Journal of Experimental Botany, Vol. 52, No. 355, pp. 309±317, February 2001
Responses of plant growth rate to nitrogen supply:
a comparison of relative addition and N interruption
treatments
Robin L. Walker1,3, Ian G. Burns1 and Jeff Moorby2
1
2
Horticulture Research International, Wellesbourne, Warwick CV35 9EF, UK
Imperial College at Wye, Wye, Ashford, Kent TN25 5AH, UK
Received 31 May 2000; Accepted 22 September 2000
Abstract
This paper investigates the effects of uptake of
nitrate and the availability of internal N reserves on
growth rate in times of restricted supply, and examines the extent to which the response is mediated by
the different pools of N (nitrate N, organic N and total
N) in the plant. Hydroponic experiments were carried
out with young lettuce plants (Lactuca sativa L.) to
compare responses to either an interruption in external N supply or the imposition of different relative N
addition rate (RAR) treatments. The resulting relationships between whole plant relative growth rate (RGR)
and N concentration varied between linear and curvilinear (or possibly bi-linear) forms depending on the
treatment conditions. The relationship was curvilinear
when the external N supply was interrupted, but linear
when N was supplied by either RAR methods or as
a supra-optimal external N supply. These differences
resulted from the ability of the plant to use external
sources of N more readily than their internal N
reserves. These results show that when sub-optimal
sources of external N were available, RGR was maintained at a rate which was dependent on the rate of
nitrate uptake by the roots. Newly acquired N was
channelled directly to the sites of highest demand,
where it was assimilated rapidly. As a result, nitrate
only tended to accumulate in plant tissues when its
supply was essentially adequate. By comparison,
plants forced to rely solely on their internal reserves
were never able to mobilize and redistribute N
between tissues quickly enough to prevent reductions in growth rate as their tissue N reserves
declined. Evidence is presented to show that the rate
3
of remobilization of N depends on the size and type of
the N pools within the plant, and that changes in their
rates of remobilization anduor transfer between pools
are the main factors influencing the form of the
relationship between RGR and N concentration.
Key words: Lettuce, N uptake, N reserves, nitrate N, organic N,
relative addition rate, relative growth rate.
Introduction
Studies of the interrelationship between N uptake, plant
N concentration and growth rate are central to an understanding of the role of N within plants. It is widely
accepted that improved information on the factors controlling the acquisition and utilization of N by crops will
help to identify the constraints to developing more effective strategies of N fertilization. These in turn could
increase the ef®ciency of N use to the bene®t of the environment. However, there is still some confusion about
the exact mechanisms controlling responses to N, which
appear to vary depending on the relative availability
of N from internal reserves and from external sources of
nitrate (Burns, 1994a).
Previous investigations of the effects of external N
supply on plant growth have focused largely on steadystate approaches, using techniques developed previously
(Ingestad and Lund, 1979; Ingestad, 1979a, b). In these,
plants in the exponential phase of growth are fed with
nitrate at a range of exponentially increasing rates, each
corresponding to a different constant relative addition
rate (RAR). After a period of equilibration, the relative
growth rate (RGR) adjusts to match the RAR, and the
To whom correspondence should be addressed at: Scottish Agricultural College, Bucksburn, Aberdeen AB21 9YA, UK. Fax: q44 1224 711293.
E-mail: [email protected]
ß Society for Experimental Biology 2001
310
Walker et al.
mean N concentration of the whole plants remains
constant. As a result, the total N contents of plants
forced to grow in this way increase linearly with their
weight (Ingestad and Kahr, 1985; Ingestad et al., 1986;
Oscarsson and Larsson, 1986; Freijsen and Otten, 1987).
This has often been interpreted as indicating that RGR is
directly controlled by the concentration of N within the
plant, and that there is little or no limitation to the rate
Ê gren, 1985).
at which N can be recycled between tissues (A
More recently, the responses of plants have been examined using non-steady-state techniques by measuring the
patterns of growth and plant N concentration following
interruption in external nitrate supply (Burns, 1994a, b).
These results showed that the relationship between RGR
and plant N concentration can be curvilinear, implying
that the rate of recycling of N within the plant declines as
the reserves become depleted. These conclusions contrast
with those from the steady-state approach. Although
the interpretation of these data was questioned recently
Ê gren, 1995), subsequent investigations
(WikstroÈm and A
have con®rmed the original observations (Burns et al.,
1997; Walker et al., 1997).
