Idea Transcript
New Phytol. (1991), 118, 63-68
On the relationship between specific root length and the rate of root proliferation: a field study using citrus rootstocks BY DAVID M. EISSENSTAT Citrus Research and Education Center, University of Florida-IFAS, 700 Experiment Station Road, Lake Alfred, Florida 33850-2299 USA {Received 6 September 1990; accepted 2 January 1991) SUMMARY
This study tested two hypotheses: (1) species with roots that have a high length to dry mass ratio or specific root length (SRL) also have the potential for high rates of root growth in small volumes of favourable soil and (2) variation in average root diameter fully accounts for variation in SRL. To minimize differences among shoots, the study used 13-year-old 'Valencia' sweet orange [Citrus sinensis (L.) Osbeck] trees budded to rootstocks representing a range of genotypes. Soil cores 7-4 cm in diameter and 14-2 cm deep were extracted from beneath the canopy, and the soil was sieved free of roots and replaced. Root length, diameter and dry weight of the roots in the disturbed soil and adjacent undisturbed soil were evaluated 5, 10, 19 and 40 weeks following soil replacement. The disturbed soil had a higher water content than the undisturbed soil for the Hrst three sampling dates. Averaged across rootstocks, root length density increased in a linear fashion in the disturbed soil and was comparable to that in the undisturbed soil by 40 weeks. Mean root diameter of the fibrous roots ( < 2 mm) declined with age. Rootstocks with the highest SRL had the most rapid rate of root proliferation (cm cm"^ wk"^) (r = 0-94) and the greatest rate of water extraction at 19 weeks (r = 0 79). Although variation in root diameter contributed to rootstock variation in SRL, the data also suggested that rootstocks of high SRL had roots with lower tissue density than those of low SRL {P < O'OS). The potential trade-offs of constructing root systems of high SRL are discussed. Key words: Root length density, root diameter, root tissue density, root proliferation, root growth.
TNTRODLCTiON The total length of a plant's root system is of fundamental importance in the acquisition of water and nutrients. Species vary widely in the amount of root length per unit root biomass or specific root length (SRL) (Barber, 1984; Fitter, 1985). If drymass density and the energetic cost of constructing a unit of root biomass are similar in species with roots of large and small diameter (Chapin, 1989), then the former would expend more energy to construct a root system of a given length. Species with a relatively small initial investment in biomass per unit root length may be at an advantage at exploiting pulses of water or nutrients in the soil, In a comparison of two arid-land grass species of similar shoot:root biomass ratio, shoot architecture and shoot photosynthetic rates, the species with the smaller diameter roots had more rapid root proliferation in soil enriched by pulses of nutrients (Eissenstat & Caldweli, 1988a; Jackson & Caldwell, 1989) and also more rapid root invasion into
disturbed soil (Eissenstat & Caldwell, 1989). How^^.^^ ^^^^ diameter by itself is not typically a good indicator of the rate of root extension of a whole root system (e.g., Christie & Moorby. 1975; Schenk & Barber, 1979), presumably because species often have large differences in their total carbon assimilation rate and shoot: root allocation patterns. This study investigated the link between SRL and the rate of root extension in disturbed pockets of soil for citrus rootstocks under field conditions. In addition, I examined w^hether variation in average root diameter fully accounted for variation in SRL. Shoot differences were minimized by using genetically identical shoots grafted to genetically diverse rootstocks representing a wide range of root architecture. ^_^^^^^^^^
^^^
METHODS
The study was conducted in a citrus rootstock trial located 7 km south-east of Avon Park, Florida, USA (27° 34' lat. 81° 28' long.). The soil was a deep, uniform, Astatula fine sand (Typic quartzipsam-
64
D. M. Eissenstat
