Epidemiology

Yield Loss in Chickpeas in Relation to Development

of Fusarium Wilt Epidemics
Juan A. Navas-Cortés, Bernhard Hau, and Rafael M. Jiménez-Díaz

First author: Departamento de Protección de Cultivos, Instituto de Agricultura Sostenible (IAS), Consejo Superior de Investigaciones
Científicas (CSIC), Apartado 4084, 14080 Córdoba, Spain; second author: Institut für Pflanzenkrankheiten und Pflanzenschutz, Uni-
versität Hannover, Herrenhäuser Str. 2, 30419 Germany; and third author: Departamento de Protección de Cultivos, IAS-CSIC, and
Escuela Técnica Superior de Ingenieros Agrónomos y de Montes, Universidad de Córdoba, Apartado 3048, 14080 Spain.
Accepted for publication 1 August 2000.

ABSTRACT

Navas-Cortés, J. A., Hau, B., and Jiménez-Díaz, R. M. 2000. Yield loss developed to relate chickpea yield to Fusarium wilt disease intensity with
in chickpeas in relation to development of Fusarium wilt epidemics. the following independent variables: time to initial symptoms (tis), time
Phytopathology 90:1269-1278. to inflection point (tip) of the disease intensity index (DII) progress curve,
final DII (DIIfinal), standardized area under DII progress curve (SAUDPC),
Development of 108 epidemics of Fusarium wilt of chickpea caused and the Richards weighted mean absolute rate of disease progression (rho).
by Fusarium oxysporum f. sp. ciceris were studied on cvs. P-2245 and Irrespective of the chickpea cultivar × pathogen race combination, the
PV-61 in field microplots artificially infested with races 0 and 5 of F. oxy- absolute and relative seed yields decreased primarily by delayed sowing.
sporum f. sp. ciceris in 1986 to 1989. Disease progression data were fitted The relative seed yield increased with the delay in tis and tip and de-
to the Richards model using nonlinear regression. The shape parameter was creased with increasing DIIfinal, SAUDPC, and rho. A response surface
influenced primarily by date of sowing and, to a lesser extent, by chick- was developed in which seed yield loss decreased in a linear relationship
pea cultivars and races of F. oxysporum f. sp. ciceris. Fusarium wilt re- with the delay in tis and increased exponentially with the increase of rho.
duced chickpea yield by decreasing both seed yield and seed weight. These
effects were related to sowing date, chickpea cultivar, and virulence of Additional keywords: Cicer arietinum, crop loss models, quantitative
the prevalent F. oxysporum f. sp. ciceris race. Regression models were epidemiology.

