Because texture is a primary driver of watermelon acceptability, the development of methods to test for small differences in texture between new cultivars would be of great utility to fruit breeding efforts. The objective was to investigate the effect of the shear-to-compressive force ratio in puncture tests on watermelon, then design new probes that would improve the test's sensitivity. A new hollow probe design of increased shear force (compactness = 11.6 mm(2)/mm(2)) was more sensitive at quantifying watermelon tissue mechanical properties when compared to the industry standard Magness-Taylor probe (compactness = 1 mm(2)/mm(2)). Compressive force applied is constant between the two. The hollow probe was more sensitive to differences between tissue types, though was not able to discriminate between cultivars, using the maximum force value. Based upon the improved performance of the hollow probe with tissue types, a high-shear 'snowflake' probe was designed and compared to the hollow and MagnessTaylor probes. The Magness-Taylor probe misclassified tissue types in 42% of samples tested, while the hollow and snowflake probes performed better, misclassifying 32% and 34% of samples, respectively. This was an improved accuracy over the Magness-Taylor, but the hollow and snowflake probes were not significantly different (alpha = 0.05) from each other. These results suggest that of the two, the hollow probe, due to its simplicity, offers an improvement over the industry standard Magness-Taylor in maximum force parameter applications. (C) 2014 Elsevier Ltd. All rights reserved.