Mepiquat Chloride Effects on
Irrigated Cotton in Arizona
E.J. Norton and J.C. Silvertooth
Abstract
A series of experiments have been conducted from 1988 to 1999 at various
locations across the cotton producing regions of Arizona to evaluate mepiquat
chloride (MC) applications in terms of plant growth and yield. These
experiments were designed to evaluate MC under three application regimes.
These regimes included low rate multiple applications, late season applications,
and a feedback vs. scheduled management of MC and nitrogen (N) applications.
The objective of this summary (including a total of 31 site-years) is to determine
which of these three application regimes offer the greatest opportunity for a
positive lint yield response to MC. Stability analysis was conducted by
regressing the treatment mean lint yield against the environmental mean for
each application regime. Results from the stability analyses revealed that the
most viable method of application is a feedback approach for both MC and
fertilizer N. The most reliable technique associated with plant assessment in a
feedback approach was the height to node ratio (HNR) to indicate vegetative
tendencies for determining the appropriate rate and timing of MC applications.
Introduction
The growth and development of crop plants can usually be characterized by a period of vegetative growth followed
by a reproductive growth phase. During these periods of growth, photosynthetic products are partitioned into
various plant parts. The degree to which plants allocate photosynthates between vegetative and reproductive
structures determines in large part the growth rate and yield of many economically important crops (Brown, 1984).
The cotton (Gossypium spp.) plant is somewhat unique by having a perennial nature and indeterminate growth
pattern in comparison to many other more determinate crop plants. Ramifications of this growth habit are that
vegetative and reproductive growth occur simultaneously on the cotton plant and thus compete for photosynthetic
products (Brown, 1984). This presents specific challenges when managing the cotton plant as an annual in a crop
production system. Excessive vegetative tendencies in cotton have been shown to lead to losses in reproductive
structures (squares, flowers, and bolls) (Gausman, et al., 1979; York, 1983b; Fletcher et al., 1994). The loss of
reproductive structures (carbohydrate sinks) can shift energy from reproductive to vegetative portions of the plant,
resulting in a rapid proliferation of main stem growth (Mauney, 1986). Self-shading may also contribute to the loss
of reproductive structures in some cases due to the fact that the major portion of the assimilate supply for these
structures is obtained from the subtending leaf to the reproductive structure (Ashley, 1972; Benedict and Kohel,
1975). When these leaves are shaded by excess vegetative growth, assimilate supply is depleted due to decreased
photosynthetic ability and the abortion of fruiting structures can result (Mauney, 1986; York, 1983b). The most
likely cause of this reproductive structure loss is not one single event, however, but combinations of several factors
and processes working in concert. Collectively, these physiological processes can serve to initiate a process of
reproductive structure abortion and a decline in overall fruit retention (FR) on the plant (Guinn, 1982).
Maintaining an optimum fruit load in relation to vegetative structure is often difficult. Research efforts have been
directed to developing methods and indices to monitor cotton plants on an in-season basis. Indices that have been
This is part of the 2000 Arizona Cotton Report, The University of Arizona College of Agriculture, index at
http://ag.arizona.edu/pubs/crops/az1170/
developed for cotton include a height to node ratio (HNR), fruit retention (FR), main stem node numbers (MSN),
growth rate (GR), and nodes above top white flower (NAWF) (Bourland et al., 1992; Kerby et al., 1997; Kerby et
al., 1998). These indices are more effective in terms of management when they are made region- and area-specific.
A set of indices have been developed for Arizona (Silvertooth et al., 1993b; Silvertooth, 1994; Silvertooth and
Norton, 1998c). The use of these indices to monitor the plant in-season has many practical implications. Foremost
among these is the fact that it can alert the grower to any change in the reproductive to vegetative balance in the
plant that may prove detrimental to yield. It may then be decided if corrective measures are needed.
Mepiquat chloride (MC) is a plant growth regulator that has been used in cotton production for several decades as a
management tool in controlling vegetative growth. Mepiquat chloride is a gibberellic acid suppressant that is
absorbed by the green portions of the plant and serves to reduce cell elongation, thus offering the potential of
decreasing leaf area and restricting additional plant height increases (York, 1983a and Kerby, 1985). Mepiquat
chloride has also been associated with enhancing earliness with regards to fruiting development (Walter, et al., 1980;
York, 1983b; Kerby, 1985).
Much research has been devoted to determining optimum rates and MC application regimes for irrigated cotton
production (Kerby, 1985; McConnell, et al., 1992; Boman and Westerman, 1994). However, application strategies
that result in consistently significant increases in lint yield from MC have yet to be identified or demonstrated.
