ON Eco~oc~ Spatial and Temporal Dynamics of Spider Mites ... · Spatial and Temporal Dynamics of Spider Mites (Acari: Tetranychidae) in ‘Thompson Seedless’ Vineyards ... canopy

POPULATION E c o ~ o c ~

Spatial and Temporal Dynamics of Spider Mites (Acari: Tetranychidae) in ‘Thompson Seedless’ Vineyards

RACHID HANNA, LLOYD T. WILSON,’ FRANK G. ZALOM, DONALD L. FLAHERTYF AND GEORGE M. LEAVITT3

Department of Entomology, University of California, Davis, CA 95616

Environ. Entomol. 25(2). 3 7 0 3 8 2 (1996) ABSTRACT Quantitative knowledge of beneficial and pest arthropod distribution patterns in relation to plant vigor and habitat structure is essential for the development of reliable and cost-effective methods for the assessment and management of arthropod populations in agro- ecosystems. The spatial and temporal distributions of the spider mites, TetrunycAus pucijii~7~5 McCregor, Pacific spider mite, and Eotetrunychus willarwttei (McGregor) were determined in 3 ‘Thompson Seedless’ grape, Vitk uin$era L,., vineyards located in the central San Joaquin Valley of California. I: pacifim.~ was the dominant species in 2 of the vineyards, whereas E . willanwttei was dominant in the 3rd vineyard. Z pacijifi~~s infestations were most severe in vineyard areas with low vine vigor compared with area.. with high vigor. In contrast, E. wil- Zumttei appeared to be less sensitive to differences in vine vigor in the vineyards where the effect of vine vigor could be tested. Spatial variations in between- and within-vine distributions of spider mites were significant during most of the growing season in all vineyards. Z p~~i j i f i cus and E. willumettei distributions overlapped considerably on the south and north zones of vine canopy but differed between top and interior zones. The majority of T. pacificus and E. willamettei were found on leaves near the base of shoots during spring, arid on midshoot leaves during summer, Aggregation indices indicated that both spider mite species were aggregated during most of the season, with ‘I: puc~fi~vus being more aggregated between vines and within vines. The biological sipificance of T ~ U C @ C U S arid E. willarrLettei spatial dynamics are discussed along with implications for population assessment.

KEY WORDS spatial distribution, aggregation

Tetrarychus pac$ficus, Eotetrurychus wilhnwtiei, Vitis vinfern, plant vigor,

TWO SPECIES OF spider mites, Tetranychus PU:$- cus McGregor, Pacific spider mite, and Eotetran- ychus willamettei (McGregor) are of major eco- nomic concem in California vineyards. Although feeding damage on leaves of grape, Vitis vinifera L., by both species similarly affects leaf photosyn- thetic rates and stomatal conductance (Welter et al. 198Ya), T. pacificus has a greater potential for reducing yield and quality of most grape cultivars (Kinn et al. 1974, Welter e t d. 1989b). The pest status of E. willumettei is more dependent on the grape cultivar and does not normally affect yield and quality of ‘Thompson Seedless’ grapes (Flah- erty and Huffaker 1970); however, high levels of this species will reduce yield and quality of ‘Chen- in blanc’ arid ‘Zinfandel’ grapes in vineyards located in the Sierra Nevada foothills (McNally and Famham 1985, Welter et al. 1989b).

‘Current address: Department of Entomology, Tcxas ACtM

2University of C:aliforiiia Cooperative Extension, Visalia, CA

3University of California Coopwative Extension, Madttrs, CA

University, College Station, TX 77843.

9329 1.

93637.

0046-225X/96/0370-OG82$02.O(MO 0 1996 Entomological Society of America

In addition to their impact on grapevines, ?: pa- cifzcus and E. willumettei may interact negatively 011 Thorripson Seedless vines (Hanna 1992). E . wil- lawtetlei is generally found on the grapevines earlier in the season and persists longer into the fall compared with T. pac$cu.s (Flaherty and Huffaker 1970). When present in spring, E . willamettei can be beneficial as alternate prey for the predatory mite Metaseiulus ( = Typhlodronzus = Gubndrom- nus) occiclentalis (Nesbitt) (Acari: Phytoseiidae), enhancing the suppression of T pac$cus later in the season (Flaherty and Huffaker 1970, Hanna 1992). Other evidence also indicates that on Zin- fandel grapevines, damage by E. willamettei in early spring may result in lower ?: pac$ms densities during the summer (Karban and b!nglish-Loeb 1990); however, this effect was not observed in Thompson Seedless vineyards (IIanna 1992).

Despite the importance of spider mites in vineyards, quantitative information on their spatial dynamics throughout the grape growing season is lacking. Flaherty and Huffaker (1970) studied the distribution of spider mites on vines but did not describe seasonal changes in distribution patterns, or possible factors affecting these patterns. Several

April 1996 HANNA ET AL..: SPATIAL A N D TEMPORAL DYNAMICS OF SPIDER MITES 371

biotic and abiotic fiactors can caiise large variations in vine vigor which may affect spider mite distributions within vineyards. In addition, vine canopy structure affects the amount of sunlight intercept- ed by leaves and can alter within-vine physical factors such as temperature, humidity, and air move- ment (Smart 1985, Grimes and Williams 1990). Variations in these physical factors can affect the abundance of spider mites greatly (Iloltzer et al. 1988). Quantitative knowledge of spider mite distribution patterns in relation to vine vigor arid vine canopy structure is essential for the development of reliable and cost-effective methods for the as- sessinerit and management of spider mite populations in vineyards.

This study presents quantitative information on within-vineyard and within-vine distributions of I: paci$c~cs arid E. willarnettei in 3 commercial Thompson Seedless vineyards. This information is iised in developing a general approach to spider mite sampling in vineyards.

Materials and Methods Growth and Development of Thompson Seed-

less Vines. Many grape cultivars are grown com- mercially throughout California, all belonging to V. wint$eru LA. Although most of these cultivars have similar growth patterns, differences in prunin training of canes, and differences in uses o har- vested grapes (i.e., table, wine, and raisin grapes), have led to cultivar-specific structure and manage- inent of vine canopy (Winklcr et d. 1975). We have restricted our investigations to Thompson Seedless grape, which is the most widely grown cultivar in the San Joacpin Valley of California.

Thompson Seedless vines used in raisin produc- tion are cane-pruned during the winter months,

=5 canes per vine and 15 biids per cane lea?$ (Win er et al. 1975). Hudbreak (or the initiation of vegetative growth) of Thompson Seedless grapes riorrnally occiirs after the accumulation of 4 6 DD (degree-days; >10"C) from 20 Fcbniary (Williams et al. 1985). After the accumulation of -200 DD from bud break in the absence of severe water stress, vine leaf area increases linearly reaching a maximum of =23 in2 per vine at 1,000 DD (Williams 198721). Total leaf area decreases during the remainder of the season because of shoot trim-

Thompson Seedless vines are normally planted in an east-west orientation and trained on a single wire trellis or a cross arm with 2.5 m between vines (within a row) and 3.5 m between rows. On mature vines with fully developed canopies, both training systems result in 4 canopy zones which receive different amounts of sunlight (Mullins et al. 1992). The top zone of the canopy is generally fully exposed to siinlight throughout the day, whereas the north and soiith zones receive varying levels of sunlight depending on time of day arid season (Mullins et 4. 1992). By contrast, the interior of

f p and

~ ming and leaf senescence.

