Crop Science Grow Your Career with CSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 24 January 2006
Published in Crop Sci 46:437-447 (2006)
© 2006 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Agricola
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Related Collections
Right arrow Crop Growth and Development
Right arrow Other Oil Crops
Right arrow Other Crops

PLANT GENETIC RESOURCES

Effect of Sowing Time on Coriander Performance in a Semiarid Mediterranean Environment

Alessandra Carrubba*,a, Raffaele la Torrea, Filippo Saianob and Giuseppe Alonzob

a Dipartimento di Agronomia Ambientale e Territoriale, Facoltà di Agraria, Universita di Palermo, Viale delle Scienze, I 90128, Palermo, Italy
b Dipartimento di Ingegneria e Tecnologie Agrarie e Forestali, Facoltà di Agraria, Università di Palermo, Viale delle Scienze, I 90128, Palermo, Italy

* Corresponding author (acarr{at}unipa.it)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In semiarid environments, time of sowing is one of the most important factors influencing seed yields. For coriander (Coriandrum sativum L.), the most commonly recommended cropping technique is spring sowing (March–April), since the optimum soil temperature for seed germination ranges between 20 and 23°C, and the crop shows a remarkable sensitivity to frost and cold. In many semiarid areas of southern Italy, however, the occurrence of prolonged dry periods in summer and spring does not allow for the scheduling of summer crops without irrigation. However, the generally mild winter temperatures and the typical rainfall distribution, which is mostly concentrated over the winter months, could allow sowing time to be moved to the winter season to take advantage of the winter rainfalls. To evaluate the effect of moving the sowing time of coriander on seed yield and plant performance in semiarid Mediterranean environments, a field trial was performed in 1998–1999, 1999–2000, and 2000–2001 at Sparacia (Cammarata, AG, Sicily). Coriander seeds were sown in rows 50 cm apart every month for 5 mo from December to April. The fruits were harvested from mid-June to mid-July. The time from sowing to harvest was greatly dependent on the sowing date; the duration was 193 to 195 d for the December sowings, and 91 to 100 d for the April sowings. In all 3 yr, the most productive sowing time was December, and sowing after this date resulted in lower yields.

Abbreviations: D, sowing date • DAS, days after sowing • E, crop emergence • GDD, growth degree days • H, harvest time • MR, multiple regression • SLR, simple linear regression • ST, growth stage • Y, year


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
CORIANDER is an annual herb belonging to the family Apiaceae (ex Umbelliferae). In Mediterranean areas, it is spontaneous and ubiquitous, and it is often found as a weed in wheat crops (Catizone et al., 1986; Pignatti, 1997). In many countries, its fruits (commonly termed "seeds") are traditionally marketed and used as a spice. In Italy, its market is rapidly increasing because of the growing presence of immigrants from Asia and the increasing possibilities of using coriander seeds outside the food industry (Berti and Schneiter, 1993; Biomatnet, 2001). Only a very small part of the total domestic supply comes from crops grown in Italy. According to recent studies (Vender, 2001), the total Italian coriander area is no larger than 10 ha. As with many medicinal and aromatic plants, much of the produce utilized by industry comes from developing countries, which are able to market their product at very competitive prices because their low labor costs keep production costs down. The imported items, however, are often of very poor quality owing to the uncontrolled use of pesticides, the residues of which are often found in significant amounts.

There is abundant scientific literature on coriander's botanical (Diederichsen, 1996; Diederichsen and Hammer, 2003) and chemical (Kurkcuoglu et al., 2003; Lawrence, 1997; Pino et al., 1996) characteristics, but there is only a small amount of work concerning agronomic practices, with the exceptions of fertilization, which has been carefully studied under many environmental conditions (Ghosh et al., 1985; Hornok, 1983 and 1986; Oliveira et al., 2003; Rahman et al., 1990), irrigation (Khashmelmous, 1984), and disease control (Dennis and Wilson, 1997; Singh et al., 2003). The effect of sowing time under irrigation has been studied in Argentina by Luayza et al. (1996), who stated that, under nonstressed environmental conditions, autumn–winter sowing should be optimal for achieving the highest yields. A field trial performed in Argentina (de la Fuente et al., 2003) suggests the possibility to introduce coriander for essential oil production as a means for the exploitation of poor soil environments.

