Structure and Function in Agroecosystem Design and Management - Chapter 18 doc

377 The Effects of FACE on the Growth and Yield of Paddy Rice.. Objectives of the Rice FACE Experiment In view of the importance of rice in the lives of a large proportion of theworld’s

CHAPTER 18

Growth and Yield of Paddy Rice

Kazuhiko Kobayashi, Mark Lieffering, and Han Yong Kim

CONTENTS

Introduction 371

Atmospheric CO2and Rice 371

Objectives of the Rice FACE Experiment 373

Growing Crops under Elevated [CO2] 374

Chamber Studies 374

FACE Systems 374

Rice FACE System: Description 375

Ring Description 376

CO2Control and Monitoring 376

Temporal and Spatial Control of [CO2] (1999) 377

The Effects of FACE on the Growth and Yield of Paddy Rice 378

Materials and Methods 378

Results 380

Discussion 387

Conclusions and Implications 390

References 392

INTRODUCTION

Atmospheric CO 2 and Rice

It is estimated that up until the industrial revolution in the eighteenth century, atmospheric CO2concentrations ([CO2]) were about 280 ppmV (parts

371

0-8493-0904-2/01/$0.00+$.50

per million by volume) Since then, the [CO2] has risen to 370 ppmV at ent and is expected to keep increasing at a rate of about 15 ppmV per decade.The increase in [CO2] is attributed to human activities such as fossil fuel burn-ing and deforestation (Houghton et al., 1996) It is predicted that the increasewill continue into the twenty-first century, resulting in a [CO2] concentrationsomewhere between 450 and 550 ppmV around the year 2050 (Houghton

pres-et al., 1996) Because CO2is a “greenhouse” gas, the increase in [CO2] is dicted to affect the global radiation energy balance and thereby climate Thepredicted changes in climate most notably include an increase in the Earth’smean surface temperature and alterations in rainfall patterns, both factorswhich strongly affect biomass production in both agricultural and naturalecosystems worldwide (Reilly, 1996)

pre-Besides the indirect effects on plant growth induced by climate change,elevated [CO2] can also directly alter plant processes, most importantly pho-tosynthesis and stomatal conductance Because photosynthesis in plants uti-lizing the C3 pathway is limited by current [CO2] levels, elevating [CO2]increases rates of carbon (C) fixation, leading to greater plant biomass pro-duction (Drake et al., 1997) Elevated [CO2] also tends to reduce stomatal con-ductance which, coupled with the increase in photosynthesis, leads to anincrease in water use efficiency

In terms of both area and tonnage harvested, rice, oryza sativa, h, is the

primary crop in Asia and is among the world’s three major crops (the other

two are wheat, Triticum aestivum L., and maize, Zea mays L.) Rice is unique in

that 95% of the world’s total production occurs in developing countries, andthe majority of that grown is consumed locally (Alexandratos, 1995) In most

of the countries where it is produced, rice provides a major part of the humandietary needs, and its production is usually a large factor in the economy.Rice production in Asia has increased almost linearly since 1960 and hadrisen by 150% by 1995 (FAOSTAT; http://apps.fao.org/) The harvested areahas increased by only 20%, hence the increased production has mostly comefrom a 100% increase in yield per unit harvested area This large yieldincrease can be ascribed to technological advances such as the breeding ofnew, high-yielding varieties, the development and expansion of irrigationsystems, increased fertilizer use and efficiency, and improved pest manage-ment (Greenland, 1997)

It has been estimated that in the next 30 years the growing population inAsia may need nearly 70% more rice (Hossain, 1997) Because the area avail-able for cultivation is predicted to decrease, yield per unit harvested areamust increase more than the growth in population However, there is evi-dence that the impressive yield increases since 1960 may be plateauing(Cassman et al., 1997), and there appears to have been little increase in poten-tial crop yields in recent times (Khush and Peng, 1996) Therefore, it is spec-ulated that further increases in yield may be achieved only by optimizing thesupply of resources limiting crop growth, such as water and nitrogen (N)(Sinclair, 1998a)

The effects of elevated [CO2] on rice growth have been studied since the1960s (e.g., Murata, 1962) In these early experiments, higher [CO2] was shown to enhance both biomass growth (Imai and Murata, 1976) and yield(Yoshida, 1973) It was also found that environmental variables such as N(Imai and Murata, 1978) and temperature (Imai and Murata, 1979) can affectgrowth enhancement due to higher [CO2] In these studies, plants were grown under higher [CO2] for only a portion of the growth duration It waslater confirmed that rice yield also increases when plants are grown underhigher [CO2] throughout the growth duration (Imai et al., 1985; Baker et al.,1990; Ziska and Teramura, 1992; Baker and Allen, 1993a, b; Seneweera et al.,1994; Kim et al., 1996a,b; Ziska et al., 1997; Moya et al., 1998).