I

I

the canopy is generally fiilly shaded from siinlight. Differences in light intensity may influencc the distribution and abundance of arthropods (ßult- man and Faeth 1988).

Field Sampling Methods. We conducted our studies during the 1986 grape growing season in 3 Thompson Seedless vineyards located in the San Joayuin Valley near Fresno (Fresno County), Ma- dera (Madera County), and Dinuba (Tiilare Coun- ty). The 3 vineyards were selected on the basis of their history of spider mite outbreaks, spider mite species composition, and the distribiition of vine vigor. In this study, vine vigor is defined qualita- tively; the majority of vines in the inoderate to high-vigor areas had dense foliage and long shoots, whereas those in low-vigor areas had relatively sparse foliage and short shoots. The Dinuba arid Fresno vineyards had areas with relatively high and low vine vigor, whereas the Madera vineyard had vines with only moderate-to-high vigor. Vines were trained on single trellis wire with canes of adjacent vines in contact at their terminal ends and forming a continuum within a row; however, their foliage rarely touched vines of' adjacent rows.

Two 0.30-ha sampling plots were established within each of the 3 vineyards, 1 plot in each of the high- and low-vigor areas of the Diniiba and Fresno vineyards, arid 2 plots (which were similar in vine vigor) in the Madera vineyard. The sampling plots were at least 30 m from row ends and were separated by 70, 60, and 40 1711 in the Madera, Dinuba, and Fresiio vineyards, respectively. Within each plot, we designated 9-15 experimental units (consisting of 3 adjacent vines per mit) from which we sampled leaves to determine between- arid within-vine distribution of I: pacijicms and E . willamettei. The 3-vine experimental unit was selected to minimize the effect of repeated leaf removal on mite densities.

We sampled all vineyards at intervals of 7-14 d beginning on 15 April ("4 wk after bud brcak) in the Fresno vineyard, 6 May in the Dinuba vineyard, and 15 April in 1 plot in the Madera vinc- yard. We initiated sampling in the 2nd plot in the Madera vineyard on 16 August. We terminated sampling foIlowing propargite (Omite) treatments on 9 July in the Fresno vineyard and 19 July in the Dinuba vineyard. Miticides were not used in the Madera vineyard, allowing sampling to continue until 13 September, when numbers of 2: pacijims and E. willumttei populations declined to near 0 levels.

On each sampling date in the Dinuba and Fres- no vineyards, we sampled fifteen .?-vine experimental units from each of the 2 sampling areas. In the Madera vineyard, we sampled 15, 3-vine experimental units from only 1 area until 6 August. On 16 August, we initiated sampling from a 2nd area in the Madera vineyard, but we reduced to 9 the number of experimental units sampled from each of 2 sampling areas in this vineyard. We restricted sampling to primary shoot leaves which

., ,. I ,

372 ENVIHONMENTAL ENTOMOLOGY Vol. 2S, no. 2

cornprisc =70% of total leaf area of Thompson Seedless vines (Williams 1987a). Ihring April, May, and the 1st half of June (the period of rapid vegetative growth of vines), wr sampled pririiary shoots originating from vine tnniks, and from l-yr- old wood iii the iiorth and south vine zones. TWO

leaves werca selected from the basal and middle portions of each shoot. As the vine canopy became fully developed (at approximately the middle of Tuiie), 2 atlditional slioots were included in the sainpling, 1 shoot from the interior and 1 shoot From the top canopy zones, for a total of 4 shoots (zones) from each experimental unit. We initiated leaf sampling from the terminal region of each shoot by mrd-June. The 3 locations on a shoot rep- resented different age classes of leaves. All S ~ I W

pled leaves were classified according to sarnpling area, vine, vine zone, and position on a shoot (leaf age). In addition, all l rmw were classified as sun- exposed or shaded at the time of sampling, which typically occurred between 0900 arid 1500 hours. All leaves were inimediately chilled arid brought to tlae laboratory, where they were stored at 3°C iiiitil they wcre exainined under a binocular microscope. We counted all life stages of spider mites arid predatory mites, arid immatiires and adnlts of the sixspotted thrips, Scolothrips sexmuculnlzrs (Per- gartde) (Thysnnoptera: Thripidae).

Statistical Analyses. We used a random effect, 4-way nested analysis of variance (ANOVA) (PROC GLM, SAS Institiite 1989), stratjfied by date, to coinpare iriean densities of 7: pnczjLicirs and E. wil- Zairwttei between the 2 sampling areas of each mneyard, aniong vines ( I st experirnental iinit, nestcd within sampling area), among zoiies (2nd ex- perirrieiital unit, nestcd within vine), and within zones ( i t a . , k i f age effect was the 3rd experimental unit, nested within zone, and was tlie residual term). ANOVAs were condiicted separately for each vineyard. All reported densities were calcii- latcd on a per-leaf basis. We used the Ronferroni procedure (Milliken arid Johnson 1984) to adjust probability vahics becai ~ s e the aiialyses were strat- ificd by sampling datc. Rcpeated-rrieasures analyses ;across dates were not appropriate because we did not sample all zones and leaves within zones on each sainpling datc. Stratification by sampliiig date allowed the estiiriation of teinporal changes in the size and proportion of the varihility of mite densities explainrd by variatioii between vineyard arcus, vines, zones, and leaves within zones (PROC VABCOMP, SAS Institiite 1989). We used Border- roiu-adjiisted linear contrasts (Milliken and John- son 1084) to coinpare mite densities in the top arid south zones witti densities in thc north and interior zoncs oi the vines. In addition, we used a 1-way ANOVA (stratified by date) to compare mean mite densities on shaded and sun-exposed leaves. We log-transforaied all data to correct for heterogeneity of variance before conducting the statistical

We used the Taylor (Taylor 1961) and Iwao (Iwao 1968) regression iriodels to estimate the degree of between- and within-vine aggregation of ?: pac$cus and E. willnnzettei. The merits and appli- cability of these techniques for the estirnation of species aggregation were reviewed by Taylor (1984) and Kiino (1991). The 2 models provide essentially the same information about the spatial patterns of organisms biit differ in the interpreta- tion of their parameters. In the Ttylor model, sample variance ( S 2 ) is relatcd to sample mean density (f) by the relationship s2. = a . f!3, This equation is typically linearized witlt logaritlirnic transforina- tion: 1n(,s2) = In(n) + b ‘ In@). The b coefficient is species-specific and is a measure of the density dependence of aggregation (Taylor 1961, Taylor et al. 1978). Both CI and b determine thc degree of aggregation of a species (Banerjee 1876, Wilson 1985, Ilanna and Wilson 1991).

