Evaluation of Four Cherry Firmness Measuring Devices
Four firmness devices with commercial potential for use with sweet cherries were evaluated and compared to a laboratory firmness instrument, the Instron. The inherent error of each device was estimated by repeated measurement of a uniform, symmetrical, and resilient rubber ball. These tests revealed that the Instron device (a firmness testing research tool) measured firmness most reliably, followed by the FirmTech1, Low Mass Impactor, and MTG. Firmness measurements of cherries from 5 harvests indicated that the FirmTech1 correlated more closely to the Instron, followed by the Low Mass Impactor, MTG, and Penetrometer. Subjective firmness sensing by compressing cherries between human fingers was found to be less accurate than all the devices tested. The overall results determined that the least variability of firmness measurement existed with the FirmTech1 device, with increasing variability, or error evident with the Low Mass Impactor, MTG, Penetrometer, and finger firmness sensing method, respectively. The commercial or potentially commercial devices were ranked from most desirable to least desirable with respect to ease and speed of operation in the following order: FirmTech1, MTG, Low Mass Impactor, and Penetrometer.
Cherry firmness correlated reasonably well with skin color, and the best correlations were with firmness as measured by the Instron, FirmTech1 and Low Mass Impactor, respectively. Cherry Firmness correlated poorly with soluble solids, but correlated positively with titratable acidity and negatively with specific gravity. Soluble solids content formed a strong relationship with specific gravity; fruit skin Hue color formed a reasonable relationship with fruit soluble solids and specific gravity. Funding for this study was provided by the California Cherry Advisory Board, 1996
Firmness is an important attribute of sweet cherries and is used by inspectors, growers, buyers, researchers, and consumers to assess fruit quality. Objective measurement of cherry firmness would reduce variability and biases associated with sensing firmness by compression of the fruit between human fingers.
Four firmness devices were evaluated, which were considered to have commercial or potential commercial application for the cherry industry, the MTG (Momentum Transfer Generator) developed at Washington State University, The FirmTech1 from Michigan State University, the yet officially unnamed Low Mass Impactor developed by UC Davis, and the UC Firmness Tester (AMETEK penetrometer). These devices were compared to a fifth device, the Instron Universal Testing Machine, which is widely accepted as a reliable compression-testing instrument for biological materials.
Momentum Transfer Generator
The MTG consists of a modified audio speaker with a plastic disk covering its cone. A cherry dropped onto the protected speaker cone generates a sinusoidal wave. The characteristics of the wave are considered to relate to cherry firmness, and the initial slope of the wave determined by the instrument electronics and used as a firmness index. Individual cherry firmness, sample average firmness, and standard deviation of the sample firmness are displayed on a computer interfaced with the instrument and can be saved to a data file. Cost of the MTG is approximately $2,000 (computer not included).
The FirmTech1 device measures firmness using a force deformation mode of action. Cherries are positioned into shallow indentures on a turntable which automatically rotates aligning each cherry periodically under a small load cell. A pre-determined force is progressively applied onto the cherry by the load cell. The rate at which the force increases is defined as firmness. For a sample of cherries, average firmness, sample maximum and minimum firmness, sample standard deviation of firmness, and a frequency distribution of firmness are presented through an interfaced computer. Firmness of individual cherries in the sample can be retrieved from a file accessed through the computer. Cost of the FirmTech1 is approximately $3,300 or $3,900 for desktop or laptop computer interfaced models, respectively (computer not included).
Low Mass Impactor
UC Davis' Low Mass Impactor senses fruit firmness by impacting a small spherical low mass object onto the cherry surface, and evaluating the acceleration response of the rebounded mass. A small accelerometer behind the mass sends the impact response to a computer where the slope of the initial acceleration curve is determined and used as a firmness index. Results of each test are displayed on a computer, and successive tests can be saved into a single file consisting of the entire sample. Estimated cost of the Low Mass Impactor is $3,000 (computer not included).
