Behaviour of liquid water in soil

Soils with different textures and compositions differ widely in the amount of water they can hold. Soil matric potential is a measure of the tension with which water is held on the surface of soil particles. In a particular soil, matric potential is a function of soil water content; the drier the soil, the greater the tension. The responses of plants and other biota, as well as many physical traits of soil, are closely related to matric potential. Since the attraction of water to the surface of a soil particle reduces its energetic potential relative to a free, pure body of water, the potential of soil water becomes more negative as a soil dries, and the units of matric potential are properly expressed as negative numbers. For convenience, however, soil water potential is sometimes expressed without the negative sign and referred to as soil water tension or suction, with the understanding that increasing (more positive) tension actually refers to decreasing (more negative) water potential.

Hydraulic conductivity is a measure of the rate at which water can move in a soil. Conductivity decreases rapidly as a soil dries. Table 1 shows an example of the relationship between matric potential and hydraulic conductivity. The relationship is exponential and can only be graphed using log scales on both axes, thus obscuring the drastic reduction in water conductivity as water content and matric potential decrease. For example, a very moist, but well-drained, soil at ‘field capacity’ (generally assumed to be around − 0.03 MPa) and the same soil at − 0.5 MPa (somewhat moist and still supporting healthy plant growth) might differ in hydraulic conductivity by a factor of 600.

Table 1 Water potential, equilibrium relative humidity (RH) and corresponding filled capillary size (based on Papendick and Campbell, Reference Papendick, Campbell, Parr, Gardner and Elliot1981), and an example of hydraulic conductivity and volumetric water content for a fine-textured Geary silt loam (Hanks, Reference Hanks1965). In a medium-textured soil, the volumetric water contents might range from 0.40 to 0.05 cm³ cm⁻³, with hydraulic conductivity at saturation of over 100 cm d⁻¹

^a Example of hydraulic conductivity and volumetric water content. Soils vary widely but always demonstrate an exponential decrease in conductivity with decreasing water content.

Attempts to model liquid imbibition

Since germination is vital in establishing crops, and is also a factor in ecological fitness in natural settings and weed ecology, many researchers have been interested in the environment around a seed and its effect on the rate of germination. Liquid water transfer from soil to seed has been assumed to be the prime path for uptake during imbibition (see references in Table 2), so much research has centred on the relationship between hydraulic conductivity (by varying water content/matric potential) and germination rate, sometimes in combination with other factors, such as the extent of seed–soil contact and bulk density of the soil. A strong relationship was anticipated between the amount of time it takes a dry seed to imbibe water and the water content and hydraulic conductivity of the soil. However, in the research reports listed in Table 2, large differences in soil water matric potential and hydraulic conductivity produced little or no difference in either the rate of water imbibition by the seed, or the time for it to complete germination and subsequent seedling emergence. These results are puzzling, given the assumption that liquid transport is the major mechanism for imbibition by seeds.

Table 2 Examples of research where little differences in timing of germination, timing of seedling emergence or water imbibition by seeds were recorded over a range of soil water matric potentials that produced very different hydraulic conductivities

^a Wheat, Triticum aestivum L.; pea, Pisum sativum L.; chick pea, Cicer arietinum L.; vetch, Vicia sativa L.; maize, Zea mays L.; rapeseed, Brassica campestris L. and Brassica napus L.; rye, Secale cereale L.; canola, Brassica napus L.

To develop a mechanistic model of liquid flow from soil to seed, not only must the rate at which liquid water can move to the seed be known, but also the extent of liquid contact between soil and a seed. Models of contact area have been developed based on theoretical ideas of what contact between seed and soil water films would be like. Collis-George and Hector (Reference Collis-George and Hector1966) estimated that, at field capacity ( − 0.03 MPa), about 1% of the surface of a seed is wetted when surrounded by soil aggregates of a diameter similar to the seed. However, they are careful to state that the assumptions and simplifications used to make the calculations are most accurate only very close to saturation ( − 0.001 MPa). Those assumptions include a lack of hysteresis, which means that a water bridge must form and be maintained, despite water content fluctuations near the seed. When a dry seed is introduced into a moist soil, the seed quickly dries the soil it touches, reducing local water films to very thin layers (Collis-George and Melville, Reference Collis-George and Melville1975). These difficulties are often handled by modelling only very wet conditions and using very coarse media for a soil analogue (Hadas and Russo, Reference Hadas and Russo1974). Given the evidence and theoretical difficulties, it is safest to assume that liquid bridges contact a very small proportion of the seed surface under many field conditions.