ment). This soil has a low cation-exchange capacity, low water-holding capacity and essentially no horizon development or soil structure. The 13 year-old 'Valencia' sweet orange [Citrus sinensis (L.) Osbeck] trees were grown in a 46 m spacing in the row and 6-2 m spacing between the rows. Rootstocks were arranged in a randomized complete block design w^ith rootstock genotypes randomly located in threetree sets within each row (block). To eliminate any potential overlap, samples were only collected from the middle tree in each rootstock set in each of eight rows. Selected rootstocks represented a broad range in root diameter, based upon observations of roots of seedlings (Graham & Syvertsen, 1985 ; W. S. Castle, unpublished data). The rootstocks were sour orange (SO) (C. aurantium L.), Cleopatra mandarin (CM) (C. reticulata Blanco), trifoliate orange (TO) {Poncirus trifoliata (L.) Raf.), Swingle citrunrielo (SC) (C. paradisi Macf. x P. trifoliata), rough lemon (RL) (C. jambhiri Lush) and Volkamer lemon (VL) (C. volkameriana Tan. & Pasq.). Using 1987-88 data (W. S. Castle, unpublished), the ranking of rootstocks based upon average tree yield were as follows: RL - VL > SO > CM > SC > T O . Based upon canopy volume, the ranking was: VL =i CM > SO ^ RL ^ SC > T O . To compare roots of known age with undisturbed, mixed-age roots, six to seven soil cores, 7-4 cm diameter and 14-2 cm deep, were extracted from beneath the canopy of each of 48 trees (six rootstocks X eight replications). Each core was taken two-thirds the distance from the trunk to the canopy dripline, in mid-September, 1988. Into the resulting hole, a plastic pipe was snugly fitted so that the top was flush with the soil surface. To permit roots to readily grow into the plastic pipes, its sides had twelve 3-6 cm diameter holes covered with a 3 mm-mesh nylon screen. The top and the bottom of the container were covered with a l-5mm-n:iesh screen. The soil extracted from beneath the canopy of the six rootstocks was sieved to remove roots greater than 3 mm in length, mixed to remove any potential nutritional and biotic differences and then replaced in the plastic containers. Two containers per tree were sampled at 5, 10 and 19 weeks following soil replacement. At 40 weeks following soil replacement, one container was sampled per tree from 16 trees (two to three trees per rootstock). In addition, two soil cores of similar size were taken from adjacent undisturbed soi! at each sampling time. Soil cores were kept in sealed plastic bags until roots were extracted, 1 to 3 d later. Roots were extracted and length of the 'fibrous ' roots ( < 2 mm diameter) was determined by the Hne-intersect technique (Newman, 1966; Tennant, 1975). Length of the 'pioneer' roots (> 2 mm diameter) was determined by direct measurement. The fibrous root system was cut in 0-5 cm segments and then 20 to 40 root segments were randomly selected and their diameter deter-
mined under a dissecting microscope at 8 x magnification with an ocular micrometer. Root biomass was determined after oven drying at 70 °C for 48 h. In addition to root measurements, soil water content in the disturbed and undisturbed soil was assessed gravimetrically on the first three sampling dates. At the 19 week sampling date, water content was measured 8 d following a heavy rain ( > 5 cm).
RESULTS
Averaged over all the rootstocks, root length density in the disturbed soil increased in a linear fashion, reaching densities comparable to that in the undisturbed soil after 40 weeks (Fig. l a ) . In the adjacent undisturbed soil, root length density was fairly constant from October to February but declined in the spring when there was relatively little rainfall {P < 0-05, by repeated measures ANOVA). Mean diameter of the fibrous roots in either the disturbed or undisturbed soil, averaged over all six rootstocks, exhibited significant declines from November to June {P < 0 05, by repeated measures ANOVA). Although at 5 weeks following disturbance, roots in the disturbed soil had substantially 6r
(a)
Undisturbed
Ol
c
Disturbed
Sep. Oct. Nov.Dec. Jan. Feb. Mar. Apr. Mav JLine 800 r (fa) 700
-
Disturbed
600
500
Undisturbed
400
300
Sep. Oct. Nov. Dec. Jan- Feb. Mar. Apr. May June Month
Figure 1. Time course of (a) root lengtb density and {b) mean root diameter of the fibrous roots, averaged across six citrus rootstocks, in disturbed and adjacent undisturbed soil. Roots were sampled at 5, 10, 19 and 40 weeks after disturbance. Bar denotes two SE.
Specific root length and root proliferation in citrus 60 r
20 -
10 -
250
500
750
1000 1250 Diameter {//m)
1500
1750
Figure 2. Relative frequency of different size categories of fibrous root diameter in Cleopatra mandarin (CM) and trifoliate orange (TO), sampled 10 weeks after disturbance. Relative frequency for each rootstock was determined from 320 0-5-cni root segments (40 segments x 8 trees).
