Integrating High-Density Orchards and Biointensive Integrated Pest Management Methods in Northeastern Apple Production

1. a. Advances in bio-intensive management of flyspeck in apple.

Flyspeck disease causes a blemish of apple. Without fungicides, growers generally will suffer flyspeck damage, but very little is known about the proper timing for fungicide applications or what alternatives might replace fungicides in flyspeck management. Several years of research have indicated that non-sprayed apple trees in some orchards develop higher levels of flyspeck than non-sprayed trees in other orchards. Furthermore, different areas within a given orchard can have significant differences in flyspeck incidence. This has led to research examining the role of various factors that might predict the potential for flyspeck disease on trees that did not receive summer fungicides. Fungicide treatments could then be adjusted for individual areas in an orchard.

From 1997-1999, physical parameters of orchards were evaluated in 48 blocks of apple trees that were of varying planting densities. These parameters were the number of borders, the distance between the trees and the closest border, the density of alternate host plants in the borders, the density of flyspeck on those host plants, the foliar density of the apple trees, the height and diameter of the apple tree canopies, the altitude and slope of the block with respect to the orchard as a whole, and the planting density of the block (no. trees/acre). Fruit were evaluated weekly from mid-July to harvest for flyspeck damage. Significant relationships between flyspeck on the fruit and physical parameters of orchards were calculated. Blocks received conventional or reduced fungicide programs.

In 1997, the most important factors were tree size, flyspeck density on the border host plants, the number of borders around the blocks of apple trees, and planting density of the apple trees. These factors, along with IPM level (amount of fungicide used) accounted for 61% of the variability in flyspeck counts at harvest. The small trees planted at high densities had less flyspeck than the larger trees planted at lower densities.

In 1998 and later years, each block of apple trees contained a sub-plot that received no fungicide after June 15. The flyspeck counts in these plots and their relationship with physical orchard parameters were the focus of the 2nd and 3rd year of the study. In 1998, the most significant site factor was the density of flyspeck on the host plants in the borders. This alone accounted for 40 % of the variability in flyspeck count at harvest. Other important factors were tree diameter, planting density, light penetration, and IPM level.

In 1999, a flyspeck risk predictive model was tested. A six point rating system was devised with the most significant physical orchard parameters of the preceding 4 years: density of alternate host plants in borders, density of flyspeck infestation on alternate hosts, number of borders around the block of apple trees, slope of the ground, apple tree size/foliar density, and relative elevation of the block in the orchard. Blocks were rated for risk and fungicide treatments were planned accordingly, with the lower risk blocks set at a 28 day spray interval of a fungicide at half the usual rate. Intervals were also adjusted during the season based on rainfall recorded by our weather stations. Half of the growers were able to follow this complicated protocol, and in their orchards the average amount of flyspeck at harvest in these blocks was reduced considerably in blocks that were moderate and high risk.

In the 1999 harvest evaluation, the factors which had the most significant relationships with flyspeck on the fruit were high density of alternate hosts in borders, low penetration of sunlight in apple canopies, low planting density, large size of apple trees, and high density of flyspeck on alternate hosts. These factors explained 84% of the variability in flyspeck on the fruit at harvest.

In 1999, additional counts were performed in 5 plots adjacent to wooded borders that had high densities of alternate hosts infested with flyspeck. These plots were 3 trees wide, extended 5 rows into a block of apple trees, and received no summer fungicide. At harvest, flyspeck incidence was much higher than in other no-summer-spray “blocks”: an average of 62% of the fruit in the trees closest to the border were infected. Counts dropped-off to a 54 % infection rate in fruit in the 5th row from the woods. This illustrated the importance of proximity to inoculum source and the ability of the inoculum to penetrate 5 rows of apple trees.

Even though blackberry is a major host, we have found flyspeck on wild grape, maple, red oak, and many other species. This study identifies factors which combine to produce an environment which supports flyspeck. Modification of this environment in a number of ways, such as summer pruning, clearing-back borders, selective removal of alternate host plants, or use of high density dwarf plantings, could reduce flyspeck pressure considerably. The most stable management plan will involve several strategies, such as border management, orchard design, pruning program, weather monitoring, and careful use of fungicides.