The pedoclimatic cropping conditions in semiarid Mediterranean areas could be optimal for coriander (Carrubba et al., 2002a); however, before beginning medium-to-large-scale cropping of coriander, it must be recognized that coriander cropping techniques are not properly defined for these environments. Most of the recommended cropping techniques for coriander are derived from experiments performed in climates very different from the semiarid Mediterranean environment, e.g., subtropical Asia, where the crop is grown under extensive conditions (Nawata et al., 1995), the temperate environments of central Europe (e.g., Hungary and the Czech Republic), where it is usually grown as a spring–summer crop (Diederichsen, 1996; Hornok, 1986), and the saline wetlands of southern Australia (Elder, 1999), where the crop has been successfully tested in field trials. In semiarid Mediterranean environments, the spring–summer period is characterized by severe drought conditions, and so a spring–summer crop needs to be irrigated. However, this is seldom possible because water is typically lacking at this time, and the limited water resources are devoted to crops with higher income potential.

The aim of this work was to evaluate, in an environment representative of the inner hilly semiarid Sicilian areas, the effects of different sowing times on coriander seed yield and plant performance and, hence, the possibility of moving the sowing time to the winter period and also to attempt to elucidate some mechanisms involved in the development and productive processes of the plant. With this objective, one coriander biotype, Coriandrum sativum L. var. microcarpum DC, was sown at five times, ranging from December to April, i.e., from the end of winter to the middle of spring.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In all trial years (1998–1999, 1999–2000 and 2000–2001), the experimental plots were in Sparacia (Cammarata, AG, Sicily; 37°38' N, 13°46' E; 415 m above sea level), on soils classed as a clayey, mixed thermic, Aridic Haploxerert (NRCS, 2003).

Coriander was sown on five sowing dates (Table 1), which were chosen to coincide with the onset of rains and suitable soil conditions and were in the first 10 d of each month from December (D1) to April (D5). The experimental plots (48 m2, 8 x 6 m) were laid out according to a randomized complete block design with three replicates. The seeds, collected from a locally grown small-seeded Coriander biotype, were manually sown in rows 0.5 m apart (16 rows for each plot) to obtain (depending on seed germinability) a plant density of 40 plants m–2; no fertilizer or chemicals were applied.


View this table:
[in this window]
[in a new window]
  		Table 1. Sparacia (Cammarata, AG, Sicily)—Sowing (S) and harvest (H) dates used in a trial on coriander in 1998–1999, 1999–2000, and 2000–2001.

 
Plant development was continuously monitored throughout the trial. Average plant height was measured every week from crop emergence to harvest (five measurements for each survey, accounting for 13 to 28 observations for each plot, depending on cycle duration), and the starting dates of the most significant development stages were determined by visual assessment of the average plot conditions. The reproductive stages (from start of flowering to harvest time) after Lawrence (1993) were named as follows: ST1, start of flowering (appearance of first flowers on the primary umbels in at least 10% of plants); ST2, nearly full flowering (appearance of flowers on the secondary umbels in at least 50% of plants); ST3, full flowering (primary umbels bearing young green fruits in at least 50% of plants); ST4, end of flowering (50% flowers and 50% fruits in at least 50% of plants); ST5, full green fruits (90% green fruits in at least 50% of the plants); ST6, brown fruits (brown fruits on the primary umbels, green fruits on the secondary umbels in at least 50% of the plants); and ST7, full ripening (primary umbels completely ripe in at least 90% of the plants).

Seeds were harvested as the plants in each plot were reaching the ST7 stage, which was between mid-June and mid-July, with the exact time varying according to the year and the sowing time (Table 1). At harvest time, to remove any border effect, a sample area was obtained by excluding the two external rows for each plot and all plants within 0.5 m of the end of each sample row (sample area 30 m2; 12 rows, each 5 m long); all plants in the sample area were manually cut at ground level. In each plot, an internal row, representative of general plot conditions, was separated from the others; a sample of 20 randomly selected plants was taken from it and used for measuring height and weight of plants and number of umbels per plant. After these measurements were taken, all seeds from the 20-plant samples were mechanically threshed, and the harvest index (%) was calculated by dividing the dry weight of seeds by the dry weight of biomass, and multiplying the obtained value by 100. All umbels on the remaining plants in the sample row were manually picked, and 50 randomly selected umbels were used for measurement of their diameter. After a short period of open-air drying, the remaining plants on the 11 sample rows were mechanically threshed, and the yield data obtained were further expressed in kg ha–1. Four samples from each replicate, each of 100 seeds, were taken and weighed and, by multiplying their averaged value by 10, the 1000-seed mass was obtained. After oven drying the seeds at 105°C for 24 h, seed weights were recorded at 0 g kg–1 moisture content (dry weight) and the moisture levels of the seeds for each harvest were calculated.