These inconsistencies in yield response have been attributed to various factors, including length of season, lack of
excess vegetative growth, and stresses (including drought, fertility, insect, and disease) (Briggs, 1980; Crawford,
1981).
A series of MC related studies have been conducted in recent years in Arizona. Low rate multiple (LRM)
application studies were conducted across the cotton producing regions of Arizona (Silvertooth et al., 1989, 1990,
1991a, 1993a, 1995b, 1997b; Hood and Silvertooth, 1993; Husman and Silvertooth, 1994). The objectives of the
studies were twofold. First, to measure any response to MC such as plant height (PH), MSN, HNR, NAWF, and
percent canopy closure (PCC). Secondly, to determine yield responses with multiple applications of MC. From
these results a final objective was to determine optimal rates and application regimes of multiple MC applications
for irrigated Upland (G. hirsutum L.) and American Pima (G. barbadense L.) cotton.
Field studies were conducted at various locations across Arizona to evaluate the effects of late-season (LS)
applications of MC on Upland cotton (Silvertooth et al., 1993a; Silvertooth and Norton, 1995b, 1997b). These
studies were designed to examine the effects of the late-season applications not only on yield but on plant
characteristics associated with earliness as indicated by NAWF, HNR, and FR.
In addition, a series of feedback versus scheduled applications of N and MC interaction studies were conducted at
various locations throughout Arizona (Fletcher et al., 1994; Silvertooth and Norton, 1995a, 1996a, 1997a, 1998a;
Norton et al., 1999). The objectives of these studies were to use plant growth parameters such as measured inseason HNR and FR in a feedback approach to N and MC applications and to evaluate this approach versus a
scheduled approach to N and MC applications. A feedback approach to crop input management, such as with MC, is
conducted in reference to established criteria for optimum or excessive vegetative growth patterns. Accordingly,
guidelines relative to FR levels and HNR have been developed for this purpose (Silvertooth et al., 1991b, 1992a;
Fletcher et al., 1994; Silvertooth et al., 1995a; Silvertooth and Norton, 1996a, 1997a, 1998a, 1998c; Norton et al.,
1999). Management guidelines for fertilizer N inputs to cotton have also developed along a similar line of rationale
regarding feedback versus scheduled approaches (Silvertooth et al., 1991c, 1992b, 1993c, 1994, 1995c; and
Silvertooth and Norton, 1996b, 1997c, 1998b, 1998c, 1999).
One of the challenges in the assessment and analysis of long term studies such as these is to provide a measure of
the effect of environment over time. Assessing environment by treatment interaction by use of conventional
analysis of variance becomes difficult as site-years increase and also due to the multiple factors influencing
environment by treatment interaction (Raun et al., 1993). Yates and Cochran (1938) first proposed stability analysis
for the interpretation of genotype by environment interaction. Their methodology held for the linear regression of
variety yield on experimental “mean yield” in order to observe varietal stability across varying environments (Finlay
and Wilkinson, 1963; Eberhart and Russell, 1966). More recently, stability analysis has been adapted for use in
comparing agronomic treatments across differing environments consisting of the linear regression of treatment mean
yield on the environmental mean (Raun et al., 1993; Boman et al., 1995; Unruh and Silvertooth, 1997).
The aggregate results of all studies involved (31 site-years) are summarized in this paper. Thus, the objectives of
this paper are to analyze the three MC application regimes and determine which regime offers the greatest
probability for a positive lint yield response.
Materials and Methods
Field studies were conducted across the cotton producing regions of Arizona from 1988 through 1999 (study
locations summarized in Table 1). Management of all studies with respect to irrigation, general fertility, and pest
control was carried out in a uniform and optimal manner across all treatments at each site. All data from these
studies were analyzed (e.g. analysis of variance) in accordance to procedures outlined by Steel and Torrie (1980)
and the SAS Institute (1997) to determine treatment differences for each site-year (Silvertooth et al., 1989, 1990,
1991a, 1993a, 1995b, 1997b; Hood and Silvertooth, 1993; Husman and Silvertooth, 1994).
The LRM MC application studies utilized a randomized complete block design (RCBD) with three or four
replications at each location. Plots at each location were a minimum of 8 rows wide and extended the full length of
the irrigation run (a minimum of 600 ft.). Applications of MC were made via ground rig with a total application
volume of 20 gallons/acre. Plant growth and development parameters were measured on 14 day intervals and
included PH, MSN, HNR, NAWF, and PCC. Treatments were initiated at or about pinhead to first matchhead
square formation. Individual applications ranged from 0.125 to 1.0 pt. MC /ac (Silvertooth et al., 1993a; Silvertooth
and Norton, 1995b, 1997b).