These results suggest that the main cause of the difference in the form of the relationship between growth rate
and plant N concentration lies in the way in which N is
partitioned or reallocated between tissues when plants are
forced to rely on either limited external or internal sources
of N. The objective of this paper is to test this hypothesis
by examining the effects of RAR and interruption
treatments on the organic and nitrate pools within the
plants.
Materials and methods
Plant culture
Two separate experiments were carried out in the same glasshouse using a ¯owing culture hydroponic system to compare the
effects of contrasting steady-state and non-steady-state N supply
regimes on the growth rate of lettuce (Lactuca sativa L. cv. Saladin
R100) plants. The ®rst experiment was conducted using the N
interruption method and the second using relative addition rate
techniques. Both experiments were conducted at the same time
of year: the interruption experiment in autumn 1996 and the relative addition rate experiment in autumn 1997. Average day and
night temperatures were maintained close to 25 8C and 10 8C
("5 8C), respectively, with the aid of supplementary heating,
and natural daylight was supplemented using 400 W Son-T
high-pressure sodium lamps to provide a 14 h day in order to
ensure that the growing conditions in the two experiments were
similar.
The lettuce plants were raised in perlite from pelleted seed
(supplied by Elsoms Seeds Ltd., Spalding, Lincolnshire,
England) and irrigated with a half-strength complete nutrient
solution for the ®rst 14 d. and a full strength solution thereafter
(Table 1). Each solution was adjusted to pH 6.0 with either
dilute Ca(OH)2or H2SO4 before use. At the four to ®ve leaf stage,
the plants were removed and any excess perlite washed from
Table 1. Concentration of major ions in the full-strength complete
nutrient solution
The solution also contained concentrations of B, Mn, Cu, and Mo
(as the sodium salt) at concentrations recommended by Hewitt
(Hewitt, 1966).
Ion
mmol l
NO3
PO4
SO4
Cl
EDTA
K
Ca
Mg
Na
Fe (III)
8.000
0.167
0.750
0.050
0.100
0.500
4.025
0.750
0.100
0.100
1
in the solution
their roots using de-ionized water, before they were transferred
to the hydroponic systems used in the two experiments.
Each hydroponic system consisted of separate units constructed from opaque grey polypropylene trays (60 3 40 3 15 cm
external dimensions) of approximately 33 l capacity. Each tray
was ®tted with a rest-on lid in which a matrix of sixty 2.5 cm
holes had been drilled. During the experiments, the lettuce plants
were suspended through these holes so that their roots were fully
submerged in the nutrient solution below. Each plant was
secured by a foam rubber plug that held the stem in an upright
position while allowing it to expand freely during growth. The
nutrient solution in each tray was aerated continuously.
Experimental design
N interruption experiment: The hydroponic system for the N
interruption experiment consisted of six separate hydroponic
units arranged in three randomized blocks with two treatments:
a control (qN) containing the same full-strength complete
nutrient solution as above (Table 1), or a zero-N nutrient
solution of identical composition except that the Ca(NO3)2
was replaced with an equivalent concentration of CaSO4.
Solution volumes and pH were adjusted every few days; with
their compositions monitored weekly for NO3 by HPLC, K by
¯ame photometer, P, Mg, Ca, SO4 -S, B, and trace elements
(Fe, Mn, Cu, and Zn) by inductively coupled plasma (ICP)
emission spectrophotometry. To avoid a major nutrient
imbalance occurring in the interim, the solutions were also
replaced every 7±14 d. Groups of ®ve randomly selected plants
were sampled destructively from each replicate 3 treatment
combination at transplanting and at intervals on 11 dates over
the subsequent 52 d.
Relative N additionl rate experiment:
In this experiment
there were a total of 16 individual hydroponic units arranged
in a four-by-four Latin Square design with four nutrient
treatments and four replicates. To facilitate imposition of the
nutrient treatments, the four trays in each treatment were
connected in parallel via input and output pipes allowing
nutrient solution to be pumped between them at a rate of
approximately 10 l min 1. This allowed the composition of the
solution in all units in the same treatment to be adjusted
virtually simultaneously.
