Translator Disclaimer
Author Affiliations +

Populations of Oligonychus perseae Tuttle, Baker and Abbatiello (Acari: Tetranychidae) were monitored for 37 weeks in an avocado (Persea americana Miller) orchard in Fallbrook California, USA. Three cultural control methods were tested to reduce vertical recolonization of avocado trees by O. perseae. Control methods examined were: (1) Tanglefoot® barriers on tree trunks. (2) Removal of ground cover and leaf litter under tree canopies, and (3) Tanglefoot® barriers combined with ground cover removal. All treatments were compared to control trees that were not subjected to any of the preceding cultural practices. No significant differences in O. perseae populations or O. perseae-induced leaf damage were found among trees receiving different treatments. However, aerial dispersal activity of O. perseae was found to increase as O. perseae populations on monitored avocado trees increased, suggesting that populations could have been augmented as new immigrants arrived from unmonitored avocado trees or other miscellaneous host plants. The phytoseiids Galendromus helveolus (Chant) and G. annectens (De Leon), Amblyseius similoides Buchelos and Pritchard, Euseius hibisci (Chant), and Neoseiulus californicus (McGregor) were also monitored during this study. Natural enemy numbers were not adversely affected by cultural control attempts targeting O. perseae. Aerial applications of abamectin with NR 415 oil for Scirtothrips perseae Nakahara (Thysanoptera: Thripidae) had an adverse effect on Galendromus spp., which are important predators of O. perseae in avocado orchards.

Oligonychus perseae Tuttle, Baker, and Abbatiello (Acari: Tetranychidae) has been a serious pest of economic importance in southern California avocado orchards since 1990 (Bender 1993). This pest inhabits undersides of leaves, where motile stages take refuge within webbed nests where they feed and reproduce. Feeding damage to leaves within nests results in localized necrosis, which manifests itself as characteristic circular brown spots (Aponte & McMurtry 1997).

Avocado trees naturally defoliate in the spring months, and leaves bearing overwintering O. perseae drop to the orchard floor. Leaves are not shed simultaneously and defoliation occurs gradually over a period of 3-4 weeks, during which new leaves begin to grow (Lovatt 1990, Liu et al. 1999). As the growing season progresses and temperatures increase, so do O. perseae populations, which tend to peak during mid to late summer months (July-September) (Kerguelen & Hoddle 1999a; Hoddle et al. 1999, Hoddle et al. 2000).

Defoliation of avocado leaves can also occur when high infestations (>500 mites/leaf) of O. perseae cause 7.5-10.0% leaf area damage (Hoddle et al. 2000, Kerguelen & Hoddle 1999a), and O. perseae will fall to the ground on leaves during defoliation events (Kerguelen & Hoddle, unpublished). Oligonychus perseae have been trapped climbing on avocado tree trunks and vertical support stakes presumably in attempts to locate viable host plants after abandoning leaf material on the orchard floor (Kerguelen & Hoddle, unpublished). Following mite-induced defoliation, the ensuing new leaf “flush” may attract other pests, such as Scirtothrips perseae Nakahara (Thysanoptera: Thripidae), which utilizes young leaves for feeding and reproduction. In addition, trees defoliated by O. perseae are at greater risk of sunburn and scalding as trunks and fruit are exposed to sun, and premature fruit drop and subsequent economic losses may result (Bender 1993).

Control of O. perseae populations is usually achieved through acaricides (e.g., refined narrow-range petroleum oils) or through augmentative releases of biological control agents. Two commercially available phytoseiids, Neoseiulus californicus (McGregor) and Galendromus helveolus (Chant) have demonstrated an ability to reduce O. perseae densities to levels comparable to those produced by acaricidal oils (Kerguelen & Hoddle 1999a; Hoddle et al. 2000). However, an unexplored control option is the use of cultural practices to mitigate vertical recolonization of trees by O. perseae from dropped leaves on the ground during periods of defoliation.

During the winter, O. perseae densities on leaves are at almost undetectable levels (Kerguelen & Hoddle 1999; Hoddle et al. 2000) and potential overwintering sites utilized by O. perseae on avocados other than leaves might occur in leaf duff under trees, cracks and crevices of branches, and bark cover on trunks. However, Berlese funnel extractions of material collected over winter have not provided evidence that O. perseae overwinters on woody twigs, bark, or ground cover (Takano-Lee, unpublished).