The LS application studies utilized a RCBD with four replications at each location. Plots at each location extended
the full length of the irrigation run (a minimum of 600 ft.). Treatments of MC were made with ground rig
applicators. Plant growth and development parameters were measured on 14 day intervals and included NAWF,
HNR, and FR. Treatments were initiated at or about cutout. Individual treatments consisted of a single application
ranging from 0.5 to 1.5 pt. MC/ac.
The feedback vs. scheduled approach to N and MC applications studies utilized a RCBD with four replications at
each location. Plots at each location were eight rows wide and extended the full length of the irrigation run (600 ft.).
Treatments of MC were made with ground rig applicators and all treatments of N were made with urea or
ammonium sulfate fertilizers with side dress applicators. Plant growth and development parameters were measured
on 14 day intervals and included PH, MSN, HNR, FR, NAWF, and PCC. Treatments consisted of all combinations
of N and MC in feedback and scheduled applications. The feedback MC treatments generally consisted of one or
two applications of MC ranging from 0.75 to 1.0 pt/ac per application as needed based upon measured growth
parameters. The scheduled MC treatments generally consisted of two applications of MC also ranging from 0.75 to
1.0 pt MC/ac per application based upon stage of growth. Both the feedback and the scheduled N treatments
consisted of split applications ranging from 40 to 50 lbs. N/ac per application. The scheduled treatments were a
more aggressive approach to N fertility and generally received a higher rate and one extra application (Fletcher et
al., 1994; Silvertooth and Norton, 1995a, 1996a, 1997a, 1998a; Norton et al., 1999).
Stability analysis was conducted utilizing the corresponding yield data from each experimental application regime.
Treatment mean lint yields were regressed on the environmental mean. Environmental mean (EM) was calculated
for each site-year by taking the mean lint yield across all treatments and replicates for each site-year (Raun, et al.,
1993).
Relative stability analysis was conducted on selected treatment pairs from the feedback vs. scheduled application
regime. Relative stability is assessed by studying the joint distribution of data pairs (i.e. mean for treatments A and
B in a given year). Slopes of the regression line are compared when the average yield ((A+B)/2) is regressed on the
yield difference (A – B) between the two treatments. A slope close to zero would indicate that the treatments
change similarly across environments and that they are equally stable. A positive slope indicates that B is more
stable than A since there is more variability in A. A negative slope would indicate that A is more stable than B. A
probability level of P<0.05 indicates that the slope from the regression equation is significantly different from zero
(Raun et al., 1993).
Results
Yield responses to MC treatments among all experiments are summarized in Table 2. We found that 15% of the MC
treatments demonstrated a positive yield response as compared to 17% of the MC treatments demonstrating a
negative yield response when compared with the control treatments (no MC applied). The HNR and FR of those
treatments that had yield responses to MC are summarized in Table 3. All references to established baselines are
based upon University of Arizona Extension guidelines (Silvertooth and Norton, 1998c). The HNR and FR
measurements provide a reasonable basis for assessing the vegetative to reproductive conditions of the crop. Also,
the HNR measurements are clearly the most reliable in this respect.
The regression of treatment mean lint yields on the EM for the LRM application regimes, both the Upland and Pima
species, are summarized and displayed in Tables 4 and 5 and Figures 1 and 2 respectively. Treatments consisted of
differing rates of total MC applied. It is significant to note in these analyses that all treatments are relatively similar
with regard to slope and intercept in both low and high yielding environments. Thus, none of the MC treatments
were consistently better that the control (treatment 1).
The regression of treatment mean lint yields on the EM for the late season application regimes are summarized and
displayed in Table 6 and Figure 3. Treatments consisted of differing rates of total MC applied near crop cut-out. It
is significant to note in these analyses that as higher yielding, longer season environments are encountered, late
season applications of MC are not beneficial in relation to the control.
The regression of treatment lint yields on the EM for the feedback vs. scheduled application regimes are
summarized and displayed in Table 7 and Figure 4. Treatments consisted of all combinations of N and MC
feedback and scheduled applications. Treatment two, which utilized a feedback approach to both N and MC, has the
highest elevation (intercept) of all treatments in both low and high yielding environments. Thus, treatment two is
the most beneficial and stable treatment across all environments.
The relative stability analyses for selected treatment pairs are displayed in Figures 5(a-d). The relative stability
diagrams further substantiate the results that were seen in the stability analysis for the feedback vs. scheduled
application regimes indicating the better stability and superiority of treatment two in contrast to other N and MC
treatments.