The Iwao linear regression model m = OL + p . R (m is the Loyd [I9671 inean crowding [m = f + s 2 E - 13 arid f is sample mean density) provides information on 2 aspects of a species aggregation. The y intercept a of the model estimates the size of the smallest unit of the population (number of individuals), whereas the dope (p ) estirnates the rate of change in the aggregation of the. srnallcst population iini t as population density increases. We used a 2-step process to estirnate betwecn-vine parameters of the Taylor and Iwao regression. In the 1st step, we calculated 7: pac$c:us and E. zuil- Zamttei mean densities per leaf for each vine, stratified by vineyard, sampling (late, aiid vineyard area. In the 2nd step, we calciilated between-vine means and variances for each sampling date and vineyard area. In total, 35 and 33 data points for 2: pac$ficx~s and E. zoillarnettei, respectively, were used in the regression analyses. Dates when spider mites were absent were not incliided in tlie analyses. To estimate within-vine aggregation indices, we calculated within-vine density nieans and variances for each mite species, stratified by vineyard, sampling date, vineyard area and vine. In total, 290 and 200 data points for ?: pacificus and E. zoilln- mettei, respectively, were used in tlae regressions. Data were analyzed with PROC REG (SAS Insti- tute 1989).

,

’

Results

Distribution within Vineyards, Largc betwecn- and within-vineyard differences were observed in the abuntlance of the 2 spider mite species during the 1986 study. T paczjLiicl~~ was the domiiiant species in the 13iniiba and Fresno vineyards, and E willarnettei was the doniinant species in the Ma- dera vineyard (Fig. l ) . For brevity, we present analyses of data only from the Dinuba and Madera vineyards (Tables I and 2) because of the similarity in the distribution and abuiidance of T paci$c~i~ and E . zoiZlnrn&-tei in the Dinuba and Fresno vineyards.

”

April 1996 HANNA ET AL.: SPATIAT. A N D TEMPORAI. DYNAMICS 0 1 ' SPII3)KH MITES 3 73

120- A. Dinuba 120- B. Fresno 120- c. Madera

%loo- 100- 100-

3 L 80: 80 - 80 - Low vigor

6o 8 60: 60 -

2 40 - ---CI-- High vigor 40 *G

:- Medium vigor 3

8 20- 5 O d l ' ' ' 2 I I . I . , I

Date of Sampling Fig. 1 . Seasonal patterns of 'I: p c ~ i j k i ~ s (A-C:) and .E. i.oihnettei (1.1-47) abmdaiice in 3

vineyards (A and D, Dinuba; 13 and E, Fresno; C iind F, Madera). L3at.a points are rneim mite de bars are standard errors. Data are plotted separately for vines with high (G), B I I ~ low (J) vigor at the 13irii11>it ;uitl Fresno vineyards (1 high and low vigor plot iit em$ vineyard), and for vines with nietliurn vigor (13) at the Madera vineyarcl ( 2 mediurn-vigor plots were saiiip1c:d).

In the Dinuba and Fresno vineyards, T ?iucijkus occurred earlier, increased more rapidly, and it occiirred at significantly higher deiisitics on low-vigor vines than on high-vigor vines (Fig. 1 A and R; Table 1). Differences in vine vigor explained 58 and 48% of the variation in T pacijims densities during the increasing phase and at peak density, respectively. In the Madera vineyard, where vines were of similar vigor in the 2 sainpling areas, 7: pncijicus densities and their seasonal patterns of growth and decline also were similar (Fig. IC; Ta- ble 2 ) . Diffcrences between the 2 sampling areas explained 54% of tlie spatial variation in T paci-

Populations of E. willnnwttei did not appear to be irifliiertced by vine vigor. E . zuillnnu?ttei densities were veiy low in the Dinuba and Fresno vineyards, with equal densities occurring on high- arid low- vigor vines (Fig. 1 D and E; Table 1). The area effect explained 112% of the spatial variation in E . willnmttei densities in the Dinuba vineyard (Table 1). E . tuillnmettui wa.. substantially ~riore abundant in both areas in the Madera vineyard (88.5 rt 12.6 [mean t SEM] initcs per leaf at pcxik abundance) compared with its abundance in the

. ficus densities (Table 2).

Dinuba and Fresno vineyards (2.8 3. 1.13 and 2.8 3- 0.97 mites per leaf at peak abiindance, respectively). The area effect expl;dned 52% of t h r h tial variation in E. wille,, ei aliiindancc~ i i i the. Madera vineyard (Fig. 1 I'; Table 2).

distribution of mites airlong vines (within ;a siiin- pling area) were gc~~ierally siiililar in all 3 viiieyards (Tables 1 and 2). Mean 'f puc@x~s densitirc in tlic Diiiubn vineyard were significantly dificrviit among vincs, diiririg t 1 1 ~ rapid incrcasca (27 Jiii ic,)

and peak (9 JliIy) ph21Ses of d)Illldilnce. Uiffbr- ences among vines also expl&ied 2445% of the spatial variation in incan 'I: pm:ijiiCris d n i n ~ l a ~ (Table I), except during thc initial incrcwsc pli, (10% of spatial variation) when T p~ciflcus deiisities were low. Similar patterns were observed in the Madera vineyard ( T d ~ k 2). T l i ~ grcatcxt effect occurred during thc lapid iiicreasc a i d peak phases of mite abiindance (29-66% of spatial variation). But the magnitiicle of tlie vine effect was negligible (54%) during tlie declining phase.

Among-vine distribution of E. willanief~ei followed ternporal patterns similar to T pczcificus, but the significance and rnagriitudca of' the cffi%ct w ( w

Distribution Among Vines. Diffcrtwc

Vol. 25, no. 2 374 ENVIRONMENTAL ENTOMOLOGY

Table 1. Results front nested ANOVA of T. pacificus and E . wiUamttei densities in relntion to siunpIlin# area, vine, canopy zone, and leaf age in the Dinuba vineyard

P valiie for 2'. padficus (do P values for E. willunwttei (df) n . I, uare.- ßetween area& Among vines Among zones Leaf agec Between Among Ainong 1,t:af

areas vines zones W('

Effect tests - 8 June 0.1.52 (1) 0.089 (28) 0.000* (60) - (90) -. ._

I8 June 0.101. (1.) 0.000 (28) 0.037 (60) - (90) 0.158 (1) 0.024 (28 ) 0 . I ) l l (60) - (90) 27 June 0.070 (1) 0.000* (28) 0.000 (89) - (237) 0.841 (1) 0.001 (28) 0.089 (89) - (237) 9 July 0.000" ( I ) 0.000* (28) 0.001 (90) - (240) 0.079 (1) 0.000 (28) 0.051 ($10) - (210)

Proportion of variation explained 1.9 July 0.080 (I) 0.000 (28) 0.008 (90) - (240) 0.016 (I) 0.000 (28) 0.000 (90) -- (240)

- - 8 June 0.02 0.10 0.68 0.1.6 - - 18 June 0.06 0.34 0.12 0.47 0.01 0.10 0.10 0.79 27 June 0.08 0.45 0.10 0.36 0.02 0.1.4 0.14 0.70

0.48 0.24 0.06 0.23 0.05 0. I4 0.08 0.73 0.11 0.34 0.15 0.40 0.12 0.21 0.15 0.52

9 July 19 July

*, P < 0.05 after ßonferroni adjustment. " Effect tests for 6 May and 23 May were not included because of the absence or near-0 abundance of spider Initc!s.

Area effect represents high versus low vine vigor in this vineyard. This is the residual term in this nested model, so P value for the leaf-age effect could not be calculated.

less than that of I: pacijicus. In the Dinuba vineyard, where E. willarnettei density was low compared with densities in the Madera vineyard, the vine effect explained 1O-21% of the spatial variation in E . willamettei abundance. In the Madera vineyard, where E. willarnettei density was very high, among-vine differences were greater (but still not significant) during the increasing and peak phases of E. willarnettei population growth (27-

33% of spatial variation), becoming negligible during the declining phase (Table 2).