Therefore, we decided to investigate the efficacy of cultural control measures that were designed to eliminate possible routes of tree re-colonization by O. perseae from the orchard floor from fallen leaves. Treatments investigated to reduce vertical recolonization rates and subsequent pest population growth were: (1) placement of sticky barriers on tree trunks to trap motile O. perseae that abandon dropped leaves under trees and walk up vertical structures (e.g., tree trunks and support stakes) to locate suitable host plants. (2) Removal of all defoliated leaves and weeds from under trees thereby removing sources of ambulatory mites. (3) Combining treatments 1 and 2, and (4) monitoring O. perseae densities on control trees that received no cultural management. In addition to monitoring O. perseae on experimental trees, we also investigated the impact of these cultural practices on resident phytoseiid populations.

Materials and Methods

Study Site and Experimental Design

A commercial 40 ha ‘Hass’ avocado orchard, located in Bonsall, California, USA served as the study site. Oligonychus perseae surveys began on April 4, 2000 (wk 0), when spring defoliation began. Thirty trees 3 m high were selected and 10 randomly selected leaves were removed from each tree and returned to the laboratory. The number of motile O. perseae was recorded by leaf for each tree.

Following these pre-counts, 24 trees were blocked into 6 trees per treatment, to produce comparable mean O. perseae densities per leaf across treatments. On April 10 (wk 1), four treatments were applied to study trees: (1) Tanglefoot® (The Tanglefoot Company, Grand Rapids, Michigan USA) was applied to both sides of a strip of double-sided sticky carpet tape (Manco Inc., Avon, Ohio, USA) that was wrapped around the trunk approximately 20 cm above soil. Leaf material was left on the orchard floor beneath trees (n = 6). (2) Ground cover (i.e., weeds, leaf litter, branches, etc.) was raked away from under experimental trees so that the ground beneath the canopy was bare dirt (n = 6). (3) Tanglefoot® on carpet tape was applied in conjunction with ground cover removal (n = 6), and (4) control trees were those to which no treatments were applied (n = 6). Oligonychus perseae populations on all experimental trees were quantified weekly for each treatment by counting the number of motile stages on each of 10 randomly chosen leaves that were returned to the laboratory. Sticky carpet tape barriers with Tanglefoot® were checked periodically to ensure adhesive properties existed and every week leaf duff was raked and removed from beneath trees in treatments 2 and 3 for the entire duration of the 37 week experimental period. Weekly removal of leaf duff was deemed acceptable because the number of leaves falling post-defoliation were few (range = 0-30).

Sampling Oligonychus perseae on Dropped Leaves

Commencing weekly from May 8, 2000 ten dropped leaves from spring defoliation from under six randomly selected trees at the study plot were removed from the ground and the percentage of leaves infested and the mean number (±SEM) of occupied nests, eggs, and motile O. perseae for 60 leaves were calculated. Seven consecutive samples were made with the last being June 12, 2000.

Statistical Analysis of O. perseae Densities

The average total number of mite days (M-D) for each treatment (i.e., the cumulative area beneath population trends graphs) was determined as:



i = number of sampling intervals

d = sampling interval in days

Mn = mean number of    mites per leaf at sampling period n

and calculated for each treatment group and compared by ANOVA. Tukey’s Studentized Range Test was conducted at 0.05 level of significance to separate means to determine treatment effects.

Assessment of Leaf Damage

To determine the previous season’s feeding damage to leaves by O. perseae, recently excised avocado leaves (n = 27) were randomly collected from all 30 trees at the study plot on April 10, 2000 (wk 1). Percentage leaf area damage caused by O. perseae feeding was measured by using automated image analysis techniques (Kerguelen & Hoddle 1999b).

From October 9, 2000 (wk 27) until monitoring ceased (December 17, 2000 wk = 37), 15 defoliated leaves per treatment were collected and percentage surface area damaged by O. perseae feeding was calculated using automated image analysis. At the conclusion of the field season, all leaves collected directly from trees in each experimental treatment were assessed for percentage leaf area damaged levels (n = 240). Percentage damage across treatments was analyzed with a fully nested ANOVA after arcsine-square root transformation. Significance was determined at the 0.05 level with Tukey’s Studentized Range test.