greater diameter than those in the undisturbed soil; they also exhibited a more rapid decline m diameter, so that by 40 weeks, root diameter was similar for the two soil conditions (Fig. 16). The decline in root diameter was found in all six rootstocks (data not shown). Visual inspection of the roots indicated considerable sloughing of epidermal cells and occasionally cortical cells in the older fibrous roots. One might question the usefulness of mean fibrous root diameter as a descriptor of the root morphology of a genotype, since root diameter varies widely for an individual plant and is affected by the order of the root, age of the root, as well as environmental
65
conditions (Fitter, 1985). At least in this study where differences in environmental conditions and root age were nninimized and a large number of roots were examined, the citrus rootstocks clearly exhibited variation in mean fibrous root diameter which was then used to contrast genotypic variation in fibrous root morphology (Fig. 2). In disturbed soil, rootstocks differed {P < 0-05) at 19 weeks in total root length density (RLD), the percentage of root length comprised of pioneer roots ( > 2 mm) and the average diameter of fibrous roots ( < 2 mm) (Table 1). Because root length increased in a linear fashion in the weeks following disturbance (Fig. 1 a), the estimated mean root age of the different rootstocks sampled at 19 weeks was fairly similar and not related to rootstock differences in root diameter (Table 1). Based upon total RLD and pioneer RLD, the rootstocks appeared to separate into two distinct groups. SC, SO and CM had the greatest average fibrous-root diameter and more pioneer roots but generally lower total RLD than RL, VL and TO. In undisturbed soil, there were few differences among rootstocks in these variables. Differences in root diameter of roots of similar age were small, w'ith the mean root diameter of CM (the thickest) only 17% larger than that of T O (the thinnest). The similarity of root diameter among these rootstocks is also illustrated in Figure 2, where a large percentage of tbe roots of CM had the same diameter as TO. If roots are cylindrical, root diameter determines the ratio of root length to root volume, which can be directly compared to the ratio of root length to root dry mass, or specific root length (SRL). The length to volume ratio of each rootstock in disturbed soil was calculated from the reciprocal of the crosssectional area [{71 x radius^)"^] of each root diameter estimate (Table 2). Based upon estimates of root
Table 1. Characteristics of roots 19 weeks after disturbance for six citrus rootstocks {8 trees each) in disturbed and adfacent undisturbed soil. Root length density {RLD) represents the sum of pioneer roots {> 2 mm) and fibrous roots {< 2 mm) Disturbed
Rootstock SC
so CM RL VL TO LSD (0-05)
-P(5,35d.f.)t
Mean* age (wks) 10-6 U-2 132 11-6 12'5 12-9 16 0-016
Undisturbed
Total RLD (cm cm"^)
Pioneer RLD (% total)
Fibrous root diam {/im)
Total RLD
Pioneer RLD
(cm cm"^) ("0 total)
Fibrous root diam {jim)
0-53 0'69 0-66 1-65 1-85 2-02 0-72 < 0 001
8'02 5'00 5-59 2-32 1-04 1-04 305 < 0-001
673 665 688 637 632 587 44 < 0-001
2-76 3-00 3-69 3-30 3-71 3-67 1-00 0'258
601 532 551 511 529 474 82 0-078
1-08 0-61 l'2O 0-51 0-24 074 0'06 0-046
* Mean root age at the 19-week sampling data was estimated for each tree by calculating the weeks following disturbance when 50% of the root length was attained, based upon the interpolation of root length estimates at 5, 10 and 19 weeks and assuming no root death. t Based upon a 2-way ANOVA (8 blocks, 6 rootstocks). S
ANP 118
D. M. Eissenstat
66
Rootstock SO CM SC RL VL TO
Length/vol (cm-n
Length/mass Mass/vol (cm g""') (gcm'^)
243 234 249 293
1240 1280 1430 1660 1750 2270 260 < 0-001
287 338 40
LSD (0'05)
P(5,35d,f.)*
< 0-001
lten
Table 2. Patterns of allocation of root length, root dry mass and root volume for six citrus rootstocks (8 trees each). Roots were sampled in soil 19 weeks after disturbance
o
0-203 0-184 0-181 0-179
2-
u 0)
ra
0167 0149 0-032 0-041
OCM 0
* Based upon a 2-way ANOVA,
T RL V SO A SC 1
2
TO
3
VL 4
Root length density (cm cm^^)
-
Figure 4. The relationship of root length density and gravimetric soil water content 19 weeks after disturbance for individual trees and (inset) averaged for each rootstock (« = 8 trees/rootstock). Sampling was 8 d following a heavy rain. The curve represents the best-fit of the 48 trees using the iteratively solved asymptotic equation, y =
r = 0-58
0-25 r
0-20
0-15 -
• * •
0-10
o O
0-05
• V A A •
RL SO SC TO VL
'VA
V^
O'OO
0-5
1.0
1-5
2-0
2-5
3.0
SRL (cm
Figure 3. The relationship of root growth rate with specific root length (SRL) for individual trees and (inset) averaged for each rootstock (n = 8 trees/rootstock). Growth rates and SRL were determined by sampling the disturbed soil at 5, 10 and 19 weeks.
volume and root mass, the mass: volume ratio, which defines the density of the root tissue, was also estimated. Compared to sour orange, trifoliate orange has about 39 % greater root length per unit root volume and 83 % greater root length per unit root dry mass (SRL), suggesting that rootstock differences in SRL were not caused entirely by differences in root diameter; trifoliate orange allocated aproximately 27 "o less dry matter to construct a unit volume of root than sour orange. The soil water content of the disturbed soil was greater than that of the undisturbed soil on the first three sampling dates (data not shown), presumably because of the lower RLD in the disturbed soil (Fig. 1 a). Thus, roots growing in the disturbed soil commonly had access to more water than those in the adjacent undisturbed soil.