The N treatments comprised four different RARs consisting
of 0.03, 0.06, 0.09, and 0.12 d 1. The treatments were imposed
on the same zero-N nutrient solution used in the N interruption experiments, to which daily supplements of nitrate
N supply, plant N concentration and growth
1
(DN mol plant ) were added (based on the methods of
Macduff et al., 1993). DN (mol plant 1) was calculated as:
DNˆNtq1
Nt ˆNt (eRAR
1)
where Ntq1 and Nt are the quantities of N in the plant after tq1
and t days, respectively, and RAR is the relative addition rate of
NO3 (d 1). Prior to the NO3 additions, the plants were maintained in zero-N solutions for 3 d to allow time for the accurate
measurement of tissue N contents needed for the calculation of
the initial nitrate addition rates for each treatment. This delay
was also useful in reducing the lag phase during which growth
rates in the 0.03 and 0.06 d 1 RAR treatments reached equilibrium. All additions of nitrate were made from a stock solution
of 1 M KNO3 at 09.30 h every day. K2SO4 was used to regulate
the K concentration of the solutions so that daily additions of K
to each treatment were identical.
The pH of all nutrient solutions were manually adjusted every
few days as above in order to maintain the pH at between 5.5
and 6.5. Solutions were monitored for nutrient composition
as for the interruption experiments, and changed completely
every 2 weeks to prevent imbalances in the ionic concentrations.
Plants were sampled at intervals over a period of 24 d from the
time the RAR treatments were imposed. At each sampling a total
of 18 plants (spread across all replicates) were randomly selected
for measurements of growth and N content.
Chemical analyses
Plants harvested on the same date from the same replicate were
pooled for each treatment. The fresh samples were divided into
root and shoot fractions and dried at a temperature of 80 8C for
a minimum of 2 d to determine their dry weight. Total plant
biomass was calculated as the sum of the two.
Organic N concentrations were measured on digests prepared
by treating 100 mg subsamples of the ®nely ground dry material
with 50 ml of a mixture of H2SO4 and H2O2 (containing 0.1%
Se) at 330 8C for 1.75 h using the indophenol blue method
(Crooke and Simpson, 1971). Nitrate concentrations were
determined on water extracts of the ground dry material
(50 ml : 100 mg) using HPLC (Hunt and Seymour, 1985). Total
N was calculated as the sum of these two nitrogen fractions.
Statistical analysis and estimation of RGR
Analysis of variance (ANOVA) was performed on the data in
order to provide a statistical comparison between the treatment
means within each experiment. In the RAR experiment, this
analysis was restricted to the data taken over the period from
day 9 to day 24 (i.e. immediately after the lag-phase of growth)
when the total N concentration of the plant tissues remained
essentially constant, and the growth of the plants appeared to
have reached a steady state.
Relative growth rates (RGR) of the plants in the N interruption experiment were estimated using the method of Schnute
(Schnute, 1981). This procedure is based on the principle of
growth acceleration, and uses analysis of variance of observed
dry weights to decide automatically which form of growth equation is the most appropriate. The procedure then ®ts the relevant
growth function to the data and calculates RGRs accordingly.
The method was implemented using a Genstat Procedure (Keen,
1991), and the model ®tted using a gamma function to compensate for increasing variability in the dry weights as the plants
grew. This procedure avoids the risk of introducing any bias in
the estimates of RGR resulting from an arbitrary choice of
Ê gren, 1995).
growth function (WikstroÈm and A
311
Comparative estimates of mean RGR of the plants grown
using the RAR technique were calculated using linear regression
analysis of log-transformed dry weight (loge) with respect to
time.
Results
Plant growth
Interruption experiment: The zero-N plants developed
paler green leaves and grew at an increasingly slower
rate than the control (qN) plants during the course of
the experiment. The roots on the zero-N plants also
tended to be proportionately larger with less branching
than those from the well-nourished control treatment
(data not shown).
Whole plant growth data for the two N treatments are
shown on a dry weight basis in Fig. 1a. These are plotted
on a logarithmic scale to compensate for the increase in
variances during the course of the experiment, allowing
least signi®cant differences to be displayed. These show
that differences in dry weights between the two treatments
became signi®cant 18 d after they were imposed. Corresponding treatment differences for the plant fresh weights
became signi®cant after only 8 d (data not shown). This
difference was due to a steady decline in the fresh to dry
weight ratio of the plants in the zero-N treatment resulting
from a fall in their tissue water contents (Fig. 1b).
The curves in Fig. 1a represent the ®t of the Schnute
model to the original growth data. In both cases, this
model identi®ed the Richards growth equation as giving
the best ®t to the data. Figures 1c and d plot predictions
of the changes in both the absolute (AGR) and relative
growth rates for the two treatments. These show that
the AGR of the zero-N treatment increased slowly for the
®rst 8±10 d of the experiment before declining steadily
thereafter (re¯ecting the steadily increasing depressive
effect of N de®ciency on growth), whereas the AGR of
the control (qN) treatment increased steadily throughout
most of the experiment. By comparison, there was
a consistent decline in RGR in both treatments, which
was particularly pronounced for zero-N plants.