Chickpea (Cicer arietinum L.) is a major source of human and Fusarium wilt of chickpea (17,36), but those reports made no
animal food and the world’s third most important pulse crop after assessment as to how factors in the pathosystem would influence
beans (Phaseolus vulgaris L.) and peas (Pisum sativum L.) (38). its efficiency. Navas-Cortés et al. (30) demonstrated that the ef-
Chickpea production is severely curtailed by Fusarium wilt caused fectiveness of sowing time as a management practice for
by Fusarium oxysporum (Schlechtend.:Fr.) f. sp. ciceris (Padwick) Fusarium wilt of chickpea may be influenced both by virulence of
Matuo & K. Sato, in most chickpea growing areas of the world the pathogen race prevalent in soil and susceptibility of the chick-
(17). In spite of its importance, there are few quantitative assess- pea cultivar (30). However, the interactive effects of sowing date,
ments of the impact of Fusarium wilt on chickpea yields. Annual chickpea cultivar, and race of F. oxysporum f. sp. ciceris on
chickpea yield losses from Fusarium wilt vary from 10 to 15% Fusarium wilt development and chickpea yield loss has not been
(17,43) but can result in total loss of the crop under specific determined. This information would provide a better under-
conditions (14,15). standing of the system and has the potential to be a valuable aid in
Fusarium wilt of chickpea has been managed primarily by the the management of this disease.
use of resistant cultivars (17), but virulent races of the pathogen Disease progress curves (25) can be described by growth curve
have undermined their importance in recent years (16,19). Seven models such as the monomolecular, Gompertz, logistic, Richards,
races, designated 0 to 6, of F. oxysporum f. sp. ciceris have been and Weibull equations (8). Progression of diseases caused by
identified (16,19). Race 0, the least virulent of the seven races, soilborne pathogens can be determined by factors other than
induces progressive foliar yellowing compared with severe leaf pathogen reproductive strategy (33). These can either be described
chlorosis, flaccidity, and early wilt induced by races 1 through 6 by the logistic model (9,40) or they do not conform to either the
(16,19). Races 1 through 4 were first described in India (19). monomolecular or the logistic model. In these cases, fits to
Later, race 0 was reported in California, Spain, and Tunisia; races disease progression data were improved with the nonsymmetrical
1 and 6 were identified in California, Morocco, and Spain; and Gompertz model (3) or the Weibull (32) and Richards models (8).
race 5 was found in California and Spain only (14,19). Disease–yield loss relationship has been determined by single-
Date of sowing is a key factor in determining yield of chickpea point, multiple-point, and integral empirical models (8,27) that
crops. In the Mediterranean region, chickpea is traditionally sown relate yield loss to disease intensity at: (i) a specific time or
in the spring, but winter sowing enables matching of various crop specific growth stage in the crop growing season; (ii) several
growth stages with optimum environmental conditions and in- points during this season; and (iii) a measurement of total disease
creases yield through better use of available water in soil (39). derived by summing disease intensities over a specific period of
Choice of sowing time has been recommended for management of crop growth (6,8,18). Increasing the understanding of crop yield
losses in relation to development of plant disease epidemics may
Corresponding author: R. M. Jiménez-Díaz; E-mail address: ag1jidir@[Link] allow better predictions of crop losses and improved disease
management practices (27). The objectives of this study were: (i)
Publication no. P-2000-0905-02R to determine the effects of sowing date, virulence of the pathogen
© 2000 The American Phytopathological Society race, and cultivar susceptibility on the development of Fusarium
Vol. 90, No. 11, 2000 1269
wilt epidemics in chickpeas; and (ii) to determine a relationship from 0 to 4 according to percentage of foliage with yellowing or
between seed yield losses and the development of disease necrosis in acropetal progression: 0 = 0%, 1 = 1 to 33%, 2 = 34 to
epidemics. 66%, 3 = 67 to 100%, and 4 = dead plant, as used previously
(19,43). Incidence (I) and severity data (S) were used to calculate
MATERIALS AND METHODS disease intensity index (DII) with DII = (I × S)/4. Disease prog-
ress curves were obtained from the accumulated DII over time in
Experimental design. A microplot experiment was conducted days from the date of sowing.
in a field with sandy loam soil (pH 8.5, 1.4% organic matter) at Disease progress curves were characterized with the Richards
the Alameda del Obispo Research Station near Córdoba (latitude model (8), which compared with monomolecular, Gompertz, and
38° north, longitude 5° east) in three consecutive seasons (harvest logistic models, gave the best fit to disease progress data from
years 1987, 1988, and 1989). This field had not been sown to 3 years. The Richards model can be written as
chickpeas in the previous 10 years. The microplots (1.25 × 1.25 m,
DII(t) = K [1 – B exp(–r × t)]1/(1–m) when m < 1
50 cm deep) were established after fumigating the field with
methyl bromide plus chloropicrin (80 g/m2) on 30 October 1986 and
by burying thin pieces of concrete 50 cm deep into soil and raised DII(t) = K [1 + B exp(–r × t)]1/(1–m) when m > 1
0.3 m above ground. Microplots were fertilized prior to sowing
in which DII = disease intensity index, K = asymptote parameter,
each season with 35 g of a 8-15-15 (N-P-K) commercial fertilizer
B = constant of integration, r = rate parameter, m = shape
Microplots were artificially infested with three inoculum rates of
parameter, and t = time of disease assessment in days after the
each of F. oxysporum f. sp. ciceris races 0 (Foc-0) and 5 (Foc-5)
date of sowing. For this model, the disease intensity at the
before sowing in the first year of the experiment on 14 December
inflection point (DII(tip)) is given by DII(tip) = Km1/(1–m), and the
1986 (early winter sowing date), 18 February 1987 (late winter
time to reach this level (tip) is given by (26)
sowing date), and 30 March 1987 (early spring sowing date) or
maintained as uninfested controls. The experimental design con- tip = [ln(B) – ln(1 – m)]r–1 when m < 1
sisted of a randomized, split-split plot design arranged in four and
replicated blocks. The treatments comprised three levels of sowing
tip = [ln(B) – ln(m – 1)]r–1 when m > 1
date (main factor), two chickpea cultivars (subplot), three initial
inoculum rates (sub-subplot) of each of the two races of F. The least-squares program for nonlinear models (NLIN)
oxysporum f. sp. ciceris, and one noninfested control. For each procedure with Marquardt’s compromise method was used to
sowing date, cultivars were randomly allocated to a set of micro- obtain estimates of parameters in the model (version 6.08, SAS
plots and pathogen races and inoculum rates were randomized Institute Inc., Cary, NC). For analyses, nonzero points for DII in
within these microplots. experimental units and the average DII values of the four blocks
Microplots were sown on 16 December 1986 and 20 February were used. The coefficient of determination (R2), the mean square
and 2 April 1987 in year 1, 21 December 1987 and 3 February and error, the standard errors associated with the parameter estimates,
21 March 1988 in year 2, and 15 December 1988 and 31 January confidence intervals of predicted values, and pattern of the
and 16 March 1989 in year 3. Cvs. P-2245 and PV-61 are small- standardized residuals plotted against either predicted values or
seeded ‘kabuli’ (ram head shaped, beige seeds) chickpeas with the the independent variable were used to evaluate the appro-
first highly susceptible to both Foc-0 and Foc-5 and the latter priateness of the model to describe the data (8,26). Because
moderately susceptible to the two races (19). Seeds were treated disease progress curves fitted to versions of the Richards model
with tridemorph (11% Calixin WP, 0.66 g a.i./kg of seed; BASF with different values of the asymptote and shape parameters can-
Española S.A., Barcelona) and captan (85% Captan WP, not be compared directly, a common weighted mean rate para-
3 g a.i./kg of seed; Argos, Valencia, Spain) fungicides to eradicate meter (rho) (8,26,31) was calculated for each epidemic in the
infections by Didymella rabiei and to control Pythium seed rot study. This new rate parameter, named the weighted mean abso-
and preemergence damping-off, respectively (22,23). In each lute rate of disease increase (8), is defined by rho = Kr / (2m + 2),
microplot there were three rows 0.4 m apart and 0.2 m from the in which K is the asymptote parameter for DII, r is the rate
closest microplot edge barrier (25 seeds per row). In year 3, an ad- parameter, and m is the shape parameter in the Richards function
ditional large-seeded ‘kabuli’ chickpea cv. PV-60, more suscep- adjusted to a data set.
tible to Foc-0 and Foc-5 than cv. PV-61 but less susceptible than Disease development in a microplot was characterized by five
cv. P-2245, was used. For the experiment in year 3, each micro- variables associated to disease progress curves (30): (i) tis = the
plot was sown to cvs. P-2245, PV-61, and PV-60, one row each. time in days to initial symptoms, estimated as the number of days
Weeds in the microplots were removed by hand and dimethoate to reach a DII level of 0.05; (ii) tip = the time needed to reach the
(Romefos 40, 60 ml a.i./ha; Agrocrós S.A., Madrid) insecticide curve inflection point; (iii) final disease intensity (DIIfinal) = DII
was applied for control of leaf miner (Hylemiya sp.) (12) as observed at the final date of disease assessment; (iv) the
needed. Daily mean temperature and rainfall data were recorded at standardized area under disease progress curve (SAUDPC) calcu-
a weather station near the experimental site. lated by trapezoidal integration method standardized by time in
Inoculum of F. oxysporum. f. sp. ciceris isolates 7802 (race 0) days (8); and (v) rho = weighted mean absolute rate of disease