Summary and Conclusions
It was determined that MC produced significant positive yield responses in only 15% and negative yield responses
in 17% of cases. Therefore, it is important to understand what crop conditions may have contributed to or predicted
these yield responses. An analysis of plant growth parameters at the time of application revealed that all positive
yield responses occurred when the HNR was above the UA baseline levels and FR was usually above baselines.
Conversely, all negative yield responses occurred when HNR was below baselines and FR was at or above baseline
levels. These results indicate that the greatest potential for a positive yield increase from MC applications is present
when crop conditions are tending toward vegetative conditions, as indicated by the high HNR parameter relative to
baselines. This agrees in principle with Kerby’s findings (Kerby, 1985) that the greatest response from MC is
achieved when final control plant height is 130 cm which indicates vegetative growth tendencies by use of indices
for the San Joaquin Valley of CA.
Another issue to consider is the best application regime to maximize the probability for a positive yield response
with respect to MC applications. The analysis of the LRM application and the LS application regimes of MC
indicated that these methods offered no benefit in terms of consistent yield increases relative to the control plots.
The analysis of the feedback vs. scheduled application experiments indicated that the feedback approach to
management of inputs, in this case combinations of MC and N, resulted in the greatest yield benefits across all
environments. Thus indicating that the feedback approach to MC applications is the most viable method.
The use of the HNR growth parameter to indicate vegetative tendencies and to indicate the optimal timing of MC
applications in a feedback approach to management appears to be the most probable method of realizing a positive
lint yield response.
The utilization of a feedback approach to any input offers many positive features. Perhaps the most significant of
these is of that being able to improve upon the efficiencies associated with crop inputs such as MC. In a crop
production system, such as that of irrigated crops in the desert Southwest where inputs are high, it is a good exercise
for producers to evaluate inputs in a feedback approach to management in an effort to optimize any benefit from
those inputs. Also, the use of a feedback approach allows the grower to respond to in-season changes in crop
conditions. Therefore, the use of a feedback approach, with well established, regionally-specific baselines, provides
flexibility and improves the ability of the grower to better determine when a positive yield response to MC (or other
inputs) can be realized.
References
Ashley, D.A. 1972.
14
C-labeled photosynthate translocation and utilization in cotton plants. Crop Sci. 12:69-74.
Benedict, C.R. and R.J. Kohel. 1975. Export of 14C-assimilates in cotton leaves. Crop Sci. 15:367-372.
Boman, R.K. and R.L. Westerman. 1994. Nitrogen and mepiquat chloride effects on the production of nonrank,
irrigated, short-season cotton. J. Prod. Agric. 7:70-75.
Boman, R.K., W.R. Raun, R.L. Westerman, and J.C. Banks. 1995. Nitrogen by environment interactions in longterm cotton production. p. 1300-1302. In D.A. Richter (ed.) Proc. Beltwide Cotton Prod. Res. Conf., San
Antonio, TX. 4-7 Jan. 1995. National Cotton Council of America, Memphis, TN.
Bourland, F.M., D.M. Oosterhuis, and N.P. Tugwell. 1992. Concept for monitoring the growth and development of
cotton plants using mainstem node counts. J. Prod. Agric. 5:532-538.
Briggs, R.E. 1980. Effect of the plant regulator Pix on cotton in Arizona. p. 32. In J.M. Brown (ed.) Proc.
Beltwide Cotton Prod. Res. Conf., St. Louis, MO. 6-10 Jan. 1980. National Cotton Council of America,
Memphis, TN.
Brown, R.H. 1984. Growth of the green plant. p. 153-174. In M.B. Tesar (ed.) Physiological Basis of Crop
Growth and Development. American Society of Agronomy – Crop Science Society of America, Madison,
WI.
Crawford, S.H. 1981. Effects of mepiquat chloride on cotton in Northeast Louisiana. p. 45-46. In J.M. Brown
(ed.) Proc. Beltwide Cotton Prod. Res. Conf., New Orleans, LA. 4-8 Jan. 1981. National Cotton Council
of America, Memphis, TN.
Eberthart, S.A. and W.A. Russell. 1966. Stability parameters for comparing varieties. Crop Sci. 6:36-40.
Finlay, K.W. and G.N. Wilkinson. 1963. The analysis of adaptation in a plant-breeding programme. Aust. J. Agric.
Res. 14:742-754.
Fletcher, D.C., J.C. Silvertooth, E.R. Norton, B.L. Unruh, and E.A. Lewis. 1994. Evaluation of a feedback
approach vs. scheduled approach to Pix application. Cotton, A College of Agriculture Report. University
of Arizona, Series P-96:108-118.