Our ability to predict flyspeck incidence on the basis of physical factors in the orchard, and by using disease incidence ratings on alternate hosts, has increased each year. We are continuing to review data from this work, and will be testing a revised model in 2001.

1.b. Advances in bio-intensive management of apple maggot fly (AMF) in apple blocks of low, medium, and high planting densities.

Apple maggot flies (AMF) build to high populations on millions of unmanaged wild apple and hawthorn trees from which they invade orchards in July, August and September. Several previous years of research have shown that surrounding large blocks of medium-size apple trees with odor-baited sticky red spheres (5 yards apart) to intercept immigrating AMF can provide control nearly (but not quite) equal to control provided by 3 insecticide sprays. If traps are to become a feasible alternative to sprays for AMF control, we need to take into better account a range of factors that can affect trap performance (e.g. tree size, fruit load, tree cultivar, distance between traps) and substitute something (e.g. pesticide-treated spheres) for sticky spheres as traps. We report here on progress toward these ends. For further detail, see articles in "Resource" section.

Effect of orchard tree size on sphere performance. The year 1999 was the concluding year of a 3-year study in 48 commercial orchard blocks on the influence of orchard tree size on performance of odor-baited sticky spheres in controlling AMF. Each year, traps were baited with the synthetic fruit odor attractant butyl hexanoate and placed 5 yards apart on perimeter trees of 8 blocks each of large (M.7), medium (M.26) and small (M.9) trees. Each block measured approximately 40 yards wide by 40 yards deep. An equal number of blocks that received 3 insecticide sprays to control AMF served as controls.

Results across the 3 years (1997-1999) showed that numbers of AMF captured per perimeter trap were about equal across blocks of all tree sizes. As judged by captures of AMF on unbaited spheres positioned near the center of each block (for purposes of monitoring penetration of immigrating AMF) and by percentage of sampled fruit injured by AMF, traps surrounding blocks of small trees were more effective in controlling AMF (relative to control obtained by insecticide sprays in comparison blocks) than were traps surrounding blocks of large trees. In fact, each year injury to fruit was slightly less in trapped than sprayed blocks of small trees but greater in trapped than sprayed blocks of large trees.

These findings indicate that baited spheres for controlling AMF should perform very well in high density blocks of M.9 trees but may not perform so well in blocks of low-density M.7 trees, with blocks of medium-density M.26 trees intermediate in performance expectations.

Effect of fruit load and cultivar on sphere performance. Juan Rull (UMASS Entomolgy graduate student) has been investigating effect of crop size and tree cultivar on ability of baited sticky red spheres to capture AMF. In 1999 studies in commercial and unsprayed orchards, he found that as apples mature and become larger and redder as the growing season progresses, they become increasingly competitive with sticky red spheres for the attention of AMF. Also, the greater the fruit load (i.e. number of apples per tree), the greater the degree of competition with sticky spheres. Together, these results suggest that more than one sticky sphere per baited tree may be needed to overcome visual competition from natural fruit. Fortunately, the amount of attractive odor emitted by fruit of even a highly attractive cultivar (e.g. Jersey Mac, Gala, Red Delicious) does not compete with the large amount of attractive synthetic fruit odor (butyl hexanoate) released from a single polyethylene vial hung in association with a sticky sphere. These combined findings suggest that control of AMF by traps placed on perimeter apple trees may be achieved best by grouping several baited spheres onto the same perimeter tree and then allowing several successive perimeter trees to go unbaited before again grouping several baited spheres on another perimeter tree.

Performance of pesticide-treated spheres. From 1997-1999, we compared wooden pesticide-treated spheres and biodegradable pesticide-treated spheres with sticky-coated spheres for ability to control AMF in blocks of 49 trees per treatment in each of 8 commercial orchards. Each perimeter tree of each block received a baited sphere. A 4th block in each orchard received 3 summer insecticide sprays but no spheres for AMF control. The pesticide-treated spheres (PTS) received a coating of 70% latex paint, 20% sucrose to stimulate AMF feeding, and 10% Provado (yielding 2% a.i. imidacloprid). To replenish sucrose lost during rainfall, wooden PTS were capped with a disc comprised of wax and hardened sucrose that seeped onto the sphere surface, whereas sugar/flour PTS received sucrose that seeped from the interior of spheres onto the surface.