RAR experiment: All plants started to show signs of N
de®ciency (pale-green leaves) during the initial 3 d period
when no N was supplied, and the symptoms persisted in
these tissues for the remainder of the experiment, even
at the highest RAR rate. However, after the N addition
rate treatments were established all newly produced
leaves were healthy and green, even for plants at the
lowest RAR.
Figure 2a shows that the log transformed dry weights
of the whole plants increased linearly with time in all
treatments (corresponding to exponential growth in each
case). The relative rate of dry matter accumulation
increased with RAR over the four treatments with the
312
Walker et al.
Fig. 1. (a) Whole plant dry weights (loge scale); (b) fresh:dry weight ratios; and (c) and (d) the relative growth rates (RGR) and absolute growth rates
(AGR) with respect to time from treatment imposition in the interruption experiment: In (a) and (b), (m) represents the control (qN) treatment and
(.) represents the zero-N treatment. (c) Data for the control treatment, with RGR and AGR represented by (m) and (k), respectively; (d) data for the
zero-N treatment, with RGR and AGR represented by (.) and (\), respectively. The curves were ®tted to the data by the Schnute model (Schnute,
1981). Error bars (P -0.05): LSD1 between N treatment means at the same harvest; LSD2 between harvest means for the same N treatment.
result that the highest RAR treatment produced the
largest plants by the end of the experiment. Estimated
mean RGRs for each of the treatments are summarized in
Table 2, with the regression lines plotted in Fig. 2a.
Corresponding data for the accumulation of plant fresh
weights again showed a similar pattern of response,
although differences between treatments were larger at
lower RARs (data not shown), where the effects on tissue
water content were greater (Fig. 2b).
Plant nitrogen concentration
Interruption experiment: Plant tissue concentrations of
both forms of N (nitrate N and organic N) and of total
N declined over the experimental period in both treatments (Fig. 3). However, in each case these reductions
were much more dramatic in the zero-N plants than
those in the well nourished control treatment, where N
concentrations declined only slowly with time.
Of the three N forms studied, nitrate N showed the
steepest decline in its tissue concentration, especially in
the zero-N plants, from which it had almost completely
disappeared by about day 8 (Fig. 3b). By comparison,
changes in organic N concentration of the zero-N plants
occurred much more slowly (Fig. 3a), causing the total N
concentration of these plants to decline at an intermediate
rate (Fig. 3c).
RAR experiment: The organic N, nitrate N and total N
concentrations in the whole plants all varied during the
®rst 8 d following the imposition of the relative N addition treatments, whilst growth gradually equilibrated.
However, once growth reached steady-state (from day 9
onwards), the concentrations remained essentially constant (Fig. 4), except possibly at the highest addition rate,
where concentrations appeared to increase slightly.
Figure 4a shows that the mean organic N concentration increased with RAR from less than 3% at 0.03 d 1 to
around 6% at 0.12 d 1. Tissue nitrate N concentrations
followed a similar pattern but, by comparison, were extremely low in all treatments (Fig. 4b), ranging from less
than 0.03% in the 0.03, 0.06 and 0.09 d 1 RAR treatments
to about 0.09% in the 0.12 d 1 treatment. These values
are tiny compared with the corresponding organic N data,
with the result that total N concentrations were almost
identical to those for organic N (Fig. 4c).
Relative growth rate versus plant N concentration
relationships
Interruption experiment: The relationships between the
RGR and the concentrations of the various N forms for
the two treatments are shown in Fig. 5. These appear to
be approximately linear for the well-nourished control
plants, whereas those for the zero-N treatment are
N supply, plant N concentration and growth
313
clearly curvilinear (or possibly bi-linear) for each of the
N forms. The latter are similar to those observed by
Burns earlier (Burns, 1994a, b) and suggests a two-phase
pattern to the response.
In each case, the shape of the graphs for the zero-N
plants showed a slow decline in RGR during the early
stages of N de®ciency (i.e. while the concentration of each
of the N forms was still high), with the slope of the curve
increasing markedly as the de®ciency became more acute.
The transition between the upper and lower `limbs' of the
curves occurred at about the fourth sampling (7±8 d after
imposing the treatments) and, in each case, coincided with
the date at which nitrate disappeared from the plant
tissues. Thereafter the decline in RGR was more rapid
Fig. 2. (a) Mean dry weights of the whole plant plotted on loge scale;
and (b) plant fresh:dry weight ratios with respect to time after start of
RAR. Symbols: (m) 0.03, (j) 0.06, (m) 0.09, and (.) 0.12 d 1 RAR
treatments. Error bars (P -0.05): LSD1 between treatment means at the
same harvest; LSD2 between harvest means for the same N treatment.