Distribution Among Vine Zones. I: pnc$cus and E. willarnettei showed dynamic distribiition patterns within vines or among zones in all 3 vineyards. Complete analyses are presented for both species in the Diniiba and Madera vineyards (Ta- bles l and 2). For brevity, we present graphs (as mean density per leaf for each species on 4 vine

Table 2. Results from nested ANOVA of T. pacificus and E . willamttei densities in relation to sumpling area, vine, canopy zone, and leaf age in the Madera vineyard

Date('

27 June 09 July 19 July 29 July 08 Aug. 16 A L I ~ 23 Aug. 29 Aug. 06 Sept. 13 Sept.

P value for T. p a o i j h s (dt) P value for E. wihrw l td (df)

Between Among Among . Leaf Between Among Among are& vines mnes age': areas vines zollc?s

Effect tests - 0.000' (14) 0.017 (45) - (120) - 0.000 (14) 0.000 (45) - 0.000* (14) 0.061 (45) - (120) - 0.000 (14) 0.007 (45) - 0.000* (14) 0.419 (45) - (120) - 0.001 (14) 0.000 (45) - 0.000 (1.4) 0.340 (45) - (120) - 0.000 (14) 0.316 (15)

0.386 (I) 0.156 (16) 0.000 (54) - (144) 0.811 (1) 0.436 (16) 0.000 (54) 0.575 (I) 0.476 (16) 0.000 (52) - (1.42) 0.985 (1.) 0.052 (16) 0.002 (52)

0.342 (1) 0.589 (15) 0.000 (51) - (136) 0.360 (1) 0.622 (1.5) 0.000 (51) 0339 (1) 0.033 (1.6) 0.170 (54) - ( 144) 0.529 ( I ) 0.372 (1.6) 0.000 (54)

- 0.774 (14) 0.000 (45) - (120) - 0.026 (14) 0.014 (45)

0.064 ( 1 ) 0.662 (16) 0.000 (54) - (144) 0.608 (1) 0.015 (16) 0.010 (54)

1,t:d age

- ( I 20) - ( 120) - (120) -- ( 1 20) - (120) - (144) - (142) -- (144) - (136) - (144)

Proportion of variation explained 27 June - 0.66 0.06 0.28 - 0.33 0.29 0.38 09 July - 0.43 0.08 0.49 - 0.27 0.16 0.57 19 July - 0.57 0.01 0.43 I 0.27 0.28 0.45 29 July - 0.29 0.02 0.69 - 0.29 0.03 0.68

16 Aug. 0.01 0.06 0.38 0.55 0.01 0.01. 0.38 0.60 23 Aug. 0.01 0.05 0.35 0.59 0.02 0.10 0. I9 0.69

06 Sept. 0.01. 0.02 0.34 0.63 0.00 0.03 0.38 0.59 13 Sept. 0.00 0.08 0.07 0.85 0.00 0.02 0.30 0.68

08 Aug. - 0.05 0.43 0.52 - 0.12 0.16 0.72

29 Arrg. 0.04 0.04 0.35 0.59 0.02 0. I 2 0.15 0.7 I

*, P < 0.05 after ßonferroni adjustment. Effect tests for dates 15 April, 6 May, 23 May, 8 June, and 18 June were not included because-: of the absence or near-0 numhcrs

Only 1. area was sampled before 16 August: both areas sampled in this vineyard had vines of medium vigor. This is the residual term in this nested model, so P value for the Id-age effect could not bc! c.dcnlated.

of spider mites.

375 April 1996 HANNA ET AL.: SPATIAL A N D TEMPORAL DYNAMICS OF SPI13ER Ml'mS

' f i b 4 0 0 h' / I /t

B. 27June / 60

45 30

15

0

Fig. 2. Population distribution of T pacijiws on 3 leaf positions in 4 vine canopy zones in the Dinuba vineyard. (A) Initial increase phase, 8 June. (B) Rapid increase phase, 27 June. (C) Peak phase, 9 July. (0) Declining phase, 19 Jdy.

zones and 3 leaf ages) for selected phases of population growth: initial, rapid, peak, and declinin phases; for 2: pac$cus in the Dinuba vineyar (Fig. 2) and the Madera vineyard (Fig. 3); and for E. willarnettei in the Madera vineyard (Fig. 4).

Tetranychus pacijiws densities in the Dinuba vineyard differed significantly among 3 vine zones on 8 June (Fig. 2A; Table 1). The greatest proportion of explained variation (68%) in mite densities occurred during the initial phase of population growth, declining to 515% of the total variation in mite abundance during the rapid increase, peak, and declining phases (Table 1). During the initial increase phase, T. pacijicus occurred in greater numbers on top and south zones compared with the north zone (Fig. 2A); ( B = 0.03, df = 90, P < 0.05) ( B is the Bonferroni least significant difference for comparing zones; interior shoots were not sampled in the Dinuba vineyard during this phase). During the rapid increase, peak, and declining phases of population growth, ?: pacijicus displayed a strong preference for both south and top shoots which harbored 71-88% of 7: pac$cus densities, compared with north and interior shoots which harbored 12-29% (Fig. 2 B-D); (B = 0.30-

! 0.75; df = 90, 240; P < 0.05). Densities were generally not significantly different on top compared with south zones, and on north cornpared with interior zones ( P > 0.05).

Tetranychus pacificus distributions in the Ma- dera vineyard with respect to vine zone di€€ered slightly from the Dinuba vineyard (Fig. 3). In the Madera vineyard, the zone effect explained a non- significant and negligible proportion of the variation in T. pacijicus abundance during the initial and rapid increase phases of growth (Table 2). During the initial increase phase, 65% of ?: pacificus were found on top and south zones and 35% 011 north and interior zones (Fig. 3A). During the rapid increase phase, 60% of Z: pac+us werc found on top arid south zones and 40% were found on north and interior zones (Fig. 3B). These differences in T. pacijicus distributions during the initial and rapid increase phases were not statistically significant ( B = 0.20-0.95; df = 120, 144; Y > 0.05). The zone effect became more pronounced later in the season when it ex lained 3543% of spatial variation in 2: pacijicus Iensities during the (extended) peak phase of growth (Fig. 1C). The contribution of the zone effect became negligible

376 E NVI RO N ME NTAI , E NTOM 01 ,OGY Vol. 25, no. 2

c 0 a, L- a, n v) a, .- .- L

r-:

E Q, zn a,

h

30 25 20 15 10 5 0

30 25

30 20 25 15 20 10 15 5 10 0 5 0 8

Fig. 3. Population distribution of T. pncijijiczis on 3 leaf positions in 4 vine canopy zones in the Madera vineyard. ( A ) initial increase phase, 27 June. (13) Hapid increase phase, 19 July. ( (2 ) Peak phase, 16 August. (1)) I.)eclining phase, 1.3 September.

