| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. de Producción Vegetal, Fac. de Agronoma, Univ. de Buenos Aires, Av. San Martín 4453, Buenos Aires (C1417DSE), Argentina
b Dep. of Agronomy, Iowa State Univ., 1563 Agronomy Hall, Ames, IA 50011-1010
* Corresponding author (otegui{at}agro.uba.ar)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: An, pollen-shedding plants per square meter on Day n after anthesis • ASI, anthesis-silking interval • ASIip, ASI of individual plants • ASIpp, ASI of the plant population • DAA, days after anthesis • En, ear number n • PGMn, pollen grains per square meter on Day n
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Though pollen production does not limit kernel set (Bassetti and Westgate, 1994; Otegui et al., 1995), the amount of pollen produced per plant could become a limiting factor for kernel number if the reduction in tassel size persists. Pollen production could be particularly important in certain specific production systems, like the seed industry and the high-oil maize, where only a small proportion of plants (usually less than 20%) are used as pollinators. In these situations, knowledge of pollen production dynamics becomes essential for assessing the proportion of pollinating plants in the population needed for maximum kernel set.
Pollen quantification in field conditions is not easy in maize, with its airborne pollen, and few data are available. Hall et al. (1982) described pollen production of plants grown in pots under different water regimes. In their experiment, tassels were bagged for pollen collection, and a subsample was taken for quantifying the number of pollen grains. In a more recent attempt to quantify pollen production, Struik and Makonnen (1992) excised the tassels of plants in the field, and grew them on water in a greenhouse at 18/12°C day/night temperature. They also bagged the tassels and collected pollen every 2 d. The amount of pollen was expressed in terms of weight, with no reference to the number of pollen grains per unit area or per plant. The main problem in both studies is that tassels were subjected to traumatic (e.g., bagging, cutting) or artificial (e.g., constant temperatures) conditions, which could have decreased pollen production relative to the natural field environment, raising concerns as to the value of the data for predictive purposes (e.g., modeling). Basetti and Westgate (1994), alternatively, used pollen traps of the type described by Sadras et al. (1985) for pollen collection. This method does not affect normal tassel development and gives information on pollen availability per unit land area. The main restriction of Bassetti and Westgate's (1994) data is that the only source of variation in pollen production was related to sowing dates, and differences may also be introduced by genotypes (Hall et al., 1982; Struik et al., 1986) and plant densities (Struik and Makonnen, 1992).
In this paper, we characterized pollen production per unit land area of several maize hybrids, grown at contrasting plant densities in two distinct environments (cool and temperate climates at 45°35' N and 34°33' S, respectively). Pollen traps were used to avoid the possible confounding effects of tassel manipulation on pollen production. Our objectives were (i) to estimate the threshold pollen availability per exposed silk that did not impair kernel set, (ii) to define parameters for modeling pollen production per plant, and (iii) to evaluate secondary traits that may help improve screening procedures used in maize breeding programs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
At Morris, five delayed sowing dates (19, 23, 27, and 31 May, and 4 June) were used to obtain a gradual reduction in pollen availability per silked apical ear. These sowings were the sub-subplots in the lowest plant density. The 40 rows in the subplot were sown on 14 May, and the five later sowings were planted within the earliest one after first removing the plants from specific rows. Each delayed sowing was two rows wide, and was separated from the other late plantings by four rows of the earliest sowing date. All plants in the five delayed sowings were detasseled, in order to have a single source of pollen (plants sown on 14 May) and in this way to reduce gradually the amount of pollen available per exposed silk of plants sown in the delayed sowings. Late sowings were done only at the lowest plant density to determine exclusively the response of kernel set to pollen availability, independently of the effects of high plant density on final kernel number (Andrade et al., 1999).
At Salto, a nondetasseled control and two levels of detasseling (50 and 75%) were included as sub-subplots to reduce pollen supply in all plant densities. The control consisted of 10 rows, and each detasseling treatment consisted of 20 rows. Plants were detasseled in one out of two rows (50%) or in three out of four rows (75%). In all experiments, detasseling was performed when tassels appeared within the whorl of the uppermost leaves and always before anthesis to minimize leaf removal and plant damage. As with the delayed sowings at Morris, the aim of the detasseling treatments was to obtain a range in pollen availability per silk, and not to determine the exact response of pollen grain number per unit land area to the number of pollen-shedding plants. For this reason, no special care was taken to avoid contamination completely from other treatments of the same experiment, and detasseling levels were not fully isolated from each other at the recommended (Luna et al., 2001) minimum distance (about 300 m). Pollen traps (described below) were placed next to plants of the central most detasseled row of each sub-subplot, and ears for kernel number determination were collected from plants of the same row and next to the trap to relate kernel set to pollen availability for each harvested ear.