and 8012 (race 5) was increased in a cornmeal-sand mixture (43) increase. The time needed to reach the inflection point and rho
incubated at 25°C for 2 weeks. The upper 15-cm soil layer in a were obtained by the estimates of parameters of the Richards
microplot was excavated and mixed thoroughly with desired model fitted to DII progress data.
volumes of infested cornmeal-sand with a cement mixer to es- Chickpea yield in microplots was determined by the end of June
tablish three inoculum densities, and the soil was placed back in every crop season, ≈28, 21, and 14 weeks after chickpea sowing in
the corresponding microplots. The three rates of inoculum were early winter, late winter, and early spring, respectively. Seed yield
25.0 (low), 50.0 (intermediate), and 100.0 g (high) for Foc-0 and, (total seed weight per row in microplot) and 100-seed weight were
6.25 (low), 12.5 (intermediate), and 25.0 g (high) for Foc-5. determined for each microplot. Actual loss of seed yield was
Similar rates of noninfested cornmeal-sand were used for micro- determined as the difference between the total seed weight per
plots that served as controls. row in a microplot and the average seed weight per row in the
Assessment of disease and yield and data analyses. Disease corresponding control microplots. Also, seed yield in a microplot
incidence and severity were assessed at 7- to 10-day intervals. for a treatment was expressed in relative units as relative yield
Severity of symptoms on individual plants were rated on a scale (RY) by dividing seed weight per row (actual yield) by the average
1270 PHYTOPATHOLOGY
seed weight per row in the corresponding, uninfested control to 1988 experiment and was lowest in 1988 to 1989. The year also
microplots (attainable yield). influenced 100-seed weight but to a lesser degree than for seed
The effect of sowing date, host cultivar, and race of F. yield to sowing date (Table 1).
oxysporum f. sp. ciceris on chickpea yield was determined by Both the seed yield and 100-seed weight were affected by the
analysis of variance. Yield data for each season were analysed time of disease onset and development, but the effect on seed
with the general linear model (GLM) procedure of SAS. The three yield was larger than that on 100-seed weight in all treatments
initial inoculum levels showed variance homogeneity as deter- (Table 1). For each sowing date, seed yield loss was determined
mined by the Barlett’s test for equal variances. Therefore, initial primarily by virulence of the F. oxysporum f. sp. ciceris race and
inoculum rates and replications within them were considered ran- to a lesser extent by susceptibility of the chickpea cultivar. Seed
dom sources of variation. Separate analyses were performed for yield loss was higher with the highly virulent Foc-5 than with the
each year of experiment × chickpea cultivar combination. Mean less virulent Foc-0. This loss in seed yield was greater in cv. P-
comparisons among sowing dates within a cultivar were per- 2245 than in cvs. PV-60 and PV-61 (Table 1). Seed yield loss
formed according to Fisher’s protected least significant difference caused by Foc-5 averaged over sowing dates in the 3 years of
test at P = 0.05. The Dunnett’s two-tailed t test was used to test study was highest (99.7%) in the most susceptible cv. P-2245 and
significant differences between infested and noninfested control lowest (81.9%) in the least susceptible cv. PV-61. Similarly, seed
microplots at P = 0.05. yield loss caused by Foc-0 ranged from 65.6 to 30.6% for cvs. P-
Yield loss–disease intensity relationships were determined by 2245 and PV-61, respectively. Yield reduction by Fusarium wilt
regression analyses with the five disease progress curves- was also associated with poor seed size and quality, as indicated
associated variables as independent variables. For each year of by the correlation between seed yield loss and 100-seed weight
experiments, regression analyses were performed by data pooled (r = –0.697, P < 0.001).
over sowing dates and initial inoculum rates. Therefore, a single The quantitative relationship between the decrease in relative
model was produced for each chickpea cultivar and F. oxysporum chickpea seed yield and Fusarium wilt development was ex-
f. sp. ciceris race combination, except for the 1986 to 1987 experi- plained by a linear regression model that includes tis, tip, or DIIfinal
ment. For this latter experiment, no disease developed in Foc-0- as an independent variable and by an exponential regression
infested microplots, and available data for Foc-5 from both model that includes SAUDPC or rho as an independent variable.
chickpea cultivars were included in a single model. Two models Because yield was 0 in microplots infested with Foc-5, regression
for the decrease in relative yield with disease development were analyses could not be performed for treatments involving this
fitted to data race. Regression models were compared by the rate of relative
RY = a + rd X (1) seed yield decrease (rd). This rate was influenced primarily by the
race of the pathogen in soil and secondly by the chickpea cultivar
RY = exp[1 – exp(rd X)] (2)
used (Table 2).
in which RY is the relative yield, a is an intercept, rd is a rate of The relative seed yield was significantly (P < 0.05) correlated
RY decrease, and X is either tis, tip, DIIfinal, SAUDPC, or rho. A with tis and tip in the 1987 to 1988 and 1988 to 1989 experiments,
total of 60 regression analyses were performed by the least- but not in 1986 to 1987. In 1987 to 1988 and 1988 to 1989, the RY
squares procedure for linear (equation 1) and nonlinear models increased when disease onset (tis) and subsequent development
(equation 2). Coefficient of determination (R2), the mean square (tip) were delayed. Comparisons of rd estimates and parallelism
error, the asymptotic standard error associated with the estimated analyses (to test if the same rd value could be used for each chick-
parameter, and the pattern of the standardized residuals plotted pea cultivar × pathogen race combination) indicated that rd was
against either predicted values or the independent variable were not influenced by chickpea cultivar or pathogen race, but a com-
used to evaluate the appropriateness of a model to describe the mon rd for all cultivar × race combinations was detected (Table 2;
data. The standard errors of parameters obtained from regression Fig. 2). However, for similar tis or tip values and for Foc-0 and
analyses were used to compare the effects of chickpea cultivars, Foc-5 the reduction in RY was greater in cv. P-2245 than in cvs.
as well as, the effect of pathogen races on relative yield (8). PV-60 and PV-61 (Fig. 2). The parallel-lines model predicted a RY
increase of 0.60 and 0.37% per each 1-day delay in tis and of
RESULTS 0.55 and 0.33% per each 1-day delay in tip in 1987 to 1988 and
1988 to 1989, respectively (Fig. 2).
Analysis of disease progression. Over all three seasons, a total Relative yield decreased significantly (P < 0.05) with the
of 108 Fusarium wilt epidemics were available for analysis in this increase in the DIIfinal (Table 2; Fig. 3). The selected model
study. The Richards model adequately described all epidemics in explained 72 and 96% of the total RY variation in 1986 to 1987
the study and improved fit of disease progression data to model. and 1988 to 1989, respectively, and predicted 1.0% RY reduction
DII progress curves for 1988 to 1989 are illustrated in Figure 1. for each 1% increase in the DIIfinal (Fig. 3). In the 1988 to 1989
Mean parameter values estimated either with the fixed-shape or experiment, the rd in cv. P-2245 × Foc-0 interaction was signifi-
the Richards models yielded similar values (data not shown). A cantly (P < 0.05) greater than that in other cultivar × race com-
range of values for the Richards model shape parameter (m) were binations (Table 2), and the data set from this interaction showed a
obtained. Of the 108 DII progress curves in the study analyzed strong deviation from those of other interactions. Parallelism
with the Richards model, 11 had no inflection point (m = 0) and analyses of regression lines for interactions between cvs. PV-60
97 had an inflection point but were asymmetrical. Out of these 97 and PV-61 and Foc-0 or Foc-5 indicated a common rd among
DII progress curves, 93 had a positive skewness (0 < m < 2, i.e., them. The parallel-lines model explained 98% of the total RY
increase in the absolute rate was faster as the rate approached the variation and predicted a RY reduction of 1.2% for each 1% in-
inflection point) and only four had a negative skewness (m > 2) crease in the DIIfinal (Fig. 3).
(Fig. 1). Relative yield decreased according to a negative exponential
Yield loss–disease intensity relationships. Chickpea seed model with both increase in the weighted mean absolute rate of
yield (gram per microplot row) and 100-seed weight in uninfested DII increase (rho) and the SAUDPC. The rate of reduction in RY
control microplots were influenced by sowing date and varied (rd) was highest with low SAUDPC or rho and decreased as
among cultivars and years of experiment (Table 1). Overall, there SAUDPC or rho increased (Figs. 4 and 5). In 1987 to 1988 and
was a trend for seed yield to decrease as sowing date was delayed 1988 to 1989, rd was significantly higher (P < 0.05) with Foc-5
from early winter to early spring in the 1986 to 1987 and 1988 to than with Foc-0, irrespective of chickpea cultivar and independent
1989 experiments (P < 0.05). Seed yield was highest in the 1987 variables (Table 2). Comparisons of data fit to separate and
Vol. 90, No. 11, 2000 1271
parallel-curve models indicated that the fitted curves had no rd was higher (P < 0.05) for cv. P-2245 than for cvs. PV-60 and PV-
parameter in common (data not shown), except for regressions 61, irrespective of the F. oxysporum f. sp. ciceris race. These
using SAUDPC data from 1987 to 1988 for which data fitted best models explained 62 to 97% of the total RY variation (Table 2;
to the single-curve model (Fig. 4). For the remaining models, rd Figs. 4 and 5).