(7%) on 13 September during the declining phase (Fig. 3D; Table 2). T pacijicus occurred in greater niimbers on top and south zones during the peak phase of growth compared with north and interior zones (Fig. 3 C ) ; ( B = 0.18, df = 144, P < 0.05). The top zono siipported greater numbers of T pa- c$ficu~ coinpar(2d with the south zone ( P < 0.05), and the north zone supported higher numbers conipar~d with the interior zone ( P < 0.05). Jhr - ing the declining phase of the popiilation cycle (Fig. 3D), 2: pac~ f i~us numbers remained higher on top and south zones (70% of T pacijkus numbers) compared w'th north and interior shoots ( B = 0.82, df = 144, P < 0.05). The top zone supported similar mite densities compared with the south zone ( P > 0.05), biit the north zone supported significantly greater mite densities compared with the interior zone ( P < 0.05).

Eotetranychus willamettei also showed differences in abiindance among vine zones in the Ma- dera vineyard (Fig. 4). In the Dinuba and Fresno vineyards, where E. willnmettei densities were low, its abundance was slightly higher on the north and intcrior zones compared with the top arid south zones (data not shown). This preference was not

significant at any time during the season, and the zone effect did not accoiint for > 15% of the spatial variation in mite abundance (Table 1; analysis not shown for the Fresno vineyard). However, E. willamettei abundance was not detectably different among the 4 vine zones in the Madera vineyard after the Bonferroni adjiistinent on any of the 10 sampling dates (Table 2). Durin the initial increase phase, vine zone explainecf 29% of spatial variation in E. toillanwttci abundance, with greater abundance occurring on the south zone (Table 2; Fig. 4A); ( B = 0.18, df = 120, P < 0.05). During the rapid increase phase, E. willairmttei was more abundant on interior arid north zones (75% of E. willarnettei niiinbers) compared with south and top zones (Fig. 4B); ( B = 2.06, df = 120, P < 0.05). Despite this preference for north and interior zones, the zone effect explained only 16% of the spatial variation in E. willninettei abiindance during the same period. Differences among zones declined during the peak and declining phases; however, E. willamettei remained in greater abiindance on the north and interior zones compared with the top and south zones (Fig. 4 C and D; L3 = 0.73- 1.01; df = 142, 144; P < 0.05).

April 1996 HANNA ET AL,.: SPATIAL, AND TEMPORAL DYNAMICS 01.' SPIDER MI'I'ES 377

10 8 6 4 2 0

200 150

100 50

0

B. 8August 125 100 75 50 25 0

125 100 75 50 25 0 B

.

Termin, al\f Interior

40

30

20

10

0

Fig. 4.. Population distrihtion of E . willanwttei on 3 leaf positions in 4 vine canopy zones in the Madela vineyard. (A) Initial increase phase, 1.9 July. ( U ) Rapid increase phase, 8 August. (C) Peak phase, 23 August. (1)) I)ecbhring phase, 6 September.

Distribution Among Leaves Within Zones. ?: pac@us and E. willawttei occiirred in greater abundance 011 specific lcaves within zones. Similar spatial and temporal patterns of the within-zone effect were observed in id1 3 vineyards. Signifi- cance levels of the leaf' age effect (and interaction between zones arid leaves) could not be calculated because the leaf position (or age) effect was the residiid in the nested model. It was possible, however, to calculate the proportion of spatial variation in mite abundance explained by the leaf age cffect. Leaf age effect in the Dinuba vineyard explained a large proportion of the spatial variation in T. PU-

cijiccws abundance, except during the initial growth phase when tlie leaf' position effect explained only 16% of the spatial variation in T. pac$cus abundance. During the remaining growth phases, the leaf age effect explained 36, 23, and 40% of the spatial variation in T pac@x~s abundance during the rapid increase, wak, and declining phases, respectively (Table 1 f . In general, T. pnc$cus was inore abundant on midshoot leaves, with basal and terminal leaves supporting approximately similar mite densities during most of the population cycle (Fig. 2). These differences became negligible after the rapid decline in ?: pnc$cu.s densities. ?: pnci-

fic1~s distribution patterns 011 leavcs witliin m i w s in the Madera vineyard were similar to those observed in the Dinuba vineyard (Fig. 3 ) . The lcaf age cffect explained a large percentage of thc s p a - tial variation in ?: pir$cus abundance throiighout most of the population cycle (Tab1e 2). Eotetrmychus willnmetttei distributioii patterns 011

specific leaf positions within xones wew siniilw to those of 'I: pciajicus. In tlie Dimitxi virieyar-tl, wlic-rci E. tuillmwttei deiisities were very low, the lcaf p s i -

in mite abuiidance (Table 1). E. zuillairwttei wit< found primarily on basal leavw during the initid increase phase of pop1 dation growth, and ill eqilal numbers on basal arid middle leaves diii-ing tlw r ~ - inainder of the sampling period E. zoiZZr~~u~t/c~i w x ~ found rarely on tc~~nirial leaves i n the Dinuba vi iw yard, where its abundance was very low (Fig. 1 U). In the Madera vineyard, E. willnnettei was fbiirid in greater abundance on mid-shoot leaves, whereas basal and terrninal leaves siipported equivalent ptwxsilt- ages of E. wiZZaw&?ttei densities (Fig, 4). Tliesc. among-leaf differences were most proiioiiiicctl during the rapid increase and peak phxm of abundance (Fig. 4 l3 and C). Differences lxtween Icavc~ declined later in the season, when only slightly Iiighcr

tion effect explained 52-79% of tlie sp~itid vari a t ' 1011

Vol. 25, no. 2 378 ENVIRONMENTAL ENTOMOLOGY

B. Madera 120

100

80 60

40

20

0

A. Dinuba . - Exposed __e_ Shaded

120- 3 CJ 'Z l o o r

1 . 1 . I

C.

4

3 :r T

100

80 60

40

20

0

Date of sampling Fig. 5. Popidation distributions o f T. p a c i j ~ u s (A and B) and E . will"& (C and U) on sun-exposed (E) and

shaded (J) leaves in the Dinuba and the Madera vineyards. Data points are mean mite densities per leaf; error bars are standard errors.

densities of E. willurwttei occurred on middle leaves during the declining phase of abundance (Fig. 4D).

Distribution in Relation to Sunlight. At the time oiir saim les were taken, all selected leaves wcre classifieias sim-exposed or shaded. The resiilts indicated that the frequency of leaf exposure to sunlight in the 4 vine zones changed through the season. The most consistent patterns were ob- tained for the top and interior zones which were, respectively, the most and the least likely to be sun-exposed ($1 = 3.0-79.3, P < 0.05). Leaves in the soiith and north zones received a variable amount of sunlight but were equally likely to be sun-exposed through mid-July ($1 = 0-1.8, P > 0.05). Later in the season, leaves in south zones

also tested for the association between leaf exposure to sunlight and spider mite densities across dates and by date for each of the 3 vineyards. i? pacijicus occurred in reater densities on exposed compared with shade i leaves across dates in each of the 3 vineyards (t = 7.14-15.4; df = 1,674- 2,536; P < 0.05); however, this pattern was observed only during the peak and declining phases of abundance (Fig. 5 A arid B), (t = 3.77-7.81; df = 178-214; P € 0.05). In contrast, E . willamettei occiirred in greater abundance on shaded than on exposed leaves only in tho Madera vineyard (across dates: t = 7.82; df = 2,536, P < 0.05). These differences were caused primarily by the greater abundance on shaded than on exposed leaves dur-

were inore likely to be sun-exposed than those in the north zones = 5.9-17.3, P < 0.05). We

ing the declining phase (Fig. 5 6 ; t = 3.443.88; df = 202-214; P < 0.05).