Results showed that across the 3 years, percent fruit injury from AMF averaged 4.2% in blocks having wooden PTS, 1.4% in blocks having sugar/flour PTS, 1.3% in blocks having sticky spheres and 0.9% in blocks receiving 3 insecticide sprays. Versions of PTS used in 1999 performed better than versions used in 1997 and 1998. Even so, caps of sugar atop wooden PTS required replacement twice during 1999 to ensure a sufficient amount of sugar on the sphere surface, and the majority of sugar/flour PTS in 1999 was eaten in part or whole by rodents, requiring replacement once or twice by new spheres.

In 1999, we also evaluated performance of Actara (thiamethoxam) as a potential substitute for Provado (imidacloprid) as insecticide on the sphere surface. Actara showed excellent kill of AMF alighting on wooden PTS under low to moderate rainfall conditions, but lost residual activity faster than Provado under high rainfall. Provado thus remains the best and most durable insecticide yet tested for combining with latex paint to coat the surface of PTS.

As an indication of future commercial application, we note that a company, "Fruit Sphere, Incorporated", was formed in Illinois in November 1999 to manufacture sugar/flour PTS for widespread testing in apple, cherry, walnut and blueberry plantings in eastern, central and western states in 2000 to control various species of fruit flies.

1. c. Advances in bio-intensive management of plum curculio (PC) in apple blocks of low, medium, and high planting densities.

Among key insect pests that attack pome and stone fruit in North America, plum curculio (PC) is the only one for which there does not yet exist an effective monitoring trap. In 1995, we began in earnest to conduct research aimed at developing such a trap. Here, we present a summary of findings from 3 principal lines of field research conducted in 1999. These were based on research conducted in 1997 and 1998. Full reports on each of these aspects, plus findings from lab studies, are given in Fruit Notes 64(3):1-10, as noted in "Resource" section.

Commercial-orchard trials of traps for monitoring PC. In the aforementioned 48 blocks of trees in 8 commercial orchards, we evaluated 3 types of traps: (a) black pyramid traps (24 inches wide at base x 48 inches tall) placed on the ground next to apple tree trunks, (b) black cylinder traps (3 inches diameter x 12 inches tall) fixed vertically onto horizontal branches within apple tree canopies, and (c) aluminum-screen "circle" traps (developed in Oklahoma for pecan weevils) and wrapped tightly around ascending tree limbs, designed to intercept PC adults walking upward. Traps were placed in 6 blocks of apple trees in each orchard. Each block consisted of 7 perimeter trees. Each tree (save one) contained 1 unbaited and 1 baited trap of the above types. The bait consisted of a combination of 1 polyethylene vial containing limonene and 2 polyethylene vials containing ethyl isovalerate (components of host fruit odor found to be attractive to PCs in 1998 studies) plus 1 rubber septum impregnated with grandisoic acid (attractive male-produced pheromone of PC). Vials were attached to the exterior of traps at mid height and the septum was placed inside the inverted wire-screen funnel (boll weevil trap top) that capped each trap and captured responding PCs. All traps were deployed at bloom and were examined for captured PCs every 3-4 days for 6 weeks thereafter. At each trap examination, 15 fruit on each of the seven trees per block were examined for PC oviposition scars. All blocks received two grower-applied sprays of azinphosmethyl to control PC.

Significantly more (about 3 times more) total PCs were captured by pyramid traps than by cylinder traps, with circle traps capturing no PCs. There was no significant difference in captures between unbaited and baited traps of any type. Disappointingly, none of the three types of baited or unbaited traps yielded captures whose amounts or phenology (pattern of occurrence over time) reflected even in a very minimal way the amount or phenology of egglaying injury to fruit caused by PC. If there were a perfect relationship, a value termed R2 would equal 1.00. Here, the R2 value describing the relationship between abundance of PCs in traps and amount of injury never exceeded 0.14 for any unbaited or baited trap, and the R2 value describing relationship between time of capture of PCs in traps and time of injury did not exceed 0.06 for any type of unbaited or baited traps.