Fig. 1. Fusarium wilt progress on chickpea cvs. P-2245, PV-60, and PV-61 sown at three dates in microplots infested with three initial inoculum rates (IR) of
6.25 g (IR1), 12.5 g (IR2), and 25.0 g (IR3) per kg of soil for race 5, and 25.0 g (IR1), 50.0 g (IR2), and 100.0 g (IR3) per kg of soil for race 0 of Fusarium
oxysporum f. sp. ciceris races 0 and 5 for 1988 to 1989. Each point represents the mean disease intensity index of four replications. Solid lines represent the
predicted disease progress curves calculated by the Richards equation. Chickpeas were sown on: 15 December 1988 (early winter), 31 January 1989 (late
winter), and 16 March 1989 (early spring).

1272 PHYTOPATHOLOGY
TABLE 1. Effect of combinations of races of Fusarium oxysporum f. sp. ciceris, chickpea cultivars, and sowing dates on seed yield loss (YL) and yield componentsz
Seed yield (g/microplot row) 100-Seed weight (g)
Race 0 Race 5 Race 0 Race 5
Cultivar Sowing date Control Infected YL (%) Infected YL (%) Control Infected YL (%) Infected YL (%)
1986–1987
P-2245 Early winter 154.2 b … … 9.5 b* 94.7 ± 5.5 18.9 a … … 10.5 a* 45.5 ± 32.7
Late winter 184.2 a … … 39.9 a* 78.3 ± 11.0 18.2 a … … 15.0 a 17.1 ± 10.5
Early spring 112.7 c … … 30.1 a* 73.4 ± 15.1 18.0 a … … 16.2 a* 10.8 ± 2.8
PV-61 Early winter 133.3 a … … 78.7 b* 43.0 ± 15.1 27.9 a … … 25.0 b* 10.5 ± 2.7
Late winter 140.9 a … … 150.3 a –7.1 ± 6.5 25.4 b … … 27.8 a –9.2 ± 2.7
Early spring 103.0 b … … 91.7 b 11.1 ± 6.7 23.8 b … … 24.0 b –0.6 ± 6.5
1987–1988
P-2245 Early winter 279.2 a 233.1 a* 15.5 ± 19.5 2.4 a* 99.2 ± 1.0 18.3 b 18.5 a –1.1 ± 3.0 8.0 a* 56.2 ± 16.6
Late winter 210.3 b 172.1 b* 22.3 ± 12.8 0.0 b* 100 ± 0.0 18.5 b 17.5 a 7.3 ± 4.4 0.0 b* 100 ± 0.0
Early spring 180.1 c 25.5 c* 86.0 ± 5.2 0.9 a* 99.5 ± 0.8 22.6 a 11.0 b* 51.0 ± 19.0 18.3 a* 76.3 ± 23.2
PV-61 Early winter 349.7 a 319.0 a 8.8 ± 10.8 207.9 a* 40.4 ± 23.5 26.0 b 25.6 a 1.3 ± 0.2 21.5 a 17.5 ± 14.5
Late winter 248.4 b 232.3 b 6.6 ± 2.0 36.2 b* 85.5 ± 7.7 27.4 ab 25.7 a 5.7 ± 6.0 14.4 a* 47.4 ± 14.9
Early spring 177.2 c 93.8 c* 47.1 ± 20.6 16.2 b* 90.9 ± 11.8 29.4 a 23.5 a 20.2 ± 27.7 15.3 a* 47.8 ± 18.5
1988–1989
P-2245 Early winter 40.0 a 7.1 a* 83.8 ± 14.4 0.4 a* 99.1 ± 1.3 15.1 a 7.4 a* 50.3 ± 23.7 1.1 a* 92.3 ± 7.3
Late winter 37.9 a 2.4 a* 94.5 ± 9.3 0.0 b* 100 ± 0.0 19.5 a 1.9 a* 90.3 ± 12.5 0.0 b* 100 ± 0.0
Early spring 40.5 a 9.0 a* 77.4 ± 15.0 0.0 b* 100 ± 0.0 15.2 a 4.6 a* 68.1 ± 20.9 0.0 b* 100 ± 0.0
PV-61 Early winter 185.8 a 130.4 a 32.3 ± 19.4 39.4 a* 78.8 ± 1.7 23.4 a 22.6 a 1.7 ± 23.8 7.3 a* 68.5 ± 4.3
Late winter 149.2 ab 100.5 a 34.2 ± 29.8 3.8 b* 97.3 ± 2.5 28.3 a 21.5 a 23.6 ± 21.4 6.9 a* 74.6 ± 19.1
Early spring 107.3 b 48.0 a* 54.6 ± 8.9 1.7 b* 98.4 ± 1.5 22.4 a 13.2 b* 41.2 ± 17.7 5.2 a* 75.2 ± 15.6
PV-60 Early winter 158.0 a 114.4 a 26.1 ± 18.9 33.2 a* 79.0 ± 7.6 56.9 a 48.6 a 14.0 ± 15.5 23.8 a* 58.7 ± 15.9
Late winter 158.7 a 82.2 a* 46.4 ± 15.4 2.5 b* 98.5 ± 1.3 62.0 a 53.2 a 13.3 ± 14.3 10.3 a* 83.9 ± 8.1
Early spring 99.3 a 34.8 b* 64.3 ± 16.3 1.4 b* 98.5 ± 0.5 51.1 a 25.5 b* 50.2 ± 26.4 8.9 a* 83.3 ± 10.5
z Fumigated field soil in microplots were artificially infested with F. oxysporum f. sp. ciceris races 0 and 5 grown in cornmeal-sand or mixed with noninfested
substrate (Control). The infested substrate was mixed thoroughly with the upper 15-cm layer of soil. Microplots were sown on 16 December 1986 and 20
February and 2 April 1987 for 1986 to 1987; 21 December 1987 and 3 February and 21 March 1988 for 1987 to 1988; and 15 December 1988 and 31 January
and 16 March 1989 for 1988 to 1989. Data are the average of four blocks (microplots, 75 plants each) with three initial inoculum rates per block. Seed yield
and 100-seed weight were determined for each microplot row. YL was determined as the difference between the total seed weight per row in a microplot and
the average seed weight per row in the corresponding control microplots. Means in a column followed by the same letter for each cultivar within an experi-
mental period are not significantly different according to Fisher’s protected LSD (P = 0.05). Means in a row followed by an asterisk are significantly smaller
(P < 0.05) than the mean for the corresponding uninfested control according to Dunnett’s test.