April 1996 HANNA ET AL.: SPATIAL A N D TEMPORAL DYNAMICS OF SPIDER MITES :379

Table 3. Between- and within-vine aggregation indices of 2'. pacificus and E . wiUumctei

13ehveen vines Within vines

n Intercept t SEM" Slope t SEMb rs! n Intercept t SEM' Slope i: SEMb 9 Species

~- ~~

Taylor power law 1..76 t 0.02 0.97 T. pncij5~ur.v 35 0.64 t 0.08 1.57 rt 0.07 0.93 290 1.59 t 0.05

E. roillanmtttei 33 0.58 t 0.05 1.33 t 0.05 0.96 200 1.70 t 0.07 1..58 t 0.02 0.96

'I: ~Jacijiclls 35 11.8 t 5.4 1.72 t 0.21 0.68 290 12.7 t 0.1 2.78 t 0.06 0.86 E. willorrwttsi 33 4.36 i: 1.1 1 .L2 -+ 0.04 0.97 200 1.2.9 rt 3.5 1.88 rt 0.08 0.71

Iwao patchiness regression

'1 Intercepts are In ri Li)r Taylor power law and a for Iwao patchiness regression. Slopes are b for Taylor power law and /3 for Iwao patchiness regression.

Aggregation Indices. Aggregation indices based on the Taylor (1961) arid Iwao (1968) regression techniques were calculated for both I: pucifcus and E. willumettei. Data from all 3 vineyards were pooled in the analyses. The Taylor power law pro- vided a highly significant fit of the relationship between variances and the means of between- and within-vine densities of both species (Table 3). Both a and b coefficients indicated that I: pacijk7~7 and E. willamettei were aggregated in their distributions between and within vines [ln(ap,, u ,, u , , , ~ ) > 0 and hl,t,, b,,, b,,,,:, b,,, > 1, P < 0.0ff. The p and w subscripts denote the Coefficients for I: pucizfc~~s and E. willunwttei, respectively, and the e and g subscripts denote between- and within- vine coefficients. T pucijijicus was more aggregated (at the densities recorded in this study) than E. willumettei in both between- and within-vine distributions (hr,, > h,,, t = 2.75, df = 64, P < 0.01 and hi,, > b,L,g, t = 6.06, df = 486, P < 0.01; between-vine and within-vine u coefficients did not differ between species, P > 0.05). Both species were also significantly more aggregated within vines than between vines [In(ul,,) > ln(u!,,), t = 2.06, df = 331, 0.01 < P < O.OS;\n(n,,) > ln(uto,), t = 2.38, df 229, P < 0.01; hpg > b,,, t = 3.37, df = 331, P < 0.01; b,, > b,,,, t = 4.62, df = 229, P < 0.01).

The Iwao regression model also indicated that both 2: pncijic7~s and E. willunzettei displayed aggregated distributions (Table 3). This model indicated that the basic unit of T pucijicus and E. wil- lurnettei populations is based on groups of individiials [ln(ur.uIg: n,,,,, uw,) > 0, P < 0.051, and the size o f t lese asic pophation units did not change with regard to between- arid within-vine distributions [ln(a,,) = ln(ui>g), t = 0.18, df = 331, P > 0.05 and ln(a,,,,) = ln(a,), t = 1.16, df = 229, P > O.OS]. The basic-unit size of I: puciLficus was also similar to that of E. tvillumettei in both distribiitions [ln(up,) = ln(atos), t = 1.67, df = 64, P > 0.05; In(al,J = ln(u,,,J, t = 0.07, df = 486, P > 0.05). The Iwao regression also indicated that these basic units of I: pacifcus and E. willumettei populations were highly aggregated (b,,, b,, b,,, hlLg > 1, P < ().OS), and that both species were more aggregated within vines than between vines (b,,g > h),,,, t = 3.04, df = 331, P < 0.01 and b,Lg

> b,,, t = 2.97, df = 229, P < 0.05). T pacijicus was more aggregated than E. wil1nrncttec.i between vines > b,,, t = 2.95, df = 64, P < 0.05) and within vines (b,, > b, t = 7.95, df = 486, P < 0.05).

Discussion

Our results indicated that the distributions of '1: puc$cus and E. willarnettei in Thompson Seedless vineyards followed predictable and dynamic spatial patterns. Differences in apparent vine vigor (as- sociated with areas within vineyards) showed a substantial effect on within-vineyard distribution of 2: pucijic7~s. This species increased more rapidly and reached greater densities on low-vigor than on high-vigor vines in the Dinuba and Fresno vineyards (Fig. 1). In the Madera vineyard, where vines in both sainpling areas were of moderate to high vigor, differences in I: pucifcus densities between the 2 sampling areas were considerably less than we observed in the other 2 vineyards. Although these observations clearly suggested an association between vine vigor and relative size of 1T: pnc$cus infestations, the low number of replications does not allow 11s to make strong statistical inferences regarding the effect of vine vigor on mite abundance. In addition, we do not know the uncierlying factors causing these differences in mite abundance on vines of different vine vigor. It is known that reduced plant vigor can be the result of water and nutritional (and biotic) stresses which, at some levels, can lead to spider mite outbreaks (Jepson et al. 1975, Kliewer et al. 1983, Mattsori and Maack 1987, Holtier et al. 1988, Younginan et al. 1988). It is also possible that higher mite abundance on weak vines could be caused partly by mite concentration on the lower total leaf area on weak compared to vigorous vines.

In contrast to the distribution of I: pacijic7~s in relation to apparent vine vigor, E. willnrrwttei densities were similar in both sampling areas of the Diniiba and Fresno vineyards (Fig. 1). In the Ma- dera vineyard, E. willumnettei densities increased rapidly during the middle of tlie summer, but there was little difference between the 2 sampling areas where the vines were of similar apparent vigor. He- cause of the low abundance of E. willurwttei i i i

380 Vol. 25, no. 2

the Dinuba and Fresno vineyards, it was not possible to siiggest R relatioiisliip between vine vigor and the abiindance of E. willnmettei Very little is I<nowii about the response of E. wil luwwttp i to water and nutritional stresses of' its host plants.

The tiistribiitioiis of 1: pncijificr~s arid E. toilla- ~ v ~ e t f e i showed strong h i t opposite tendencies to- ward spccific vine zones and leaf ages during cer- tain phases of growth. In general, 1: pacijicus occurred in greater abundance in the soiltli and top zones than in the north and interior zones (Figs. 2 and 3) . In contrast, E. coillarrgttei occurred in greater ab~indance in the north and interior zones than in tlie top and south zones (Fig. 4); except tlurin the initial growth phase in tlie Ma- dera vineyari, where E. willuvrwttei occurred primarily in the south vine zone (Fig. 4A).

The result that tlie abiindaiice of the 2 spider mite species differed between vine zones was not unexpected. Flalierty and Hiiffaker ( I 970) intlicat- ed that 7: puc~j ic i i s preferred the sontli zones whercas E zoillanwttei prcfcrred the north zones of vines, but these authors did not provide details on vine zoiies and seasonal changes in the distribution patterns of the mites. Although our data confirins their observations regarding T pacijicus

he south zone, we did not find .any rences in E. toillnn~ttei densities

between the iiortli and south zones. We observed trends in the patterns of E. willnrrwttei distributions on viiies only when we added the top and interior zones.