The durations of the growth stages were measured in days, and the corresponding thermal sums, expressed in growth degree days (GDD; °C), were calculated from the following formula:

Formula
where Tmax is the highest day temperature (ceiling value: 30°C); Tmin is the lowest day temperature (when it is lower than Tbase, then Tbase is used for calculation); and Tbase is the base temperature (0°C). The thermal parameters (Tbase = 0°C and ceiling Tmax = 30°C) were selected after previous experiments (Carrubba and la Torre, 2002, unpublished data).

The volatile components of fruits were analyzed at the Dipartimento di Ingegneria e Tecnologie Agrarie e Forestali (DITAF- University of Palermo) by following the guidelines already successfully used with coriander in similar trials (Di Prima et al., 2000; Carrubba et al., 2002b).

All the data were statistically analyzed; tests included analysis of variance (ANOVA), simple linear regression (SLR), and multiple regression (MR). The ANOVA, based on a fixed linear model, was used on pooled data to detect the occurrence of any systematic variation attributable to the two factors under study [year (Y) and sowing date (D)]. Y was considered as a fixed factor because the strong climatic variability over years, which is a characteristic of the experimental area, does not allow a 3-yr period to be considered as a statistically representative sample of the overall environmental conditions. When the F test indicated statistical significance at the P ≤ 0.05 level, Tukey's honestly significant difference test was used to separate the means.

The SLR was first used to determine the degree of association between seed yield (kg ha–1) and all yield parameters for which an a priori cause–effect relationship could have been previously stated, namely mass of 1000 seeds (g of dry matter), number of umbels per plant, umbel diameter (cm), average plant height (cm), total cycle length (days after sowing) and total rainfall amount throughout the plant cycle (mm).

The SLR was also applied to the duration of the plants' development substages, defined as previously mentioned and expressed in days after sowing, total cycle duration and accumulated rainfall for each substage. The same statistical procedure was applied (data not shown) to another pooled dataset including the most important climatic parameters, as well as thermal sums, rainfall and daylength; for each substage, we took into consideration the number of days characterized by (according to literature) physiologically crucial temperature levels (0, 4, 15, and 30°C). This step was performed to initiate analyses of individual parameters to select the variables to be used in the MR analysis, and so avoid redundancy and multicollinearity.

For each parameter studied in MR, analyses of residual plots and exploratory data plots (data not shown) were used to guide potential improvements in the regression fit. F tests were used to evaluate the significance of additional predictor variables, and coefficients of determination (R2) were examined to determine the amount of variability explained by the best fitting models (Gomez and Gomez, 1984; Norusis, 1994; Steel and Torrie, 1980).

Climatic Details
In all trial years, the major climatic parameters (rainfall, global radiation, and minimum and maximum daily temperatures) were recorded throughout the growth cycle (Fig. 1 ) by means of a weather station on the experimental farm.


Figure 1
View larger version (32K):
[in this window]
[in a new window]
  		Fig. 1. Ten-day values of maximum ( Tmax) and minimum ( Tmin) daily air temperature and rainfall (mm) recorded in Sparacia (Cammarata, AG, Sicily) in a trial about coriander sown from 1998–1999 to 2000–2001 in five different sowing times (from D1, December, to D5, April).

 
As shown in the graphs, the total rainfall amount from November to June, as is typical for the trial environment, was highly variable from year to year. It was about 200 mm in 1998–1999, 420 mm in 1999–2000 and 505 mm in 2000–2001, with most of total rainfall (69, 72, and 84%, respectively) being recorded between November and February. In the first two trial years, the rainfall amount was lower (–57% and –11% respectively) than the 20-yr (1987–1998) average of 471 mm, whereas the recorded value was the 7% higher in 2000–2001.

