measurement or tion from root length andvolume

calcula-Amount of secondary

thickened root

Investment in system infrastructure Measurement of weight or

lengthBranching pattern Intensity of soil exploitation Measurement of lengths or num-

ber of roots of different ordersLongevity Potential for rapid adjustment to root

lengthPlasticityFlux of carbon to rhizosphere

Cohort analysis of rhizotron orminirhizotron images

Production Overall potential for soil exploitation

Ability to increase system length

Sequential soil core estimates

or rhizotron/minirhizotronmeasurements

Vertical distribution Physical stability

Depth of soil exploitedPotential for resource use

Soil core samplingProfile walls

Horizontal

distribution

StabilityInteraction with other species

Profile wallsExcavation

II METHODS OF STUDYING ROOT SYSTEMS

IN THE LABORATORY

Laboratory studies are normally concerned with assessing the effect of either asingle or a limited range of soil physical properties on plant and root performance.Single-factor studies usually relate to soil temperature, water potential, osmoticpotential, or aeration, whereas studies of the effect of soil impedance often involvesimultaneous changes in other factors For example, when bulk density waschanged from 1.24 to 1.52 Mg m⫺3(Fig 1), there was also a reduction in the vol-ume of pores that were filled with air at field capacity and an increase in thevolume of pores from which water would be unavailable to plants (Schuurman,1965) Because the results of such experiments are likely to be influenced by thetypes of containers and media used, these are also briefly reviewed

A Containers

1 General Factors

Container methods permit the isolation of individual environmental factors thatwill normally interact with other characteristics and influence root growth in thefield Replication and ‘‘management’’ are easier than in the field, although con-tainer effects on root growth may be unnatural because of the restricted space andabsence of soil organisms (bacteria, fungi, soil arthropods, etc.) Container meth-ods are best suited for studying plants with small root systems or for investigatingthe early stages of plant development

2 Container Types

The size of a container determines the total volume of soil available to a plant.

Conventional plastic or clay plant pots, Mitscherlich pots, glass pots, petri dishes,tubes made from glass or plastic, and cardboard cartons have all been successfullyused as containers, but their limited volume frequently results in roots concentrat-ing near the walls of the vessel and around its bottom As a consequence of mois-ture and temperature differentials, the concentration of roots between the wall of

a pot and the soil tends to be greater in a porous clay pot than in a plastic pot

In container experiments designed principally to study root growth and tribution, the depth of the container needs to be large because of the root’s ten-dency to grow downward when restricted by the walls of a container (Bohm,1979) Boxes 80 cm high, made from metal, wood, or plastic, have been used inthis type of study, but cylindrical tubes are more common Iron, clay, asbestos,plastic, acrylic, and glass have all been used Where tubes are of a transparentmaterial, or where boxes have glass windows set into their sides, roots can bedirectly observed and measured (Fig 2) Containers of all dimensions can be

dis-Fig 2 Use of acrylic cylinders to observe differences in root form, density, and bution in spring barley To calculate information, such as total or average root lengths, therelationship between length at the observation surface and in the whole soil volume must

distri-be known (From Atkinson, 1987.)

modified and constructed to have a viewing window or sides to permit repeatedobservations to be made of root growth (Mackie-Dawson et al., 1995a) Tubesmay be buried in the soil or in insulated boxes to prevent the establishment ofunrealistic temperature differentials (Mackie-Dawson et al., 1995a) and video re-cording equipment can be used to record root growth.

Root boxes vary in size in relation to the type of plant being investigated.For studies of M.1 apple rootstocks, boxes 60⫻ 17.5 ⫻ 42.5 cm high were used(Rogers, 1939a), while for studies of maize, Walker and Barber (1961) usedsmaller boxes Such boxes with windows may be used simply as a means of ob-serving the response of roots receiving particular treatments or as a means of as-sessing the uptake of radioisotopes incorporated into the soil adjacent to the soil-observation interface The latter method has been used to observe the uptake of

86Rb from the soil around individual roots (Walker and Barber, 1961)

Such boxes have been used to assess the effect of soil moisture and soiltemperature on the root growth of grass and clover species (Garwood, 1968) Us-ing the observation windows, it was possible to assess treatment effects on rootsystem length, the elongation of individual roots, root diameter, and root number.All these parameters were affected by soil temperature

Where plants are grown in containers that allow observation of the rootsystem, the possible effects of light on root growth must be considered, althoughfew studies have addressed this issue A comparison of the effect of a range ofdifferent light exposures, varying from total darkness to total light, on the growth

of apple rootstocks showed that while continuous illumination severely checkedgrowth, increased suberization, and reduced the development of lateral roots, theshort periods of 3⫻ 20 minutes per week or 2 hours per 2 weeks needed forobservation had little obvious effect Effects were greatest during periods of maxi-mum root growth, with length being more affected than root number (Rogers,1939a) Given the paucity of data on the effects of light, it is prudent to reduceunnecessary exposure to a minimum

3 Filling Containers

Care must be taken when filling containers (both tubes and boxes), particularlywhen physical conditions (e.g., bulk density) are being controlled (Schuurman,1965) Soil should be sieved when nearly air-dry, the appropriate fertilizer added,and the soil then moistened to a friable condition and mixed It should then be put

in the containers and compressed, layer by layer, several cm at a time, with the topfew mm of each layer loosened before adding the next layer, to avoid stratifica-tion All containers should normally be watered and allowed to stand for a periodfrom a week to a month before the experiment begins, to allow the soil to settle.Where short-lived radioisotopes are being used, however, this may not be possible(Walker and Barber, 1961) These procedures are needed both for tubes and boxes