Fig. 3. Mean N concentrations as a percentage of whole plant dry
weight with respect to time from treatment imposition in the N
interruption experiment for (a) organic N; (b) nitrate N; and (c) total N;
(m) the control (qN) treatment, (.) zero-N treatment. Error bars
(P-0.05): LSD1 between N treatment means at the same harvest; LSD2
between harvest means at the same N treatment.
Table 2. Estimated RGRs of whole plants for each RAR treatment based on the slope of the linear regression of log weight against time
Estimated
RGR
Adjusted r2(%)
RAR treatment (d 1)
0.03
0.06
0.09
0.12
0.0372"0.0031
97.3
0.0615"0.0059
96.4
0.0786"0.0050
98.4
0.0970"0.0012
99.9
314
Walker et al.
Fig. 4. N concentration as a percentage of whole plant dry weight for
(a) organic N; (b) nitrate N; and (c) total N with respect to time after
imposition of RAR treatments: (m) 0.03, (j) 0.06, (m) 0.09, and (.)
0.12 d 1. Error bars (P -0.05): LSD1 between treatment means at the
same harvest; LSD2 between harvest means for the same N treatment.
until growth eventually stopped, at which point both the
organic N and total N concentrations were about 0.6%.
RAR experiment: Figure 6 shows the relationships for
both RAR and RGR against the corresponding total N
concentrations in the whole plants. Both graphs appear
to be linear but the average slope of the line for RGR is
less than that for RAR, mainly because of lower than
expected RGRs at the higher addition rates (Table 2).
This may have been an indication that growth in these
treatments was limited by reduced light levels caused by
self-shading (Hirose, 1988). This would be consistent with
the slight increase in plant nitrate concentration in the
larger plants at the highest RAR. Nevertheless, despite
these differences, the minimum concentration at which
growth occurs (i.e. the intercept on the abscissa in
Fig. 6b) was c. 0.8% which compares well with the
value of c. 0.6% from the interruption experiments.
Fig. 5. Relative growth rates with respect to (a) organic N; (b) nitrate N;
and (c) total N concentrations as a percentage of whole plant dry weight
in the interruption experiment: (m) control (qN) treatment; (.) zero-N
treatment.
Discussion
Plant growth
Interruption experiment: The coef®cients of variation for
the ®tted growth curves show the latter provided a good
description of the actual data (Fig. 1). The sigmoidal
shape of the Richards growth equation (selected for
both treatments by the Schnute ®tting procedure) is evident from the decline in AGR towards the end of the
experiment. Thus even in the control (qN) treatment,
plant growth rates started to decline by the ®nal harvest. As this was not due to N de®ciency, the most
N supply, plant N concentration and growth
315
tissues produced after the steady-state had been established showed no new signs of de®ciency. Others using
this approach have also shown that the effects on
existing tissues are irreversible once symptoms appear
(Oscarsson and Larsson, 1986; Macduff et al., 1993).
Although the RGRs of each treatment were constant
from about day 9, their values did not match the corresponding RARs in the two higher addition rate treatments. Because the plant N concentration data remained
relatively constant for each treatment, this implies that
these plants were unable to take up all of the nitrate provided on each day. However, it is not clear whether this
was caused by some physical limitation to nitrate in¯ux at
these higher rates of supply or whether some other factor
(e.g. self-shading) also restricted growth rate in these
faster growing plants, creating an associated reduction in
N demand.
Nitrogen concentration
Fig. 6. (a) Relative addition rates; and (b) relative growth rates with
respect to total N concentration as a percentage of whole plant dry
weight.
likely cause was from the result of self-shading, both of
the youngest leaves (as the plant began to heart up) and
of the older leaves (by the fully expanded middle ones),
much as has been observed previously (Hirose, 1988).
By comparison, the ®tted growth curve for the zero-N
treatment was more strongly sigmoidal in shape, with
the AGR declining earlier, and the ®nal weight of the
plants signi®cantly suppressed as a result of the acute N
de®ciency, compared with the control (qN) plants. This
was also re¯ected in the stunted appearance and pale
green leaves of these plants, caused by their need to distribute the limited amount of N present at the start
between an increasing number of leaves. N de®cient
plants are known to have higher speci®c leaf weights,
shortened vegetative growth stages and often a tendency
to mature earlier, compared with well-nourished plants
(Koch et al., 1988; Mengel and Kirkby, 1978).