Specific root length of the rootstocks was significantly correlated to average rate of root extension, based upon either the mean tree response (Fig. 3; n — 48; /* < 0-0001) or mean rootstock response (Fig. 3, inset; n = b, P = 0-018). A 77 "o increase in mean rootstock SRL from 1-3 to 2 3 cm mg"' was associated with a 250"o increase in average root extension rate. Consequently, rates of root growth in the disturbed soil, based upon mass (mg cm~^ wk"^) rather than length (Fig, 3), were also greatest for rootstocks of greatest average SRL (r = 0-74; w = 6; P < 005). The positive correlation of root proliferation in disturbed soil with SRL could not be related to root growth patterns in undisturbed soil, since five of the six rootstocks exhibited a decline in root length density in undisturbed soil frotn October to February (data not shown) with trifoliate orange exhibiting the greatest decline (25%). Differences in RLD in disturbed soil were associated with differences in water extraction (Fig. 4). Eight days following a heavy rain, water content of the disturbed soil was strongly correlated to R L D in the disturbed soil. Rootstocks with greater RLD and higher SRL extracted significantly more water than those with less RLD and lower SRL (Fig. 4, inset; P< 0-05).
DISCUSSION
In this field study using genetically identical shoots, genotypic variation in SRL was associated with genotypic variation in rates of root proliferation in disturbed soil. Rootstocks of high SRL not only
specific root length and root proliferation in citrus invaded and proliferated the relatively moist disturbed soil on a length basis, they also exhibited greater opportunistic biomass allocation than rootstocks of low SRL. Furthermore, rootstocks with high SRL and a greater root length density in the disturbed soil were able to extract water more rapidly than those of low SRL. Similar results have been reported in a series of studies that compared two arid-land tussock grasses of different root diameter (Eissenstat & Caldwell, 1988G, 6, 1989; Jackson & Caldwell, 1989). Rootstock variation in SRL was not entirely explained by differences in root diameter. A close relationship of SRL with diameter only occurs if the density of the root tissue is constant. In this study, estimates of root volume, root length and root mass indicate that there were important differences in the density of the tissue as well (Table 2). Anatomical observations at 19 weeks revealed a considerably more lignified hypodermis in sour orange than in trifoliate orange (Achor & Eissenstat, unpublished data), which may contribute to the apparent differences in their tissue density (Table 2). The positive correlation of SRL with the ability to develop rapidly high RLD in fibrous roots is in contrast to the relationship of SRL or root diameter with root extension of individual pioneer roots with diameters greater than 2 mm. Using root-observation boxes or trenches in which individual 'pioneer" roots in the soil were viewed with glass windows, several investigators have indicated that pioneer roots of greater diameter have more rapid extension rates than those of smaller diameter (Mason, Bhar & Hitton, 1970; Kozlowski, 1971; Head, 1973). Presumably, the root tips of large-diameter roots ( > 2 mm) are strong sinks for available photosynthates. Clearly, the relationship of SRL with root extension is very different for a population of fine fibrous roots versus individual coarse roots, which exhibit radial growth and extend much greater distances horizontally and vertically in the soil. This study involved pruning roots at the time of disturbance, which may have contributed to differences in root proliferation found among rootstocks. A more plastic root system that more readily reallocates root length in favourable locations by building energetically less expensive root length also might be more responsive to root pruning (analogous to herbivory) than root systems which invest heavily in more biomass and possibly more defense compounds (e.g., lignins) in construction of root length. Root proliferation in soil patches rich in water and nutrients is likely to aid in the competition for limited soil resources. Plants that construct roots with high SRL may be more successful in this regard. However, substantial costs may also be associated with construction of roots of high SRL. Limited comparisons of cold desert shrubs (Caldwell & Camp, 1974; Fernandez & Caldwell, 1975),
67
chaparral shrubs (Kummerow, Krause & Jow, 1978) and tundra graminoids (Shaver & Billings, 1975} indicate that plants with roots of small diameter tend to have shorter life spans and higher root turnover rates than those with roots of large diameter. In addition. Grime, Crick & Rincon (1986) suggest that plants adapted for high morphological plasticity are likely to have high turnover rates of their tissues. In conclusion, low root biomass allocation for the production of root length (high SRL) was associated with relatively high rates of root proliferation in disturbed soil. Opportunistic root growth may be a characteristic of plants with high SRL. Both variation in root diameter and variation in tissue density apparently contributed to differences in SRL among citrus rootstocks.