Experiments were isolated from other maize fields with similar anthesis periods by at least 200 m. At Morris, 100 kg N ha-1, 50 kg P ha-1, and 50 kg K ha-1 were applied before seeding, and additional 100 kg N ha-1 were side-dressed 30 d after sowing of the main plot. At Salto, 150 kg of N ha-1 were side-dressed at the eight-ligulated leaf stage. Weeds were controlled with a preemergence application of 2.2 kg a.i. ha-1 alachlor (2-chloro-N-(2,6-diethylphenyl)-N-(methoxymethyl)acetamide) and 3.3 kg a.i. ha-1 cyanazine (2-(4-chloro-6-ethylamino-1,3,5-triazin-2-ylamino)-2-methylpropionitrile), and by hand weeding after the crop was established. Water stress was prevented with sprinkler irrigation, with soil water content near field capacity throughout the growing season.
Flowering Dynamics
At least 30 plants were tagged at random within each plot before tasseling, and the date of silking (i.e., at least one silk visible after extruded from the husks; Bolaños and Edmeades, 1993) of the apical (E1) and subapical (E2) ears was registered for each tagged plant. Anthesis (i.e., at least one extruded anther visible at the tassel; Bolaños and Edmeades, 1993) date was also registered for the 30 plants tagged in the nondetasseled plots. The ASI of each plant was calculated as the difference in days between anthesis and silking, and could yield positive (protandry) or negative (protogyny) values (Struik and Makonnen, 1992). For each hybrid x plant density treatment, the average of these values described the ASI of individual plants (ASIip). At the plant population level, the cumulative total of tagged plants that had anthesed or silked were recorded daily, and the dates of 50% anthesis and 50% silking were registered and used to calculate the ASI of the plant population (ASIpp), as usually expressed in literature (Struik and Makonnen, 1992; Bolaños and Edmeades, 1993). Average dates of anthesis and silking and ASIip allowed an estimation of interplant variability for these flowering parameters, which is not possible with dates to 50% of the event at a plant population level.
Silks and Pollen Quantification
The number of silks exposed daily was determined on the apical ears of at least 10 tagged plants in the nondetasseled plots, as described by Cárcova et al. (2000). Briefly, exposed silks were cut every 1 to 2 d after silking at Morris and 2 to 5 d after silking at Salto. Silk tissues removed from the apical ear were stored in 700 g kg-1 ethanol until counting. Apical ears used for silk-number determination were harvested for a final counting of exposed silks 9 d after silking.
Pollen shed per unit land area was monitored daily from the onset of anthesis with pollen traps of the type described by Bassetti and Westgate (1994). One trap per plot was placed in the middle of the interrow at ear height and changed daily at about 1400 h or twice daily (during peak pollen shed) at about 1100 and about 1700 h. Pollen grains were counted on three 1-cm2 areas on each trap, by a scanned image and an image measurement software (SigmaScan, Jandel Scientific, La Jolla, CA). Scanned images were calibrated on the basis of eye-countings of pollen grains from predetermined 1-cm2 areas marked on the traps. These countings were made with an enhancer (Leica MZ6, Leica AG Heerbrugg, Switzerland).
On the basis of pollen production per unit land area, the dynamics of daily pollen production of an individual tassel was estimated. Because all plants in the population did not start anthesis on the same date, the daily pollen production of an individual tassel could not be obtained simply as the quotient between pollen production per unit land area and plant density, and was estimated as follows:
 | [1] |
 | [2] |
 | [3] |
 |
 | [4] |
 |
 |
Where, PGTn = pollen grains produced per tassel on Day n; An = pollen-shedding plants per square meter that started anthesis on Day n; PGMn = pollen grains per square meter on Day n. Thus, PGT1 estimated in Eq. [1] is a constant value across subsequent calculations, as PGT2 estimated in Eq. [2], etc. The application of these calculations for the estimation of daily pollen production per tassel assumes that all tassels in the population have the same dynamic of pollen production per day as those that start the process (i.e., A1). Moreover, the estimation requires very uniform environmental conditions (e.g., no rainfall or strong winds) during the whole anthesis period, because it does not allow for a reduction in daily pollen production along the upward part of the curve (i.e., before the maximum daily pollen production of a tassel) because of, for example, adverse weather conditions.