The trend in minimum and maximum temperatures was substantially in agreement with the dominant climatic conditions of the trial environment: each year, Tmin and Tmax reached their lowest levels in January–February; Tmin rarely fell below 0°C except for some short winter periods, whereas Tmax rapidly increased after March–April, with values exceeding 30°C for the entire period from the end of May to harvest time.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Seed Yield and its Determinants
The biometrical and yield data collected during the trial years, along with the results of ANOVA, are reported in Table 2. The average seed yield did not show any significant interaction effect between the two factors under study (Y x D), which indicates a substantially additive behavior for these two sources of variability; it is therefore possible to examine them as an average of the main effects. In the first trial year, overall productivity was very low (less than 900 kg ha–1), but production almost doubled in the second and third trial years. Moreover, over all years, the effect of sowing date was significant and productivity sharply decreased as the sowing time was postponed, reaching its lowest in the treatments sown in April. The maximum yield (3305 kg ha–1), which was statistically different from all the others, occurred with the earliest sowing, whereas the yields from the other sowing times ranged from 1722 to less than 400 kg ha–1. In fact, partitioning of the treatment sum of squares (SS) revealed that D was the major source of variation in seed yield in these experiments, and accounted for 57% of total experimental variability.


View this table:
[in this window]
[in a new window]
  		Table 2. Main yield and biometrical characters collected at Sparacia (Cammarata, AG, Sicily) in 1998–1999, 1999–2000, and 2000–2001 on a coriander biotype sown in five different sowing time from December to April (D1 to D5), partitioning of the treatments sums of squares and significance of the F ratio calculated for each parameter.

 
The possible reasons for this high variability in yield are many and diverse. One of these is the climatic pattern. As well recognized by literature (Arnon, 1992; Mahdi et al., 1998), rainfall variability is a major factor influencing rainfed crop production in the semiarid regions. In our experiment, total rainfall from November to June, as typical for the trial environment, was highly variable from year to year. Simple linear regression (Table 3) showed that rainfall was strongly associated with yield (r = 0.93), and this may explain the lower yields gained in the first trial year. Moreover, in each year the diverse sowing times allowed the plants to take advantage of different rainfall events, which induced a parallel variation in yields: the lower the former, the lower the latter.


View this table:
[in this window]
[in a new window]
  		Table 3. Sparacia (Cammarata, AG, Sicily)—Correlation coefficients among seed yield, yield determinants, cycle length, and total rainfall amount in Coriander sown in 5 different sowing dates in 1998–1999, 1999–2000, and 2000–2001. n = 15.

 
This simple result may be due to various interrelated factors, and simple linear regression may offer some interpretative key. Seed yield was directly correlated with most of the examined parameters, there being complex relationships involving the combined effect of number of umbels per plant, average diameter of umbels and 1000-seed mass. The number of umbels per plant (Fig. 2 ) in our work ranged from 2.8 (1998–1999, D5) to 63.7 (2000–2001, D1); each year the lowest value was reached in D5, whereas the remaining treatments showed a rather irregular distribution. The date of sowing accounted for the largest share (41%) of total variability, almost twice that due to the climatic effect (Y, 23%). Although the occurrence of a significant Y x D interaction, which accounted for 35% of total variability, does not allow any general conclusion, there was a general tendency toward a decrease in this parameter in the later sowing times. The analysis of simple linear regression confirmed the close association claimed by Diederichsen (1996) of number of umbels per plant with seed yield (r = 0.94), and also highlighted its consistently positive association with all the other examined traits.


Figure 2
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 2. Sparacia (Cammarata, AG, Sicily)—Number of umbels per plant counted on Coriander plants sown from 1998–1999 to 2000–2001 in five different sowing times (from D1, December, to D5, April). Each value is the average of three replications. Vertical lines indicate the mean standard deviation.

 
The 1000-seed mass showed the lowest values (less than 7 g) in crops sown in April and March, whereas significantly higher values (7.7–7.9 g) were shown by earlier sowings. The ANOVA revealed a highly significant effect (P ≤ 0.001) of the sowing date, whereas neither Y nor Y x D showed any significant effect. The partitioning of the SS confirmed this trend, with a 59.5% of total variability explained by sowing date, whereas only 3.4% was attributable to the year and 37.1% to the interaction of both factors.

The umbel diameter showed a similar trend, with values decreasing (from 4.1–3.3 cm) from D1 to D5. The variability between years showed up as the major determinant (60%) of total field variability. The ANOVA revealed that the sowing date was highly significant (P ≤ 0.01) and was able to explain almost one third of total variability.

