Date of Award

Spring 2022

Project Type


Program or Major

Biological Sciences

Degree Name

Master of Science

First Advisor

Iago Hale

Second Advisor

Rebecca Sideman

Third Advisor

Eleni Pliakoni


Maturity at harvest is an important determinant of fruit quality in kiwiberry [Actinidia arguta (Siebold & Zucc.) Planch. ex Miq.,], a climacteric fruit that is harvested after reaching physiological maturity but not yet ready-to-eat ripeness. Although the recommended cultivar for commercial kiwiberry producers in the northeast United States is ‘Geneva 3’, no published research exists regarding recommended harvest and postharvest practices for that variety. In this study, conducted across two seasons, ‘Geneva 3’ kiwiberries were harvested at a range of mean maturities (6.5, 8.0, and 10.0 °Brix), held in cold storage (CS, 0-1 °C, >90% relative humidity) for various durations (4, 6, and 8 weeks), and then ripened at room temperature (RT). At regular time points during ripening (0, 3, 6, 9, and 12 days), visual quality was assessed and measurements taken of soluble solids content, dry matter content, and firmness as a means of quantifying overall fruit quality. Results show that berries harvested at 6.5 °Brix largely became visually unacceptable under CS conditions and resulted in low overall quality fruit. Harvesting at 8.0 °Brix resulted in high-quality fruit amenable to CS, and such quality was not enhanced by delaying harvest to 10.0 °Brix. Specifically, fruit harvested at 8.0 °Brix was found to be acceptable for consumption based on Brix measurements for, on average, a six-day window after ripening at RT for three days. After six weeks in CS, the consumability window shortened to approximately four days, starting after one day of ripening at RT. After eight weeks in CS, the fruit was found to be largely visually unacceptable for fresh eating. In summary, the results indicate that harvesting ‘Geneva 3’ kiwiberries at 8.0 °Brix produces berries with the greatest storability (up to six weeks in CS), longest window of peak consumability, and highest overall quality, while mitigating the risks associated with leaving physiologically mature fruit to ripen further in the field. Spectrographic signatures were captured using a near-infrared (NIR) produce quality meter (Felix F-750) to build predictive models of internal fruit quality for ‘Geneva 3’ kiwiberries that had been held under CS conditions and that had not been in CS. The model for fruit held in CS was constructed with non-invasive NIR scans of 133 ‘Geneva 3’ kiwiberries that had been in cold storage for 4-6 weeks and predicts post-CS soluble solids content (SSC) using the calibration range of 729-975 nm. 507 ‘Geneva 3’ kiwiberries fresh from the vine were used to construct a predictive model for the SSC of non-CS fruit using the same calibration range of 729-975 nm. In all cases, model performance was confirmed using split-half cross validation, a method where half of the model’s training set is used to predict quality metrics of the other half, which resulted in a predicted R2 = 0.87 for the CS model and R2 = 0.95 for the non-CS model. The models were then used to predict SSC of kiwiberries from CS conditions for which the models were not designed. The non-CS model predicted the SSC of CS fruit with a 95% confidence interval of 1.6 ± 2.4 °Brix while the CS model predicted SSC of non-CS fruit with a larger error of 4.5 ± 4.0 °Brix. The difference in errors between models implies that the non-CS model is more robust than the CS model, although separate models are better able to predict the SSC of fruit that had experienced storage conditions which matched the models’ respective training set. NIR spectroscopy appears to be a promising tool for predicting SSC ‘Geneva 3’ kiwiberries and two separate models are needed for the highest prediction accuracy of fruit that had been held under cold storage conditions and that had not been in cold storage.