Back to top

Grapes 101

Manipulating Cluster Size at Bloom

Grapes 101 is a series of brief articles highlighting the fundamentals of cool climate grape and wine production.

By Tim Martinson and Chrislyn Particka

Grape Clusters
Effect of leaf removal at trace bloom on Riesling clusters in 2014.  
Photo by Meredith Persico.

Bloom is a critical time for grapevines.  Before bloom, early shoot growth in grapevines is supported almost entirely by vine reserves. By bloom, reserves are depleted and actively-growing shoot tips serve as a strong ‘sink’ for photosynthates produced by newly-formed leaves.  At the same time, leaves close to the florets supply the flower cluster with photosynthate to support bloom and fruit set.  Furthermore, bloom is also when next year’s buds form at the lower nodes of the shoots. (see Grapes 101 Sources and Sinks: Allocation of Photosynthates during the Growing Season for a more complete description.)

Manipulating cluster compactness.  This competing set of physiological ‘sinks’ (shoot growth, bloom and fruit set, new bud formation) often affects fruit set and growth.  Cool, cloudy weather during bloom reduces photosynthesis and can extend the bloom period, which leads to irregular fruit set, for example.  But more recently, researchers have been looking at intentionally limiting fruit set to reduce ‘cluster compactness’ in tight-clustered varieties prone to fruit rots such as Riesling, Chardonnay, and Pinot noir.  The thought was that by producing looser clusters, growers could limit severity and incidence of Botrytis fruit rots and improve quality.  Looser clusters dry out faster after rainfall, and Botrytis spreads more rapidly when berries are touching.

Bryan Hed and Jim Travis at Penn State University worked on several techniques to reduce cluster compactness including early leaf removal and application of gibberellic acid. In a six-year study at North East, Pennsylvania (near Lake Erie in northwestern Pennsylvania) Hed and Travis removed the first five to seven leaves in the cluster zone.  They found that leaf removal at trace bloom reduced cluster-compactness and Botrytis, and also had a minimal effect on final yield and ‘return bloom’. 

Leaf removal OR Shoot Tipping: In a Riesling block here at the NYS Agricultural Experiment Station, we were interested in asking a related question:  How much can we influence cluster architecture by either cluster-zone leaf removal OR by removing the shoot tips?  Leaf removal limits the flow of photosynthates and therefore should reduce the number of berries that set, making the clusters smaller and less compact.  Shoot-tipping removes the apical meristem (a major ‘sink’), interrupting shoot growth and resulting in temporary re-allocation of photosynthates into the fruit clusters. 

Methods: We had eight treatments.  Six involved either leaf removal (five basal leaves) or shoot tipping at three different timings:  trace bloom, full bloom, and fruit set.  The other two treatments were 1) ‘Control’ (standard grower practice) with partial cluster-zone leaf removal on the east side after fruit set; 2) Total leaf removal in the fruit zone at veraison on August 17. We then measured cluster weight, rachis length, berry weight and fruit composition [soluble solids (brix), pH, and titratable acidity] at harvest. We also rated the vines and clusters for Botrytis incidence and severity in the field and on the harvested clusters.

Expected results:  We expect that by removing 60-90% of leaf area via early leaf pulling and retaining shoot tips we would produce smaller, looser clusters by temporarily starving clusters of photosynthate.  By removing shoot tips, we would increase cluster compactness by improving set.  The late leaf removal treatment in August would not change cluster compactness, but would increase light interception and thereby could reduce incidence and severity of Botrytis on clusters.

1. Cluster Weight. Leaf removal at either trace bloom or full bloom reduced cluster weight by about 19% compared to the control.  Shoot tipping increased cluster weight by about 7%.  However, clusters on the late leaf pull treatment were also 9% smaller than controls.

Figure 1 Grapes 101

2. Rachis Length. Leaf removal reduced rachis length by 16% (trace bloom) or 20% (full bloom).  Leaf removal one-week post bloom had no effect, and shoot tipping (all timings) had no effect.

Figure 2

3. Berries per cluster.  Leaf removal at trace and full bloom reduced the number of berries per cluster by 21%, from about 70 to 55 berries per cluster.  Shoot tipping at trace and full bloom increased berry number by 11% from 70 to 78 per cluster.

Figure 3

4. Cluster Compactness.  Cluster compactness can be expressed as either the cluster weight per centimeter of rachis or the number of berries per centimeter of rachis length.  Leaf pulling at trace bloom reduced the cluster weight per centimeter by 9%.  When berry number was factored in (bottom), both the trace and full bloom treatments had one less berry per centimeter than the control, but the one week post-bloom treatment showed no effect.  Shoot tipping increased cluster weight per centimeter of rachis by 13% across all timings, and the trace bloom and full-bloom shoot tipping treatments increased the berries per centimeter by two compared to the control (the final treatment increased berries/cm by about one). 