The studies cited above have identified some common factors whichresult in the increase in yield with elevated [CO2] Individual leaf area and thenumber of leaves per stem are usually decreased but a greater tiller numberresults in an increase in leaf area per plant (Imai, 1995) Photosynthesis perunit leaf area is usually increased with elevated [CO2], though rates maydecrease as the leaf matures (photosynthetic acclimation) (Imai and Murata,1978b) The net result is an increase in photosynthesis per plant, resulting ingreater carbohydrate accumulation and dry matter production (Rowland-Bamford et al., 1990; Baker et al., 1993) Frequently, the increase in root dryweight (d.wt) with elevated [CO2] is greater than the increase in shoot d.wt(Imai et al., 1985) The greater tiller number leads to an increase in the pro-duction of panicles, an important determinant of grain yield (e.g., Ziska et al.,1997) Increased carbohydrate supply leads to an increase in both grain num-ber per panicle and the percentage of mature grains that develop (Yoshida,1981) Elevated [CO2] rarely increases individual grain weight because of thephysical limitations imposed by the grain and husk characteristics (Yoshida,1981)

Objectives of the Rice FACE Experiment

In view of the importance of rice in the lives of a large proportion of theworld’s population and the anticipated decreases in per capita yield, there is

a need to determine the effects of the predicted elevated [CO2] on rice growthand yield under field conditions An important question is by how much willelevated [CO2] increase rice yields under field conditions and to what extentwill these increases satisfy the predicted demand? Also, will there be interac-tions between elevated [CO2] and the other factors that limit rice yields, and

if so how can these be utilized to maximize yields? In this chapter we brieflyreview the techniques that have been used in past research efforts on the effects of elevated [CO2] on rice growth and yield All these studies havegrown rice in some kind of enclosure fumigated with air containing elevated[CO2] We highlight some of the drawbacks in using enclosures to growplants and then introduce the free-air CO enhancement (FACE) technique as

a method to grow large areas of crops under elevated [CO2] We then presentsome results from the first experiment to grow rice using the FACE techniqueand discuss their implications.

GROWING CROPS UNDER ELEVATED [CO 2 ]

Chamber Studies

Much information on the response of rice to elevated [CO2] has come fromexperiments conducted using chambers or enclosures which were fumigatedwith either ambient or CO2enriched air The types of enclosures that havebeen used include temperature gradient chambers (TGCs); (Kim, 1996a,b),soil-plant-air research (SPAR) units (e.g., Baker et al., 1992; Gesch et al., 1998)and open-topped chambers (OTCs) (e.g., Moya et al., 1998) However, to iso-late the effects of elevated [CO2] on plant growth, it is important that theexperimental system imparts minimal effects on other abiotic environmentalparameters that may influence growth In many of the early experiments con-ducted in fumigated glasshouses (e.g., Imai et al., 1985), plants were grown

in pots The soil environment in pots differs markedly from that under fieldconditions, with differences in factors such as nutrient availability, waterdrainage, and soil temperature In fact, the response of plants to elevated[CO2] has been shown to decrease with decreasing pot size (Arp, 1991).Chambers and enclosures can affect abiotic environmental factors such

as temperature, solar radiation, humidity, and wind (McLeod and Long,1999) Frequently, compared to outside conditions, within the chamber there

is less light, the air is drier, and temperatures are higher These differences canaffect plant growth (commonly called a “chamber effect”) to as large anextent as the effect of the elevated [CO2] (e.g., Knapp et al., 1994) The cham-ber effect can influence many aspects of the response of plants and crops toelevated [CO2], including photosynthesis, metabolism, biomass production,and crop water and energy balances (McLeod and Long, 1999), making thetranslation of results to outside conditions difficult For example, Van Oijen

et al., (1999) found that the response of wheat grain yield to elevated [CO2]was less in OTCs cooled to very close to the ambient temperature compared

to uncooled OTCs However, the yield of plants in the cooled OTCs was stilldifferent from those grown outside, suggesting that abiotic factors other thantemperature also contributed to the chamber effect