Where an observation surface is to be viewed, it is essential that care be taken toprevent smearing and the formation of voids.

to be taken during transportation so that artificial voids are not created

Sand. Sand is often used in nutrient experiments because of its low bufferingcapacity and the ease with which it can be manipulated It is also relatively easy

to wash from the roots, although some may still firmly adhere (Atkinson, 1987).The physical properties of the medium can be altered by varying the particle size.Fine sand gives a higher water-holding capacity, while coarse sand is more freelydraining and better aerated Root systems obtained from experimental media ofthis type are generally similar to those grown in solution culture (Bohm, 1979)

Other Solid Media. Perlite, which is composed of expanded volcanic rock ments, is uniform and inert and so is suitable for studies of germination or seedling

frag-development However, it is less suitable for long-term growth experiments andfor studies of physical effects on roots Perlite seems to result in root system de-velopment similar to that in solution culture (Bohm, 1979) It is a good mediumwhere roots are to be used for studies of ultrastructure Particles of perlite embed-ded in roots cause less damage to knives used in sectioning than do sand grains.However, roots can penetrate perlite, from which they are difficult to extract Thediffering penetrating abilities of roots have been studied using agar, paraffin, andwax materials of differing hardness (Taylor and Gardner, 1960a, b; Yu et al.,1995) A wax mixture layer has been used to assess the root penetration abilityamong rice roots (Yu et al., 1995) Vermiculite has also been used as a growingmedium and seems to give growth comparable to that of roots grown in soil(Bohm, 1979).

2 Solution Culture

The major advantages of solution culture are that the ionic composition of the rootenvironment can be defined, measured, and manipulated with precision and thatthe entire medium can be held under standard conditions (e.g., of temperature andaeration) (Atkinson, 1986) It can either be applied as a solution or as a mist ap-plication Because of ion uptake/efflux by plant roots, however, nutrient solutionsare liable to rapid changes in composition, and so require more routine mainte-nance than is needed for soil-based systems

Solution culture is only of limited use in soil physical studies The uniformmedium, lack of physical resistance, and the absence of soil flora and fauna make

it difficult to compare root growth in solution culture with that in soil However,this approach has applications in studies of impeded aeration, water stress, andtemperature Roots can be maintained at precise temperatures by flowing the so-lutions through a refrigeration or heating unit before entry into the plant growthcontainers (Bhat, 1980), or by immersing plant-growth containers in thermostati-cally controlled water baths Mist chambers have been used in the study of wateruptake, using an applied dye, sulphorhodamine (Varney and Canny, 1993)

3 Special Techniques

Split-Root Techniques. Spatial variability in nutrient supply can be controlledusing split-root containers, in which isolated parts of the root system receive dif-ferent nutrient supplies, either in solution culture or in solid media Individualroots can also be separated out to study specific effects

Water Stress Control. Osmotic control of plant water stress can be obtained insolution culture using sodium chloride, polyethylene glycol 4000 (PEG), or arange of other chemicals This method allows plant water stress to be accuratelymaintained and more easily reproduced than is possible in soil However, the

stresses brought about by the two methods have different physiological bases Thewater stress to which roots are normally exposed in soils is primarily due to asubstantial negative matric potential, while in a PEG-modified solution the stressresults from a substantial negative osmotic potential Although stresses due toboth matric and osmotic potential have been shown to produce similar effects onplant growth, it must be remembered that the matric potential at the root surfacecan be considerably less than that measured in bulk soil.

Pressure Control. Many workers have studied effects of mechanical stress bygrowing roots in pressure cells through which aerated nutrient solution is circu-lated The cell walls consist of flexible impervious polyester membranes A knownhydrostatic pressure is applied by suspending the cells in water-filled vessels, towhich an external pressure is applied (Barley, 1962; Abdalla et al., 1969) Details

of construction, use, and the types of measurements that can be made are given byGoss (1977) A technique that combines a pressurized wall with time-lapse videoanalysis has been used to study pea lateral root emergence (Gordon et al., 1992).Pressure has also been applied to roots grown behind thin rubber diaphragmsforced against the root by gas under controlled pressure (Gill and Miller, 1956).However, the actual pressure experienced by the root is uncertain

C Measurement of Roots in Laboratory Media

1 Measurement in Soil

Impregnated Sections. Roots can be studied in undisturbed samples by nating the soil with wax or resin, using samples collected from pot experiments

impreg-or field plots The method involves removal of soil water in exchange fimpreg-or a solvent