Aerial Dispersal

Aerial dispersal of O. perseae was monitored in the study plot. Four wooden stakes (1.5 m high with two 1.5 m cross-arms) were established at the north, south, east, and west directions of the experimental plot. Each cross-arm held four white sticky cards (19.5 × 16 cm) facing in each cardinal direction for a total of 16 sticky cards in the study plot. Every sampling interval, cards were removed from cross-arms, placed in clear plastic bags, and returned to the laboratory and the number of O. perseae within a central square (10 cm × 10 cm) of each sticky card was counted. Sticky cards were replaced each sampling period. The directional tendencies of O. perseae dispersal were calculated by square root (x + 1) transforming O. perseae counts on cards. Transformed data by cardinal direction were compared with ANOVA and Tukey’s Studentized Range test for means separation at the 0.05 level.

Monitoring Phytoseiid Populations

All phytoseiid predators on sampled leaves were removed, mounted in Hoyer’s medium, and identified to species. Cumulative mean predator days by treatment at each sampling interval were calculated (see mite days calculations) and compared with ANOVA and Tukey’s Studentized Range test was used for means separation at the 0.05 level.

Scirtothrips perseae Management

On April 15, 2000, the entire orchard was helicopter-sprayed with abamectin (Agri-Mek, 0.15 EC, Novartis Crop Protection, Inc., Greensboro, North California, USA) to control S. perseae. Abamectin was applied at 0.25 l/ha, delivered with 2% Leffingwell Supreme 415 Oil (Pace International, Seattle, Washington, USA) and 0.02-0.1% CMR Silicone Surfactant (Monterey Chemical Company, Fresno, California, USA) with water at a rate of 153 l/ha.

Temperature and Humidity Monitoring

Temperature and relative humidities in the orchard were recorded during the 37 wk of the experiment. Measurements were made at 30 min intervals with a HOBO data logger (Onset®, Bourne, Massachusetts, USA) attached to a wooden stake 1.5 m above the ground and positioned in the center of the experimental plot. Mean weekly maximum and minimum temperatures (°C) and maximum and minimum weekly percentage relative humidities were calculated for the duration of the experiment.


Oligonychus perseae on Dropped Leaves

All life stages of O. perseae were found on spring defoliated leaves. Percentage of defoliated leaves with either occupied nests, eggs, or motile mites, ranged 0-28.33%, 0-23.33%, and 0-26.67% respectively. The mean number of nests, eggs, and motiles per leaf ranged from 0-19.01 ± 7.65, 0-41.20 ± 23.33, 0-39.1 ± 21.65, respectively.

Assessment of Treatment Efficacy

Oligonychus perseae populations began to increase 12 wk after the study was initiated and densities peaked in October before declining steadily thereafter (Fig. 1A). Mean weekly population counts were consistently lower for ground cover removal trees (Treatment 2), than other treatments (Fig. 1A). Percentage leaf infestation began to increase steadily after June 19 (wk 11) and by September 4 (wk 22), 100% of collected leaves across all treatments were infested with ≥1 motile O. perseae (Fig. 1B). Percentage of leaf infestation did not begin to decline until December 4 (wk 35) and remained high (≈85%) until the conclusion of the experiment and no differences among treatments were discernible. Comparison of mean cumulative mite days across treatments did not reveal significant differences between treatments (F = 0.62; df = 3, 20; P = 0.61) indicating that the use of Tanglefoot® barriers, ground cover removal, and a combination of both these tactics had no effect on O. perseae population dynamics on trees receiving these treatments (Fig. 2).

Mean Percentage Leaf Area Damaged

Mean percentage leaf area damaged by O. perseae feeding for leaves remaining on all 30 experimental trees at the end of the previous season (1999) before this experiment was initiated was 30.87 ± 3.28%. In contrast, leaf damage at the end of the 2000 season was only 6.68 ± 0.03% for leaves remaining on trees and 9.04 ± 0.37% for defoliated leaves collected off the orchard floor. There were no significant differences between treatment groups at any time interval for leaves collected from the orchard floor that were analyzed for percentage area leaf damage (F = 1.237, df = 18, P = 0.229) (Fig. 3). Oligonychus perseae feeding damage differed significantly between sampling dates (F = 10.283, df = 5, P < 0.005), a probable reflection of population trends. However, for leaves collected directly from trees at the conclusion of the experiment, there were no significant differences in mean percentage area leaf damaged across treatments (F = 0.226, df = 3, P = 0.878) (data not shown).