Unsprayed-orchard trials of traps for monitoring PC. In 3 small unsprayed orchards, we evaluated unbaited and baited above-type pyramid and cylinder traps as well as clear Plexiglas squares (2 feet x 2 feet) fastened vertically 5 feet above ground to wooden poles seated in the ground. One side of each Plexiglas square was coated with Tangletrap to capture alighting PCs. Plexiglas traps were positioned with sticky-side facing woods either 6 feet from the edge of the woods or 1 foot outside of perimeter foliage of apple trees. Traps were placed in 4 blocks of apple trees in each orchard. Each block consisted of 6 perimeter trees. Each tree contained 1 unbaited and 1 baited trap (above type bait) of each trap type. Each block in 2 of the orchards also received 1 unbaited and 1 baited clear Plexiglas trap placed at the edge of the woods. All traps were emplaced at bloom. Every 3-4 days thereafter for 6 weeks, traps were examined for captured PCs and fruit were examined for PC scars. No insecticide was applied to any of the blocks.

Significantly more (about 8 times more) PCs were captured by pyramid traps than by cylinder traps, with clear Plexiglas traps positioned next to apple trees capturing slightly but not significantly more PCs than cylinder traps. Captures by unbaited versus baited traps did not differ significantly for any of these 3 trap types. However, baited clear Plexiglas traps placed at the edge of woods did capture significantly more PCs (about 14 times more) than similarly positioned unbaited traps. In contrast to above findings in commercial orchards, R2 values describing relationship between abundance of PCs in traps and amount of injury ranged between 0.56-0.80 for unbaited and baited pyramid traps and clear Plexiglas traps placed next to perimeter apple trees. Less encouraging, however, were R2 values describing relationship between time of capture of PCs in traps and time of injury, which did not exceed 0.05 for any type of unbaited or baited trap.

We feel more encouraged by results of these studies in unsprayed orchards than by results from commercial-orchard tests. For some unknown reason, there was a much better correlation between amount of captured PCs by pyramid traps and amount of fruit injury in unsprayed than in commercial orchards. We found this to be true also in 1998. Because it is in commercial orchards (not unsprayed orchards) that PC traps will have major practical use, we can not feel satisfied until traps perform better in commercial orchards.

We feel quite encouraged by the finding that PCs responded positively to clear Plexiglas traps placed next to woods. In the future, a simpler and more attractive version of this type of baited trap could be very useful for monitoring the beginning, peak and (most importantly) the end of immigration of overwintered PCs from woods or hedgerows into orchards.

Evaluation of attractiveness of different compounds from host odor. To date, 57 compounds have been identified as components of odor of plum or apple fruit at the most attractive stage to PC (2 weeks after bloom). Last year, we presented results of 1998 tests evaluating 16 of these 57 compounds for attractiveness to PC. In 1999, we re-evaluated these 16 compounds plus an additional 14 compounds: the 30 compounds were the most readily available from a commercial source and least expensive to purchase.

Each compound was introduced into a 2-dram polyethylene vial and assessed at 3 different rates of odor release, so as to create a low, moderate or high dose of odor concentration in the surrounding air. Release rates were varied either by adding mineral oil to the contents of a vial to reduce release rate or drilling small holes in a vial just beneath the cap to increase release rate. Intended release rates for each compound were 3, 12, and 48 milligrams of odor per day, but it was not always possible to achieve intended precision with each compound.

Compounds were assayed in association with yellow-green boll weevil traps placed on the ground beneath perimeters of unsprayed apple tree canopies in Massachusetts and Ohio. PCs frequently drop from host tree canopies to the ground and thus may encounter odor from a nearby baited trap. Each trap was baited either with a vial containing one compound or an empty vial. Vials were suspended vertically by wire attached to the base of the screen funnel top of the trap. Over a 7-week period from early May to late June, 360 traps were deployed in Ohio and another 360 in Massachusetts for compound evaluation. Traps were examined for captured PCs and rotated in position daily or every other day.

To measure attractiveness of a particular release rate of a particular compound, a Response Index (RI) was created. If a baited trap were 2 times as attractive as an unbaited trap, RI = 32 If 3 times, RI = 50. If 4 times, RI = 60.

Results showed that 13 of the 30 compounds had RI values of 32 or greater at the most attractive release rate. In descending order of attractiveness, these were E-2-hexenal (RI=90), hexyl acetate (67), decanal (64), limonene (64), geranyl propionate (59), 1-pentanol (59), benzaldehyde (46), benzyl alcohol (44), ethyl isovalerate (40), 2-pentanol (35), 2-hexanol (32), phenylacetaldehyde (32), and 2-propanol (32).