in which the concentration of resin is gradually raised After addition of an erator or hardener, the cured soil blocks are sectioned for examination (Alte-muller, 1986) Several combinations of fixative and impregnating resin have beenused, including a mixture of acetic acid–formaldehyde and ethanol and a polyes-ter resin for impregnation (Lund and Beals, 1965), acetone instead of ethanol inthe preceding procedure (Altemuller, 1986), and glutaraldehyde –acetone, with aresin for impregnation (Darbyshire et al., 1985) The best methods can preservethe form of delicate biological materials, such as root tissues and protozoa cells(Altemuller, 1986) Staining roots in the blocks with methylene blue and basicfuchsin followed by sectioning can lead to good identification of the detailedstructure of preserved materials Fluorochromes, such as acidine orange (Darby-shire et al., 1985), can be used to increase the natural root fluorescence Glutaral-dehyde impregnation has been used to study the soil pore network available toprotozoa and roots Although the method is expensive and labor-intensive, it al-lows detailed examination of the soil-root interface Root –soil interactions, at amore detailed level, can be assessed using the scanning electron microscope(Fig 3)

accel-Nuclear Magnetic Resonance Imaging. Nuclear magnetic resonance (NMR)has been used as a noninvasive tool for studying roots in situ (Bottomley et al.,1986) However, because the images are based on H detection, soil moisture levelshave to be kept low The technique has been used to obtain images of root systemsgrown in a range of soil types, vermiculite, sand, perlite, fritted clay, potting soil,and ‘‘peatlite,’’ but the clarity of image varies according to the magnetic properties

of the medium examined To observe relatively dry soil, hence to optimize theroot image, measurements were made at the end of the watering cycles Recentdevelopments have used both 2-D and 3-D images and have been able to distin-guish plant vasculature from surrounding parenchymal tissue (MacFall and John-son, 1994) Images can now be measured for root surface area, volume, and ori-entation (MacFall and Johnson, 1996) However, there are still limits to theresolution at which it can operate

X-Ray Computer-Aided Tomography. This is a non-destructive x-ray techniquethat can separate out features such as roots, soil pores, and cracks, due to their lowabsorbing properties (Tollner et al., 1989) Currently, its resolution is approxi-mately 1 mm, which limits its application for detailed root studies Also, as thex-ray absorption of each pixel is a function of the water content, the dry bulk

Fig 3 A root of spring barley and the soil attached to it as seen with the Stereoscanelectron microscope

density, and chemical properties of the soil (Tollner et al., 1991), it has limitationsfor use in soil physical investigations However, it is possible to calibrate CT mea-surements with root length density measurements from core samples, which al-lows simultaneous nondestructive measurement of both root length density andwater removal from the same soil core.

Neutron Radiography. Neutron radiography has been used to produce dimensional images of plant root systems (Willat et al., 1978) Plant roots grown

two-in narrow (2.5 –5.0 cm wide) boxes, with neutron-sensitive back plates, were radiated with thermal neutrons, and photographic images were obtained fromthese plates Roots were identified because of preferential neutron scattering bythe roots In this way the elongation rates of soybean and maize roots were ob-tained from sequential radiographs Lateral roots (⬍ 0.33 mm) were poorly vis-ible There is a need for improved resolution before the method can be regarded

ir-as a practical means of producing quantitative data Neutron radiography, CT, andNMR are rapidly developing techniques that are constantly improving in imageresolution quality and have the advantage of being nondestructive and usable withsoils under relatively natural conditions However, they have the disadvantage ofnot being available to the majority of researchers

Autoradiography. Radioisotopes have been used in a variety of ways to observeroots or give a measure of root activity in soil (Walker and Barber, 1961) If twospecies grown in mixed culture are injected, one with32P and one with33P, theroots of the two species can be subsequently identified in a section of a soil blockcontaining the cut ends of the roots of both species Mixing radioisotopes into thesoil and assessing depletion around roots has been used by a number of workers

as a means of assessing root activity in soil boxes (see also Sec III.D.3) Thetechnique could also be used to assess the effects of soil impedance on uptake

2 Measurements of Root Parameters

Number. The number of roots can be counted in samples of washed roots or insitu (e.g., glass-faced columns) The number of root tips per unit volume of soilhas been used as an indicator of root distribution in soil (Weller, 1971) It has beensuggested that root number is closely related to leaf number (Richards and Rowe,1976) Image analysis methods now make this easier to determine (Smucker,1993)

Mass. For the determination of root mass, roots are washed free of soil, thenoven-dried (usually at about 80⬚C) for 24 – 48 hours Mass can be measured in allmethods that permit roots to be destructively sampled Mass characterizes the totalamount of root but is not a good indicator of absorbing potential It is less sensitive

to soil factors than root length Treatments with a major effect on root length may

have no effect on mass Measurements of mass are important to the production ofcarbon budgets (Gansert, 1994; Thomas et al., 1996).

Surface Area. Surface area can be related to water and nutrient uptake An mate of root surface area can be obtained from root length and diameter, if roothairs and the extramatricular hyphae of mycorrhizas are ignored Direct methodsused to estimate the root surface of roots washed free of soil include photoelectricattenuation (Morrison and Armson, 1968), dye adsorption, and the retention ofcalcium on the external surface of the root, following a brief immersion in

esti-a concentresti-ated solution of cesti-alcium nitresti-ate esti-and centrifugesti-ation (Cesti-arley esti-and Westi-at-son, 1966) Surface area is now most commonly estimated by image analysis(Smucker, 1993)

Wat-Strength. The tensile strength of single washed roots has been obtained from theforce required to break 5 –10 cm lengths of root of known diameter (Parlychenko,1942) The buckling stress of clamped, excised roots may be used to characterizeroot elasticity Weights are hung from one end of a 10 mm length of root clamped

at its other end Elasticity is related to the deflection caused by a known weight(Goss et al., 1987)