RAR experiment: The data from this experiment
showed that growth had essentially reached steady state
approximately 9 d after the imposition of treatments.
Thereafter growth remained exponential (i.e. graphs of
log weight versus time were linear) for the rest of the
experiment (Fig. 2). In contrast to Ingestad (Ingestad,
1979a), N de®ciency symptoms which developed in existing leaves during the lag phase did not disappear over
time, even from the highest RAR treatment, although
Interruption experiment: The declines in tissue nitrate N,
organic N and total N concentrations with time were
most dramatic in the zero-N treatments (Fig. 3). As no
more N entered the zero-N plants from external sources,
this was due largely to the dilution effects of growth.
However, the rate of decline varied between the different N forms. The most rapid changes were in the concentration of nitrate which had almost completely
disappeared after about 8 d. This was caused by its conversion into organic forms of N, re¯ecting a reallocation
of N from storage to metabolic pools within the plant.
Once this nitrate had been completely assimilated, the
change in total N concentration per unit change in dry
weight decreased, much as had been observed earlier
(Burns, 1994a, b). By comparison, the much smaller
decline in tissue N concentrations in the control (qN)
plants was not caused by increasing N de®ciency, but
rather by an increase in the amount of structural
material (with an inherently lower N content) needed to
support the plants as they became larger (Caloin and
Yu, 1984).
RAR experiment: A major consequence of steady-state
growth (where RGR normally adjusts to match RAR)
is that the nutrient concentration in the plant remains
Ê gren, 1992;
constant (Ingestad, 1979a; Ingestad and A
Macduff et al., 1993; Hellgren and Ingestad, 1996). The
N data from this experiment (Fig. 4) con®rm that this
occurred in the three lower RAR treatments. In contrast,
the concentrations of all forms of N in the plants in the
0.12 d 1 RAR treatment increased marginally over the
same period. These plants were clearly unable to sustain
the exponential rate of growth needed to convert all of
the N taken up by the roots into new dry matter
(Ingestad, 1982). This may have been due to the combined
effects of self-shading in these larger plants (Hirose,
316
Walker et al.
1988), and to the increase in structural material as the
plants grew bigger (Caloin and Yu, 1984), as discussed
above.
Despite this, the nitrate N concentrations in all four
RAR treatments were generally extremely low. As nitrate
was the sole source of N to these plants, this implies that
once uptake had occurred, the nitrate was almost immediately reduced to organic N (presumably in the shoots),
avoiding any excessive accumulation in plant tissues. This
is consistent with the results of other authors (Mattsson
et al., 1991; Smolders and Merckx, 1992; Macduff et al.,
1993) who also found extremely low concentrations of
tissue nitrate at RARs below approximately 0.12 d 1 (the
highest RAR treatment used in the present study). The
small increases in tissue nitrate concentrations observed
at the highest RAR were presumably linked to luxury
consumption when supply exceeded demand. Even
greater accumulation of nitrate occurred when the plants
were fed on constant supra-optimal external nutrient
concentrations, as in the control (qN) treatments in the
N interruption experiments. The current results therefore
con®rm that lettuce plants growing under near steadystate conditions using RAR techniques do not accumulate
luxury amounts of nitrate at RARs below 0.12 d 1,
because virtually all of it is readily converted to organic
forms.
Growth and N relationships
Interruption experiment: Regression of RGR on the
plant N concentration data for the control (qN) plants
showed these relationships were approximately linear
(Fig. 5). This is in accordance with the results from ®eld
trials (Greenwood et al., 1990), from sand culture experiments (Dapoigny et al., 1996) and from well-nourished
control treatments (Burns et al., 1997; Walker et al.,
1997).
Corresponding regressions for the zero-N plants showed
conclusively that curvilinear (or possibly bi-linear) relationships occurred for all three N forms, with the two
`limbs' of each relationship corresponding to the situations when nitrate was either present (upper limb) or
absent (lower limb) within the plant. As soon as these
plants were subjected to the zero-N treatment, RGR
began to decline with respect to N concentration showing
that plant N could not be recycled quickly enough to
avoid reductions in growth, even in the early stages of
de®ciency. However, while nitrate N was still present in
the tissues, RGR declined more slowly than when it had
all been utilized. Thus as nitrate is reduced, the amount of
organic N (incorporated in the metabolically active plant
components) increases, ensuring that a relatively higher
assimilation rate is maintained until the nitrate reserves
are fully depleted. Thereafter, RGR decreases rapidly as
both the C and N use ef®ciencies of the plants decrease,
mostly due to dilution effects on the organic N concentration and to reductions in leaf area expansion rates.
