There is a desperate need to re-evaluate our perspective of the finest roots of all species (Pregitzer, 2002). Trewavas (1986) suggested that ‘In biology we deal with the most complex systems known … complex situations are surely going to require complex explanations if the explanation is to be accurate. Or if we retain simple views should we not honestly admit that they may achieve little when faced with biological complexity?’. Because we, as investigators of roots, have conveniently adopted the ‘a root is a root’ concept (in effect lumping all roots into the same mortar), we have achieved relatively little in root research when compared to advances with the shoot portion of the plant. I am in close agreement with Pregitzer's conclusion that ‘… a root is not necessarily a root’. We need to revise some of our basic assumptions if we are to address the critical questions before us effectively.
Good questions, poor assumptions
Pregitzer asked two very pertinent questions in his commentary (Pregitzer, 2002): what are fine roots?; and which fine roots really die?
These questions remain unanswered principally because of four assumptions: there are few significant morphological differences between roots of similar or differing ontological levels (i.e. the root is a root hypothesis); a root system consists of a smooth (relatively random?) size distribution of root diameters down to the finest roots (i.e. no distinct diameter classes); root ontogeny is an inherently acropetal developmental process (top down); and fine root life spans are in the order of two or more months.
The first two assumptions together preclude ‘splitting’ roots into logical groupings rather than the usual ‘arbitrary classes’ for assessment of mass, density, C : N ratio, etc. This forces us into the ‘pick and weigh’ (Craine et al., 2002; Pregitzer, 2002; Pregitzer et al., 2002) method of root analysis. The third assumption suggests that the smallest roots are always at the terminal end of major roots and are functionally similar to each other. Finally, the fourth assumption suggests that nondestructive analyses like minirhizotron images or rhizotron tracings, with monthly sampling cycles, will adequately identify the first existence and then the disappearance of every root within the field of view. I contend that the answers to Pregitzer's questions can be found only if we dispose of these four assumptions.
What are fine roots? Assumption 1
Margaret McCully finds many subtle differences between roots. One difference applicable to fine roots is the observation that the finest roots of corn (Zea mays) have a dramatically reduced anatomical development (McCully, 1987). Much of the reduction is in a cortex reduced to an epidermis, hypodermis and endodermis, no large xylem vessels and, apparently, no pericycle. Thaler & Pages (1996–Hevea brasiliensis) and Freixes et al. (2002–Arabidopsis thaliana) suggest that apical diameter and developmental pattern, including branching, are controlled by the availability of assimilate. Assimilate is undoubtedly restricted towards the apex of these anatomically rudimentary roots. Questions of minimum size and mechanism of size control need to be addressed by future research. Other research by McCully's group has pointed out additional significant anatomical differences between larger roots (St. Aubin et al., 1986–Z. mays; Kevekordes et al., 1988–Glycine max). We need to look much more closely at our root materials to determine if apparent developmental or functional differences are based on classifiable anatomical differences.
What are fine roots? Assumption 2
The second assumption also fails. Zobel (2003), arguing from results acquired with image analysis of root length and diameter of several different species (Dactylis glomerata; Chicorium intybus; Trifolium repens; and Panicum virgatum) (using pixel sizes < 25% the diameter of the smallest roots), suggests that Varney et al. (1991–Z. mays) had it right: that there are distinct diameter classes amongst roots < 1 mm in diameter. Ryser (1998–D. glomerata) and Wells & Eissenstat (2001–Malus) provide relatively fine resolution root diameter data that suggest distinct root diameter classes among their finest roots. Assuming discrete diameter classes within a broad diameter class (i.e. < 1 mm diameter), important differences in longevity, function and developmental patterns among these discrete classes will be obscured by a lack of precision in diameter class definition. For instance, do 0.3 mm diameter roots live twice as long as 0.15 mm diameter roots? Should Zobel's hypothesis (Zobel, 2003) prove true, there is, indeed, a large body of research left to do before we can begin adequately to respond to Pregitzer's lament that ‘One of the most remarkable gaps in our knowledge is that we still do not know which fine roots on the branching root system die and what controls the mortality of individual roots’. Not to mention individual root functionality and the physiological bases for the observed integrated functionality of a root system.