Gausman, H.W., L.N. Namken, M.D. Heilman, H. Walter, and F.R. Rittig. 1979. Physiological effects of a growth
regulator (PixTM) on the cotton plant. p. 51-52. In J.M. Brown (ed.) Proc. Beltwide Cotton Production Res.
Conf., Phoenix, Ariz. 7-11 Jan. 1979. National Cotton Council of America, Memphis, Tenn.
Guinn, G. 1982. Causes of square and boll shedding in cotton. U.S.D.A. Tech. Bull. No. 1672.
Hood, L.R. and J.C. Silvertooth. 1993. High rate Pix use on Upland cotton. Cotton, A College of Agriculture
Report. University of Arizona. Series P-94:112-116.
Husman, S.H. and J.C. Silvertooth. 1994. The effects of Pix application timing of Upland cotton lint yield and
growth and development parameters. Cotton, A College of Agriculture Report. University of Arizona,
Series P-96:119-127.
Kerby, T.A. 1985. Cotton response to mepiquat chloride. Agron. J. 77:515-518.
Kerby, T.A., R.E. Plant, and R.D. Horrocks. 1997. Height-to-node ratio as an index of early season cotton growth.
J. Prod. Agric. 10:80-83.
Kerby, T.A., R.E. Plant, S. Johnson-Hake, and R.D. Horrocks. 1998. Environmental and cultivar effects on heightto-node ratio and growth rate in Acala cotton. J. Prod. Agric. 11:420-427.
Mauney, J.R. 1986. Vegetative growth and development of fruiting sites. p. 11-28. In J.R. Mauney and J. Stewart
(eds.) Cotton Physiology, Number One, The Cotton Foundation Reference Book Series. The Cotton
Foundation, Memphis, Tenn.
McConnell, J.S., W.H. Baker, B.S. Frizzell, and J.J. Varvil. 1992. Response of cotton to nitrogen fertilization and
early multiple applications of mepiquat chloride. J. Plnt. Nutr. 15:457-468.
Norton, E.J., J.C. Silvertooth, and E.R. Norton. 1999. Evaluation of a feedback approach to nitrogen and Pix
applications, 1998. Cotton, A College of Agriculture Report. University of Arizona, Series P-119:73-80.
Raun, W.R., H.J. Barreto, and R.L. Westerman. 1993. Use of stability analysis for long-term soil fertility
experiments. Agron. J. 85:159-167.
SAS Institute. 1997. The SAS system for Windows. Release 7.0. SAS Inst., Cary, NC.
Silvertooth, J.C., D.R. Howell, C.R. Farr, and J.E. Malcuit. 1989. Evaluation of Pix multiple application treatments
on Upland and Pima cotton in Arizona, 1988. Cotton, A College of Agriculture Report. University of
Arizona. Series P-77:104-109.
Silvertooth, J.C., J.E. Malcuit, D.R. Howell, and C.R. Farr. 1990. Pix multiple application evaluations in Arizona
on Upland and Pima cotton, 1989. Cotton, A College of Agriculture Report. University of Arizona. Series
P-81:50-56.
Silvertooth, J.C., J.E. Malcuit, S.H. Husman, W.S. Winans, and L. Hood. 1991a. Cotton response to multiple
applications of Pix, 1990. Cotton, A College of Agriculture Report. University of Arizona. Series P87:55-58.
Silvertooth, J.C., P.W. Brown, and J.E. Malcuit. 1991b. Basic crop development patterns. Cotton, A College of
Agriculture Report. University of Arizona. Series P-87:209-221.
Silvertooth, J.C., L.J. Clark, J.E. Malcuit, E.W. Carpenter, T.A. Doerge, and J.E. Watson. 1991c. Nitrogen
management experiments for Upland and Pima cotton, 1990. Cotton, a College of Agriculture Report.
University of Arizona. Series P-87:209-221.
Silvertooth, J.C., P.W. Brown, and J.E. Malcuit. 1992a. Cotton crop growth and development patterns. Cotton, A
College of Agriculture Report. University of Arizona. Series P-91:9-24.
Silvertooth, J.C., L.J. Clark, J.E. Malcuit, E.W. Carpenter, T.A. Doerge, and J.E. Watson. 1992b. Nitrogen
management experiments for Upland and Pima cotton, 1991. Cotton, a College of Agriculture Report.
University of Arizona. Series P-91:183-198.
Silvertooth, J.C., J.E. Malcuit, L. Hood, and S.H. Husman. 1993a. Cotton response to applications of Pix, 1992.