FACE Systems

To overcome the limitations of chamber methods, the FACE method wasdeveloped in the mid 1980s (Lewin et al., 1994) The first full scale field exper-iment was established at Maricopa (Arizona, U.S.) using cotton as the crop

(Nagy et al., 1994) Generally, FACE systems involve fumigating a circulararea of vegetation with pure CO2or CO2/air mixtures, thereby generating azone having [CO2] higher than that of the surrounding ambient atmosphere.The CO2is usually emitted from a structure (sometimes referred to as a ring)constructed from pipes or tubes that surrounds the crop The CO2is emittedfrom the upwind direction of the ring, relying on the wind to mix and dis-perse it over the whole ring The target [CO2] in the fumigated zone may beeither static (e.g., a constant 500 ppmV) or dynamic, whereby the target is set

at a certain level (e.g., 200 ppmV) above the real-time ambient [CO2] A trol system regulates the amounts of CO2emitted by monitoring and inte-grating wind speed and direction together with [CO2] levels at ring center Thesystem must be able to deal with short-term changes in the weather, mostnotably differences in wind speed and direction, both of which may changeover very short periods of time The control system must also be able to copewith longer term temporal variations in [CO2], which may be caused by fac-tors such as diurnal and seasonal changes in the relative amounts of crop photosynthesis and respiration

con-The FACE method has been successfully used to study the effects of vated [CO2] on a variety of vegetation types These include agriculturallyimportant crops such as cotton (Lewin et al., 1994), wheat (Kimball et al.,1995), and pastures (Hebeisen et al., 1997), as well as harvestable tree species(Hendrey et al., 1999) FACE has also been used in more natural vegetationtypes such as desert vegetation (Jordan et al., 1999)

ele-The most important advantage of FACE systems over other methods ofgrowing vegetation under elevated [CO2] is that the vegetation is not undulyinfluenced by the effects of enclosures on environmental factors such as solarradiation, temperature, and wind (McLeod and Long, 1999) Also, relativelylarge areas of vegetation can be treated, meaning that a large number of sam-ples can be collected for analyses and a range of experiments can be con-ducted in one season The major disadvantage of FACE systems is theirrelatively high cost, both to build and run, the latter due primarily to the largeamounts of CO2required for fumigation However, expressed on the basis ofcost per usable fumigated crop area, FACE systems can be more cost effectivethan other methods of growing plants under elevated CO2(Kimball, 1992)

Rice FACE System: Description

The Rice FACE project was established in 1996 to study the effects of vated [CO2] concentrations on rice crop growth, yield, and ecosystemprocesses It is the first FACE experiment to be conducted on rice Afterdesign trials in 1997, a facility consisting of four FACE rings and their associ-ated ambient (control) plots was constructed for use during the 1998–2000rice growing seasons

ele-Ring Description

Each Rice FACE ring consists of a CO2emission structure, a CO2toring system, and a computerized control system In order to minimizeatmospheric contamination of the control plots, there is at least 90 m betweenthe controls and the nearest ring Each emission structure consists of a 12-mdiameter octagon made of eight 5-m long, 3.8-cm diameter polyethylenetubes Each tube is horizontally supported by a 5-m long, 2.2-cm diametergalvanized steel pipe, which is supported at each end by similarly sized,upright pipes dug 40 cm into the soil The polyethylene tubes have 0.6–0.9-

moni-mm diameter CO2-release holes located approximately every 4 cm on the sidefacing into the crop The height of the emission tubes above ground level isset at approximately 50 cm above the canopy Liquid CO2contained in a hold-ing tank passes through a vaporizer, and the CO2gas is delivered to the emis-sion tubes via valves to the emission tubes Pure CO2at a maximum pressure

of 0.13 MPa is “sprayed” from the tubes; preliminary simulation studies haveshown that, depending on wind speed and emission pressure, concentrationsdrop from 100% to 2000 ppmV within 20 cm of the emission tube (M.Yoshimoto, pers comm.) The use of pure CO2in the Rice FACE experiment

is different from that used in many other FACE designs, which emit a

CO2/air mixture into the ring using blowers Under some circumstances thiscan influence the microclimate within the FACE ring (“blower effects”)(McLeod and Long, 1999), and the control plots must have blowers installed

to cancel out the blower effects There is no such problem with the pure-CO2FACE

The total area within each FACE ring is approximately 120 m2.Walkways, situated approximately 15 cm above the paddy water level,extend from one of the surrounding earth dikes to the ring center and pro-vided access to the crop and monitoring equipment Preliminary studiesindicate that canopy microclimate such as wind and canopy temperature donot appear to be affected by the presence of the ring structures (M.Yoshimoto, personal communication)