What are fine roots? Assumption 3
The third assumption is true on average, and with seedlings or young large roots (> 1 mm diameter), but false in terms of the life cycle of a root system. Observation of published images such as those of Pregitzer et al. (1997, 2002) supports the concept that the smaller diameter roots tend to be at the distal end of larger roots. A careful examination would suggest, however, that short, thin, roots do exist throughout the root system, even towards the base of larger roots. Lateral roots on corn nodal roots are a mixture of two determinate diameter classes (0.15 mm diameter – 1 cm long and 0.3 mm diameter – 2 cm long) and a number of indeterminate roots (R. Zobel, unpublished; Varney et al., 1991). The presence of fine diameter roots at positions other than the distal end of the parent root suggests that rules for classification based on general developmental patterns may be insufficient for investigating the life cycle(s) and functionality of roots. Jones (1943) mentioned that small roots on the main axis of alfalfa (Medicago sativa) die back and are replaced periodically during the life of the plant. Subsequently, Lyford (1975), Reynolds (1975) and Pregitzer et al. (2002), among others, have described fine root stubs and scars on larger roots. Paollilo & Zobel (2002), investigating the root scar issue, have found root scars and root stubs along the larger roots of a wide range of dicotyledonous species, including Acer and Daucus spp. They conclude, from detailed anatomical studies of these scar regions, that these scars represent sites where a previous root had died back and one to several rounds of adventitious root had subsequently developed and died back at the same site. This provides a mechanism for extensive morphological plasticity of the root system where it can explore and re-explore the same volume of the soil over an extended period of time. Are these adventitious roots functionally identical to their lateral counterparts? I originally termed these roots ‘collateral’ to suggest their functional similarity to lateral roots, but their functionality has yet to be adequately investigated.
Which fine roots really die? Assumption 4
The fourth assumption is as problematic as the second. Before the technology encompassed by rhizotrons and minirhizotrons was developed, it was necessary to assess root longevity by measuring total root length and mass as it changed over time. This has led to the assumption of longevities in the order of months or longer. Most published minirhizotron studies have a sampling frequency of between 2 wk and one month, precluding accurate assessments of turnover rates of less than 4 wk. Cahn et al. (1989–Z. mays) suggest that the fine lateral roots of maize have a life duration of only two weeks under his conditions, and Huck et al. (1987) present data for soybean that suggests similar life spans. Some of our recent pasture research suggests similar life spans for smaller roots of pasture grasses (Lolium arundinaceum; D. glomerata), dicots (C. intybus) and legumes (T. repens). Without minirhizotron sampling frequencies of less than 7 d, it is impossible to determine whether a given set of roots has shorter or longer turnover rates than those suggested by Huck et al. (1987) and Cahn et al. (1989). An interesting question also arises: are these roots replaced by other lateral roots or by adventitious roots? The mechanism for adventitious root replacement is rather straightforward, but does the pericycle allow for development of new lateral roots adjacent to dead or dying lateral roots?
Integrating it all
Cahn et al. (1989) reported that maize roots less than 0.6 mm diameter were determinate and only grew to 20 mm in length, while roots larger than 0.6 mm were inherently indeterminate in growth habit. Fusseder (1987) and McCully (1987) both described the smallest maize roots as determinate in that they lose their apex (after, or during, cessation of growth at around 1–2 cm) and gradually die back to a stub or scar on the parent root. Tippett (1982) has described the formation of abscission layers under short roots in balsam fir (Abies balsamea), hemlock (Tsuga canadensis) and white pine (Pinus strobus) and traced their development to the eventual shedding of the short roots. Pregitzer (2002) points to the research of King et al. (2002) to suggest that nutrient availability and mycorrhizal infection control/modulate the longevity of fine roots. This conclusion is in agreement with the results of Thaler & Pages (1996) and Freixes et al. (2002), who argue that nutrient availability is the controlling influence. If roots smaller than 0.6 mm have inherently restricted phloem, the assimilate restriction hypothesis and abscission layer development are good candidates for determination of longevity.
Longevity of roots is critical to the overall functionality of the root system. Short longevity requires continual new root formation to replace lost roots and maintain absorptive length. Concomitant new root initiation and root death provides the plant with temporal root length stability (Huck et al., 1987). Variations in the ratio between fine root death and initiation will give rise to increases or decreases in total root length. The demonstration of cyclical adventitious root development as replacement roots adds further support to this concept. The concept of continuous root initiation coupled with a comparable rate of root death provides a mechanism for a plant root system to change the pattern of spatial coverage when environmental conditions change. If the presence of nutrient patches increases the flow of assimilate to the fine roots in the patch, this increased flow should extend the life of individual roots (vis à vis the discussion in King et al. (2002)), resulting in a dramatic increase in the total root mass within the nutrient patch. This is one of many possible scenarios.
Perspectives – and a good assumption
As Pregitzer (2002) suggests, ‘… a root is not necessarily a root’. To me this means that roots may differ in shape, size, anatomical complexity and functionality. We, as root researchers, need finally to admit that roots and root systems are far more complex than we wish them to be. My laboratory has definitively demonstrated this for primary and secondary roots (Zobel, 1975–Lycopersicum esculentum; Bushamuka & Zobel, 1998a,b–Z. mays and G. max), and the above discussion suggests that there are similar differentiations between roots of the finest diameters. I propose a new assumption to replace the four flawed assumptions: a plant root system is best described as an integration of multiple genetically and anatomically determined functional root classes. By adopting this assumption, without diameter prejudice, and increasing the frequency of nondestructive sampling to < 1 wk, we should be able, ultimately, to answer both of Pregitzer's primary questions: ‘Which fine roots really die?’ and more importantly, ‘What is a fine root?’.