There was a wide variation in plant height (Fig. 3 ), from little more than 30 cm (1998–1999, D5) to 115.5 cm (2000–2001, D2). Values were highest with the earliest sowing and, regardless of sowing time, in the second and third year. The ANOVA showed a very high level of significance (P ≤ 0.001) of both tested sources of variability (Y and D) and of their interaction on the variation of this parameter, but the partitioning of SS indicated that sowing date had a much higher importance, accounting for the 68% of total experimental variability. Even though it may not be considered strictly as a "yield factor," plant height was tightly associated with yield and with all yield determinants; the correlation coefficient (r = 0.81) between plant height and number of umbels per plant was highly significant, which confirms some of the data in the literature (Diederichsen, 1996; El-Ballal and Abou El-Nasr, 1987) and shows that, under our experimental conditions as well, healthy plants have higher productivity.


Figure 3
View larger version (19K):
[in this window]
[in a new window]
  		Fig. 3. Sparacia (Cammarata, AG, Sicily)—Average plant height measured on Coriander sown from 1998–1999 to 2000–2001 in five different sowing times (from D1, December, to D5, April). Each value is the average of three replications. Vertical lines indicate the mean standard deviation.

 
The importance of the vegetative part of the plant on the assessment of yield is confirmed by the examination of harvest index (Fig. 4 ). This value was calculated to obtain some information about the partitioning of photosynthates between the vegetative and reproductive parts of plant, i.e., the efficiency of plants in using the available resources for producing seeds (Arnon, 1992; Sinclair, 1998).


Figure 4
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 4. Sparacia (Cammarata, AG, Sicily)—Values of harvest index calculated on Coriander plants sown from 1998–1999 to 2000–2001 in five different sowing times (from D1, December, to D5, April). Each value is the average of three replications. Vertical lines indicate the mean standard deviation.

 
In our experiment, harvest index values ranged from 18.5 (1998–1999, D4) to 66.0 (1999–2000, D2). Although the ANOVA highlighted a strong Y x D interaction effect, it explained only 2.3% of the total variability in the experiment. The highest explanatory contribution (73.8%) belonged to the sowing date, which is the cause of the evident decreasing trend (with the only exception of D3) from the earlier to the later sowing time. As stated for other seed crops, the increase in harvest index is related to the occurrence of more appropriate environmental conditions (Kang et al., 2002), which cause an enhancement in both seed yield and plant biomass, but with an increase in seed yield more than proportional to the enhancement in plant biomass, a situation that Donald and Hamblin (1976) define as "typical of responses to water." It may be concluded that a close association exists between the productivity levels of coriander and its aboveground biomass, and this association seems to be linked to the overall superior vigor of plants in the most productive crop; generally, bigger plants bear more umbels per plant, have a higher 1000-seed mass and, therefore, allow higher yields.

Cycle Length and Growth Stages
The graph in Fig. 5 shows the trend of the development stages of coriander in all the trial years and for each sowing date. As can be seen, postponing the sowing time significantly shortened the days to harvest, which reached 193 to 195 DAS in the D1 and 91 to 98 DAS in the D5 treatments. This result is consistent with that obtained by Luayza et al. (1996); and the close association between yield and cycle duration found by these authors supports our findings (correlation coefficient, r = 0.79). The same trend is evidenced by the correlation coefficients calculated between the cycle duration and each yield parameter, which confirm the strong influence of cycle duration on the general health of plants and, therefore, on yields.


Figure 5
View larger version (54K):
[in this window]
[in a new window]
  		Fig. 5. Sparacia (Cammarata, AG, Sicily)—Trend of the development stages in Coriander, according to year of cultivation (1998–1999 to 2000–2001) and sowing time (D1, December, to D5, April). E, crop emergence; ST1 to ST6, reproductive stages according to Lawrence (1993); H, harvest time.

 
The vegetative phase (S–ST1) lasted 40 to 146 DAS, i.e., always more than 50% of total cycle duration, but the length of this phase decreased (from 71.4–53.6%) from the earliest to the latest treatment. Emergence occurred 10 to 41 DAS, but the following substage (E–ST1) lasted from 33 to 115 d and there was a strong correlation (r = 0.94) between the length of the substage E–ST1 and the total duration of the crop cycle.

Within the treatments, the beginning of the ST1 stage was observed over a period ranging from the first days of April (1999–2000, D1 and D2) to the first days of June (2000–2001, D5). Time of flowering seemed to be less affected by sowing date (from D1 to D5) than was plant development. This trend was especially evident in the first cropping year; the D1 treatment entered the ST1 stage when the crop had reached an average height of 90 cm and there was full ground cover, whereas, in the D5 treatment, the first flowers began to appear when the average height of the crop was less than 20 cm and the plants scarcely covered 10% of the ground.