TABLE 2. Relationship between relative seed yield of chickpea cvs. P-2245, PV-60, and PV-61 and disease progress curve-associated variables of Fusarium wilt
epidemics developed in microplots artificially infested with races 0 and 5 of Fusarium oxysporum f. sp. cicerisy
Disease progress curve-associated variablez
tis tip DIIfinal SAUDPC rho
Race Cultivar rd ± SE R2 rd ± SE R2 rd ± SE R2 rd ± SE R2 rd ± SE R2
1986–1987
Foc-5 P-2245, PV-61 NC … NC … –1.04 ± 0.16 0.72 2.15 ± 0.28 0.71 87.44 ± 18.88 0.36
1987–1988
Foc-5 P-2245 … … … … … … … … … …
PV-61 0.0074 ± 0.0011 0.86 0.0055 ± 0.0009 0.84 –1.22 ± 0.15 0.90 2.06 ± 0.27** 0.63 53.57 ± 3.30** 0.94
Foc-0 P-2245 0.0069 ± 0.0008 0.93 0.0066 ± 0.0009 0.90 –0.99 ± 0.09 0.96 1.99 ± 0.30 0.90 34.95 ± 5.38 0.90
PV-61 0.0042 ± 0.0010 0.73 0.0041 ± 0.0010 0.72 –0.95 ± 0.04 0.99 1.46 ± 0.07 0.97 29.96 ± 2.26 0.93
1988–1989
Foc-5 P-2245 … … … … … … … … … …
PV-60 0.0034 ± 0.0006 0.82 0.0031 ± 0.0006 0.82 –1.37 ± 0.12 0.95 2.28 ± 0.09** 0.85 53.84 ± 1.19** 0.96
PV-61 0.0032 ± 0.0003 0.94 0.0030 ± 0.0003 0.92 –1.33 ± 0.14 0.93 2.00 ± 0.09** 0.78 52.97 ± 3.10** 0.70
Foc-0 P-2245 0.0031 ± 0.0012 0.48 0.0029 ± 0.0010 0.53 –2.14 ± 0.30* 0.88 2.19 ± 0.15* 0.63 58.38 ± 4.21* 0.62
PV-60 0.0047 ± 0.0011 0.71 0.0039 ± 0.0009 0.72 –1.15 ± 0.13 0.92 1.66 ± 0.11 0.81 39.36 ± 3.45 0.65
PV-61 0.0038 ± 0.0015 0.48 0.0037 ± 0.0013 0.55 –1.12 ± 0.15 0.89 1.22 ± 0.11 0.68 31.51 ± 2.95 0.67
y Fumigated field soil in microplots artificially infested with F. oxysporum f. sp. ciceris races 0 (Foc-0) and 5 (Foc-5) grown in cornmeal-sand or mixed with
noninfested substrate (Control). The infested substrate was mixed thoroughly with the upper 15-cm layer of soil. Microplots were sown on 16 December 1986
and 20 February and 2 April 1987 for 1986 to 1987; 21 December 1987 and 3 February and 21 March 1988 for 1987 to 1988; and 15 December 1988 and 31
January and 16 March 1989 for 1988 to 1989.
z tis = Time in days to initial symptoms, estimated as the number of days to reach a disease intensity index (DII) level of 0.05; tip = the time needed to reach the
curve inflection point; DIIfinal = disease intensity index observed at the final date of disease assessment; SAUDPC = area under disease progress curve
standardized by time in days; rho = weighted mean absolute rate of disease progression. The time needed to reach the inflection point and rho were obtained
by the estimates of parameters of the Richards model fitted to DII progress data. For each year of experiments, regression models were performed by pooling
data over sowing dates and initial inoculum rates. rd = relative rate of seed yield decrease; SE = standard error; and R2 = coefficient of determination. The
standard errors of the rd obtained from regression analyses were used to compare the effects of experimental treatments. NC indicates the relative seed yield
was not significantly correlated with tis and tip (P = 0.05). No regression analyses were performed on Foc-5 cv. P-2245 because severe disease in microplots
resulted in no seed yield for most cases. For each experimental period and F. oxysporum f. sp. ciceris race, the rd value for a cultivar in a column followed by
an asterisk is significantly higher (P < 0.05) than the corresponding values for other cultivars. For each experimental period and chickpea cultivar, ** indicates
the rd value for a pathogen race in a column is significantly higher (P < 0.05) than the corresponding rd at the other pathogen race.

Vol. 90, No. 11, 2000 1273

 Based on the single regression analyses, a response surface for onset was at least 120 days after sowing or rho < 0.002. Con-
RY as a function of both tis and rho was developed with original versely, RY reduction of the same cultivars caused by the highly
data from 1987 to 1988 and 1988 to 1989 and the equation virulent Foc-5 ranged from 75 to 100% for epidemics with early on-
RY (tis, rho) = (c1 + c2tis) × {exp[1 – exp(c3rho)]} (3)
set, i.e., tis < 40 days after sowing, because rho was >0.016 (Fig. 6).

In this equation, RY is the relative seed yield, c1 is an intercept, DISCUSSION

and c2 and c3 are rates of RY decrease. Parameters (c1 – c3) in that
equation were simultaneously estimated for each chickpea culti- Fusarium wilt of chickpeas can be managed by adjusting the
var × F. oxysporum f. sp. ciceris race combination. Because the sowing date (17,36) but the efficiency of this disease management
single model analysis indicated no effects of tis on rd estimates practice may be influenced by factors in the pathosystem deter-
(Table 2), regression analyses were carried out with common c2 mining the disease development (30). No prior quantitative studies
parameter for each year of the study, and the effects on c1 and c3 have been carried out on the relationship between Fusarium wilt
were separated. The reduction in RY was influenced both by the development and chickpea yield loss. The aim of this work was
nature of the F. oxysporum f. sp. ciceris race and susceptibility of primarily to relate seed yield losses to Fusarium wilt development,
the chickpea cultivar (Fig. 6). In the cultivar × race combination as influenced by several components in the pathosystem, i.e., viru-
most conducive for disease development (cv. P-2245 × Foc-5), lence of the pathogen race, susceptibility of the host cultivar and
100% seed yield loss (maximum RY value) occurred even with a environment as determined by sowing date.
very late epidemic onset, i.e., 90 days after sowing or very low In our study, the amount of Fusarium wilt varied considerably
rho of 0.012. On the contrary, reduction in RY of cvs. PV-60 and among the 3 years of the study. The differences in the levels of
PV-61 caused by the less virulent Foc-0 was small when disease disease intensity may be due in part to differences in weather

Fig. 2. Relationship between time in days to initial symptoms and the Fig. 3. Relationship between final disease intensity index and relative seed
relative yield of chickpea cvs. P-2245 (black circles), PV-60 (light gray yield of chickpea cvs. P-2245 (black circles), PV-60 (light gray squares), and
squares), and PV-61 (dark gray triangles) grown in microplots artificially PV-61 (dark gray triangles) grown in microplots artificially infested with
infested with Fusarium oxysporum f. sp. ciceris races 0 (Foc-0, dashed line Fusarium oxysporum f. sp. ciceris races 0 (Foc-0, dashed line and open
and open symbols) and 5 (Foc-5, solid line and solid symbols) for three symbols) and 5 (Foc-5, solid line and solid symbols) for three experimental
experimental periods. Each point is the mean of data from four microplots periods. Each point is the mean of data from four microplots (75 plants
(75 plants each). For each year of experiments, regression analyses were each). For each year of experiments, regression analyses were performed by
performed by pooling data over sowing dates and initial inoculum rates. pooling data over sowing dates and initial inoculum rates. Solid and dashed
Solid and dashed lines represent the predicted models with common rate of lines represent the predicted model with common rate of relative yield
relative yield decrease according to parallel-lines analyses. decrease estimated by parallel-lines regression analyses.