Diameter. The diameter of newly washed root samples from soil cores or tion culture, or from roots in impregnated soil blocks, can be measured directlyand used to estimate either root surface area or length where volume is known(Bhat, 1983) Large numbers of measurements are needed to characterize diameteraccurately The effects of external pressure on root diameter have been studied inbeds of ballotini (Goss, 1977) By varying the size of the glass beads, interactionsbetween pore size distribution and root diameter can be assessed Root diameter

solu-is usually measured using a microscope micrometer eyepiece Diameter can also

be measured on images captured using a minirhizotron system By using pixelcounts, the direct measurement of the average root diameter can be made (Lebo-witz, 1988)

Length. Other than mass, this is probably the most common single ment made

measure-basic calculations Length can be measured directly using calipers andsamples of wet root in a water-filled dish, or by placing roots on graph paper andcounting squares This method is time-consuming, but for large roots (⬎ 5 mmdiameter) it is often the most practical (Atkinson et al., 1976) For samples ofgreater root length, measurement requires some type of sampling method, such ascounting the number of intersections between roots and a random or regular pat-

tern of lines Total root length R can be estimated by

AN

2H

where R is the total length of roots, A is the area of the field of view, and N is the

number of intersections between the roots and a set of randomly oriented straight

lines whose total length is H (Newman, 1966).

Newman (1966) applied this principle to a system where a number of fields

of view were examined using a microscope with a hairline in the eyepiece, whichwas randomly reoriented before each new examination Using this method, thetime for root length measurement was reduced to less than half that required bydirect measurement (e.g., 24 min to measure 3.4 m of root with a CV of 4.3%versus 67 min by direct measurement)

Since Eq 1 requires only that the orientation between roots and a set oflines be random, this equation can be used equally well for regular arrangements

of lines such as parallel straight lines or a grid, provided there is no preferredorientation of the roots in relation to these lines Furthermore, where the line spac-ing or the spacing of a square grid is equal to the unit in which root length is

measured, it is easy to show that A /H is 1 (for lines), and thus

where k is the grid spacing The theory has been used by Marsh (1971) and

Ten-nant (1975) A similar result was obtained empirically by Head (1966) With thismodification of Newman’s (1966) method, there is no necessity for the roots to beuniformly spread out over the counting area A procedure that enables a single set

of measurements to be obtained within 6 minutes, and with a coefficient of tion of 5% or less, has been given by Tennant (1975) It is appropriate to match

varia-grid size and the length of root to be measured and to keep N between about 100

and 500 Time can also be saved because all organic debris does not have to becarefully removed from the roots This method was used in the study of Oosterhuisand Zhao (1994), who calculated root length as intercepts (with a 1⫻ 1 cm grid)

⫻ 0.7857

automated methods The theory presented above underpins systemswhere the scanning of the root sample or the counting of the number of timesroots intersect a regular pattern are automated In one system (Richards et al.,1979; commercially available from Commonwealth Aircraft Corp Ltd., Port Mel-bourne, Victoria 3207, Australia), roots are spread out on a transparent rotatingturntable, which is traversed by a light beam and detector The number of timesthat roots interrupt the beam is converted to give a direct readout in meters Thismachine works best with samples of 20 – 40 m total root length Greater root

lengths tend to be underestimated because of overlapping, while lengths less than

20 m tend to be overestimated The measurement time is about 10 min per sample,with CVs of about 3% for root lengths⬎ 10 m

Length can also be measured using a high-resolution scanning camera ris, 1986; Smucker, 1993) and an image-analyzing computer (such as the Quanti-met) A computerized scanning system is commercially available from Delta-T-Devices Ltd (Burwell, Cambridge CB5 0EJ, UK) When intersect systems havebeen used to process samples scanned with computer scanners (Kirchoff, 1992),video cameras (Harris and Campbell, 1989), or flatbed scanners (Lebowitz, 1988),care has to be taken with line intercept methods due to root overlap, and correctionfactors have been added to compensate for this (Sackville-Hamilton et al., 1991b;Kirchoff, 1992) Also, roots have to be well spread out to ensure random orienta-tion Alternatively, pixel-count methods can be used on a root skeleton and havebeen shown to be more accurate and more precise than the line intercept method(Lebowitz, 1988; Ewing and Kaspar, 1995)

(Har-Limitations to the size of the field of view mean that only small samples can

be measured, and organic debris must be carefully separated from the root sample.Staining roots with dyes such as methyl violet can aid detection with some instru-ments Use of fluorescent dyes permits roots to be distinguished from debris whenilluminated with ultraviolet light (McGowan et al., 1983) Farrell et al (1993)have shown that mean CVs between repeated measurements are normally greaterfor manual than for digitized methods

indirect methods Root length can be calculated from counts of rootnumbers found in sectioned impregnated blocks of soil Formulas are available toconvert root number to root length for both random (Melhuish and Lang, 1968)and strongly anisotropic (Lang and Melhuish, 1970; Melhuish and Lang, 1971)root distributions For the former, provided a reasonably large sample of randomlyorientated sections are taken, the assumptions inherent in the calculations are met.The equation for random root distributions is

where LTis the total length of root per unit volume of soil (cm cm⫺3and N is the

number of roots intersecting a plane of unit surface area (per cm2)