The phases after start of flowering (ST1) proceeded rather quickly, and their duration was 50 to 55 d, irrespective of sowing time. The graphs in Fig. 6 show, for each year and sowing time, the relationships between the time when each crop stage began (days, on x axis) and the corresponding cumulated thermal sum (GDD, on y axis). There was a rather large variability in all thermal sums, with most of the variability being due to sowing time and cropping year. The onset of the reproductive phase, further split into its substages, occurred after reaching a thermal sum (from sowing time), which depended on the year and sowing time and ranged from 687 to 1524 GDD. Almost all of the crop became suitable for harvest shortly after receiving a thermal sum of 2000 GDD (°C); in 2 of the 3 yr, this value was not reached with the last sowing time (D5).


Figure 6
View larger version (35K):
[in this window]
[in a new window]
  		Fig. 6. Sparacia (Cammarata, AG, Sicily)—Development stages of Coriander in relation to the respective GDDs, according to year of cultivation (1998–99 to 2000–01) and sowing time (D1, December, to D5, April). E = crop emergence; ST1 to ST6, reproductive stages according to Lawrence (1993); H, harvest time.

 
Thermal sums models different from the above (data not shown) gave similar results, which is in agreement with the suggestion that coriander floral induction should be subjected to other environmental factors, acting in addition to, or in interaction with, thermal sums. An attractive hypothesis (Nawata et al., 1995; Elder, 1999) is that coriander should be a quantitative long day plant. To verify this hypothesis with the tested genotype, data related to crop development (i.e., the average plant heights measured for each year and treatment throughout the trial, along with the respective flowering dates) were linked to the average daylength (Fig. 7 ). In this case, crop response did not seem to be univocal, and the rather good uniformity in flowering dates in 1998–1999, in which the above stage was reached shortly after reaching a 13-h daylength, did not match the early flowering shown by the crop in 1999–2000.


Figure 7
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 7. Sparacia (Cammarata, AG, Sicily)—Trend of average plant heights measured on Coriander in relation to time and measured daylength (dotted line), according to year of cultivation (1998–1999 to 2000–2001) and sowing time (D1, December, to D5, April). The circles on each line represent the flowering date.

 
The lack of any specific and univocal correlation between the measured parameters and the onset of flowering suggests that the beginning of the reproductive stage could be controlled by many factors acting simultaneously. An exploratory MR analysis was therefore performed; first, by inserting as independent variables all climatic data thought to possess some biological and/or physiological meaning, and then performing the backward elimination of all nonexplanatory variables. The analysis (Table 4) was performed setting as dependent variables the number of days from sowing time to emergence (S–E), from emergence to the start of flowering (E–ST1) and from the start of flowering until harvest time (ST1–H). The three regression models obtained satisfactorily explain (R2adj = 0.89; R2adj = 0.99; R2adj = 0.96, respectively) the variability observed in the duration of each development stage. Concerning the S–E duration, our analyses indicated that four independent variables, all of them linked to the thermal and moisture environmental conditions, could be included in the regression model. With a beta value of 1.21, the thermal sum (GDDS–E) seems to be the best predictor for the variations in S–E length (P ≤ 0.001). The analysis also highlighted the direct influence of exposure of germinating seeds to minimum temperatures lower than 0°C; under our specific experimental conditions, for each additional low-temperature day, the emergence date was delayed by 0.8 d, whereas each additional day with a minimum temperature higher than 4°C exerted the opposite action, pushing the plants to an earlier emergence time.


View this table:
[in this window]
[in a new window]
  		Table 4. Results of multiple regression (MR) analyses on coriander data obtained in 1998–1999, 1999–2000 and 2000–2001 in Sparacia (Cammarata, AG, Sicily). Response variables are the durations (d) of the substages from sowing to crop emergence (S-E), from crop emergence to the onset of flowering (E-ST1) and from the onset of flowering to harvest time (ST1–H). Independent variables shown are those included in the regression models after the backwise elimination procedure. n = 15.

 
The same analysis performed for the number of days elapsed from emergence to flowering time (E–ST1) saved only one independent variable, namely, the number of days in which the daily Tmin values were lower than 15°C. Therefore, the onset of flowering should be induced by increasing Tmin to levels higher than 15°C, and somehow ensuring that the temperature is maintained at such levels.