A Gaussian equation of the type described in Eq. [5] was fitted to the estimated values of daily pollen production per tassel.
 | [5] |
 |
In this equation, a (in pollen grains tassel-1) is the area under the curve; b is the number of days between those two points of the curve for which pollen production is half of maximum daily pollen production; DAA is days after start of anthesis; and c is the DAA when maximum daily pollen production of a tassel is reached.
Kernel Set and Pollen Availability
For each hybrid, plants at the lowest plant density were pooled by silking date, including plants from different sowing dates at Morris. Kernel number in the apical ear was calculated as the average of each silking date group. Kernel set of each group was then computed as the quotient between its kernel number and the largest average kernel number value of each hybrid. Kernel set of each group of plants with the same silking date could be related to the corresponding value of pollen production per unit land area (i.e., kernel set of plants that silked on Day n and pollen grains per square centimeter on Day n). This relationship allowed the estimation of the daily pollen shed intensity threshold that may limit kernel set in late-silking plants.
Pollen availability per exposed silk was estimated on the basis of the volume occupied by the silks after their extrusion from the husks. Because silks are not exposed to pollen on a horizontal area parallel to the ground, their surface cannot be matched to a given area covered by pollen. It was assumed that once pollen was shed from the tassels (i) it was uniformly distributed on a horizontal plane that moved downwards across the canopy, (ii) the amount that reached the ear was represented by captures made with the adhesive traps, and (iii) the amount of pollen in that plane was also uniformly distributed in a volume with height equal to the average length of an exposed silk. Thus, daily pollen counts made on an area basis (i.e., pollen grains per square centimeter) at ear height could be expressed in volume units (i.e., pollen grains cm-3), and matched to the number of exposed silks in a given silk volume to obtain the number of pollen grains per exposed silk. The volume of an individual silk was estimated on the basis of the average exposed length and diameter of a silk, and multiplied by the number of receptive exposed silks to obtain a daily total of silk volume. The estimate of the number of receptive exposed silks was based on silk counts described above (for details see Cárcova et al., 2000) and a silk receptivity of 6 d (Bassetti and Westgate, 1993). Because pollen viability in the field is drastically reduced a few hours after shedding from the tassel (Jones and Newell, 1948), pollen availability per exposed silk was based on daily pollen counts and daily values were not accumulated for the computation of pollen grains per exposed silk.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
The ASI did not differ significantly (P < 0.05) between ASIpp and ASIip (Table 2). The rise in ASIip promoted by increased plant density was usually matched by an increased interplant variability in this parameter, except for AP9191 between 2.5 and 7.5 plants m-2.
Pollen Production per Unit Land Area
Significant differences (P < 0.05) were detected between treatments in total pollen production per unit land area (Table 3) . At Morris, a threefold increase in plant density (from 2.5–7.5 plants m-2) was accompanied by a 37% (AP9191) to 42% (AP162) increase in total pollen production. When the number of plants was further augmented to 12.5 plants m-2, only AP9191 exhibited a significant (P < 0.05) additional 11% increase in pollen production. For AP162, the total amount of pollen supplied by each additional plant between 7.5 and 12.5 plants m-2 was almost compensated by the reduction in pollen production per plant.
|
For hybrids cropped in the cool (mean temperature of the pollen shed period = 20.5 ± 2.4°C) environment of Morris, the period for which pollen was detected in the traps ranged between 12 and 14 d, and maximum pollen production was registered between 4 (AP162 at 2.5 plants m-2) and 7 d (AP9191 at 12.5 plants m-2) after the first pollen collection. For hybrids grown at the slightly warmer environment of Salto (mean temperature of the pollen shed period = 22.1 ± 2.2°C), this stage always lasted 8 d, and peak pollen production was usually earlier than at Morris (Fig. 1) . The highest daily pollen production (>900 pollen grains cm-2 d-1) corresponded to the nondetasseled plots of the Argentine hybrids cropped at 9 plants m-2. At Morris, the highest value (about 700 pollen grains cm-2 d-1) was recorded for both AP162 at 7.5 plants m-2 and AP9191 at 12.5 plants m-2.
