|
Description
|
Supplementary Table S5. Model statistics for the best models and post-hoc contrasts. The effect of month, cladode size, cladode age cladode order from soil, number of daughter cladodes above a given cladode on the concentrations of 4-Hydroxybenzoic acid (4-HBA), Chlorogenic acid (CGA) and Quercetin (QUE) in female and hermaphrodite cladodes of individuals of Opuntia robusta from Central Mexico.
To analyze how concentrations of the different substances were related to cladode length, width, thickness, months, age, cladodes order from soil, and cladodes above a given cladode, we used generalized linear models (GLMs) created in R (R Core Team, 2020). Since many concentrations were null, we analyzed only the positive concentrations. The full model contained the interaction between each individual variable and sex. For each response variable, we used different error distribution and link functions that gave the best fit of the models. To analyze the effect of months and age on 4-HBA concentrations, we used the Gamma error distribution (GED) and the inverse link function. For the effect of months on CGA concentrations, we used GED and log link functions. For the effect of cladode age on CGA concentrations, we used GED and inverse link functions. For the effect of months on QUE concentration, we used GED and log link functions. For the effect of cladode age on QUE concentration, we used GED and inverse link function. For the models that related substances with cladode length, width and thickness, we used different models: for 4-HBA concentrations we used a GED with square root link function. For CGA concentrations, we used GED and identity link function; for QUE concentrations, we used GED and inverse link function. To test the effect of cladodes order from soil on the concentrations of the three substances, we used GED and square root link function. To test the effect of the number of cladodes above a given cladode on the concentrations of the three substances, we used GED and inverse link function. We checked the goodness of fit of the linear models (GLM or GLMM) by plotting the standardized residuals against fitted values and by normal QQ‐plots, and revised the assumptions of the homoscedasticity, proper distribution used, and independence. We chose the best GLM model used in this study according to the lowest AIC (Akaike, 1987) following the recommendations of Zuur et al. (2010).
Table 1. The effect of months on the concentrations of 4-Hydroxybenzoic acid (4-HBA) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 2. The effect of months on the concentration of Chlorogenic acid (CGA) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 3. The effect of months on the concentration of Quercetin (QUE) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 4. The effect of cladode age class determined by the color of spines on the concentration of Hydroxybenzoic acid (4-HBA) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 5. The effect of cladode age class determined by the color of spines on the concentration of Chlorogenic acid (CGA) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 6. The effect of cladode age class determined by the color of spines on the concentration of Quercetin (QUE) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 7. The effect of cladode length, width and thickness on the concentrations of 4-Hydroxybenzoic acid (4-HBA), Chlorogenic acid (CGA) and Quercetin (QUE) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 8. The effect of cladode order from soil on the concentrations of 4-Hydroxybenzoic acid (4-HBA), Chlorogenic acid (CGA) and Quercetin (QUE) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts.
Table 9. The effect of the number of levels of daughter cladodes above a given cladode on the concentrations of 4-Hydroxybenzoic acid (4-HBA), Chlorogenic acid (CGA) and Quercetin (QUE) in female and hermaphrodite cladodes of the individuals of Opuntia robusta from San Nicolás Tecoaco (trioecious population), Central Mexico. Model statistics and post-hoc contrasts. |