Several new compounds thus proved as attractive to PC as the most attractive compound (limonene) found in 1998 tests. Just as important, these results give us good insight into the amount of compound (release rate) that may be most attractive. To humans, a scent may be undetectable at too low a concentration, and repellent at too high a concentration. The same is true for insects. In fact, these 1999 results suggest that the amount of ethyl isovalerate we used in aforementioned trap comparisons in commercial and unsprayed orchards was probably too great and therefore repellent, canceling out the attractiveness of the other host odor used (limonene).

Overall, we are very encouraged by these 1999 findings of PC response to host odor compounds. There is now much promise that one or more of these attractive compounds alone (or together in a blend) at an appropriate release rate can be applied to one or more of the above trap types to substantially enhance trap effectiveness.

1. d. Bio-control of pest mites (ERM) by predatory mites (T. pyri) in apple blocks of low, medium, and high planting densities.

Pest mites are completely controlled by predatory mites on apple trees that receive no insecticide or fungicide. Some commonly used orchard pesticides kill or otherwise harm predatory mites, leading to pest mite outbreaks and need for frequent miticide application. In Massachusetts, the predatory mite Amblyseius fallacis is present in 90% of commercial orchards but it does not usually build to substantial numbers until mid-July or later, too late for early- and mid-season biocontrol of pest mites. Hence, in cooperation with Jan Nyrop of the Geneva lab of Cornell University, we embarked on a program of seeding third-level IPM blocks of apple trees with the predatory mite Typhlodromus pyri. This predator can provide effective pest mite biocontrol during early and mid season, provided certain harmful pesticides (e.g. pyrethroid and carbamate insecticides and EBDC fungicides) are not used. Until 1995, when we seeded it in some second-level IPM blocks, it was present in less than 10% of Massachusetts orchards.

In May of 1997, each of the third-level IPM blocks in commercial orchards received 100 blossom clusters containing T. pyri predatory mites. The clusters were sent to us by Jan Nyrop. All 100 clusters were fastened to the centermost tree of each 49-tree third-level block. No T. pyri were released in the first-level blocks. Every 2 weeks during July and August, we sampled leaves from the center tree, the 2 outermost trees in the center row and the center trees in the 2 outermost rows to determine degree of establishment and rate of spread of T. pyri in each third-level block. Comparable samples were taken from each first-level block. In all, nearly 13,000 leaves were sampled in 1997, more than 17,000 were sampled in 1998, and more than 17,000 in 1999. All samples were sent to Jan Nyrop for counting of numbers of pest and predatory mites in each sample. Identification of predatory mites to species requires highly specialized expertise not currently available at UMass.

Results on establishment and spread of T. pyri for all 3 years (1997, 1998 and 1999) show good establishment of T. pyri in 1997 on the trees on which they were released, and this establishment was maintained at about the same level during 1998 and 1999. There was very little spread of T. pyri in 1997 to the most distant trees up and down the row in which they were released, some up and down row spread (especially in blocks of small trees) by 1998, and excellent up and down row spread in blocks of all tree sizes by 1999. There was no spread whatsoever of T. pyri in 1997 to the most distant trees across row from which they were released, very slight across-row spread in 1998 (and only in blocks of small trees), and considerable across-row spread in 1999 (especially in blocks of small trees). T. pyri were essentially absent in 1997 and 1998 from blocks in which they were not released but were detectable in several such blocks (albeit in very low numbers) by 1999, indicating some spread of T. pyri by 1999 beyond the confines of blocks in which they were released. For additional detail see article in "Resource" section.

Data on presence of A. fallacis mite predators (taken from 1998 and 1999) show somewhat lesser abundance of A. fallacis in blocks where T. pyri were released than where T. pyri were not released, with no detectable influence of tree size or location of sample site within blocks on abundance of A. fallacis.

Data on abundance of pest European red mites (taken from 1998 and 1999) show a powerful effect of T. pyri in suppressing pest mites in 1999 in blocks of all tree sizes in which T. pyri were released. The suppressing effect was not detectable in 1998. In 1999, it was greatest on trees on which T. pyri had been released and was least on the most distant trees across rows. Also, in 1999 it was comparatively greatest in blocks of large trees.

