Non-parametric Statistical Formulas for Factors of Safety of Plant Stems
β Scribed by Karl J. Niklas; Jans Varna; Lars A. Berglund
- Publisher
- Elsevier Science
- Year
- 1999
- Tongue
- English
- Weight
- 162 KB
- Volume
- 197
- Category
- Article
- ISSN
- 0022-5193
No coin nor oath required. For personal study only.
β¦ Synopsis
A previously proposed statistical approach for computing factors of safety (i.e. numerical measures of mechanical reliability) for any load bearing structure, like a vertical plant stem, is here extended to cope with organic structures whose morphological or mechanical properties have Weibull frequency distributions. This approach is illustrated using the actual length L and critical buckling length L cr of flower stalks (peduncles) collected from isogenic garlic (Allium sativum) populations grown under windy field and protected glasshouse conditions. Our analyses of the data indicate that L and L cr of peduncles harvested from both populations have Weibull frequency distributions, that the factor of safety for glasshouse grown peduncles is very near unity (i.e. S = 1.03), and that the factor of safety of field grown peduncles is 73% higher than that of glasshouse grown plants (i.e. S = 1.73). Comparisons between the S-values computed on the basis of our formulas and on the basis of the quotient of the mean values of L cr and L for each of the two populations indicate that the statistical method gives biologically realistic S-values and that the difference in the S-values for stems grown under protected and unprotected environmental conditions likely reflects the effects of chronic mechanical perturbation (due to wind-induced drag) on normal stem growth and development.
π SIMILAR VOLUMES
## Abstract The significance level of a nonβparametric linkage (NPL) statistic is often found by simulation since the distribution of the test statistic is complex and unknown. Ideally, simulation occurs by randomly assigning founder genotypes and then simulating meiotic events for the descendants
Counting processes and their compensators are introduced at a heuristic level. The martingale property of stochastic integrals with respect to a compensated counting process leads to moment estimates and asymptotic normal distributions for statistics arising in multiple state, non-parametric and sem