Elongation Rate. Root elongation rates have been studied using pressurizedballotini-filled cells Elongation is most commonly recorded as the difference inlength between successive measurements made directly on still film shots or usingtime-lapse cinematography (Atkinson and Lewis, 1979) Time-lapse photographycan also be used on observation units It allows the quantification of detailed re-actions between root and soil The method has been used to study root nutation(Head, 1965), variation in root –soil contact, and variation in root diameter (Huck

et al., 1970; Atkinson and Lewis, 1979)

Root Age. Root color and morphology are the commonest criteria for identifyingroot age For apple, anatomical changes with age have been described in detail(MacKenzie, 1979) The fluorescence of roots has been shown to decrease withroot age, disappearing when suberization begins A positive correlation has beenfound between the intensity of fluorescence and rate of new root growth (Dyerand Brown, 1983) However, studies have shown that UV fluorescence cannot beused as a universal indicator of root age or functionality, but in some species itcan be used to separate roots from the background, using image analysis tech-niques (Smit and Zuin, 1996) Root activity can also be assessed using stains such

as tetrazolium blue, fluorescine diacetate, acridine orange, or pH indicators likebromocresol purple Aging, by visual identification in situ, is normally performedwhen roots are visible through a glass face (Head, 1966, 1968)

Species Identification. It is possible to identify anatomical differences betweenplant roots (Schwaar, 1971) In observation units, different species can be distin-guished on the basis of characteristics such as color (which can vary from trans-lucent white to pale brown), diameter, branching pattern, root hair development,and UV fluorescence In general, grass species show the highest levels of fluores-cence (Smit and Zuin, 1996)

Distribution. Root distribution can be studied in situ (Schumacher et al., 1969)

or in impregnated blocks of soil (Melhuish and Lang, 1968) Nuclear magneticresonance techniques can also be used for qualitative assessments in situ of rootand water distribution in relation to soil physical and chemical factors

Branching. In solution culture, or with an easily washed-off potting medium,the entire root system of a plant can be extracted and the main and lateral rootsmeasured Hackett and Bartlett (1971) have given a detailed description of thedensity of branching and of changes in lateral length along axes, for plants grown

in solution culture Hooker et al (1992) have described the impact of mycorrhizalinfection on similar developments for a perennial species Describing the branch-ing of root systems and characterizing the amount of branching, the pattern ofbranching, and the different orders of branching has proved to be difficult Fitter(1985, 1991) and Berntson (1995) have approached the problem by using a topo-logical (‘‘tree’’) system of analysis This type of analysis provides a way of as-sessing the morphological effects of differences in soil physical properties How-ever, it remains unclear as to how well the system works on root systems extractedfrom soil Root branching can also be studied in situ using observation chambers

or a soil impregnation technique (e.g., Pages and Serra, 1994)

Volume. Root system volume can be calculated from measurements of lengthand mean root diameter Asamoah (1984) measured root volume directly to anaccuracy of⫾0.025 mL using a meteorological micrometer gauge This fast, non-destructive method can be used to make nondestructive time sequence measure-ments on roots grown in solution culture

III FIELD METHODS

A Introduction

There are no field methods that allow roots to be observed directly without eithertheir removal from the soil or the establishment in the soil of an in situ observationsurface Where soil sampling is used, because observations cannot be repeated forthe same volume of soil, temporal and spatial variation become confused in datafrom sequential samplings Where spatial variation is very high, as is normal(Waddington, 1971), it prevents the detection of temporal variation The tech-niques that can be used to assess root growth in the field divide into three maingroups

Root System Removal. The complete root system (excavation) or part of the rootsystem (soil monolith, soil cores, needleboards) are either removed from the soil

or assessed in situ (profile wall)

Observation Methods. A viewing surface is inserted into the soil, either a smallobservation window (Asamoah, 1984), a minirhizotron (Hendrick and Pregitzer,1992a), or a large walk-in facility (Rogers, 1939b, 1969; Rogers and Head, 1963;Fogel and Lussenhop, 1991)

Indirect Methods. The presence of the root system is inferred from activity (e.g.,the removal of soil water or the uptake of radioisotopes) (Newbould et al., 1971;McGowan, 1974) Indirect methods do not, however, predict what may happenunder other conditions Low root activity may result from few roots being present

or from root inactivity

This section deals with each of these groups of techniques in turn Methodsare illustrated by reference to a small number of selected papers, often the earliestpublished Listing of all published variants of the basic methods is not possible.Most methods used in the field are simple in concept, and so emphasis has beenplaced on the interpretation of results, the situations in which the methods havebeen used, and the factors that may limit their use in studies of the effects of soilphysical conditions

B Root System Removal

1 Excavation

For large plants, excavation involves the removal of more soil than is the case forany other sampling method (Atkinson, 1972, 1981; Tamasi, 1986) Total excava-tions are useful for determining the mass and distribution of large roots (Fig 4).The large loss of fine roots during excavation means that excavation is unsuitable

as a means of estimating root length The method can be useful for studying theeffects of soil type or soil features, such as impeded drainage or depth of indura-tion, on the development of a whole root system (Rogers and Vyvyan, 1934)

Total Excavation Methods. A major study of apple root systems has involved theexcavation of 26 mature apple trees grown on either a light sand, a sandy loam, or aheavy clay loam with about 1000 Mg of soil having to be moved by hand (Rogersand Vyvyan, 1934) An entry trench was dug beyond the rooting volume of the treeunder investigation Beginning from this trench, soil was removed in 50 cm sec-tions, moving systematically across the ground occupied by the root system understudy Soil was brushed away from the side of the trench with a small hand fork,leaving the root system exposed Following such an excavation, the root system iskept as entire as possible to allow later reconstruction (Fig 4).