Figure 4a

Figure 4b

Summary:  Leaf removal at trace and full bloom reduced cluster weight and rachis length, and reduced the cluster compactness by about one berry per centimeter of rachis. Leaf removal one week after fruit set did not affect rachis length, cluster weight or cluster compactness.  Shoot tipping at trace bloom and full bloom increased average cluster weight, but not rachis length.  It measurably increased cluster compactness, as measured by cluster weight per centimeter of rachis length (all three timings) or berries per cluster (two berries per centimeter during the first two timings).  So the range of cluster compactness was 11 to 15 berries per centimeter of cluster – a 40% difference.  By weight per centimeter of rachis, we observed a 21% difference between the trace-bloom treatment and the control treatment.

But did it reduce Botrytis?  We rated the vines in the field by examining 20 clusters per vine and estimating the percent (from 0 to 100) of each cluster injured by Botrytis. To show the trend, we repeated this measurement three times (9/28, 10/2, and 10/9).

Incidence.  'Botrytis incidence' measures what percent of clusters have any amount of Botrytis infection.  How it changes over time measures the extent of Botrytis spreading from cluster to cluster.  Note how the leaf removal or shoot tipping treatments compare to the control at left.  The first two timings (yellow and blue bars) show a modest increase, but the final (red bar) measurement showed a significant jump in incidence across all treatments, including the control.  Average Botrytis incidence appeared to be lower in the leaf-pulled treatments and the first shoot-tip treatment.  The other shoot tipping treatments look about the same as the control.  But the standout is the late (August 17) cluster-zone leaf pulling, which left clusters totally exposed.  In this treatment, incidence did not increase as much over time–indicating that cluster-to-cluster spread was less of a factor with the complete late-season light exposure this treatment provided.

Figure 5

Severity. Severity measures the overall percent of the cluster area affected by Botrytis. In the control, this went from 5% to 17% in the final rating – a huge jump.  The three leaf removal treatments were 6-7% - as did the ‘shoot-tipping at trace bloom’ treatment.  The other two shoot-tipping timings (full bloom, and one week post-bloom), had 14-16% Botrytis in the final sample.  But note that the late leaf removal (August 17) showed a pattern equivalent to the leaf removal treatments, ending up at 7% Botrytis severity.  

Figure 6

Interpretation:  The leaf removal treatments were designed to reduce cluster compactness – and succeeded in doing so.  Shoot-tipping may have increased cluster compactness (through better fruit set and more resources available to support clusters at bloom), but may have affected the light environment via growth of additional lateral shoots. The overall spread in cluster compactness ranged from about -20% to +10%, compared to the control – an overall 30% difference.  

Leaf removal around bloom appeared to be successful in reducing the incidence and severity of Botrytis.  This is nothing new; Cornell plant pathologists Wayne Wilcox and Stella Zitter found the same results when they snipped berries from Pinot noir clusters to produce looser clusters in the early 2000s.  In this experiment, the mid-August leaf removal treatment also decreased the incidence of Botrytis at harvest – presumably by increasing light exposure and air movement in the cluster zone.

Practicality:  In our experiment, we removed by hand the first five basal leaves on each vine shoot, or removed the tip of every shoot.  Either of these procedures is cost-prohibitive for commercial producers – for us, labor requirements ran in excess of five minutes per vine, which would amount to 4000 minutes or 66 hours per acre. 

Mechanization might make this process more feasible.  Growers already use leaf-removal equipment after fruit set to open up the grape cluster zone.  But this equipment typically has not been used earlier at trace bloom, when young clusters are much more vulnerable to being removed along with the leaves. 

With older suction-and-blade type leaf removal machines, doing this so early would likely result in significant damage to grape florets and clusters early in the season.  But many recently-introduced leaf removal machines use compressed air pulses to remove leaves, and with proper adjustments, these machines might be capable of removing the appropriate cluster-zone leaf area while minimizing damage to clusters at bloom.  That’s what we would like to try next.

Bottom line:  By manipulating ‘sources’ (by leaf removal) and ‘sinks’ (by shoot tipping) between trace and full bloom, we were able to change the compactness of grape clusters ranging by about 30% (reduce by 20% or increase by 10% relative to control) in average cluster compactness.  After bloom, however, these manipulations had a much less dramatic effect. 


Hed, B., H. Ngugi and J. Travis. 2015. Short- and Long-Term Effects of Leaf Removal and Gibberellin on Chardonnay Grapes in the Lake Erie Region of Pennsylvania.  Am. J. Enol. Vitic. 66: 22-29.

Martinson, T. 2010 Sources and Sinks: Allocation of Photosynthates during the Growing Season.  Appellation Cornell, Issue #4, November 2010.

Zitter, S. and W. Wilcox. The biology and control of botrytis bunch rot of grapes.  Cornell College of Agriculture and Life Sciences Impact Statement. 

Tim Martinson is a senior extension associate in the Section of Horticulture, based at the New York State Agricultural Experiment Station in Geneva, NY.

Chrislyn Particka is an extension support specialist in the Section of Horticulture.