RAR experiment: Both RAR and RGR increased linearly with plant N concentration once the plants had
reached steady-state (Fig. 6), much as has been observed
previously (Ingestad, 1979a). However, in these experiments the slope of the graph for RGR was less than
that for RAR, possibly suggesting that factors such as
self-shading may have restricted growth (see above).
These results are not atypical as RGRs lower than RARs
have also been found in similar experiments (Mattsson
et al., 1991; Stadt et al., 1992).
It has been argued that the linear nature of the relationship between RGR and plant N concentration is evidence that all N above a minimum concentration can be
freely recycled within the plants to help sustain growth in
Ê gren, 1985). Howtimes of restricted external supply (A
ever, this argument depends on the assumption that it is
plant N concentration which drives growth in these
experiments. In reality, once steady-state is obtained,
growth is driven by the rate of N uptake, which is (in
turn) controlled by the rate of addition of nitrate to the
external nutrient solution. Thus, all of the N taken up on
any one day is used for the production of new dry matter
(e.g. new leaves), irrespective of the prevailing average
N concentration in existing tissues, which remains
unchanged. Therefore, an increase in RAR will cause
a proportionate and parallel increase in both the concentration of N within the plant and in RGR. From this it
is clear that the actual concentration of N within the plant
is a consequence of the chosen RAR and not a direct cause
of the observed RGR in these treatments. This explains
why there can be a different form of relationship between
RGR and plant N concentration from that observed in
the interruption experiments, where growth is directly
controlled by the availability of N within the plant.
Conclusions
The relationship between RGR and N concentration of
lettuce plants can be either linear or curvilinear (possibly
bi-linear), depending on the availability of N from
external and internal sources. Where external N is available (albeit in restricted amounts), this relationship tends
to be linear, particularly under steady-state growing conditions, when growth rate is controlled by the rate of
external supply. Under these circumstances the amount
of N taken up by the plant increases linearly with its
dry weight, so that plant N concentration remains constant, with all of the nitrate converted into organic forms
of N virtually as soon as it is taken up and transported to
the shoots. However, when the external N supply runs
out, the plants are forced to re-utilize their existing N,
and this becomes increasingly dif®cult as the more
N supply, plant N concentration and growth
readily-available forms are utilized. Under these circumstances a curvilinear (or bi-linear) relationship between
RGR and plant N concentration predominates. The
shape of this relationship appears to depend on the
availability of nitrate in the plant tissues, because the rate
of decline of RGR with plant N concentration increases
dramatically once all of its nitrate has been assimilated.
This shows that the ef®ciency of N re-use by plants
decreases as their stored N becomes increasingly depleted,
and that reduction of tissue nitrate is more effective than
the recycling of organic forms of N for maintaining
growth when mineral N is no longer available for uptake.
Acknowledgements
The authors wish to thank Mr Andrew Mead for statistical
advice. The ®nancial support of the UK Ministry of Agriculture,
Fisheries and Food is gratefully acknowledged.
References
Ê gren G. 1985. Theory for the growth of plants derived from the
A
N productivity concept. Physiologia Plantarum 64, 17±28.
Burns IG. 1994a. Studies of the relationship between the growth
rate of young plants and their total-N concentration using
nutrient interruption techniques: theory and experiments.
Annals of Botany 74, 143±157.
Burns IG. 1994b. A mechanistic theory for the relationship
between growth rate and the concentration of nitrate-N and
organic-N in young plants derived from nutrient interruption
experiments. Annals of Botany 74, 159±172.
Burns IG, Walker RL, Moorby J. 1997. How do nutrients drive
growth? Plant and Soil 196, 321±325.
Caloin M, Yu O. 1984. Analysis of the time-course of change in
nitrogen content in Dactylis glomerata L. using a model of
plant growth. Annals of Botany 54, 69±76.
Crooke WM, Simpson WE. 1971. Determination of ammonium
in Kjeldahl digests of crops by an automated procedure.
Journal of the Science of Food and Agriculture 22, 9±10.
Dapoigny L, Robin P, Raynal-Lacroix C, Fleury A. 1996. Study
of the relationship between relative growth rate and nitrogen
content during lettuce (Lactuca sativa L.) growth: effects of
nutrient and radiation levels. Agronomie 16, 529±539.
Freijsen AHJ, Otten H. 1987. A comparison of the responses of
two Plantago species to nitrate availability in culture experiments with exponential addition. Oecologia 74, 389±395.