Using an excavation method, it is possible to compare horizontal and cal root distribution, the total amount of root, and the uniformity of distributionfor different soil types (Table 3) In this study, it was found that a single quadrant(25% of the soil volume) could contain as little as 6% or as much as 51% of totalroot weight This has major implications for the choice of sampling methods andthe accuracy of the data they will produce Although the time needed to sample asingle individual tree will tend to limit the use of excavation, given the spatialvariation inherent in tree root systems and the limited number of large roots whichrepresent most of the root weight, excavation is probably the only method thatwill give a reliable estimate of total biomass This method is easier to apply tosmall trees (Atkinson et al., 1976)

verti-Fig 4 Root system of 26-year-old apple tree (Fortune/M9) excavated by the skeletonmethod (From Atkinson, 1972.)

Partial Excavation. Partial excavations have been used for trees (Coker, 1958).This involves either the excavation of a section (usually one-quarter of the rootingvolume) or of a combination of stump-pulling and root excavation It has beenestimated that this latter procedure removes about 38% of root weight (Atkinson

et al., 1976) Partial excavation has been used to compare the effect of soil type

on root weight and distribution (Coker, 1958; Tamasi, 1986)

2 Profile Wall Method

General Considerations. In this technique, a trench or pit is dug to expose avertical soil profile from which records of partially exposed roots can be made.Horizontal areas, at different soil depths, can be prepared in the same way Unlikepinboard and soil coring methods, this technique can be used on stony soils Themethod is, however, labor-intensive and time-consuming, and it leads to extensivesoil disturbance; moreover it can be difficult to obtain a statistically meaningfulnumber of replicates Nevertheless, it has been suggested that this method givesamong the most favorable ratios of information gained to labor expended (Ward

et al., 1978) The trench, initially dug at a distance from the crop, can be cut backserially toward the plant if information on lateral distribution is needed For a rowcrop, the trench is usually dug across the rows The trench can be dug by hand or us-ing a mechanical digger, and it should be positioned so that a further layer of soil(⬃30 –50 mm) can later be removed from the trench face so as to avoid damagingroots The size of the trench required will depend on factors such as crop type

The Spiral Trench Method. Special considerations apply to tree crops, for whichthe use of a logarithmic spiral trench has been suggested (Huguet, 1973) This

Table 3 The Effect of Soil Type on an Apple Root System (Lanes Prince Albert /M1)

Parameter

Soil texture (series)

Sandy(Wisley)

Sandy loam(Malling)

Clay loam(WisboroughGreen)

Percentage of root weight in 1 m2of

Lowest percentage of root in 25% of

Percentage of root weight at⬎50 cm

Ratio of stem to root weight 0.92 2.36 2.18

Source: Rogers and Vyvyan, 1934.

attempts to weight the intensity of sampling relative to the amounts of soil at adistance from the tree trunk by using a trench in the form of a part of a spiral(Fig 5a) At the periphery of the root system, root density tends to be at a mini-mum (Fig 5b), but soil at this distance, nevertheless, contributes a very largeproportion of the total soil volume With the spiral trench method, such soil is

Fig 5a A spiral trench normal to the plantation line, in a wide herbicide strip.● treetrunk

relatively heavily sampled Comparisons between estimates derived from the rithmic spiral trench and more conventional straight trenches give a higher aver-age root density (0.017 roots cm⫺2soil face area) for the spiral than for the straighttrench (0.011 cm⫺2) (Gurung, 1979) Where the soil cover is not uniform, forexample, where trees are grown in herbicide-treated strips in grass, the orientation

loga-of the spiral trench will influence the density loga-of roots detected Where the tation of the trench is such that it samples soil between tree and grass alley (normal

orien-to row), the estimated root density is higher than when the sampling is principally

of soil between trees in the row

Face Preparation and Measurement. When the trench, of whatever form, hasbeen dug, the working face of the soil profile is prepared using a profile knife to

Fig 5b The number of roots per 500 cm2of trench wall with distance from the trunk oftrees grown under overall herbicide (䡩) or herbicide strip management (●) (Reproducedfrom Gurung, 1979.)

remove a layer of around 10 to 20 mm In stony soils, the preparation is best donewith a spade, trowel, and knives The roots exposed against the wall are cut offwith scissors Starting at this surface, the face is cut back by a further⬃3–5 mm,but usually 10 mm for trees, so as to expose the root system Water and air underpressure have been used to remove this soil layer A frame containing a grid (themesh size depending on root size and sampling strategy) is positioned over theprepared face and the root system recorded.