In summary, this 3-year study of the rate and spread of T. pyri among trees in blocks of different tree sizes shows that by the third year after release T. pyri can spread effectively as far as 3 trees away up and down row and 3 trees away across rows, with spread fastest and greatest in blocks of small tree size. Also, by the third year after release, T. pyri is able to very effectively suppress pest mites in parts of blocks where it has become firmly established.

2. To evaluate the influence of apple tree architecture on abiotic factors that significantly impact pest managemnent (see also article in "Resource" section: Do Planting Density and Tree Size Effect Canopy Microclimate, Spray Penetration, and the Severity of Summer Diseases in Apples? 1997-1999 Results. submitted to Fruit Notes.)

Canopy microclimate was measured repeatedly in all 48 blocks of the experiment. During 1997 and 1998, 5 trees per block were spot-checked 10-13 times during the growing season. Ventilated ambient temperature was measured with the dry bulb of a psychrometer, relative humidity was measured with the same psychromenter, plant surface temperatures were measured with an infrared thermometer, wind speed was measured with a hot-wire anemometer, and light penetration was measured with an integrating radiometer/photometer. In 1997 spray penetration was measured by placing water sensitive cards at strategic points in the canopies, spraying the trees, and then performing digital image analysis on the cards. In 1998, light penetration was also measured with fish-eye lens photography. In 1999, we repeated the photometer readings, but with an increased sample size of 15 trees in each of the 48 blocks.

In 1997, we found that canopy microclimate and plant surface temperatures were warmer and relative humidity was lower as planting density increased and tree-size decreased. In 1998, temperature data showed the same differences but humidity data did not show significant differences due to planting density. Wind speed did not differ significantly among canopies of the three planting densities. Spray penetration was greater in high-density canopies than in the canopies of the larger trees, indicating more efficient spray coverage. Light penetration data taken with a photometer, was not significant in 1997 or 1998, but when we increased the sample size in 1999, the canopies of small, high density trees were more porous to light. Light penetration, as measured by fisheye lens photography, did not yield significant results.
These differences in microclimate readings and spray penetration indicate that the small apple trees in high density plantings have canopies that are generally warmer, drier, and more open to light and spray penetration, than larger trees in lower density plantings. These small trees are less susceptible to economic injury by summer diseases and can be sprayed less with comparable results and significant savings for the grower. The increased warmth and light should also increase fruit quality.

3. To evaluate the influence of apple tree architecture on fruit quality. (see also article in "Resource" section: Effects of Planting Density and IPM Level on Apple Fruit Quality and Crop Density. Fruit Notes 64(4):1-3.)

Fruit quality characteristics, crop density, and yield were measured in 48 blocks of apple trees in MA in 1997 and 1999. Blocks were 7 rows by 7 trees and were managed with either 1st-level IPM or 3rd-level bio-intensive IPM strategies. The blocks were evenly divided among high density, medium density, and low density plantings.

Fruit quality was evaluated by determining weight, sweetness, % red of surface, and firmness of flesh of 50 apples per block at harvest. The high density plantings (the smallest trees) had heavier fruit than the low density plantings. Planting density did not affect sweetness, color, or firmness. The fruit in the 1st-level IPM plots was slightly redder (65% red) than the fruit in the 3rd-level IPM plots (62%), but was of equal weight, sweetness, and firmness.

To estimate yield, the fruit on 5 trees per block was counted at harvest and the number of trees per acre was calculated. The high density plantings had fewer fruit per tree than the medium density trees, which had fewer fruit than the low density trees. Calculated on a per acre basis, however, there was no difference among planting densities. The 3rd-level IPM blocks had similar numbers of apples per tree as 1st-level blocks, but significantly greater estimated yields per acre than 1st-level blocks.

To measure crop density, 20 limbs per block were evaluated by the limb circumference count method. Crop density was slightly greater for low density plantings than for high density plantings, but was not affected by IPM level.

These results indicate that by following a bio-intensive advanced IPM approach, a grower can harvest a crop that is of equal or better quality and yield than by following a chemically-based 1st-level IPM approach. The move to smaller trees and higher density plantings may also give the grower heavier fruit and better per acre yields.