Cotton, A College of Agriculture Report. University of Arizona. Series P-94:84-99.
Silvertooth, J.C., P.W. Brown, and J.E. Malcuit. 1993b. The development and delivery of a crop monitoring
program for Upland and Pima cotton. Cotton, A College of Agriculture Report. University of Arizona.
Series P-94:17-26.
Silvertooth, J.C., L.J. Clark, J.E. Malcuit, E.W. Carpenter. 1993c. Nitrogen management experiments for Upland
and Pima cotton, 1992. Cotton, a College of Agriculture Report. University of Arizona. Series P-94:297311.
Silvertooth, J.C. 1994. Practical uses of crop monitoring for Arizona cotton. Cotton, A College of Agriculture
Report. University of Arizona. Series P-96:18-23
Silvertooth, J.C., E.R. Norton, B.L. Unruh, J.A. Navarro, L.J. Clark, and E.W. Carpenter. 1994. Nitrogen
management experiments for Upland and Pima cotton, 1993. Cotton, A College of Agriculture Report.
University of Arizona. Series P-96:378-397.
Silvertooth, J.C. and E.R. Norton. 1995a. Evaluation of a feedback approach to nitrogen and Pix application.
Cotton, A College of Agriculture Report. University of Arizona. Series P-99:327-335.
Silvertooth, J.C. and E.R. Norton. 1995b. Evaluation of late-season PixTM applications. Cotton, A College of
Agriculture Report. University of Arizona. Series P-99:79-82.
Silvertooth, J.C., E.R. Norton, B.L. Unruh, J.A. Navarro, L.J. Clark, and E.W. Carpenter. 1995c. Nitrogen
management experiments for Upland and Pima cotton, 1994. Cotton, A College of Agriculture Report.
University of Arizona. Series P-99:311-326.
Silvertooth, J.C. and E.R. Norton. 1996a. Evaluation of a feedback approach to nitrogen and Pix applications.
Cotton, A College of Agriculture Report. University of Arizona. Series P-103:333-340.
Silvertooth, J.C., E.R. Norton. 1996b. Nitrogen management experiments for Upland and Pima cotton, 1995.
Cotton, A College of Agriculture Report. University of Arizona. Series P-103:317:332.
Silvertooth, J.C. and E.R. Norton. 1997a. Evaluation of a feedback approach to nitrogen and Pix application.
Cotton, A College of Agriculture Report. University of Arizona. Series P-108:107-113.
Silvertooth, J.C. and E.R. Norton. 1997b. Evaluation of late-season PixTM applications. Cotton, A College of
Agriculture Report. University of Arizona. Series P-108:114-116.
Silvertooth, J.C., E.R. Norton. 1997c. Nitrogen management experiments for Upland and Pima cotton, 1996.
Cotton, A College of Agriculture Report. University of Arizona. Series P-108:389-401.
Silvertooth, J.C. and E.R. Norton. 1998a. Evaluation of a feedback approach to nitrogen and Pix applications,
1997. Cotton, A College of Agriculture Report. University of Arizona. Series P-112:469-475.
Silvertooth, J.C. and E.R. Norton. 1998b. Nitrogen management experiments for Upland and Pima cotton, 1997.
Cotton, A College of Agriculture Report. University of Arizona. Series P-112:461-468.
Silvertooth, J.C. and E.R. Norton. 1998c. Cotton monitoring and management system. Publication no. AZ1049,
University of Arizona, College of Agriculture, Tucson, AZ.
Silvertooth, J.C. and E.R. Norton. 1999. Nitrogen management experiments for Upland and Pima cotton, 1998.
Cotton, A College of Agriculture Report. University of Arizona. Series P-116:213-220.
Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics. McGraw-Hill, New York.
Unruh, B.L. and J.C. Silvertooth. 1997. Planting and irrigation timing effects on the yield of upland and pima
cotton. J. Prod. Agric. 10:74-79.
Walter, H., H.W. Gausman, F.R. Rittig, L.N. Namken, D.E. Escobar, and R.R. Rodriguez. 1980. Effect of
mepiquat chloride on cotton plant leaf and canopy structure and dry weights of its components. p. 32-35.
In J.M. Brown (ed.) Proc. Beltwide Cotton Prod. Res. Conf., St. Louis, MO. 6-10 Jan. 1980. National
Cotton Council of America. Memphis, TN.
Yates, F., and W.G. Cochran. 1938. The analysis of a group of experiments. J. Agric. Sci. 28:556-580.
York, A.C. 1983a. Cotton cultivar response to mepiquat chloride. Agron. J. 75:663-667.