The main objective of the Rice FACE experiment is to determine the ence of elevated [CO2] on various crop and ecosystem processes It is there-fore crucial to have control over the amounts of CO2applied and also to knowwith confidence what the level of [CO2] is at any time and location within thering over the duration of the experiment

influ-Because a dynamic target (200 ppmV above ambient) is being used, bothambient and ring [CO2] levels must be monitored Ambient [CO2] concentra-tions are measured at the center of the two distal control plots using infrared

CO2 gas analyzers [CO2] in the FACE rings is monitored at ring center,together with wind speed and direction, which are measured every second

Table 18.1 Mean CO 2 above ambient (target  200 ppmV) at ring center

and 2.5 m (average of 4 locations) and 5 m (average of 8 locations) from the center for the 4 Rice FACE rings from May 21 until August 20, 1999 Ambient [CO 2 ] concentration during the same time period was 391.1 ppmV.

CO 2 concentrations above ambient (ppmV)

Because a number of different experiments are conducted in various plots within each FACE ring, it is important to know what the [CO2] levels are

sub-at these sites over the season For each ring a separsub-ate infrared CO2analyzersamples the atmosphere at canopy height at 13 locations Sampling tubes arelocated at the center and equidistantly spaced in two concentric circles 2.5 m(4 locations) and 5 m (8 locations) from the center [CO2] levels at any locationwithin the ring can be estimated by interpolating the actual [CO2] at each ofthe sampling locations

The ability of the FACE system to control [CO2] can be assessed by paring the actual and target [CO2] at any location for a given time period.Performance can be expressed as the average [CO2] concentration aboveambient for the time period, or the percentage of the time that all the actualvalues were within 10 and 20% of the target can be calculated During the firsthalf the 1999 season (up to the time of writing), [CO2] levels at ring center were between 185 and 213 ppmV above ambient (Table 18.1) and about 55 and 85% of the samples were within 10 and 20% of the target, respectively(data not shown) Within 2.5 m of the center, [CO2] averaged 220 ppmV aboveambient, and around 50% of the samples were within 10% of the target At 5

com-m frocom-mthe center, [CO2] averaged 280 ppmV above ambient, and only about30% of the samples were within 10% of the target (A different control algo-rithm was used in 1998 which resulted in less satisfactory performance, with

CO2levels averaging 224 and 340 ppmV above ambient at the center and lying yield plots respectively.)

out-CO2levels at 2.5 and 5 m from the center were around 3.5 and 13% greaterthan at the center (Table 18.1), resulting in a “bowl shaped” [CO2] distribu-tion pattern CO2is released from the peripheries of the ring and dispersedtowards the center, and, as long as wind speeds and directions are evenly dis-tributed over the season, such a distribution pattern is typical for FACE rings.The size and shape of the [CO2] gradient from the ring edge to the center willdepend on factors such as ring architecture, the force of CO2emission, windspeed, and the control algorithm used

The CO2control performance of the Rice FACE in terms of the age of observations that were within 10% of the target at ring center was not

percent-as good percent-as those reported for other FACE systems of similar size For ple, for the Maricopa FACE experiment, [CO2] levels at ring center were within10% of the target 90% of the time (Nagy et al., 1994), compared to only 55% for the Rice FACE experiment This difference in performance can be partiallyattributed to differences in the wind characteristics of the two sites Greateraverage wind speeds result in better CO2 distribution and mixing At theMaricopa site, average daily wind speeds were about 1.7 ms1(Nagy et al.,1994) with calm periods ( 0.4 ms1) occurring about 19% of the fumigationtime (Nagy et al., 1992) In contrast, at the Rice FACE site, average daily windspeed ranged from 1.1 ms1in June to 0.5 ms1in September (season average

exam-of 0.7 ms1), while calm ( 0.3 ms1) periods ranged from 30% of the time early in the season to nearly 60% near the end (season average 45%) This lower average wind speed and greater calm percentage makes effective tem-poral control and uniform spatial distribution difficult and is probably a major reason for the differences in CO2performance between the Rice FACEand other FACE experiments