UC Firmness Tester
The UC Firmness Tester is a penetrometer and has the force sensoring component housed in a frame (drill press stand). A lever fixed to the frame is used to depress the penetrometer tip (3 mm diameter) into the cherry flesh, and a firmness value is given by a gauge on the device. The maximum force applied to the fruit, occurring momentarily prior to collapse of the flesh beneath the penetrometer tip, is defined as the firmness value. This was the only device which caused obvious permanent fruit damage. In this study, a small area of skin was removed prior to penetrometer use. This device can be used with the skin intact. Cost of the UC Firmness Tester is approximately $800.
Instron Universal Testing Machine
An Instron device was utilized to compare each of the above instruments for measurement of firmness reliability. The Instron is a commercially available tool specifically used to determine firmness and strength of materials, usually in a research environment. In this study, a force was progressively applied to a cherry balancing on three small ball bearings fixed to a plate, by a fourth descending ball bearing secured to a load cell. Fruit contact with the four ball bearings during a compression cycle is minimal, thereby reducing the influence of fruit diameter or surface irregularities on firmness measurements. The slope of the resulting force deformation curve relates to cherry firmness. Cost of the Instron model used in this study is approximately $30,000.
Evaluate four devices with commercial potential for assessing 'Bing' cherry fruit firmness with respect to a laboratory based firmness testing instrument with proven reliability, and to firmness sensing using human fingers.
Determine the relationship between fruit firmness, as measured by each type of firmness sensor, and fruit maturity.
Bing cherries were harvested on five occasions between May 15 and May 29, 1996 (commercial harvesting period) from either of two commercial properties in the Stockton area. Neither property received gibberellin treatments, although one property received an application of calcium. At each harvest, approximately 20 pounds of fruit were harvested from the mid to outer canopy area of trees, up to a height of about 6 feet. Emphasis was focused on acquiring fruit of mixed maturity to ensure the collection of a full range of fruit firmness.
Cherries were immediately transported to UC Davis where they were sorted to remove blemished and irregular fruit. They were then divided into 6 groups of 100 fruit each for the first harvest, and 6 groups of 50 fruit each for subsequent harvests. Each group was assigned for measurement by one of the 5 firmness devices, while the sixth group was subjected to firmness measurement by all 5 devices.
Firmness measurement was conducted on the fruit in a temperature controlled environment of 68 °F. The order in which each device was used to measure firmness on the common group of cherries subjected to measurement by all devices was rotated across the 5 harvests. This was a precaution to allow us to determine if any device damaged the fruit and therefore affected subsequent firmness measurements with other devices. Penetrometer measurements were always conducted last as they were destructive.
Following firmness measurements from the 4 non-destructive devices, each cherry was evaluated for skin color (Minolta Chroma Meter), width (dimension across checks), weight (with and without stem attached), and specific gravity by the relationship of stemless weight and fruit volume, the latter determined by water displacement). For penetrometer firmness measurements, a small section of skin was removed on the side of the cherry, as a site for the penetrometer measurement. Individual fruit were then juiced to evaluate total soluble solids (refractrometry) and acidity (titration) .
An additional sample of cherries was utilized as a means of comparing firmness measured by each device with firmness subjectively measured by human fingers. Sample sizes of 50 fruits were measured on 6 occasions, some after being held in cold storage for a short period. Firmness sensoring using human fingers was undertaken by one individual using a soft and hard rubber ball representing the range of a 5-point scale between which cherries were rated for firmness.
A rubber ball of typical mature cherry firmness was used to test for instrument variability by conducting multiple firmness measurements from each device on the object. Twelve measurements from each device, except the destructive penetrometer, were made on the rubber ball.