Greenwood DJ, Stone DA, Draycott A. 1990. Weather, nitrogensupply and growth of ®eld vegetables. Plant and Soil 124,
297±301.
Hellgren O, Ingestad T. 1996. A comparison between methods
used to control nutrient supply. Journal of Experimental Botany
47, 117±122.
Hewitt EJ. 1966. Sand and water culture methods used in the
study of plant nutrition. Technical communication No 22.
Farnham Royal: Commonwealth Bureau of Horticulture and
Plantation Crops, 421±432.
Hirose T. 1988. Modelling the relative growth rate as a function of
plant nitrogen concentration. Physiologia Plantarum 72,
185±189.
317
Hunt J, Seymour DJ. 1985. Method for measuring nitratenitrogen in vegetables using anion-exchange high-performance
liquid chromatography. Analyst 110, 131±133.
Ingestad T. 1979a. N stress in birch seedlings. II. N, K, P, Ca
and Mg nutrition. Physiologia Plantarum 45, 149±157.
Ingestad T. 1979b. A de®nition of optimum nutrient requirements in birch seedlings. III. In¯uence of pH and temperature
of nutrient solution. Physiologia Plantarum 46, 31±35.
Ingestad T. 1982. Relative addition rate and external concentration: driving variables used in plant nutrition research. Plant,
Cell and Environment 5, 443± 453.
Ê gren GI. 1992. Theories and methods on plant
Ingestad T, A
nutrition and growth. Physiologia Plantarum 84, 177±184.
Ingestad T, Arveby AS, KaÈhr M. 1986. The in¯uence of ectomycorrhiza on N nutrition and growth of Pinus sylvestris.
Physiolgia Plantarum 68, 575±582.
Ingestad T, KaÈhr M. 1985. Nutrition and growth of coniferous
seedlings at varied relative addition rate. Physiologia
Plantarum 65, 109±116.
Ingestad T, Lund A-B. 1979. Nitrogen stress in birch seedlings.
I. Growth technique and growth. Physiologia Plantarum 52,
454±466.
Keen A. 1991. Procedure FITSCHNUTE. In: Payne RW,
Arnold GM, Morgan GW, eds. GENSTAT 5 procedure library
manual release 2[2]. Oxford, UK: NAG Ltd., 104±108.
Koch GW, Schulze E-D, Percival F, Mooney HA, Chu C. 1988.
The N balance of Raphanus sativus 3 Raphanistrum plants. II
Growth, N redistribution and photosynthesis under NO3
deprivation. Plant, Cell and Environment 11, 755±767.
Macduff JH, Jarvis SC, Larsson C-M, Oscarson P. 1993. Plant
growth in relation to the supply and uptake of NO3 : a
comparison between relative addition rate and external
concentration as driving variables. Journal of Experimental
Botany 44, 1475±1484.
Mattsson M, Johansson E, Lundborg T, Larsson M, Larsson C-M.
1991. Nitrogen utilization in N-limited barley during vegetative and generative growth. I. Growth and nitrate uptake
kinetics in vegetative cultures grown at different relative
addition rates of nitrate-N. Journal of Experimental Botany
42, 197±205.
Mengel K, Kirkby EA. 1978. N fertiliser application and crop
production. In: Principles of plant nutrition. Berne:
International Potash Institute, 318±328.
Oscarsson P, Larsson C-M. 1986. Relations between uptake and
utilization of NO3 in Pisum growing exponentially under
nitrogen limitation. Physiologia Plantarum 67, 109±117.
Schnute J. 1981. A versatile growth model with statistically stable
parameters. Canadian Journal of Fisheries and Aquatic Science
38, 1128±1140.
Smolders E, Merckx R. 1992. Growth and shoot : root partitioning of spinach plants as affected by nitrogen supply.
Plant, Cell and Environment 15, 795±807.
Stadt KJ, Taylor GJ, Dale MRT. 1992. Control of relative
growth rate by application of the relative addition rate
technique to a traditional solution culture system. Plant and
Soil 142, 113±122.
Walker RL, Burns IG, Moorby J. 1997. Variations in the form of
relative growth rate (RGR) and nitrogen concentration
relationships. Fertilizer for sustainable plant production and
soil fertility, Vol. II. Proceedings of the 11th International
World Fertilizer Congress, September 7±13, 1997, Gent,
Belgium, 382±388.
Ê gren GI. 1995. The relationship between the
WikstroÈm F, A
growth rate of young plants and their total-N concentration is
unique and simple: a comment. Annals of Botany 75, 541±544.