It is possible that the level of solar radiation and resulting differences in leaf temperature may be irnportarit factors for explaining the differences in spatial distributions o i the spider mite species witliin vines. Although we did not measure the level of solar radiation reaching each canopy zone, we knew that different canopy zones arid leaves within zones rcprcsented a continiiui~i of exposure to sunlight. IMft~reiices in the distrihiition of T pac~jicus and E. toilluimftei with respect to the level of leaf exposure to sunlight may be related to differences in tlie sensitivity of the 2 mites to the physical and cheniical qiialities of leaves. Shading leaves froin sunlight generally lowers leaf temperature arid al- ters its chemical and physical traits, which may in turn lead to variations in herbivore abuntlance within plants (Schiiltz 1983). Herbivorous artliro- pods rnay also respond differently to leaf shading (Bultman arid Faeth 1988). The differences in abundance of T. pacijiais and E. willnmettei on shaded compared with exposed leaves were prob- ably riot affected by differences in predator abun- ciance, bccaiise predators were nearly absent in all of our study vineyards except during the declining phase in the Dinul)a and Madera vineyards.

The distributions of T paefictcs and E. toilla- mettei on vines wcre strongly affected by leaf position (or leaf age). Grapevine shoots do not have a tcrnminlil bud aiid continue to produce leaves when conditions are favorable (Winkl~r et al.

1975). Therefore, at any time during tlie season, there will be a range of leaf ages which may influencc spider mite distributions. Sampling leaves from the basal, middle, and tcrmiiial regions of' shoots allowed us to cornpare a range of leaf agcs from the oldest (basal) to the youngest (tcmiinal) leaves. That thc distrihitiori of spider mites 011

vines varied with leaf age i s not surprising. Leaf age can be an iinportant deterininant of herbivore distribution (Raiipp and Denno 1983). Photosyn- thetic activity and total nitrogen concentration of grapevine leaves declrne with age (Williarris 1987b). Studying 'I: pac$ms on grapevines under greenhouse conditions, Wilson et 211. (1988) found that iinmature development and fecundity of ?: pacijicus are nonlinearly related to total leaf nitrogen. A sirnilar nonlinear effect of plant nitrogen on fecundity has been obscrved for Tetranychus ur- ticae (Koch) (Acari: Tetranychidae) on bean plants (English-Loeb 1989). Under field conditions, 2: pucijicus fecundity was highest on niidshoot leaves, lowest on tlie oldest (basal) leaves, and interme- diate on the youngest (tcrminal) leaves (K.H, L.T.W., and F.G.Z., iinpiiblished data). These data help to explain the distribution of T pacijicus along vine shoots, but little is known about the effect of leaf age and nutrition on the life history parameters of E. zuillamettei. Based on the sinaller variation in within-vine distribution of E. willnrrgttei cornpared with 1: puc,$kw, it is possible that the former is less sensitive to variation in leaf quality compared with the latter.

According to the Taylor and Iwao models, both 'r: pac@Lfic1is arid E. willamettei were more aggregated in their distribiitions within vines compared with between vines; T pae+$ms was more aggregated than E. zoillamettti, as indicated by differences in the h and ß coefficients of the Taylor and Iwno models, respectively (Table 3). The positive value of the Iwao U also indicated that botli 7: pu- c$&s and E. willunaettei lived in groups of i id i - viduals, which were similar in size for the 2 spc- cies. Taylor b coefficients for 1: pacifia~s and E willaivwttei on grapes cornpare well with those oh- tained on other crops for several Tetraniychus arid Panonychus spp. (reviewed hy Jones [ 19901).

Our article reports resiilts froin the data for 1 year in 3 vineyards diiririg the 1986 growing season. Conceivably, different results could be ob- tained under weather conditions that are different froin conditions observed in 1986. Examination of weather data near each of our study sites indicated that monthly low and high temperatures were within 1°C from May through July (compared with 30-yr averages), but were 14°C higher for August arid 13°C lower for September (University o f Cal- ifornia IPM Project, Davis). It is unlikely that these srriall temperature differences between 1986 arid 30-yr avera es woiild affect the rehtivc difference (at a given c B ensity level) i n mite distributions witliin vine ards, arnong vincs, or witliin vines as cx- arriinc Y in our study.

The major objectives of oiir stiidy were to determine how spider mites are distributed aniong diffcwnt areas within vineyards, and the spatial heterogeneity in their distributions among vines and within vines. We were able to show consistent differences in spider mite densities in relation to vine vigor. T. pacijicus outbreaks were likcly to start in vineyard areas with low vine vigor. Sub- stantial savin s may be realized by concentrating

ation in mite densitics also indicated that among- vine variations were greater thm within-vine variations at low densities, with thc opposite observed at high mite deiisities. To achieve a greater accic- racy of popillation assessmcnt as mite population sizes increase, the number of leaves selected from each vine should be increased with a correspond- ing decrease in number of vines sampled. That the 2 species were more aggregated within-vines compared with among-vine distributions also empha- sizes the need for selecting several leaves from each vine to reduce sample variability. Saviiigs in sampling efforts with a concomitant increase in ac- curacy shoiild be achieved hy selecting leaves from the interior, south, and top vine zones, which to- gether harbored greater percentages of mite cien- sities, thereby reducing travel time that would oth- erwise be needed to sninple the north zone of the sample vines.

monitoring e 8 ‘forts in those areas. The spatial vari-

Ac knowledgnents

We thank James R. Carey (University of Chlifornia, I . h i s ) , C . k1. Pickett (California l~epir tnient of F’ood and Agricdture), and Stephen C . Welter ( IJniversity of C.aliforni;i, Berkeley) for their review of an early draft of the manuscript. Special. thanks are also extended to M. Radrojian (grower, Fresno, C A ) , C. Ndson (grower, Ma- clera, CA), and E. Shannon (grower, Dinuba, CA) for per- mission to use their vineyirds as stricly sites; ‘r. I%ites (University of (hlifornia, Davis) and J. Srnilanick (Uiii- versity of California, Davis) for their technical assistance; S. Randall (Pest Control Advisor) tor his help in locating the Dinuba vineyard, and C. Summers (University of Cal- ifornia, Davis) and the staff of the Keerney Agricultural Center for providing logistical sr.ippoit. This research was supported in part by a grant to L.T.W and L3.L.F from the University of California Statewide IPM Project, ;ind by grants to EG.Z and H.13 from the C:alifornia Raisin Advisory Hoard and the California Tdde Grape Coininis- sion.

References Cited

Banerjee, B. 1976. Variance to inem ratio and the spatial tlstrjbutjon of animals. Experientia 32: 993- 994.

Bultman, T. L., and S. H. Faeth. 1988. Abiintlance aiid mortality of leaf miners on artificially sliaded Emory oak. Ecol. Entomol. 13: 131-l42.

English-I,oeb, G. M. 1989. Nonlinear response of the two-spotted spider mite to nitrogen stress in bean plants (Acari. Tetranychidac) Jnt J Acarol. 16. ‘77- 83.

Flaherty, D. L., and C. B. Huffaker. 1970. 13iological control of ‘I: ,ncici~%is and E. willolnet(.ei in Snri Joa- quin Valley vineyards: role of Metr~seiu1rc.r (JMkh l-Iilgar&iii 40: 267-308.