Instrument variability of firmness was evident from multiple firmness measurements made on a rubber ball. Evidence of the ball's resilient properties and negligible source of variability itself are apparent from testing with the Instron, where average firmness variability from 12 measurements was 0.28%. Noting that some variability would exist with the Instron device measurements, it can be concluded that firmness variability of the rubber ball must be less than 0.28%, and can be considered negligible. Measurements from the remaining devices (excluding penetrometer as destructive measurement on a rubber ball was not possible) which exhibited much larger variability can be interpreted as being largely attributed to device variability or error. Successive testing on the rubber ball for device accuracy revealed that the Instron was the most reliable instrument followed by the FirmTech1, Low Mass Impactor, and MTG. MTG variability may not be entirely due to device error, as the instrument operator likely has some influence on a firmness result. However, this factor must be acknowledged as a possible source of variation for this device.
Evaluation of device accuracy was also possible by comparing firmness measurements from each of the commercially applicable instruments on a common sample of cherries from each harvest, against Instron measurements on the same samples. The comparison would reflect a combination of:
- Variability or error of measurement from the cherry firmness devices themselves
- Any relative firmness inconsistencies from the individual fruit at positions around their circumference
- Measurement of firmness variability caused by changes in fruit contact area with the firmness devices, either due to fruit size or shape irregularities (negligible for the Instron because of the use of ball bearings).
Variability of firmness due to Instron device measurement, already found to be small, would be consistent when correlated against the other devices on a common sample of fruit; and therefore does not have to be considered in any comparison. Figure 1 illustrates firmness measurements made by each device on the same sample of cherries, correlated against firmness measurements made by the Instron. The results are presented in histograms detailing firmness variability as a percentage against the number of cherry measurements featuring such variability. The correlation trends were consistent from each harvest. Therefore, firmness variability from the 5 harvests has been combined, totaling 300 measurements from each device. True fruit firmness inconsistency, measurement variability due to device/fruit contact area, and device measurement variability or error for the FirmTech1, Low Mass Impactor, MTG and Penetrometer featured standard deviations of 9.49%, 12.90%, 13.41% and 15.27%, respectively.
Estimation of firmness using a compression force between human fingers could be evaluated for accuracy when correlated against previous Instron measurements on the same fruit. Assuming Instron measurement variability or error to be negligible, the revealed variability would be a combination of firmness inconsistencies of the individual fruit and human error in sensing firmness by compression of fruit between fingers. Firmness variability from the combined samples amounted to a 16.32% standard deviation (Figure 2); greater variability than the devices used in this study. Relatively erratic results were apparent when fruit firmness estimated by compression between fingers was correlated against the remaining firmness devices. A consistent trend was not evident from the 6 samples of fruit; most likely due to the high degree of variability of finger firmness sensing coupled with the measurement variability from the devices, and inconsistency of firmness of the individual fruits.
Firmness and Maturity Indices
Maturity indices of cherry other than firmness, often considered important harvest or postharvest factors, include: skin color, percent soluble solids, percent acidity, and to a lesser extent fruit specific gravity. The extensive firmness testing on cherries from this study presented an opportunity to evaluate these indices for relationships with fruit firmness. Measurements were made on 100 or 200 fruit per harvest date.
Cherry skin color can be measured objectively by evaluating a light pulse reflected off the skin. The Minolta Chroma Meter can measure various parameters of color, including Lightness (surface shade on a scale of black to white) and Hue (attribute of colors classed as red, yellow, green, blue, or an intermediate between any adjacent pair of these colors). Both Lightness and Hue were equally correlated with fruit firmness, and the correlation of firmness and Hue is shown in Table 1. The strength of Lightness or Hue color relationships with firmness varied dramatically between harvests, generally reflected by the distribution of fruit maturity from each harvest. A large distribution of maturity expressed wide firmness and color qualities, which mathematically correlated more closely than a narrower distribution of maturities. Furthermore, the relative relationship trends between firmness values from the individual firmness devices with skin Lightness or Hue color within each harvest were not always consistent (i.e. Instron firmness measurement did not always correlate best with skin color). However, when averaging correlations across all harvests, Instron and FirmTech1 device measurements related more closely to skin color than did Low Mass Impactor, MTG or Penetrometer (Table 1). Omitting harvest 1 (exceptionally poor correlations due to narrow fruit maturity distribution) increased the average r value for each device (Table 1).