York, A.C. 1983b. Response of cotton to mepiquat chloride with varying N rates and plant populations. Agron. J.
75:667-672.
Table 1. Summary of MC related experiments, 1988 – 1999.
Year
Study Type†
Location
Elevation (ft) Soil Series
1988
LRM
Yuma Valley
100
Holtville Clay
1988
LRM
Yuma Valley
100
Holtville Clay
1988
LRM
Gila Bend
750
Brios Sandy Loam
1988
LRM
Gila Bend
750
Brios Sandy Loam
1989
LRM
Yuma Valley
100
Gadsden Clay
1989
LRM
Yuma Valley
100
Holtville Clay
1989
LRM
Gila Bend
750
Brios Sandy Loam
1989
LRM
Gila Bend
750
Brios Sandy Loam
1990
LRM
Waddell
1350
Avondale Clay Loam
1990
LRM
Gila Bend
750
Brios Sandy Loam
1990
LRM
Poston
400
Holtville Clay Loam
1991
LRM
Waddell
1350
Glenbar Loam
1991
LRM
Maricopa
1200
Casa Grande Sandy Loam
1992
LRM
Buckeye
1000
Laveen Sandy Loam
1992
LRM
Maricopa
1200
Casa Grande Sandy Loam
1992
LRM
Mohave
600
Indio Clay
1992
LRM
Gila Bend
750
Brios Sandy Loam
1992
LRM
Sunsites
4000
Tubac Sandy Clay Loam
1992
LRM
Mohave
600
Indio Clay
1992
LS
Sunsites
4000
Tubac Sandy Clay Loam
1993
LRM
Buckeye
1000
Laveen Sandy Loam
1993
LRM
Maricopa
1200
Casa Grande Sandy Loam
1993
FB
Maricopa
1200
Casa Grande Sandy Loam
1994
FB
Maricopa
1200
Casa Grande Sandy Loam
1994
LS
Maricopa
1200
Casa Grande Sandy Loam
1995
FB
Maricopa
1200
Casa Grande Sandy Loam
1996
FB
Maricopa
1200
Casa Grande Sandy Loam
1996
LS
Marana
1950
Pima Clay Loam
1997
FB
Marana
1950
Pima Clay Loam
1998
FB
Marana
1950
Pima Clay Loam
1999
FB
Marana
1950
Pima Clay Loam
†
LRM = Low Rate Multiple Application; LS = Late Season Application; FB = Feedback Approach
Table 2. Summary of yield responses to MC experiments, 1988 – 1999.
Positive Yield Response†
15%
†
Responses significant at the P≤0.05 level.
n=148
Negative Yield Response
17%
No Yield Response
68%
Table 3. Summary of HNR and FR with respect to baselines in treatments responding to MC.
Positive Yield Responses
HNR
100% above
Negative Yield Responses
FR
61% above
39% below
HNR
100% below
FR
100% above
Table 4. Linear regression equations of treatment mean yield on the environmental mean, low rate multiple
application regimes (Upland), 1988 – 1993.
Treatment
1
2
3
4
5
pt. MC/ac
0
0.375 – 0.75
1.0 – 1.25
1.5 – 2.25
2.5 – 4.0
Intercept
Standard Error
Estimate
Slope
Standard Error
Estimate
-58.028
-25.346
61.762
88.446
14.931
177.613
84.604
72.536
45.004
586.894
1.062
1.014
0.974
0.939
0.887
0.123
0.052
0.049
0.032
0.458
C.V.
r2
%
8.0
2.4
3.2
3.7
7.3
0.86
0.96
0.97
0.96
0.43
Table 5. Linear regression equations of treatment mean yield on the environmental mean, low rate multiple
application regimes (American Pima), 1988 – 1993.
Treatment
1
2
3
pt. MC/ac
0
0.375 – 0.75
1.0 – 2.5
Intercept
Standard Error
Estimate
Slope
Standard Error
Estimate
62.745
-1.445
9.387
112.410
55.276
32.950
0.936
1.009
0.960
0.111
0.051
0.039
C.V.
r2
%
7.3
4.3
5.4
0.95
0.94
0.98
Table 6. Linear regression equations of treatment mean yield on the environmental mean, late season
application regimes (Upland), 1992 – 1996.
Treatment
1
2
3
pt. MC/ac
0
0.5 – 0.75
1.0 – 1.5
Intercept
Standard Error
Estimate
Slope
Standard Error
Estimate
-105.313
134.448
-40.896
57.759
47.685
68.821
1.106
0.860
1.043
0.052
0.043
0.062
C.V.
r2
%
1.1
0.89
1.3
0.99
0.99
0.99
Table 7. Linear regression equations of treatment mean yield on the environmental mean, feedback vs.
scheduled application regimes (Upland), 1993 – 1999.