ACKNOWLEDGEMENTS
Research was supported hy project LAL-03056 of the Florida Agricultural Experiment Station. I wish to thank K. Smolarz and C. Zickefoose for technical assistance; W. S. Castle and C. E. Crews, Sr. for use of the rnotstock trial; and the helpful reviews by A. H. Fitter, M. P. Coutts and an anonymous reviewer. Florida Agricultural Experiment Station Journal Series no. R-00969.
REFERENCES B.'iRBER, S. A. (1984), Soil Nutrient Btoaiailability. York,
Wiley, New
CALDWELL, M . M . & CAMP, L . B. (1974). Belowground pro-
ductivity of two cool destTt communities. Oecologia 17, 123-130. CH.'^PiN, F. S. i l l . (1989), The cost of tundra plant structures: evaluation of concepts and currencies. American Naturalist 133. 1-19. CHRISTIE, E . K . & MOORHY, J, (1975). Physiological responses of semiarid grasses. I. The influence of phosphorus supply on growth and phosphorus absorption. Australian Journal of Agricultural Research 26, 423-436. EiSSENSTAT, D. M. & CALDWELL, M . M . (!988fl). Seasonal timing of root growth in favorable microsites. Ecology 69. 870-873, EISSENSTAT, D , M . & CALDWELL, M M . (1988&). Competitive
ahility is linked to rates of water extraction: a Beld study of two aridland tussock grasses. Oecologia 75, 1-7, EISSENSTAT. D , M , & CAI.DWELL, M , M . (1989), In\'asive root growth into disturbed soil of two tussock grasses that differ in competitive effectiveness. Functional Ecology 3, 345-353, EERNANDEZ, O . A, & CALDWTLI., M , M , (1975), Phenology and
dynamics of root growth of three coo! semi-desert shruhs under field conditions. Journal uf Ecology 63, 703-714, FITTER, A, H, (1985), Functional significance of root morphology and root system architecture. In: Ecological Interactions in Soil: Plant, Microbes, and Animals (F^d. by .\. H, Eitter, D, Atkinson, D, J. Read & M - B , Usht-r). pp. 87-106, Blackwell Scientific Publications, London. CiRAHAM, J, H, & SYVERTSEN. J, P, (1985), Host determinants of mycorrhizal dependency of citrus rootstock seedlings, Neu' Phytologisi 101, bbl-bib. GRIME, J. P., CRICK, J, C. & RINCON, J, E, (1986), The ecological
significance of plasticity. In: Plasticity in Plants (Ed, by D. H, Jennings & A, J, Trewavas), pp, 5-29, Cambridge t'niversity Press, New York. HEAD, G . C , (1973), Shedding of roots. In: Shedding of Plants Parts (Ed, by T, T, Kozlowski), pp, 237-293, Academic Press. New York. JACKSON, R , B , & CALDWELL, M , M , (1989), The timing and S-2
68
D. M. Eissenstat
degree of root proliferation in fertile-soil microsites for three cold-desert perennials. Oecologia 81, 149-153. Kozi.OWSKi, T. T. (1971). Gromh and Det-elopment of Trees, vol. 2, Cambial Growth, Root Growth, and Reproductive Growth. Academic Press. New York. Kl.!MMERow, J., KRAUSE, D . & Jow, W. (1978). Seasonal changes in fine root density in the southern California chaparral. OecologtaZl. 201-212. MASON, G . F . , BH.AR, D . S. & HITTON, R. J. (1970). Root growth
studies on Mugho pine. Canadian Journal of Botany 48, 43—47.
NEWMAN, E . I. (1966). A method of estimating the total length of root in a sample. Journal of Applied Ecology 3, 139-145. ScHENK, M. K. & BARBER, S. A. (1979). Phosphate uptake by corn as affected by soi! characteristics and root morphology. Soil Science Society of America Journal 43, 880-883. SHAVER, G . R . & BILLINGS, W . D . (1975). Root production and root turnover in a wet tundra ecosystem. Barrow, Alaska. Ecology 56.401-409. TENNANT, D . (1975). A test of a modified line intersect method of estimating root length. Journal of Ecology 63, 995-1001.