Behaviour of water vapour in soil

Water vapour content in soil air remains very near 100% relative humidity over the entire range of soil water contents at which plants and other organisms thrive (Table 1). The question is, how quickly can air at 100% relative humidity supply water to an imbibing seed? Convection is limited in soil, so diffusion is the predominant transport mechanism of soil gases. Diffusion is the process of random relocation of molecules through their individual motions. Gaseous molecules travel at high speed (about 630 m s^− 1 at room temperature), but have frequent collisions with other molecules. As a result, it takes very little time for molecules to move short distances, but a considerable time to move longer distances. For example, diffusion of water in air is about ten times faster over a distance of 1 mm compared to 4 mm, and 100 times faster over 0.1 mm compared to 1 mm (Denny, Reference Denny1993).

Water films covering soil particles are the source of vapour for diffusion to the seed. Seeds planted in soil experience very short soil-to-seed distances. Therefore, when attempts are made to estimate the capacity of seeds to imbibe water vapour, very short diffusion distances must be taken into account.

The problem of diffusion distance appears to have affected several attempts to estimate the capacity of seeds to imbibe water vapour. The commonly quoted dogma is that a sealed vessel containing liquid and gaseous phases will attain an equilibrium of saturated vapour above the liquid. This, theoretically, is correct, but it is difficult to verify. When a dry seed is introduced into the air above the liquid, it is often assumed that the seed will be surrounded by an unlimited supply of vapour-saturated air. But because diffusion is quite slow over long distances, without significant convection currents there will be a gradient between a dry seed, at perhaps − 100 MPa matric potential, and the ‘vapour-saturated air’ at nearly 0 MPa. As a result, a seed will germinate faster the closer it is to a water surface (Fig. 1).

Figure 1 Wheat seeds suspended above water in sealed test tubes. A seed germinates faster (in 3 d) when it is closer to the liquid surface, demonstrating that diffusion of water through air is very dependent on distance. Distances range from c. 1 to 10 mm.

In some studies, fairly short distances were used between the liquid water supply and imbibing seeds. In others, there were distances of from one to several centimetres, or the distance was not reported. Despite different methods and measurements, a comparison of research reports indicates a relationship between distance and estimated imbibition rate. Owen (Reference Owen1952) used a 1-mm air gap and reported that wheat germination was completed in 48 h at 20°C, about twice the time required in contact with liquid. Bruckler (Reference Bruckler1983) used a 1- to 6-mm air gap, and concluded that full imbibition of maize took three times longer in vapour than in liquid. Collis-George and Melville (Reference Collis-George and Melville1978) compared distances of 1 and 3 cm and measured 60 and 80 h for wheat germination to be completed. And, finally, Schneider and Renault (Reference Schneider and Renault1997) placed maize grains 2 cm above water and concluded that it took ten times longer for them to imbibe in the presence of vapour than in liquid water. In all of these studies, diffusion distances were much greater than exist for a seed planted in soil, and therefore potential vapour flow was underestimated by many orders of magnitude.

Can vapour really be sufficient?

Few experiments have been conducted in which seeds have been intentionally germinated in vapour alone, although several reports exist where this has occurred while pursuing other goals. As mentioned earlier, Owen (Reference Owen1952) determined that wheat imbibed with vapour alone germinated in 48 h at 20°C, and Collis-George and Melville (Reference Collis-George and Melville1978) observed germination in vapour in 60–80 h. In addition, Etherington and Evans (Reference Etherington and Evans1986) used sealed Petri dishes packed with soil to germinate various native seeds in contact with soil of different water contents. They found that a piece of fibreglass filter paper placed between the soil and a seed made it easier to observe the seed, and caused no noticeable delay in germination. These experiments demonstrate that liquid water is not required for timely germination.

Experiments designed to compare imbibition with and without seed–soil contact

In experiments where the behaviour of liquid was not producing an adequate explanation for water uptake, the involvement of vapour was offered as an explanation (Harper and Benton, Reference Harper and Benton1966; Choudhary and Baker, Reference Choudhary and Baker1982; Martin and Thrailkill, Reference Martin and Thrailkill1993). Rogers and Dubetz (Reference Rogers and Dubetz1980) even concluded that vapour was apparently more important than soil texture, water content, bulk density and seed–soil contact in this phenomenon. The challenge is to demonstrate unequivocally that seeds at normal seed-to-soil distances, but without the possibility of liquid contact, can imbibe water at the same or nearly the same rate as seeds with soil contact.

An initial attempt to achieve this (Wuest et al., Reference Wuest, Albrecht and Skirvin1999) involved a set-up very similar to that of Etherington and Evans (Reference Etherington and Evans1986). Sealed Petri dishes, filled with soil at different moisture contents, were incubated at a range of temperatures. In some dishes, seed–soil contact was prevented by a piece of coarsely woven fibreglass cloth. This, necessarily, also increased seed-to-soil distance, but still there were comparable germination times with or without soil contact for wheat, barley, mustard and pea seeds (Fig. 2). This experimental set-up worked well for wheat and mustard, but barley appeared to suffer from compression of the lid in the soil treatment. The pea seeds were large and depleted the soil moisture in the Petri dishes to a greater extent than the smaller seeds. More refined methods should be used if this experiment is to be repeated, but the results demonstrate that seed–soil contact is not essential, or even important, for timely germination. When there is less water in the soil immediately surrounding the seed, it takes longer for seeds to imbibe the water they require, because it must be accumulated from a greater volume of soil; but it does not take nearly as long as the decrease in hydraulic conductivity would indicate (Table 1).