Aerial Dispersal

Aerial dispersal by O. perseae peaked during the week of November 6 (wk 31), which corresponded with peak O. perseae populations on leaves (Fig. 4). Aerial movement by O. perseae tended to be northerly, but no significant differences between cardinal directions were detected (F = 0.72; df = 3, 60; P = 0.55) (Fig. 5). Mite life stage and gender were not determined, due to the difficulty of removing intact specimens from the sticky medium, for purposes of microscopic examination.

Monitoring of Phytoseiid Populations

Four different species of phytoseiid predators were recovered from experimental trees. Euseius hibisci (Chant) dominated the fauna comprising 79.66% (n = 1731) of predators recovered. Galendromus spp. (19.88%) (n = 432), Neoseiulus californicus (0.28%) (n = 6), and Amblyseius similoides Buchelos and Pritchard (0.18%) (n = 4) were all recovered with lesser frequency, and 0.8% (n = 17) of phytoseiids were found within O. perseae nests. Of these 17 phytoseiids, 6% (n = 1) were E. hibisci, 6% (n = 1) were N. californicus, and 88% (n = 15) were Galendromus spp.

Galendromus spp. were immediately reduced in number after the abamectin application for S. perseae control. Galendromus spp. densities began to increase after September 18 in response to increasing O. perseae densities (Fig. 6A). Euseius hibisci populations did not appear to be affected by the abamectin application as abrupt declines in predator density were not observed (Fig. 6B). There was no significant difference in the overall total number of phytoseiids recovered when pooled by treatment group (F = 0.19; df = 3, 112; P = 0.90).

Comparison of predator densities across treatments over time using predator days did not yield significant differences for either Galendromus spp. (F = 0.52, df = 3, 20, P = 0.6706) or E. hibisci (F = 0.10, df = 3, 20, P = 0.9575) (Fig. 7).

Temperature and Humidity Monitoring

Maximum weekly temperatures peaked during the first two weeks of September (Fig. 8). Temperature peaks did not coincide with peak O. perseae populations but did coincide with O. perseae dispersal activity. Maximum weekly relative humidities were consistently 100% until late October (Fig. 8).


Removing leaf litter and blocking entry to the tree canopy via the trunks with Tanglefoot® on carpet tape did not significantly reduce subsequent O. perseae population development following defoliation of leaves in spring in comparison to control trees not subjected to cultural management. This suggests that the main route of reinfestation of trees by O. perseae may occur in ways other than vertical recolonization from the ground after leaf drop. Possible recolonization mechanisms of new leaves in late spring could be: (1) O. perseae motiles are not dropping from trees on senescent leaves in sufficient numbers during periods of spring defoliation to permit effective recolonization from the ground, and the majority of the population remains on trees. Consequently, overwintering by O. perseae may take place within cracks and fissures on twigs or bark on trunks or large branches. However, Berlese funnel extractions of these substrates for O. perseae have yielded negative results and an alternative sampling technique (e.g., alcohol washes) may be required to detect overwintering O. perseae motiles. (2) Oligonychus perseae populations are initiated in spring by resident mite populations on leaves bearing O. perseae that do not defoliate. Retention rates of previous season’s leaves have not been determined and this potential route of reinfestation requires further investigation. (3) Resident O. perseae populations on avocado leaves are eliminated when defoliation occurs and mites die on the orchard floor as leaf quality deteriorates or ambulatory mites that abandoned leaves are unable to locate nutritive resources. Consequently, subsequent recolonization of avocado trees could occur via aerial dispersal of motiles from other host plants in or around orchards. In California, O. perseae has been recorded from a variety of alternate host plants such as acacia (Acacia senegal L.), annatto (Bixa orellano L.), apricot (Prunus armeniaca L.), camphor (Cinnamomum camphora T. Nees & C. Eberm.), carob (Ceratonia siliqua L.), citrus (Citrus spp.), sow-thistle (Sonchus spp.), grapes (Vitis spp.), lamb’s quarters (Chenopodium album L.), liquidambar (Liquidambar spp.), milkweed (Asclepias spp.), nectarine and peach (Prunus persica L. Batsch.), persimmon (Diospyros spp.), plum (Prunus spp.), rose (Rosa spp.), sumac (Rhus spp.), and willow (Salix spp.) (Baker & Tuttle 1994, Bender 1993, Aponte & McMurtry 1997). These host plants may act as reservoirs for O. perseae during periods of leaf drop in or around avocado orchards and low density aerial movement via ballooning by O. perseae from alternate host plants may be a route of re-infestation of avocados. In this study, O. perseae ballooning activity increased as avocado leaf infestation approached 100%, suggesting that aerial dispersal occurred as a function of population density and deteriorating resource quality. Such behavior may be initiated by O. perseae on non-avocado host plants also. To better understand the overwintering and recolonization ecology of O. perseae in avocado orchards, points 1-3 need further investigation.