|
|
Pollen Production and Kernel Set
Two data sets were defined for the relationship between kernel set and daily pollen production per unit land area (Fig. 3a)
: (i) early-silking plants (i.e., those that silked before maximum daily pollen production was reached), for which kernel set was independent of pollen production, and (ii) late-silking plants (i.e., those that silked after the peak of pollen production), for which a significant (P < 0.05) negative response to pollen production at silking was established at Morris. At this site, kernel set was reduced at a rate of 0.4% per pollen grain after daily pollen availability (Fig. 1) reached a threshold of less than 227 pollen grains cm-2 d-1 at the late stages of pollen production (Fig. 3a). Results from Salto followed a similar trend, but the few data from late-silking plants could not be included in the model fitted to the American hybrids. On the other hand, a single significant (P < 0.001) exponential model could be fitted to all data for the relationship between kernel set and pollen availability per exposed silk (Fig. 3b), with no distinction among hybrids or silking dates. The exponential model indicated that about two pollen grains per exposed silk could grant 95% kernel set, but the quotient rose to about three for 99% kernel set. For the early-silking set of data, there was only one observation with a pollen availability level of less than two pollen grains per silk and it reached maximum kernel set (i.e., kernel set = 1). Most data from late-silking plants, obtained exclusively from the delayed plantings (Morris) and the detasseled plots (Salto), were below these thresholds and these late-silking plants experienced a marked reduction in kernel set (Fig. 3b). Conversely, plants from the main plot at Morris (i.e., the earliest planting and only source of pollen) and those within the nondetasseled plots at Salto were always above the mentioned thresholds, with average values of 4.3, 4.8, 7.8, and 8.4 pollen grains per exposed silk for AP162, AP9191, DK664, and DK752, respectively. The amount of pollen needed to reach the threshold of two pollen grains per silk differ among hybrids, because hybrids differed in the maximum number of exposed silks (Fig. 2, and Cárcova et al., 2000). At the low plant densities included in the analysis (2.5 and 3 plants m-2), the amount of pollen required for 95% kernel set would be 1180, 1500, 1490, and 1845 pollen grains per ear for AP162, AP9191, DK664, and DK752, respectively.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Total pollen production per plant of hybrids included in our experiments suggested a reduction in this trait with respect to old cultivars. In spite of a possible underestimation in pollen production because of traumatic effects (bagged tassels) and growth restrictions (plants in pots), values obtained by Hall et al. (1982) for open pollinated varieties (20.7 x 106–42.2 x 106 grains) and a 1978 single-cross hybrid (20.9 x 106 grains) grown at 3.79 plants m-2 were always larger than those obtained in our research (9.65 x 106–11.3 x 106 grains) at lower plant densities (2.5 and 3 plants m-2). These data support the opinion that breeding for improved grain yield has promoted an increased femaleness (i.e., large female productivity at the expense of the male one) in maize (Galinat, 1992), a trend observed in the more advanced and productive races of maize but never analyzed for hybrids.
Increased plant density promoted a large reduction in total pollen production per tassel, which was overcompensated for by the number of tassels per square meter and did not cause a reduction in pollen availability per unit land area. This result contrasts with data from Struik and Makonnen (1992), who did not detect a significant reduction in the weight of pollen produced per tassel between 3.8 and 9 plants m-2. Their data, however, should be interpreted with caution, because the collection procedure (every 2 d from bagged tassels, which had been cut from the field and then grew in a greenhouse) may have affected the mass of pollen produced by the tassel and biased the results. On the other hand, the quantification of the mass of pollen produced per tassel on a daily basis could be very difficult at the start and the end of the anthesis period of a plant, when the amount shed is very small (Fig. 2). These aspects, together with (i) lack of information on the genetic variability for the mass of an individual pollen grain in maize and (ii) the fact that the number of pollen grains can be directly matched to the number of exposed silks in terms of units, support the use of pollen number rather than pollen weight as an indicator of pollen availability.
