A section of an apple blossom, enlarged about 15 times. After flowers are fully opened, the anthers pop open and release pollen. Pollen grains stick to the stigmas on the tops of the styles and germinate, sending out pollen tubes that grow down through the styles to reach the ovules. (SOURCE: Pollination and Fruit Set of Fruit Crops, Cornell Cooperative Extension Information Bulletin 237 by Roger Way.)


Horticulturists in Washington State say a new predictive model helps to take the guesswork out of chemical thinning of apples and could help reduce the need for postbloom thinning sprays.

“We’re using less petalfall and post-bloom sprays because we’re getting enough off during bloom,” Kevin Larson, production manager at Roche Fruit Company in Yakima, said during a discussion on the Pollen Tube Growth Model at Washington’s winter hort meeting.

Harold Schell, director of field services for Chelan Fruit Cooperative, said he began testing the model three years ago. He wanted to know how to better apply multiple bloom sprays. He used it on Gala, Golden Delicious, and Fuji.

His objectives were to ensure consistent return bloom, reduce the amount of hand thinning needed (though not necessarily eliminate it), and produce a consistent volume of fruit with consistent quality and size.

During bloom, pollen is deposited on the sticky flower stigmas on the tips of the styles (see figure). Pollen grains germinate on the stigmas, sending out pollen tubes that grow downward through the styles to reach the ovary of the flower.

The pollen tubes carry the male gametes (sperm) to the ovary where they fertilize the eggs, which then have the potential to develop into seeds. Seed initiation stimulates hormone production, which causes the ovary to develop into an apple.

How long it takes for a pollen tube to grow down through the style and for the eggs to be fertilized depends on the weather.

Scientists at Virginia Tech, who have spent over a decade studying pollen tube growth, developed a temperature-based, variety-specific model that calculates the time required to fertilize a blossom after pollination. Much of the work was funded by the Washington Tree Fruit Research Commission.

The model was developed, using field data from Washington collected by Research Commission staff, specifically for Washington growers. For the past five years, a number of growers and horticulturists have been beta testing the model, which is available at the Washington State University’s AgWeatherNet website (

Using the model

Users need to set up a free AgWeatherNet account. On the Pollen Tube Model page of the website, the model can be set up for each orchard block by identifying the variety and selecting which weather station the model should pull data from.

As soon as the trees begin to bloom, the grower should measure the longest style on 25 to 50 blossoms throughout the block to calculate the average style length, which is entered into the model. The AgWeatherNet Web site gives instructions on how to measure the style.

The grower should know in advance what the target yield is on a per-tree basis. For example, if the target yield is 60 bins per acre of 88 size fruit, that is 126,720 apples per acre. With a spacing of 3 feet by 12 feet (1,210 trees per acre), that works out at 105 fruit per tree.

The grower should then count open flowers on five to ten representative trees, and when there are enough blooms open on the trees to achieve the target yield (for example, 105), the model should be started. It will predict pollen tube growth (as a percentage of the average style length), based on the weather data.

The grower should check the model routinely and apply a chemical thinner about the time when the pollen tube is predicted to have grown the full length of the style. At this point, the desired number of blooms should be fertilized and the thinning product should prevent further fruit set.

When the thinner is applied, the model resets pollen growth at zero so that subsequent thinning applications can be made to prevent fertilization of blossoms that open later. Second, third, and fourth (if necessary) bloom thinning sprays should occur when the model predicts pollen tubes are no more than 75 percent of the average style length to ensure that fruit are not set in flowers with short styles.

Larson said figuring out the target crop load is very important, and that process should begin before pruning.

“Look at how many potential flowers you have and the fruit you need on each tree,” he suggested. “If your tree’s going to have 6,000 flowers on it and you want 150 apples, you’re asking a lot of the model. This process doesn’t start the day before you employ the model.”

To measure the styles, Schell plucks off all the petals from the blossoms and then cuts off the sepals with a knife so he can see the base of the styles and then measures the longest one of each flower with a ruler. He emphasized that style length can vary from year to year. “So, you may have one season when your longest style might average 10 millimeters, and the next it could be 14 mm. Just don’t assume anything.”

Adam Zediker, area manager with Washington Fruit and Produce Company, Yakima, said he uses a digital micrometer, or caliper to measure the length and recommends having two people working together, with one person measuring and the other being the scribe.

“Two people make it four times as fast,” he said.

Larson said measuring styles can seem time consuming, depending on how many blocks need to be done, but the manager needs to find people at each ranch to do it. “They pick it up right away,” he said. “It’s not difficult.”

Schell said he begins measuring styles and counting blooms as soon as bloom begins because it can progress very rapidly.

“If you want 60 apples per tree, when that 60th bloom opens up, you’d better start the model.”

No guesswork

Larson said the model takes some guesswork out of when to thin.

“The model is going to tell you when to apply because you’ve made the decision about how aggressive you want to be and how much you want to take off. You’ve already made those decisions. Once you push the button all you have to do is watch it. You’re going to know 12 to 24 hours ahead so you can tell your managers. This is a huge advantage.”

Larson makes the second and subsequent spray applications when the model shows that pollen tube growth has reached 50 to 75 percent of the length of the style to ensure that they’re not applied late and that there’s no further fruit set.

Zediker said for hard-to-thin varieties he applies the first spray at 80 percent pollen tube growth and subsequent sprays at 50 percent, for the same reason.

He said the model helps him to plan and prioritize because he’s not making last-minute judgments about when, during bloom, to apply a thinner.

“I can collect the data and be in a planning phase, rather than a crisis phase, and I can get across more acres and make better decisions,” he said.

Zediker has found that pollen tube growth can differ even in blocks that have the same conditions.

“If two blocks of different varieties bloom at the same time, using the model, I can make a decision whether to spray the Pink Lady or Gala first, based on the percentage of pollen tube growth that’s been achieved,” he said. “Once I spray, I can make second, third, and fourth applications and it tells me the time, approximately, that I will achieve the correct amount of pollen tube growth. Then I can manage my sprayers for optimal utilization by selecting the block of highest priority.”

But Zediker said the model is only as good as the information that goes into it. “Know the blocks and the desired bins per acre before you start the model,” he advised.

Schell advised growers to try it out on a small block and track what’s going on, bearing in mind that it assumes optimal bee activity and pollen availability. Some variability can also be introduced by the differences among pollenizers in their average pollen tube growth rate.

“There are no guarantees or absolutes, so you have to be able to take the information and put it to use in a way that will fit your orchard and your observations,” Schell said. “It did enable me to better time sprays.”

The model is available for Cripps Pink, Fuji, Gala, and Golden Delicious apples. Models for Honeycrisp, Granny Smith, and Red Delicious are being beta tested and should be released within the next two years, according to Tory Schmidt at the Research Commission. •