Gritnes, D. W., and L. E. Williams. 1990. Irrigation effects on plant water tioris arrtl pro&lctivity of ‘Thornpson Seedless’ gr vines. C:rop Sc:i. 30: 255.- 260.

Hanna, R. 1992. Ecolqy of spider tiritas in vineyards: role of predation and coinpetition. Hi. 1). tlissc~tation, University of Chli forni a, Ihvi s,

ilson. 1991. Pre

the role of prey aggregation. Biol. Control I : 5 1-58,

Berry, and J. (2. Hointz. 1988. environment on the tlynarnics of‘ sp tions. Exp. Appl. Acarol. 4: 247-264.

lwao, S. 1968. A new regression inethotl for analyzillg the aggrqption pattern of a i ii mal populatioii s , ~ e s , P ~ p u l . Ecol. (Kyoto)lO: 1-20.

Jepson, L. R., H . H. Keifer, and E. W. Baker. 1.075. Mites injixioiis to ecoriornic pliirlts. 1Jnivc iforilia Press, Berkeley.

Joncs, V. P. 1990. lleveloping sanipling plans for spider mites (Acari: Tetranychidae): thosc. w110 don’t r r - member the past may have t o repeat it. J. f i k o i i . ~ C I I . .

toinol. 83: 1 GS6-1.664. Knrben, H., and G . M. Phglish-l.,w:l~. 1990. A “ v x -

cination” of Willarnette spidw n nycliidae) to prevent large p p d a der niites on grupevines. J . J.:con. Entot-nol. 83: 225%- 2257.

Kinn, U. N., J. J. Joos, R. I,. Doutt, .I. T. Sorensen, and M. J. Foskett. 1974. Effect of Tetrcttt!y.:h~.r po- c$cus and irrigation practices on yield arid quality of‘ grapes in north coast vineyards of C;;iliforni;~ Kiiv Entomol. 3: 60 1-606.

Kliewer, W. M., B. M. Freeman, and C. Hoxnorn. 1983. Effect of irrigation, crop level m t l potassiriiri fertilization on Carignane vines. I . Ilegrec. of w;rtc:r stress and effect on growth a r i d yield. Am. 3 . Eiml. Vitic. 34: 186-l96.

Kunu, E. 1991. S:impling ;md :inolysis of iiisc.c.t pol)- rihtions. Arrnu. I3rv. Elitoiriol. :36: 285-:304,

Lloyd; M. 1967. “Mean Crowding.” J . A n i i n , E:c!ol. 36: 1-30.

Muttson, W. J . , and R. A. Haack. 1987. T l t t ~ role of drought i n outl)realts of pl;int-e;iting itiscxts. 13i( BIlCe 37: 1.10-1 18.

McNnlly, P. S., and 11. Fwnhain. 198.5. Ef‘fects of E. i (Acari: Tetranychidae) on ‘Chenin b h i c b ’ ndel’ gmpr vwieties. J. 1Ccon. Entoinol. SO:

947-950. Milliken, G. A., and D. E. Johnson. 1984.. Andysis

of messy data: designed experiments, vol. 1.. Vaii Nos- trancl Heinhold, New York.

Mullins, M. G., A. Bonquet, and L. E. Williarns. 1992. 13iology of the grape: vine. (hnbridge IJnivt?r- sity Press, Chmk)riclgc.

Rnupp, M. J., and R. F. Denno. 1983. L e d age I\S i i

predictor of herbivore distribirtioii and i h i i n d

pp. 91-124. Ir i R . E: Jltmno and M . S. McChre? [ Variable plants and hcrbivores in n;itiiriil and tnari:iged systems. Acudernic, New York.

SAS Institute. 1.989. SAS riser’s giiide: stutistic:s, ver- sion 6 etl. SAS Institute, Gary. NC:.

Hultzer, T. O., J. M. Norman, ’I: M. Pwring, J. S.

382 ENVIRONMENTAL. ENTOMOI~OGY Vol. 25, no. 2

Schultz, J. C. 1983. Habitat selection and foraging tac- tics of caterpillars in heterogeneous trees, p, 61-90. In H. F. 13enno and M. S. McClure [eds.f, Variable plants and herbivores in natural and managed systems. Academic, New York.

Smart, R. E. 1985. I'rinciples of vine canopy management and microclimate. Am. J . Enol. Vitic. 25: 84-91.

Taylor, L. R. 1961. Aggregation, variance and the mean. Nature (Lond.) 189: 732-735.

ng and interpreting the spatial distributions of insect populations. Annu. Rev. Entornol. 29:

Taylor, L. R., I. P. Woiwod, and J. N. Perry. 1978. The density-dependence of spatial behavior and the rarity of randomness. J. Anirn. Ecol. 47: 383-406.

Welter, S. C., D. Farnhnm, P. S. McNally, and R. Freeman. 1989a. Effect of Willamette mite and Pa- cific spider mite (Acari: Tetranychidae) on grape pho- tosynthesis and stomatal conductance. Environ. En- tomol. 18: 953-957.

Welter, S. C., P. S. McNdly, and D. S. Farnham. 1989h. Willamette mite (Acari: Tetranychidae) impact on grape productivity arid quality: a re-appraisal. Environ. Entomol. 18: 408-411.

Williams, D. W., H. L. Andris, R. €I. Beede, D. A. Luvisi, M. V. K. Norton, and L. E. Williams. 1985. Validation of a model for growth and development of the 'Thompson Seedless' grapevine: phe- nology. Am. J. Enol. Vitic. 36: 283-289.

321-57.

Williams, L. E. 1987a. Growth of' 'Thompson Seedless' grapevines: leaf area development and dry weight distribution. J. Am. Soc. Hortic. Sei. 112: 325-330.

1987h. Growth of 'Thompson Seedless' grapevines: nitrogen distribution. J. Am. Soc. I-Iortic. Sci. 11.2: 330- 333.

Wilson, I,. T. 1985. Estimating the abundance and impact of arthropod natural enemies in JPM systems, pp. 303-322. In M. A. EIoy and 13. C. Har7,og [eds.], Biological control in IPM systems. Academic, Orlan- do, FL.

Wilson, L. T., J. M. Smilanick, M. P. Hoffmann, D. L. Flnherty, and S. M. Ruuz. 1988. Leaf nitrogen and position in relation to population parameters of Pacific spider mite, Tetrunychus ~~uci$fic~u (Acari: Tet- ranychidae) on grapes. Environ. Entomol. 1.7: (364- 968.

Winkler, A. J., J. A. Cook, W. M. Kliewor, and I,. A. Lider. 1975. General viticulture. University of Cal- ifornia Press, Berkeley.

Youngman, R. R., J. P. Sanderson, and M. M. Barnes. 1988. Life history parameters of Tetmnychus pclci-

&us McGregor (Acari: Tetranychidae) 011 almonds under differential water stress. Environ. Entomol. 17: 488-495.

.

Received $1. publication 22 Murch 1995; accepted 27 November 1995.

Documents

ON Eco~oc~ Spatial and Temporal Dynamics of Spider Mites ... · Spatial and Temporal Dynamics of Spider Mites (Acari: Tetranychidae) in ‘Thompson Seedless’ Vineyards ... canopy