THE EFFECTS OF FACE ON THE GROWTH AND YIELD

OF PADDY RICE

Materials and Methods

a Site description The Rice FACE experiment is located at Shizukuishi,

Iwate Prefecture, in the northern part of Honshu, Japan (39° 38’ N, 140° 57’ E)

It is situated in a valley at an altitude of about 200 m, surrounded by 600-mhigh hills to the south, west, and north The site was chosen because it is typ-ical of the agroenvironment that grows a large proportion of the Japanese ricecrop It is also close to existing research facilities at the Tohoku NationalAgricultural Experiment Station near Morioka The climate is best described

as humid continental with a summer precipitation maximum and a cold, drywinter Over the year, daily average air temperatures range from 2.5(January) to 23.2°C (August); meteorological data from the 1998 growing

Table 18.2 Meteorological profiles of the Rice FACE site, 1998

Air temperature a Solar radiation Rainfall c

mean min max mean daily b

monthly Month (°C) (MJ m 2 d 1 ) (mm)

a Monthly average of the daily mean, minimum, and maximum air temperatures.

b Monthly average of the daily mean solar radiation.

c Monthly accumulated rainfall.

d For last 10 days of the month only.

e Season accumulated rainfall.

season is shown in Table 18.2 The soils of the site are derived from volcanicash and have been tentatively classified as humic Andosols

b Experimental design In both 1998 and 1999, the experiment was a

completely randomized block design with two levels of [CO2] (ambient [CO2](control) and elevated [CO2] within the FACE rings) replicated four times.FACE and control plots were located in eight paddies blocked by location; thefour blocks consisted of paddies with similar agronomic histories and soilcharacteristics

c Seedling establishment In both years, presoaked seeds of rice cv.

Akitakomachi (a commonly grown variety in northern Japan) were sown intoseedling trays and grown under flooded conditions Trays were placed inplastic chambers fumigated either with air containing ambient or elevated(200 ppmV) [CO2] The duration of seedling growth was 14 and 23 days ataverage air temperatures of 19.35°C and 18.25°C in 1998 and 1999, respec-tively

d Crop establishment and management Seedlings were

hand-trans-planted into either control or FACE plots on 21 May 1998 and 20 May 1999.Although most Japanese farmers use mechanical transplanters in establish-ing rice crops, hand transplanting was used in the experiment to ensure aneven number of seedlings per hill and regular hill spacing In both years,there were three seedlings per hill and 17.5 and 30 cm between hills and rows,respectively ( 19 hills m2) This spacing is commonly used by farmers inthis district Three levels of N were supplied as ammonium sulfate: 4g (low),8g (medium), and 12g N m2(high) in 1998, and 4, 9, and 15 g N m2in 1999.The medium N level is typical of the standard rate used by local farmers Inboth years N was applied as a basal dressing (63% of the total), at mid-tiller-ing (20%) and at panicle initiation (17%) Levels of phosphorus and potas-sium fertilizer were similar for all N levels and adequate for the high N

treatment Flooded paddy fields were maintained throughout the seasonexcept for a midsummer drainage conducted in mid-July in both years andfrom 10 days prior to harvest in 1998 Herbicides, insecticides, and fungicideswere applied when necessary.

e Sampling and harvesting To determine the influence of elevated

[CO2] on vegetative growth, in both years seedlings were sampled on the day

of transplanting and established plants were sampled from the medium Ntreatment of FACE and control plots at 25, 53, 81, 109, and 131 days aftertransplanting (DAT) from three locations in 1998 In addition, plants in thehigh N plots were harvested at 83 and 137 DAT, while low N plants were onlyharvested at grain maturity Plants were separated into living and dead leafblades, stems (including leaf sheath), panicles (when present), and roots;d.wt of the plant parts as determined separately The number of tillers andpanicles (when present) as determined and leaf area was measured At finalharvest, the number of spikelets per panicle was also determined To deter-mine crop N uptake, the dried plant parts were milled and total N in eachpart was determined (micro-Kjeldahl technique)