1274 PHYTOPATHOLOGY
conditions and the increase in the pathogen population in soil as a Richards model. Overall, disease progression was better described
result of successive sowings of susceptible chickpeas in micro- by a sigmoid-shape model, asymmetric or in a few cases sym-
plots. The annual variation in severity of Fusarium wilt of chick- metric. The monomolecular shape (i.e., the Richards shape para-
peas is often attributed to differences in temperature and inoculum meter close to 0) was appropriate only for epidemics on suscep-
density (13,29). For each cropping season, changes in Fusarium tible cv. P-2245, caused by the highly virulent race (Foc-5) and
wilt development are related mainly to date of sowing, cultivar the presence of conducive weather (spring sowing date). The
susceptibility, pathogen race virulence, and their interactions. For Richards model offers a continuous range of inflection points and
each chickpea cultivar × pathogen race combination, the differ- thus provides an alternative to choosing different models for dif-
ences in epidemics that developed in the three sowing dates can ferent epidemics (31). However, this model has been used in-
be explained by differences in rainfall and temperature. Delaying frequently to describe plant disease progression. In this study, it
sowing from the middle of December to the middle of March re- enabled analysis of all shapes of Fusarium wilt progress curves
sults in decreased soil moisture and increased temperature that produced. The consistently good fit provided by the Richards
favour development of Fusarium wilt (4,41). model and the well defined epidemiological concepts corres-
Progression of Fusarium wilt intensity in chickpea cultivars ponding to the parameters, make it ideal for describing disease
sown at different times in soil infested with races 0 or 5 of F. progress of Fusarium wilt epidemics in chickpeas.
oxysporum f. sp. ciceris could not be described by either the Several epidemics caused by soilborne plant pathogens have
fixed-shape models or the Richards model. On the contrary, a been appropriately described by a sigmoidal disease progress
wide range of values for the shape parameter was estimated by the model (2,10,40). Deviations of Fusarium wilt of chickpea from

Fig. 4. Relationship between the standardized area under Fusarium wilt Fig. 5. Relationship between the Richards weighted mean absolute rate of
intensity progress curve and the relative seed yield of chickpea cvs. P-2245 Fusarium wilt intensity index progress curve and the relative seed yield of
(black circles), PV-60 (light gray squares), and PV-61 (dark gray triangles) chickpea cvs. P-2245 (black circles), PV-60 (light gray squares), and PV-61
grown in microplots artificially infested with Fusarium oxysporum f. sp. (dark gray triangles) grown in microplots artificially infested with Fusarium
ciceris races 0 (Foc-0, dashed line and open symbols) and 5 (Foc-5, solid line oxysporum f. sp. ciceris races 0 (Foc-0, dashed line and open symbols) and 5
and solid symbols) for three experimental periods. Each point is the mean of (Foc-5, solid line and solid symbols) for three experimental periods. Each
data from four microplots 75 plants each). For each year of experiments, point is the mean of data from four microplots (75 plants each). For each year
regression analyses were performed by pooling data over sowing dates and of experiments, regression analyses were performed by pooling data over
initial inoculum rates. The solid and dashed lines represent the predicted sowing dates and initial inoculum rates. Solid and dashed lines represent the
model with common (1987 to 1988) or separate (1988 to 1989) rate of rela- predicted model with separate rate of relative yield decrease estimated by
tive yield decrease estimated by parallel-curves regression analyses. parallel-curves regression analyses.

Vol. 90, No. 11, 2000 1275

the monomolecular shape, which involve an acceleration of might have an influence on these modifications. Disease progress
disease during early stages of the epidemic, could be due to of a theoretical ‘monocyclic disease’ in the field can be modified
modifications of disease progression by weather, cultivar suscepti- by environment (5), host susceptibility (34), environment–host
bility, or pathogen virulence. Also, the assessment of disease by genotype interactions (40), and consequently, is poorly described
the severity of foliar symptoms compared with disease incidence, by the monomolecular function (33). Sporulation of F. oxysporum

Fig. 6. Response surfaces as a function of time to initial symptoms and the Richards weighted mean absolute rate (rho) of Fusarium wilt intensity index progress
curve for the decrease in relative seed yield of chickpea cvs. P-2245 and PV-61 grown in microplots artificially infested with Fusarium oxysporum f. sp. ciceris
races 0 (Foc-0) and 5 (Foc-5) for 1988 to 1989. Regression analyses were performed by pooling data over sowing dates and initial inoculum rates.