The duration of the stage ST1–H was scarcely affected by the photoperiod, which is evidence of the strong predictive value of thermal sums and, more significantly from a physiological point of view, of the number of days with a Tmax higher than 30°C, which seems to affect ripening time.

Qualitative Parameters
Coriander is graded by the buyer according to its aroma and appearance. Of coriander's essential oils, the components most characteristic and mainly responsible for coriander's typical scent are linalool (60–70%), {alpha}-pinene (2–3%), and limonene (1–5%) (Lawrence, 1997). Table 5 shows the average qualitative measurements of the coriander seeds obtained along the trial. The most prevalent volatile compounds were always {alpha}-pinene, {gamma}-terpinene, and linalool, and these three comprised more than 78% of total volatiles. Geranyl acetate, an important flavoring agent (EC, 1999; FAO/WHO, 2004), was detected in the last trial year only and in two of the five harvests, and a small amount of borneol was detected in the December sowing time in 1999–2000. In the trial years, the seed samples had rather low contents of linalool (29.7–38.7%), typical amounts of limonene (5.0–5.5%), and above average levels of {alpha}-pinene (22.1–24.9%).


View this table:
[in this window]
[in a new window]
  		Table 5. Identified volatile compounds in Coriander seeds obtained in 1998–1999, 1999–2000 and 2000–2001 in Sparacia (Cammarata, AG, Sicily) in five different sowing times (from December to April), medians and coefficient of variation. F values are calculated separately for the factors "Year" and "Sowing time". n = 15.

 
As stated in previous work (Carrubba et al., 2002b), {alpha}–pinene, p-cymene, and camphor are among the compounds that are most affected by modifications to cropping conditions. In our experiment, the p-cymene and camphor contents were significantly affected by the year effect, whereas only {alpha}-pinene was affected by the changes in the sowing time.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The present study suggests that coriander could be a promising crop for Mediterranean semiarid environments, where the occurrence of rather severe climatic conditions limits the establishment of many other crops. Considering the low use of technical inputs, the seed yields obtained in all trial years were satisfactory, provided the crop was sown early. The crop seemed rather tolerant to dry periods, even when prolonged, provided the dry period was not in the early phases of the cycle.

The trial showed a very high variability in yield, but it is possible to argue that the most productive treatments reached their yield value for the contemporary effect of the different yield parameters. Our work suggests that seed yield is affected by complex relationships between the number of umbels per plant, the diameter of umbels, and the individual seed weight. The number of umbels per plant and the 1000-seed weight seemed most affected by sowing date, whereas the diameter of umbels seemed most influenced by the variability between years. The climatic pattern exerted its effect on yields primarily in a direct way, but, in addition to this direct effect, an indirect effect should have played some role, expressed indirectly through the number of umbels per plant, which, in turn, was highly correlated with seed yield (r = 0.94). However, if these parameters are interdependent, the analogous findings of Luayza et al. (1996), who performed their trial under irrigation (and therefore under nonstressed conditions), allow us to suppose the involvement of other factors, which require further study.

Yield also seems to be strictly correlated with the duration of the crop development stages. Although the aim of this work was not to build a comprehensive phenological model, some interesting aspects relating to plant development were revealed and these deserve closer examination in further studies. The postponement of sowing time significantly shortened the cycle, which is linked to the progressive reduction of all the examined growth stages, especially those immediately preceding the onset of flowering. The D5 treatments, for example, started the reproductive stage in approximately one third of the days required by the D1 treatments. The start of flowering seemed to be correlated with the increase in daily minimum air temperatures to values higher than 15°C, whereas an earlier time of ripening seemed to be the result of Tmax values exceeding 30°C.

Therefore, the most advisable sowing time for coriander in Mediterranean areas similar to the experimental site seems to be early December, which would allow the crop to exploit the natural rainfall events, prolong the vegetative phase and so obtain the best possible yields.

The data collected did not show any systematic effect of sowing time on the composition of the essential oils. However, there was a significant effect of sowing time on the content of {alpha}-pinene in the seeds, although the differences do not seem to have any biological basis.


   ACKNOWLEDGMENTS
 
The authors are grateful to the anonymous referees and to the associate editor of this journal, for their critical and constructive comments on the earlier version of this paper.

Received for publication March 1, 2005.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Agricola
Right arrow Articles by Carrubba, A.
Right arrow Articles by Alonzo, G.
Related Collections
Right arrow Crop Growth and Development
Right arrow Other Oil Crops
Right arrow Other Crops

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