In order to determine flowering date, two or three locations within each[CO2] plot were investigated daily for panicle appearance in 1999, but onlyonce in 1998 Flowering date was defined as when panicles had emergedfrom 50% of the effective tillers (potential panicle bearing) The effect ofFACE on grain maturity was investigated by checking the color of the pani-cles by eye during grain filling in 1998 The date of maturity was defined by

a “yellow index” in which maturity was defined as when 90% of the panicles

at a location had greater than 80% yellow grains

For grain yield determination, subplots were set aside within both FACEand control plots and not disturbed until final harvest Plants were sampled

at grain maturity; total and fertile spikelet number per hill together withmean grain weight were determined Mean [CO2] (four replicates) in 1998 forthese grain yield plots over the season were 726 and 387 ppmV for FACE andcontrol, respectively

In this chapter we present seedling and phenological data from bothyears, but only 1998 data for crop dry matter and grain yield investigations

Results

a Seedling growth When rice crops are established by transplanting,

early seedling growth and vigor under nursery conditions are important tors in the successful establishment and eventual yield of the crops However,there is little information on the effects of elevated [CO2] on the growth of riceseedlings cultivated for transplanting using commercial agricultural condi-tions and techniques In both years, elevated [CO2] increased total and rootd.wt (Table 18.3) In 1998, leaf blade d.wt increased with elevated [CO2], whileleaf area decreased, leading to an increase in specific leaf weight (leaf d.wt

fac-Table 18.3 The effect of ambient (AMB) and elevated (ELEV)CO 2 on the

characteristics of rice seedlings used in the Rice FACE experiment in 1998 and 1999 Average air temperature and growth duration are also shown “ ,”“*”, and “**” denote significance at

the p  0.1, 0.05, and 0.01, respectively “ns” denotes not

significant.

Dry weight

LA LB CLS R T SLW Temp D Year CO 2 cm 2

hill 1 mg hill 1 mg cm 2 °C days

b Crop vegetative growth Rice crop growth consists of a vegetative

phase followed by a reproductive stage The former entails the growth ofmainstem and tiller leaves; these combine to form the crop canopy wherephotosynthesis occurs The vegetative stage is important in determininggrain yield because the number of panicles at harvest is closely related to thenumber of tillers that are produced Also, photosynthate accumulated duringthe vegetative stage can provide up to 40% of the material used for grain fill-ing (Yoshida, 1981)

Tillers are important in determining final yield of rice in two ways: theycontribute to the extent of the canopy (and hence the level of canopy photo-synthesis), and they bear panicles In the 1998 Rice FACE experiment, at allbut the first sampling (25 DAT), FACE increased tiller number by 10 to over20%, with the largest increases at around panicle initiation (Figure 18.1) Thelack of response for the first sampling may have been because CO2fumiga-tion did not commence until 10 days after transplanting Green leaf area ofthe crop was calculated by measuring individual hill green leaf area at eachsampling and multiplying this by the plant population At panicle initiation(approximately 53 DAT), FACE increased green leaf area index (LAI) by 10%(Figure 18.2) However, by flowering (81 DAT), when LAI had peaked ataround 4.0, there was no CO2effect

For rice crops in general, green leaf area reaches a peak at around ering and then decreases gradually as materials are translocated to be usedfor grain filling, and the leaves become senescent At all harvests after pani-cle initiation, FACE plants had far more dead leaf d.wt than control plants

flow-Figure 18.1 The effect of ambient [CO2] (control) and free-air CO2 enrichment

(FACE) on tiller number during the season Plant population ≈ 19 m 2 Error bars are 1 standard error of the mean.“ *”, “ **” denote significance

at the p  0.05 and 0.01 levels, respectively “ns” denotes not significant.

(data not shown; Kobayashi et al., 1999), suggesting a speeding up of leafdevelopment and senescence with FACE Under FACE, green leaf areadeclined more rapidly during grain filling compared to control, but the dif-ference was not statistically different at P 0.05 (Figure 18.2)

Total crop biomass was greater in FACE-grown plants compared to trol plants at all harvests except the first one (again, possibly due to CO2fumi-gation commencing only 10 days after transplanting) (Figure 18.2) Thegreatest d.wt response to FACE was about 20% at panicle initiation For allharvests, the d.wt of most plant parts increased with FACE, including that ofthe roots (data not shown) At all harvests the crop biomass response to FACEwas greater than that of leaf area This suggests that the increase in biomasswas due to greater crop radiation use efficiency rather than an increase inlight interception

con-e Crop reproductive growth The crop reproductive phase comprises

panicle initiation, development and heading, followed by flowering, grainfilling and finally grain maturity Grains are composed mainly of carbohy-drates which are derived from two sources: those stored in the vegetativeparts before flowering, and those produced after flowering The contribution