Treatment
Intercept
Standard Error
Estimate
Slope
Standard Error
Estimate
C.V.
%
1
NF=
159.365
221.706
0.866
0.182
5.9
2
NF MF
160.499
152.660
0.917
0.125
3.9
3
NF MS
138.532
183.554
0.906
0.151
4.8
4
NS MF
-15.788
174.847
1.034
0.143
4.5
5
NS MS
-7.308
167.009
1.029
0.137
4.3
=
NF, NS, MF, and MS refer to feedback and scheduled approaches to applications of Nitrogen fertilizers and
Mepiquat Chloride.
Table 8. Difference in slopes for selected treatment regression equation comparisons.
Comparison
t-calc†
PR > |t|†
Feedback Approach
Trmt. 2 vs. Trmt. 1
0.207
0.420
Trmt. 2 vs. Trmt. 3
0.054
0.479
Trmt. 2 vs. Trmt. 4
0.594
0.283
Trmt. 2 vs. Trmt. 5
0.583
0.286
Late Season Regime
Trmt. 1 vs. Trmt. 2
3.522
0.036
Trmt. 1 vs. Trmt. 3
0.805
0.239
Low-Rate Multiple Regime (Upland)
Trmt. 1 vs. Trmt. 2
0.348
0.365
Trmt. 1 vs. Trmt. 3
0.673
0.254
Trmt. 1 vs. Trmt. 4
1.434
0.079
Trmt. 1 vs. Trmt. 5
0.261
0.399
Low-Rate Multiple Regime (Pima)
Trmt. 1 vs. Trmt. 2
0.729
0.236
Trmt. 1 vs. Trmt. 3
0.240
0.407
†PR>|t|, probability of a greater absolute value of t. t-calc, t test of differences in slopes.
r2
0.82
0.91
0.88
0.91
0.92
Treatment Mean Lint Yield (lbs. lint/ac)
2000
1800
1600
Trmt 1
1400
Trmt 2
Trmt 3
1200
Trmt 4
1000
Trmt 5
800
1000
1200
1400
1600
1800
2000
Environmental Mean (lbs. lint/ac)
Figure 1. Treatment mean vs. environmental mean for
low rate multiple application regimes (upland), 1988 - 1993.
Treatment Mean Lint Yield (lbs. lint/ac)
1600
1400
1200
Trmt 1
Trmt 2
Trmt 3
1000
800
600
400
200
400
600
800
1000
1200
1400
Environmental Mean (lbs. lint/ac)
Figure 2. Treatment mean vs. environmental mean for
low rate multiple application regimes (pima), 1988 - 1993.
Treatment Mean Lint Yield (lbs.lint/ac)
1350
1300
Trmt 1
1250
Trmt 2
1200
1150
Trmt 3
1100
1050
1000
950
950 1000 1050 1100 1150 1200 1250 1300 1350
Environmental Mean (lbs. lint/ac)
Figure 3. Treatment mean vs. environmental mean for
late season application regimes, 1992 - 1996.
Treatment Mean Lint Yield (lbs. lint/ac)
1600
1500
1400
Trmt 1
1300
Trmt 2
1200
Trmt 3
1100
Trmt 4
1000
Trmt 5
900
900
1000
1100
1200
1300
1400
1500
Environmental Mean (lbs. lint/ac)
Figure 4. Treatment mean vs. environmental mean for
feedback vs. scheduled application regimes, 1993 - 1999.
1600
120
(A)
(B)
y = 0.0025(x) - 67.30
y = -0.0076(x) - 46.63
Yield Difference (lbs. lint/ac)
P = 0.984
80
40
0
-40
-80
-120
120
Yield Difference (lbs. lint/ac)
P = 0.955
Treatment 2 vs. Treatment 3
Treatment 2 vs. Treatment 1
y = -0.122(x) - 211.15
(C)
(D)
y = -0.115(x) - 197.83
P = 0.390
P = 0.271
80
40
0
-40
-80
-120
900
Treatment 2 vs. Treatment 4
1000
1100
1200
1300
Treatment 2 vs. Treatment 5
1400
Mean Yield (lbs. lint/ac)
1500
1600 900
1000
1100
1200
1300
1400
1500
Mean Yield (lbs. lint/ac)
Figure 5(a-d). Yield difference vs. mean yield for selected treatment pairs, feedback
application regime.
1600