The abrupt decline of Galendromus spp. populations following abamectin applications for S. perseae may have been due to the fact that these predators are Type II phytoseiids (McMurtry & Croft 1997) that attack tetranychids inhabiting webbed nests. Abamectin has known toxicity to O. perseae (J. Morse, pers. comm., 2001) and Galendromus spp. population declines could have resulted because of direct contact toxicity, poisoning through ingestion of toxic O. perseae that acquired either a lethal or sublethal dose of abamectin, or starvation because of the absence of prey. Euseius hibisci, a Type IV phytoseiid, can subsist on a diet consisting primarily of pollen, and it did not appear to be adversely affected by the abamectin application. Despite a long-term low density population of O. perseae, E. hibisci densities remained relatively high as this predator was probably able to survive on pollen and non-mite prey.

The use of abamectin for controlling S. perseae, a serious thrips pest in California avocado orchards, is the most efficacious management tool used by growers to prevent damage to fruit by feeding larvae and adults (Yee et al. 1999). The impact of abamectin applications on S. perseae and O. perseae natural enemies have not been thoroughly studied. However, Zhang and Sanderson (1990) observed that abamectin was less toxic to Phytoseiulus persimilis Anthias-Henriot (Acari: Phytoseiidae) than its prey, Tetranychus urticae Koch (Acari: Tetranychidae). Likewise, Ibrahim and Yee (2000) found that although abamectin deleteriously affected Neoseiulus longispinosus (Evans) (Acari: Phytoseiidae), this predator did not suffer the same toxic effects as T. urticae, based on comparison of growth and reproduction estimates. Earlier studies demonstrated that survivability and female fecundity were compromised among predatory mite species such as Metaseiulus occidentalis (Nesbitt), Phytoseius finitimus Ribaga, and Amblyseius gossipi E-Badry (Grafton-Cardwell and Hoy 1983, El-Banhawy and El-Bagoury 1985, Reda and El-Banhawy 1988). Development and implementation of integrated pest management programs for O. perseae and S. perseae that combine selective insecticides (e.g., abamectin) and natural enemies will require a thorough understanding of the non-target effects of these insecticides and the field ecology of target pests.


We thank J. Virzi and J. Jones for their assistance recording O. perseae and phytoseiid counts. The cooperation of D. Hedrick, who allowed us to study his avocado orchard, was greatly appreciated. Funding for this project was provided in part by the California Avocado Commission.

References Cited


O. Aponte and J. R. McMurtry . 1997. Damage on “Hass” avocado leaves, webbing and nesting behavior of Oligonychus perseae (Acari: Tetranychidae). Exp. Appl. Acarol. 21::265–272. Google Scholar


E. W. Baker and D. M. Tuttle . 1994. A Guide to the Spider Mites (Tetranychidae) of the United States. Indira Publishing House, MI.  Google Scholar


G. S. Bender 1993. A new mite problem in avocados. California Avocado Society Yearbook 77::73–77. Google Scholar


E. M. El-Banhawy and M. E. El-Bagoury . 1985. Toxicity of avermectin and fenvalerate to the eriophyid gall mite Eriophyes dioscerdis and the predacious mite Phytoseius finitimus (Acari: Eriophyidae; Phytoseiidae). Int. J. Acarol. 11::237–240. Google Scholar


E. E. Grafton-Cardwell and M. A. Hoy . 1983. Comparative toxicity of avermectin B1 to the predator Metaseiulus occidentalis (Nesbitt) (Acari: Phytoseiidae) and the spider mites Tetranychus urticae Koch and Panonychus ulmi (Koch) (Acari: Tetranychidae). J. Econ. Entomol. 76::1216–1220. Google Scholar