1276 PHYTOPATHOLOGY
f. sp. ciceris on chickpea stem surfaces under our field conditions fluenced by tis and rho. This response surface confirmed that in
has not been observed unlike F. oxysporum f. sp. lycopersici (24), the chickpea–F. oxysporum f. sp. ciceris pathosystem the relative
which could play a role as secondary inoculum. Furthermore, if seed yield loss increases in a linear relationship with increasing tis
soilborne or airborne inoculum of F. oxysporum f. sp. ciceris were and decreases in a negative exponential relationship with in-
produced, it is unlikely that they might play a role in disease creases in rho. A similar model was developed for stem rust of
progression because early infection of chickpea seedlings by the wheat caused by Puccinia graminis f. sp. tritici, where yield loss
pathogen must take place for severe symptoms to develop (20). was estimated as a function of the rate of disease increase and
Thus, a number of factors involved in the appearance of foliar growth stage of the host at time of epidemic onset (7). Models
symptoms of the disease (i.e., conduciveness of weather, cultivar represented by a response surface provide a conceptual framework
susceptibility, virulence of race prevalent in soil, and their based on a knowledge of disease epidemiology and crop physiol-
interactions) may be responsible for the observed deviations in the ogy for modelling disease-loss systems (42). Our model incor-
disease progress curves. porates the effect of chickpea cultivar susceptibility and pathogen
Results in uninfested microplots indicate that the chickpea yield virulence, thereby accounting for the effects of different disease
is determined by the time of planting, with decreasing seed yield progress curve-associated variables on seed yield and providing a
as planting is delayed from early winter to early spring. Reductions better understanding of the chickpea–F. oxysporum f. sp. ciceris
in seed yield by the disease was larger than reductions in 100-seed interaction. However, this model does not include the influence of
weight. Therefore, the overall yield loss caused by the disease other variables such as plant stand, weather, soil type, or plant
may be attributed mainly to a significant decrease in the number density, all of which relate to healthy chickpea yield potential.
of seeds per plant and to a lesser extent to reduced mean seed This would also make the model unwieldy.
weight. Similar effects were reported for field beans due to infec- In the Mediterranean region, chickpeas are traditionally planted
tions by Botrytis fabae (11). Irrespective of sowing date, seed yield in the spring and the crop develops on the residual moisture in soil
in infested microplots was determined by virulence of the F. oxy- from winter rains. As the season proceeds, the crop experiences
sporum f. sp. ciceris race and susceptibility of the chickpea cultivar. rising temperatures and increasing soil moisture stress that shorten
Our results show a significant relationship between chickpea the vegetative and reproductive periods and decrease yields
seed yield and Fusarium wilt disease progress curve-associated (37,39). Fusarium wilt incidence and severity are enhanced by
variables, by single-point models, (i.e., the time of disease onset warm, dry soils occurring in spring-sown crops (13,43,44). Planting
(tis) and development (tip), or the disease intensity index at the end in winter corresponds with optimum environmental conditions for
of the season (DIIfinal)) integral models, (i.e., SAUDPC and the chickpea growth, ensures better use of available water, and increases
weighted mean absolute rate of disease increase (rho)), or surface- yield (39). Also, under conditions in southern Spain, planting in
response models. Seed yield increased linearly with increasing tis early winter instead of early spring significantly delays epidemic
and tip, and decreased linearly with increasing DIIfinal. The rate of onset, slows the rate of development, and reduces the final amount
seed yield loss (rd) attributable to either chickpea cultivar or of disease (30). Experiments conducted in India showed that
pathogen race was not observed. However, the intercepts of the Fusarium wilt intensity decreased and chickpea seed yield in-
seed yield loss linear models developed varied for cultivar × race creased in plantings delayed until the middle of October (17,36).
combinations. Therefore, for similar tis, tip, or DIIfinal values and However, no assessments were made in these studies of the in-
for the two F. oxysporum f. sp. ciceris races included in the study, fluence of pathogen virulence and cultivar susceptibility. Our results
yield loss was higher in cv. P-2245 than in cvs. PV-60 and PV-61. confirm that planting time determines the severity of Fusarium
Work by Haware and Nene (15) in India showed that yield loss wilt and the quantum of chickpea seed yield.
caused by Fusarium wilt in ‘desi’ cultivars planted in early
October was influenced by the growth stage at which the disease ACKNOWLEDGMENTS
developed. However, no quantitative relationship was demon-
strated between yield loss and disease onset or development, and This work was supported by grants AGR89-0533-C02-01 and AGF97-
no assessment was made concerning the influence of pathogen 1479 from Comisión Interministerial de Ciencia y Tecnología (CICYT)
race virulence on that effect. In that study (15), early wilting of Spain and HA 1997-0033 from Spanish Ministerio de Educación y
reduced the seed number per plant and caused more yield loss Cultura. We thank H. F. Rapoport for the helpful comments and sug-
gestions on the manuscript and the Senior Editor and the anonymous
than late wilting, whereas the latter produced less seed and caused reviewers for editorial improvement.
substantial yield loss. One hundred-seed weight was also ad-
versely affected by wilt in all cultivars in the study even when
LITERATURE CITED
plants showed symptoms at the preharvest stage. In Verticillium
wilt of cotton, a similar pathosystem, the effect of epidemics on 1. Bejarano-Alcázar, J., Blanco-López, M. A., Melero-Vara, J. M., and
cotton yield was related to the phenological stage of plants when Jiménez-Díaz, R. M. 1997. The influence of Verticillium wilt epidemics
foliar symptoms first developed and to virulence of the on cotton yield in southern Spain. Plant Pathol. 46:168-178.
Verticillium dahliae pathotype prevalent in soil (1,35). 2. Bejarano-Alcázar, J., Melero-Vara, J. M., Blanco-López, M. A., and
Jiménez-Díaz, R. M. 1995. Influence of inoculum density of defoliating
Reduction of chickpea seed yield by Fusarium wilt develop- and nondefoliating pathotypes of Verticillium dahliae on epidemics of
ment followed a negative exponential model with the increase in Verticillium wilt of cotton in southern Spain. Phytopathology 85:1474-
SAUDPC or rho. In contrast to linear models related to tis, tip, and 1481.
DIIfinal, the rate of yield loss for models including SAUDPC or rho 3. Berger, R. D. 1981. Comparison of the Gompertz and logistic equations
depended on virulence of the F. oxysporum f. sp. ciceris race and to describe disease progress. Phytopathology 71:716-719.
4. Bhatti, M. A., and Kraft, J. M. 1992. Effects of inoculum density and
susceptibility of the chickpea cultivar. In all cases, the greatest
temperature on root rot and wilt of chickpea. Plant Dis. 76:50-54.
absolute rate of decrease in seed yield occurred at low SAUDPC 5. Bloomberg, W. J. 1979. A model of damping-off and root rot of Douglas-fir
or rho, and the absolute rate of decrease declined in absolute value seedlings caused by Fusarium oxysporum. Phytopathology 69:74-81.
as SAUDPC or rho increased. This type of model, a type I curve 6. Burleigh, J. R., Roelfs, A. P., and Eversmeyer, M. G. 1972. Estimating
(28), suggests that the remaining, apparently healthy green tissue damage to wheat caused by Puccinia recondita tritici. Phytopathology
of the host is affected by the pathogen (21). SAUDPC and 62:944-946.
7. Calpouzos, L., Roelfs, A. P., Madson, M. E., Matin, F. B., Welsh, J. R.,
AUDPC have been used as a predictor of yield losses for several and Wilcoxon, R. D. 1976. A new model to measure yield losses caused
pathosystems (8,26,45). by stem rust in spring wheat. Univ. Minn. Agric. Exp. Stn. Tech. Bull.
Results from the present study allowed the development of a 307.
response-surface function for relative seed yield decrease as in- 8. Campbell, C. L., and Madden, L. V. 1990. Introduction to Plant Disease