The reduction in pollen production promoted by increased plant density was accompanied by an estimated shortening in the duration of pollen shedding per plant. For an individual plant, this trend may result in a considerable lack of pollen for late-appearing silks (Fig. 2). These silks, on the other hand, correspond to ovaries from the tip of the ear, which usually have no chance to set kernels at high plant densities independently of pollen availability. Cárcova et al. (2000) reported kernel set in the apical ear of hybrids DK752 and DK664 was 52% (494 kernels set from 950 potential ovaries) and 58% (410 kernels set from 705 potential ovaries), respectively, at a plant density of 9 plants m-2. These reductions in kernel set were larger than the estimated proportion of silks of an individual plant that could not receive pollen at that plant density in the present work (Fig. 2). These results are in agreement with data obtained from water-stressed plants (Otegui et al., 1995), where the addition of fresh pollen to late-appearing silks did not improve kernel set.
The estimated effects of plant-density stress on pollen production per plant were different from the effects observed under a drought stress imposed immediately prior to tasseling (Hall et al., 1982). The former promoted a large reduction in the amount and duration of pollen production per plant, while the latter caused an increase in the duration of this process with a very slight reduction in the amount of pollen produced. These results could be expected, considering that at tasseling the number of pollen grains is already determined (Horner and Palmer, 1995). Plant density effects, on the other hand, can affect tassel growth at early stages (Otegui, 1997), with the concomitant reduction in pollen production per plant.
Our results support the inclusion of reduction in tassel size as a secondary trait in selection for tolerance to increased plant density, because there is yet no evidence of reductions in kernel set because of lack of pollen within a plant population of the same planting date, even after being exposed to a 50% reduction in the population of tassels shedding pollen (data not shown). On the basis of the model fitted to our data, it was determined that (i) all tested hybrids produced pollen in excess for maximum kernel set and (ii) this production could be at least halved from its present level with no expected changes in kernel set. The extent of the reduction, nevertheless, should be carefully interpreted, because the effective amount of pollen produced per silk differed among cropping systems. Pollen production per exposed silk in the Argentine experiment almost doubled (7.8 and 8.4 pollen grains per silk) the production reached in the Midwest (4.3–4.8 pollen grains per silk), a trend that could not be explained by the particular combination of the number of silks exposed from the apical ear (Cárcova et al., 2000) and total pollen production per unit land area (Table 3) of each hybrid. Moreover, values for hybrids at the Morris site were close to the threshold for effective pollen distribution among silks suggested by Sadras et al. (1985). Data from these authors indicated that an average of at least four pollen grains per silk must be met to observe one pollen grain per silk, and 20% of the silks could receive no pollen grain when mean pollen density is reduced to two pollen grains per silk. The large variability in our kernel set data for pollen densities between two and four pollen grains per silk (Fig. 3b) may be indicative of a variable proportion of silks that received no pollen grain. This proportion reached 0.24 for one observation of the DK664 that received 2.3 pollen grains per silk, and is in close agreement with predictions made by Sadras et al. (1985).
Finally, selection for reduced tassel size should not be accompanied by a reduction in the duration of pollen shedding per plant to avoid the risk of lack of pollen for late-appearing silks from the late-silking plants of the population. This negative effect could be expected if selection is based on reduced tassel branch number (Fischer et al., 1987), because pollen shed follows the pattern of tassel differentiation (Bonnett, 1966): it starts on the main branch of the tassel and continues downward to the bottommost lateral branch. In this context, results from our study support previous evidence (Bassetti and Westgate, 1994) on the importance of reaching silking before the peak of pollen shedding to avoid pollen-availability restrictions on kernel set. A threshold of 227 pollen grains cm-2 d-1 at the end of pollen shedding was estimated as the lowest limit beyond which potential kernel set could be affected.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 7, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. E. D'Andrea, M. E. Otegui, A. G. Cirilo, and G. Eyherabide Genotypic Variability in Morphological and Physiological Traits among Maize Inbred Lines--Nitrogen Responses Crop Sci., April 25, 2006; 46(3): 1266 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. E. Halsey, K. M. Remund, C. A. Davis, M. Qualls, P. J. Eppard, and S. A. Berberich Isolation of Maize from Pollen-Mediated Gene Flow by Time and Distance Crop Sci., September 23, 2005; 45(6): 2172 - 2185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. E. Westgate, J. Lizaso, and W. Batchelor Quantitative Relationships between Pollen Shed Density and Grain Yield in Maize Crop Sci., May 1, 2003; 43(3): 934 - 942. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
x e[-2(DAA-c)2/b2] fitted to pollen data, where a is the area under the curve; b is the number of days between those two points of the curve for which pollen production is half of maximum daily pollen production; DAA is days after start of anthesis; and c is the DAA when maximum daily pollen production is reached.