Measurements. Root number, length, diameter, and distribution may all be tained from profile wall measurements

ob-root number The number of exposed roots visible in every square of thegrid is recorded, either as a count or onto a prepared sheet of foil or graph paper,containing a matching array of boxes A direct record of this type can be used toderive the average number of roots per unit area of profile wall, from which esti-mates of variation may be calculated Pens of different colors can be used to dis-tinguish between new and old roots A direct count is faster than the mapping ofindividual roots (Fig 6) directly onto either graph paper or transparent sheets.However, counts do not show individual roots in their natural position relative tothe profile wall A visual representation of root counting can be used to show theeffects of treatments that result in differences in soil physical conditions Countscan be plotted versus depth (Fig 7) This method has been used to study the effects

of localized irrigation (Levin et al., 1979)

Fig 6 Root distribution map obtained using a profile wall method and illustrating thedifference in the root systems of mature apple trees, between 0 and 30 cm depth, whengrown under cultivation or herbicide management:●, roots ⬍ 2 mm diameter; 䡩, roots

⬎ 2 mm diameter (From Gurung, 1979.)

The number of roots recorded using two variants of the profile wall method,counting and drawing onto foil, have been compared (Bohm and Kopke, 1977).

In general, counts obtained with the foil method were higher than those obtainedfrom direct soil counts Counts in densely rooted areas tended to be less than thoseobtained by other field methods

root length Estimates of root length are based on the assumption thatany root present in the soil will go back into the profile for at least the depth

to which it has been exposed In one variant of this method, one root unit isset equal to a 5 mm root length for a profile that has been cut back by 5 mm.Roots 10 mm long are counted as two root length units (Bohm, 1976) If rootdistribution is assumed to be uniform, root length per unit volume can be calcu-lated Estimates of root length obtained in this way are much lower than thoseobtained by washing roots from an undisturbed block of soil (Table 4) The cause

of the underestimation is not clear, although the removal of the soil from the face of the profile wall to expose the roots may result in the loss of some of thefine roots as well as roots growing parallel to the profile face However, themethod gives a good representation of root distribution It seems to be most reli-able for plants like trees, in which most roots are horizontally distributed Ingrasses, where a large proportion of roots are vertical, many roots would be lostleading to an underestimation

sur-Fig 7 Root distribution with depth, for three crops, determined by the profile wallmethod (From Mackie-Dawson et al., 1988.)

root diameter and volume Using a small hand lens, a micrometerscrew, or calipers, the diameter of exposed roots can be measured directly, in situ.The exposed roots can be distinguished by diameter on a root map (Fig 6), or thenumber in each diameter class can be recorded directly Root volume can be cal-culated from diameter and length (Bhaskaran and Chakrabarty, 1965).

by a series of pins This latter method is necessary where it is important to knowhow the root system distribution is spatially related

Pinboard. The basic pinboard or needleboard method has been described in tail by Schuurman and Goedewaagen (1971) Monoliths 152⫻ 41 ⫻ 91 cm deepwere removed using a root extraction frame (Nelson and Allmares, 1969) that

de-could contain the root systems of four maize (Zea mays L.) plants After its

re-moval, 6 mm diameter brass pins were driven through the monolith into a board

on a 5 cm grid pattern, using a compressed air gun The monolith was then soaked,the side opposite from the pins removed, and the soil washed away The root sys-tem was photographed under water, divided, dried, and weighed This method hasbeen used to assess the effects of treatments such as straw mulching on a total rootmass and on horizontal and vertical distribution (Nelson and Allmares, 1969).Pinboards (35 cm long⫻ 20 cm deep) have been used to assess effects ofsoil physical condition on the root system of winter wheat in a comparison be-tween plowing and direct drilling (Finney and Knight, 1973) Board sampling

a volume 30⫻ 5 ⫻ 30 cm deep has been employed to study the cabbage root

Table 4 Comparison of Estimates of Root Length Density

of Maize (Zea mays) a

aObtained from profile walls and soil monoliths to 100 cm depth.

Source: Bohm, W In situ estimation of root length at natural soil profiles.

J Agric Sci Camb 87 : 365 –368 (1976).

system (Goodman and Greenwood, 1976) In this study, roots and soil were moved from the board as 36 samples of 5 cm3each, and the roots were washedfree of soil, and their length determined Photographic records can be used toindicate the effects of treatments on branching However, such records are difficult

re-to quantify

Modified Pinboard Methods. A method combining the relative ease of samplinggiven by soil coring but also providing the spatial information of the pinboardmethod has been developed (Gooderham, 1969) Samples obtained by soil coringwere encased in a perforated acrylic cylinder and the roots held in place usingnylon fishing line sewn through the holes in the cylinder with a needle Soil waswashed from the core and the remaining root system resuspended in 5% w/v gela-tin This technique allows the root system’s geometry to be related to soil physicalcharacteristics The method is more rapid and flexible than the more traditionalpinboard method However, both these methods are likely to be superseded bymore advanced spatial techniques, such as scanning NMR

Soil Coring. Soil coring is the most frequently used method of root sampling.Coring is often used for sequential sampling of an experimental plot to give esti-mates of temporal change in root length or weight, although spatial variation mayconfound such estimates When samples are taken in relation to the planting ge-ometry of the crops, information on the spatial distribution of their roots can beobtained Soil core samples can be obtained from points immediately adjacent tothe sites of soil physical measurements, or in some cases, the same cores can beused for measurements of bulk density and pore size distribution as well as rootlength The published literature on this subject is very large (Bohm, 1979, contains

an extensive listing), and the papers quoted have been selected only to illustratesome of the variations in technique that have been used and to indicate the types

of information that can be obtained

Welbank et al (1974) described the use of a powered soil coring system toassess the growth and development of cereal root systems; samples were takenusing coring tubes fitted with hardened cutting tips (Fig 8) Variations of thismethod have been used by many workers subsequently