Figure 2 Number of days required for 80% seed germination of mustard, pea, barley and wheat placed in contact with soil (solid symbols), or separated from soil by a layer of fibreglass cloth (open symbols). Seeds and soil were sealed in Petri dishes and held at 16°C. Data for wheat are from Wuest et al. (Reference Wuest, Albrecht and Skirvin1999); data for other species are previously unpublished.

In the experiment, there were rare cases of water condensation on the inside surface of the lid of a Petri dish, but the condensation never touched a seed. Whenever condensation was observed, a circular patch of dry lid, several millimetres in diameter larger than the seed, surrounded where the seed contacted the lid. This probably indicates a zone around the seed where the vapour pressure gradient was steep enough to maintain less-than-saturated vapour conditions: a visual demonstration of the diffusion-limited vapour supply when liquid water is several millimetres from the seed.

A limitation of the above research design is the difficulty in proving that the fibreglass supplied absolutely no liquid water films to the seed. While fibreglass did not gain weight during the experiment or feel wet to the touch, nor did it pick up dyed water or transfer nitrate from the soil to the seed, microscopic liquid water films might have existed.

Thus, a different experimental design was devised (Wuest, Reference Wuest2002). A block of moist soil was formed containing many holes of different diameters to provide different seed-to-soil distances. Wheat grains were glued to plastic sticks (as in Fig. 1) and suspended in the holes. The soil contained a soluble dye, so it could be determined if a seed touched the soil, and the extent of seed-to-soil contact could be estimated. The diameter of the holes ranged from 2 to 13 mm. Since the wheat grain was approximately 3 mm in diameter, the seed-to-soil distance varied from a maximum of 5 mm to zero. Two positions in the soil block were without any hole, so the seed was forced directly into the block with intimate seed-to-soil contact. After 24 h, there was only a modest difference (about 15%) in water imbibition when comparing a seed-to-soil distance of 5 mm and intimate seed-to-soil contact (Fig. 3). Under similar conditions, wheat germinated in 40 h with seed–soil contact, compared to 55 h when suspended in the middle of an 11-mm hole. The approximately 30% faster germination might have been due to the difference in vapour diffusion rate over a 4 mm distance, compared to hundredths of a millimetre in the case of seed–soil contact. Nevertheless, contact with liquid water films was not essential for imbibition or germination, and lack of seed-to-soil contact only resulted in moderately delayed germination.

Figure 3 Average water gain and dye coverage (percent of seed surface) of wheat grains suspended in holes of different diameters for 24 h (solid lines). Error bars, where larger than the symbols, show standard errors of the mean (n = 42). There were two zero-diameter (no hole) positions in each soil block. Dotted lines indicate averages for the five experiments with the greatest soil water potential ( − 0.07 MPa) and for the four driest ( − 1.3 MPa). Average soil water potential for all 42 experiments was − 0.16 MPa. From Wuest (Reference Wuest2002); reproduced with permission of the Soil Science Society of America Journal.

In laboratory set-ups such as that depicted in Fig. 1, or seed priming set-ups using osmotically controlled medium to partially imbibe seeds, conditions may be carefully controlled, but difficult to describe or verify accurately. Theoretical relationships between relative humidity and matric or osmotic potential are temperature dependent, and temperature gradients are likely. Collis-George and Melville (Reference Collis-George and Melville1978) measured a 2–3°C surface temperature rise for wheat grains, attributed to the heat of wetting and latent heat of condensation. (This is an alternative explanation for a lack of condensation on the Petri dish lids near a seed.) They concluded that unless this heat was dissipated, it would slow imbibition of vapour.

As permanent wilting point and critical water content for germination are approached, isothermal equilibrium relative humidity departs from 0.999 (Table 1). Given the effect of distance on gaseous diffusion and difficulty in detecting small temperature and humidity gradients, it may prove difficult to reconcile different germination results produced under different experimental methods when working near critical water contents.

Diurnal temperature fluctuations likely cause the soil atmosphere to fluctuate in relative humidity between supersaturated while cooling, and less than 100% when warming. This might result in liquid water condensation on the seed during cooling cycles. In terms of transport mechanisms, however, this would still be gaseous flux, independent of seed–soil contact. Lindstrom et al. (Reference Lindstrom, Papendick and Koehler1976) compared constant temperature to 10°C diurnal fluctuation and found no difference in wheat emergence, so apparently diurnal temperature fluctuation produces no net increase of vapour to the seed.