Relationships of fruit firmness with fruit soluble solids, titratable acidity and specific gravity were poor or indistinguishable. No trend or pattern was observable for firmness and soluble solids correlations between harvests, or between firmness devices within harvests (data not shown). Fruit specific gravity showed a poor relationship with firmness, although the trend indicated that specific gravity increased as fruit firmness decreased. Fruit titratable acidity established a better relationship with firmness, although this was still poor. Discrimination between each device's performance relative to titratable acidity was not apparent due to poor correlations. Average correlation coefficients for the harvests were approximately 0.30 for each device.
A strong positive correlation was found between fruit soluble solids content and specific gravity for each harvest. Correlation coefficients varied between 0.74 and 0.87, averaging 0.79. Correlations were also evident when soluble solids content and soluble solids/acidity ratio were each correlated with Hue skin color; average correlation coefficients across the harvests equated to 0.56 and 0.50, respectively. Specific gravity and Hue color also correlated reasonably well with an average correlation coefficient of 0.60. Titratable acidity featured poor associations with fruit soluble solids, skin color and specific gravity.
Challenges to obtaining an accurate estimate of cherry firmness include true firmness inconsistencies of the fruit around its circumference and measurement variability or error of the method used to test firmness. Reduction in the influence of the latter factor can be controlled by selecting a method which predicts firmness with increased consistency. Firmness measurements of the rubber ball with each device gave a direct comparison of the extent of variability of each device. This revealed that when measuring firmness of a uniform, symmetrical object, the Instron was most consistent, followed by FirmTech1, Low Mass Impactor and MTG. Instron measurements of firmness on the ball amounted to average inconsistencies of only 0.28%, thereby verifying the reliability of this research instrument as a suitable tool to evaluate the remaining more commercially applicable devices.
By testing all the firmness devices on a single sample of cherries from each harvest, a direct comparison could be made between the devices. The Instron provided a reliable measurement of firmness on which to compare the commercial devices. Any small variability of firmness registered by the Instron due to measurement error, or fruit size and shape irregularities, would be expressed through correlations with the remaining devices, thereby causing no relative effect. Such a comparison was possible by conversion, using regression analysis, of firmness measurements from each device, to the exact scale of measurement used by the Instron. The firmness variability from each device relative to the Instron could then be calculated and given on a common scale. These differences proportionally agree well with the device average error as determined with the rubber ball.
Sensing fruit firmness with human fingers, with constant referral to a soft and hard rubber ball (analogous to a very soft and hard cherry) for calibration purposes, produced more variability than either of the 4 firmness devices. This result would be expected to vary considerably depending on the skill and experience of the individual conducting such an evaluation.
Skin color formed the strongest relationship with fruit firmness; correlation coefficients averaged across harvests ranged between 0.62 and 0.40 for the Instron and Penetrometer, respectively. Contending with the added variability of fruit skin color, such correlations revealed the extent of device accuracy relative to each other. Ranking of color/firmness correlations for each device agreed with the ranking of device accuracy for the rubber ball and cherry firmness evaluations. Soluble solids correlated poorly with fruit firmness while the correlations of firmness with titratable acidity and specific gravity were significant for many harvest dates.
Correlations between the maturity indices other than firmness were good. Not surprisingly, fruit soluble solids content correlated strongly with fruit specific gravity (correlation coefficients ranged across harvests between 0.87 and 0.74). Advancing maturity results in an increase in soluble sugars and pectins thereby increasing fruit density or specific gravity. Soluble solids content and specific gravity also correlated reasonably well with fruit skin Hue color. Both of these indices increased as maturity advanced with increasing skin color development.
Elizabeth Mitcham, Murray Clayton, Bill Biasi, And Steve Southwick
Department of Pomology,
University of California, Davis
Davis, CA 95616
13th Annual Postharvest Conference