M. S. Hoddle, O. Aponte, V. Kerguelen, and J. Heraty . 1999. Biological control of Oligonychus perseae (Acari: Tetranychidae) on Avocado: I. Evaluating release timings, recovery and efficacy of six commercially available phytoseiids. Internat. J. Acarol. 25::211–219. Google Scholar


M. S. Hoddle, L. Robinson, and J. Virzi . 2000. Biological control of Oligonychus perseae (Acari: Tetranychidae) on avocado: III. Evaluating the efficacy of varying release rates and release frequency of Neoseiulus californicus (Acari: Phytoseiidae). Internat. J. Acarol. 26::203–214. Google Scholar


Y. B. Ibrahim and T. S. Lee . 2000. Influence of sublethal exposure to abamectin on the biological performance of Neoseiulus longispinosus (Acari: Phytoseiidae). J. Econ. Entomol. 93::1085–1089. Google Scholar


V. Kerguelen and M. S. Hoddle . 1999a. Biological control of Oligonychus perseae (Acari: Tetranychidae) on avocado: II. Evaluating the efficacy of Galendromus helveolus and Neoseiulus californicus (Acari: Phytoseiidae). Internat. J. Acarol. 25::221–229. Google Scholar


V. Kerguelen and M. S. Hoddle . 1999b. Measuring mite feeding damage on avocado leaves with automated image analysis of software. Florida Entomol. 82::119–122. Google Scholar


X. Liu, P. W. Robinson, M. A. Madore, G. W. Witney, and M. L. Arpaia . 1999. ‘Hass’ avocado carbohydrate fluctuations. I. Growth and phenology. J. Amer. Soc. Hort. Sci. 124::671–675. Google Scholar


C. J. Lovatt 1990. Factors affecting fruit set and early drop in avocado. California Avocado Society Yearbook 74::193–199. Google Scholar


J. A. McMurtry and B. A. Croft . 1997. Life-styles of phytoseiid mites and their roles in biological control. Annu. Rev. Entomol. 42::291–321. Google Scholar


A. S. Reda and M. El-Banhawy . 1988. Effect of avermectin and dicofol on the immatures of the predacious mite Amblyseius gossipi with a special reference to the secondary poisoning effect on the adult female [Acari: Phytoseiidae]. Entomophaga 33::349–355. Google Scholar


W. L. Yee, P. A. Phillips, B. A. Faber, J. G. Morse, and M. S. Hoddle . 1999. Control of avocado thrips using aerial applications of insecticides. California Avocado Society Yearbook 83::141–155. Google Scholar


Z-Q. Zhang and J. P. Sanderson . 1990. Relative toxicity of abamectin to the predatory mite Phytoseiulus persimilis (Acari: Phytoseiidae) and twospotted spider mite (Acari: Tetranychidae). J. Econ. Entomol. 83::1783–1790. Google Scholar


Fig. 1.

Mean number of Oligonychus perseae collected per leaf from April-December, 2000 (A) and percentage of collected leaves infested with Oligonychus perseae from April-December, 2000 (B).


Fig. 2.

Average total mite days for each treatment calculated over 37 weeks of observation. Means with same letters were not significantly different at the 0.05 level (ANOVA).


Fig. 3.

Mean percentage of leaf area damaged by Oligonychus perseae feeding activity across treatments at six different sampling intervals on defoliated leaves. Means with same letters were not significantly different at the 0.05 level (ANOVA).


Fig. 4.

Mean number (±SEM) of Oligonychus perseae caught per white sticky card at each sampling interval.


Fig. 5.

Mean (±SEM) total number of Oligonychus perseae captured per sticky card in each cardinal direction. Means with same letters were not significantly different at the 0.05 level (ANOVA).


Fig. 6.

Mean number of phytoseiids recovered for each 10 leaf sample by treatment over time; Galendromus spp. (A), Euseius hibisci (B)


Fig. 7.

Average total predator mite days for each treatment calculated over 37 weeks of observation for Galendromus spp. (A) and Euseius hibisci (B). Means with same letters were not significantly different at the 0.05 level (ANOVA).


Fig. 8.

Mean weekly maximum and minimum temperature (°C) and percentage relative humidities recorded in the study plot at Bonsall, California over April-December, 2000.

Published: 1 March 2002

Back to Top