Measurement of Root Growth. To facilitate removal of the soil cores, the tubesare fitted with split liners To avoid compaction of the soil core within the liners,the internal diameter of the cutting tip is manufactured slightly narrower than that

of the liners The coring tube is driven into the ground using a portable powered motor hammer (Fig 9) A depth of 1 m can be reached in moist, rela-tively stone-free sandy loam soil in around 30 seconds In dry soils penetration todepth takes much longer The presence of small stones does not cause a problem,but the corer will not penetrate large stones Tubes are removed from the soil using

gasoline-a tripod gasoline-and chgasoline-ain hoist After extrgasoline-action, the soil cores gasoline-are divided into sectionscorresponding to different soil layers or soil depths

In spite of using the narrower cutting tip, some compaction of the soil coreinevitably occurs Welbank et al (1974) found that the bulk density of coresamples usually showed increases of 1–5% compared with undisturbed samplesfor a range of depths, but in some soils, especially when wet, the compaction in

a 1 m core could be as high as 25% In this situation, allowance must be made forcompaction before the cores are cut up Injections of paint to known depths havebeen used to assess the distribution of soil compaction within the core and to allowfor its correction A simplified, low-cost version of the soil coring system has beendeveloped (Prior and Rogers, 1994) using styrofoam plugs to allow the collection

of multiple core samples within a plastic liner This can reduce the time spent onindividual sample collection

After coring, soil can be separated from the roots by washing on sievesmanually or automatically (Cahoon and Morton, 1961; Bohm, 1979) A range

of washing and flotation techniques and the use of chemical dispersal agents,e.g., sodium pyrophosphate (Schuurman and Goedewaagen, 1971), have been

Fig 8 A design for a soil coring tube (From Welbank et al., 1974.)

Fig 9 Soil coring, using motor-driven hammer.

reviewed by Bohm (1979) The use of dispersal agents is complicated if samplesare later to be used for chemical analysis A modification of the basic method hasbeen described (Smucker et al., 1982) as a hydropneumatic elutriation system.This combines the kinetic energy of pressurized spray jets and the low energy ofair flotation Air and water are used to isolate and deposit roots on a submergedsieve Washing times vary from 3 to 10 min per sample and are a function of soiltype, plant species, the concentration of dispersing agent, sieve size, and soakingtime Using this equipment, nearly 100% root recovery was achieved in around

2 min for a sandy soil, 6 min for a loam, and 10 min for clay (Smucker et al.,1982; Smucker, 1993) These units are available commercially from Gillison(Benzonia, MI 49616, U.S.A.) at a cost in 1997 of $6600 for an eight-chamberunit and $4250 for a four-chamber unit An alternative four-bucket model, based

on the same principle, was available from Delta T, (Burwell, Cambridge CB5 0EJ,UK) for £1815 in 1997

Welbank et al (1974) used root-washing cans (Fig 10) coupled with theintersection method for measuring root length, to compare the effects of plant typeand nutrition on root length, root weight, and specific root length (length per unitweight; seeTable 5) Because specific root length can vary with crop variety, age,nutrition, depth, and soil physical conditions, use of a given value of specific rootlength to convert root mass to length is liable to systematic error (Table 5) and isinadvisable except where these values are obtained from representative subsam-ples taken from the actual samples being assessed Using the extreme values given

by Welbank et al (1974), 1 g dry weight of roots could have a length as low as

33 m or as high as 199 m

Core sampling can result in large errors in the assessment of tional roots To overcome this problem, Ward et al (1978) placed soil cores inbags made from 100 mesh (0.15 mm opening) cloth, which allowed clay, silt, andfine sand to be washed out but retained roots, organic residues, and coarse sandwhen agitated in water Washed roots and organic materials were separated fromsand by flotation, stained with 1% Congo red, and saturated with 95% ethanol.Under these conditions, living but not dead roots stained dark pink to bright redand were then measured by the grid intersect method (Sec II.C.2)

living/func-These workers found that after staining by the method above, the percentagelength of a sample, measured using a grid intersection method, increased by 18%compared with an assessment made by laying the root system out on a dark back-ground The increase was assumed to be due to an increase in the visibility ofsmall rootlets A comparison of the staining of different species indicated thatmonocotyledons stained more than dicotyledons Some dicot species (e.g., sugarbeet) stained very poorly Storage of root samples can affect the quality of stain-ing For example, storage at 15⬚C for 35 days reduced the stainable root length ofwheat roots by 65%, whereas at 5⬚C the decrease was only 13% (Ward et al.,1978) Other dyes have also been used Ottman and Timm (1984) used trypan

blue to identify the length of viable roots using an image analysis system Colorand texture can also be used to discriminate between living and dead roots.Root data obtained through field sampling normally has a high spatial vari-ability associated with the soil physical, chemical, and biological variability Aknowledge of the variability can help both in the design of subsequent experi-ments and in the interpretation of results The coefficient of variation is often used

to describe the variability and is useful provided it is based on a large enough set

of individual measurements

Fig 10 Root washing can for separating roots from soil (From Welbank et al., 1974.)