Vol. 90, No. 11, 2000 1277

 Epidemiology. John Wiley & Sons, New York. 27. Madden, L. V., and Nutter, F. W., Jr. 1995. Modeling crop loss at the
9. Campbell, C. L., Pennypacker, S. P., and Madden, L. V. 1980. Progression field scale. Can. J. Plant Pathol. 17:124-137.
of dynamics of hypocotyl rot of snapbean. Phytopathology 70:487-494. 28. Mundford, J. D., and Norton, G. A. 1987. Economics of integrated pest
10. Colbach, N., Lucas, P., and Meynard, J.-M. 1997. Influence of crop control. Pages 191-200 in: Crop Loss Assessment and Pest Management.
management on take-all development and disease cycles on winter P. S. Teng, ed. The American Phytopathological Society, St. Paul, MN.
wheat. Phytopathology 87:26-32. 29. Navas-Cortés, J. A., Alcalá-Jiménez, A. R., Hau, B., and Jiménez-Díaz,
11. Griffiths, E., and Amim, S. M. 1977. Effects of Botrytis fabae infection R. M. 2000. Influence of inoculum density of races 0 and 5 of Fusarium
and mechanical defoliation on seed yield of field beans (Vicia faba). oxysporum f. sp. ciceris on development of Fusarium wilt in chickpea
Ann. Appl. Biol. 86:359-367. cultivars. Eur. J. Plant Pathol. 106:135-146.
12. Guerrero, A. 1990. Garbanzos. Pages 504-514 in: Cultivos Herbáceos 30. Navas-Cortés, J. A., Hau, B., and Jiménez-Díaz, R. M. 1998. Effect of
Extensivos. A. Guerrero, ed. Ediciones Mundi-Prensa, Madrid. sowing date, host cultivar, and race of Fusarium oxysporum f. sp. ciceris
13. Gupta, O., Kotasthane, S. R., and Khare, M. N. 1987. Factors in- on development of Fusarium wilt of chickpea. Phytopathology 88:1338-
fluencing epidemiology of vascular wilt of chickpea. Proc. Natl. Acad. 1346.
Sci. India 57:86-91. 31. Park, E. W., and Lim, S. M. 1985. Empirical estimation of the asymp-
14. Halila, M. H., and Strange, R. N. 1996. Identification of the causal agent totes of disease progress curves and the use of the Richards generalized
of wilt of chickpea in Tunisia as Fusarium oxysporum f. sp. ciceri race 0. rate parameters for describing disease progress. Phytopathology 75:786-
Phytopathol. Mediterr. 35:67-74. 791.
15. Haware, M. P., and Nene, Y. L. 1980. Influence of wilt at different stages 32. Pennypacker, S. P., Knoble, H. D., Antle, C. E., and Madden, L. V. 1980.
on the yield loss in chickpea. Trop. Grain Legume Bull. 19:38-40. A flexible model for studying plant disease progression. Phytopathology
16. Haware, M. P., and Nene, Y. L. 1982. Races of Fusarium oxysporum f. 70:232-235.
sp. ciceri. Plant Dis. 66:809-810. 33. Pfender, W. F. 1982. Monocyclic or polycyclic root diseases: Dis-
17. Jalali, B. L., and Chand, H. 1992. Chickpea wilt. Pages 429-444 in: tinguishing between the nature of the disease cycle and the shape of the
Plant Diseases of International Importance. Vol. 1. Diseases of Cereals disease progress curve. Phytopathology 72:31-32.
and Pulses. U. S. Singh, A. N. Mukhopadhayay, J. Kumar, and H. S. 34. Populer, C. 1978. Changes in host susceptibility with time. Pages 239-
Chaube, eds. Prentice Hall, Englewood Cliffs, NJ. 262 in: Plant Disease an Advanced Treatise, Vol. 2. J. G. Horsfall and E.
18. James, W. C. 1974. Assessment of plant diseases and losses. Annu. Rev. B. Cowling, eds. Academic Press, New York.
Phytopathol. 12:27-48. 35. Pullman, G. S., and DeVay, J. E. 1982. Epidemiology of Verticillium wilt
19. Jiménez-Díaz, R. M., Alcalá-Jiménez, A. R., Hervás, A., and Trapero- of cotton: Effects of disease development on plant phenology and lint
Casas, J. L. 1993. Pathogenic variability and host resistance in the yield. Phytopathology 72:554-559.
Fusarium oxysporum f. sp. ciceris/Cicer arietinum pathosystem. Pages 36. Saraf, C. S. 1974. Agronomic management to reduce gram wilt inci-
87-94 in: Fusarium Mycotoxins, Taxonomy, Pathogenicity, Host Resis- dence. Indian J. Genet. Plant Breed. 34:263-266.
tance. Proc. Eur. Seminar, 3rd ed. E. Arseniuk and T. Goral, eds. Plant 37. Saxena, M. C. 1987. Agronomy of chickpea. Pages 207-232 in: The
Breed. Acclim. Inst., Radzikóv, Poland. Chickpea. M. C. Saxena and K. B. Singh, eds. CAB International,
20. Jiménez-Díaz, R. M., Basallote-Ureba, M. J., and Rapoport, H. 1989. Wallingford, U.K.
Colonization and pathogenesis in chickpeas infected by races of Fu- 38. Saxena, M. C. 1990. Problems and potential of chickpea production in
sarium oxysporum f. sp. ciceri. Pages 113-121 in: Vascular Wilt Diseases the nineties. Pages 13-27 in: Chickpea in the Nineties. Proc. Int.
of Plants. NATO ASI Series, Vol. H28. E. C. Tjamos and C. Beckman, Workshop Chickpea Improvement, 2nd. International Crops Research
eds. Springer-Verlag, Berlin. Institute for the Semi-Arid Tropics, Andhra Pradesh, India.
21. Johnson, K. B. 1987. Defoliation, disease, and growth: A reply. 39. Singh, K. B., and Saxena, M. C. 1996. Winter chickpea in mediter-
Phytopathology 77:1495-1497. ranean-type environments. International Center for Agricultural Re-
22. Kaiser, W. J. 1984. Control of Ascochyta blight of chickpea through search in Dry Areas, Aleppo, Syria.
clean seed. Pages 117-122 in: Ascochyta Blight and Winter Sowing of 40. Stack, R. W. 1980. Disease progression in common root rot of spring
Chickpeas. M. C. Saxena and K. B. Singh, eds. Martinus-Nijhoff wheat and barley. Can. J. Plant Pathol. 2:187-193.
Publishing, Dordrecht, Netherlands. 41. Sugha, S. K., Kapoor, S. K., and Singh, B. M. 1994. Factors influencing
23. Kaiser, W. J., and Hannan, R. M. 1983. Etiology and control of seed Fusarium wilt of chickpea (Cicer arietinum L.). Indian J. Mycol. Plant
decay and preemergence damping-off of chickpea by Pythium ultimum. Pathol. 24:97-102.
Plant Dis. 67:77-81. 42. Teng, P. S., and Gaunt, R. E. 1981. Modelling systems of disease and
24. Katan, T., Shlevin, E., and Katan, J. 1997. Sporulation of Fusarium yield loss in cereals. Agric. Syst. 6:131-154.
oxysporum f. sp. lycopersici on stem surfaces of tomato plants and aerial 43. Trapero-Casas, A., and Jiménez-Díaz, R. M. 1985. Fungal wilt and root
dissemination of inoculum. Phytopathology 87:712-719. rot diseases of chickpea in southern Spain. Phytopathology 75:1146-1151.
25. Kranz, J. 1974. Comparison of epidemics. Annu. Rev. Phytopathol. 44. Westerlund, F. V., Jr., Campbell, R. N., and Kimble, K. A. 1973. Fungal
12:355-374. root rots and wilt of chickpea in California. Phytopathology 63:432-436.
26. Madden, L. V., and Campbell, C. L. 1990. Nonlinear disease progress 45. Zhou, Y., Fitt, B. D. L., Welman, S. J., Gladders, P., Sansford, C. E., and
curves. Pages 181-229 in: Epidemics of Plant Diseases, Mathematical West, J. S. 1999. Effects of severity and timing of stem canker
Analysis, and Modeling. 2nd ed. Ecological Studies 13. J. Kranz, ed. (Leptosphaeria maculans) symptoms on yield of winter oilseed rape
Springer-Verlag, Berlin. (Brassica napus) in the UK. Eur. J. Plant Pathol. 105:432-436.

1278 PHYTOPATHOLOGY